Osmotic laxatives (oral medications) ^*

A prospective case series [EL=3] showed that PEG 3350 plus electrolytes administered orally in hospital for up to 7 days was effective in achieving disimpaction in constipated children.

One prospective case series [EL=3] showed that a solution of PEG 3350 without electrolytes, at an initial dose ~1g/kg body weight per day (14 ml/kg/day solution) given in two divided doses for 8 weeks, was effective in decreasing soiling frequency, painful defecation, fear of defecation or stool withholding, faecal rectal impaction and dilated rectal vault after 6 weeks of treatment. It was also effective in resolving completely abdominal rectal masses after treatment.

One RCT [EL=1-] showed that PEG 3350 administered orally in four different doses (0.25, 0.5, 1.0 and 1.5 g/kg/day) for 3 consecutive days was effective in achieving disimpaction in constipated children. It also showed that higher doses of PEG 3350 were more effective than lower doses in achieving disimpaction in constipated children.

One RCT [EL=1-] showed that a pineapple isotonic intestinal lavage solution containing PEG 3350 administered during 2 consecutive days was more effective than mineral oil administered as 2 to 8 tablespoons in two divided doses for 2 days in producing the first bowel movement and in increasing bowel movements after treatment but less effective in resolving palpable abdominal masses. PEG 3350 also showed better compliance and fewer side effects in children taking mineral oil compared to children taking lavage solution.

One multicentre retrospective cohort study [EL=2-] showed that PEG 3350 plus electrolytes was more effective in achieving disimpaction within 5 days in children with constipation when compared to children who received enemas and suppositories and those who underwent manual evacuation of the bowel under anaesthesia.

Faecal softeners

One RCT [EL=1-] showed that a pineapple isotonic intestinal lavage solution containing PEG 3350 administered during 2 consecutive days was more effective than mineral oil administered as 2 to 8 tablespoons in two divided doses for 2 days in producing the first bowel movement and in increasing bowel movements after treatment but less effective in resolving palpable abdominal masses. It also showed better compliance and fewer side effects in children taking mineral oil compared to children taking lavage solution.

Osmotic laxatives and stimulant laxatives (rectal medications)

One multicentre retrospective cohort study [EL=2-] showed that enemas and suppositories were less effective in achieving disimpaction within 5 days in children with constipation when compared to children who received PEG 3350 plus electrolytes.

Manual evacuation of the bowel under general anaesthesia

One multicentre retrospective cohort study [EL=2-] showed that manual evacuation of the bowel under general anaesthesia was less effective in achieving disimpaction within 5 days in children with constipation when compared to children who received PEG 3350 plus electrolytes. It also showed that children who underwent manual evacuation of the bowel under general anaesthesia experienced more vomiting when compared to children who received macrogol 3350 plus electrolytes and those who received enemas and suppositories.

Stimulant laxatives (oral medications)

There is no evidence for the effectiveness of stimulant laxatives (oral medications) for treating disimpaction in children with constipation.

Laxatives versus laxatives

Osmotic laxatives versus osmotic laxatives: polyethylene glycol (PEG) versus lactulose

One meta-analysis of four RCTs comparing polyethylene glycol (PEG) versus lactulose showed that treatment success was significantly higher for PEG compared to lactulose (see figure 5.1).

Figure 5.1

Polyethylene glycol (PEG) versus lactulose in the ongoing treatment/maintenance of idiopathic constipation in children.

It should be noted that different types of PEG, different definitions of treatment success and different assessment points were used in the studies.

A double-blind RCT^** conducted in the UK⁶⁴ (2006) [EL=1+] assessed the efficacy of PEG 3350 plus electrolytes (PEG 3350+E;) as oral monotherapy in the treatment of faecal impaction in children and to compare PEG 3350+E with lactulose as maintenance therapy in a randomised trial. The RCT included 65 children with intractable constipation that had failed to respond to conventional treatment (mean age 5.7 years, 56% children 5 to 11 years, 68% boys). All children received PEG 3350+E (13.8 g powder dissolved in at least 125 ml water per sachet) administered orally in hospital according to an escalating dosing regimen until disimpaction was achieved (up to 7 days). Fifty-eight children (67% boys, mean age 5.7 ± 2.6 years, range 2 to 11 years) entered phase 2 of the study and were randomised to receive PEG 3350+E (13.8 g powder dissolved in at least 125 ml water per sachet) or lactulose (10 g powder dissolved in at least 125 ml water) for 12 weeks. For both medications children received oral maintenance doses starting with half of the numbers of sachets required for disimpaction per day. Additional laxative treatment with senna was allowed as rescue medication if the response to a single agent alone was judged inadequate by the investigator.

There were no significant differences at baseline between the two treatment groups regarding age, sex, height and weight. Children taking PEG 3350+E (n=27) had significantly more successful defecations per week (last on-treatment value) compared to children taking lactulose (n=26) (9.4, SD 4.56, range 2 to 24 versus 5.9, SD 4.29, range 2 to 23 with difference in means 3.5, 95% CI 1.0 to 6.0; P = 0.007). No children taking PEG 3350+E reimpacted whereas seven children taking lactulose did (23%; P = 0.011). No children taking PEG 3350+E needed to use senna as rescue medication whereas eight children taking lactulose did (31%; P = 0.002). The mean number of sachets used each day for children taking PEG 3350+E (n=27) was 0.91 (SD 0.41) whereas for children taking lactulose (n=26) it was 2.41 (SD 0.91). There were no significant post-treatment differences in mean values per patient between the two groups with respect to: predominant bowel movement form, pain, straining, soiling and overall assessment of treatment. Sixty-four percent of children on PEG 3350+E (n=27) experienced adverse effects compared to 83% of children on lactulose. There was a similar incidence of adverse effects in each age group. The most commonly reported events were gastrointestinal and resolved during the study. No clinically significant abnormal values were observed in urine and plasma electrolytes after 12 weeks of maintenance therapy in either group.

A double-blind RCT conducted in France⁶⁶ (2005) [EL=1+] assessed the safety of a PEG 4000 laxative without additional salts in paediatric patients. The RCT included 96 children aged 6 months to 3 years (51 male) with constipation despite their usual dietary treatment for at least 1 month. Children were randomised to receive either PEG 4000 (starting dose one sachet [4 g] and one placebo to be taken at breakfast) or lactulose (starting dose 1 sachet [3.33 g] and one placebo to be taken at breakfast) for 3 months. For both drugs, the dose could be doubled if it was ineffective in children aged 13 months to 3 years. If the maximum authorised dose was unsuccessful, one micro-enema of glycerol per day could be prescribed for a maximum of 3 consecutive days. If the child did not produce stools after treatment, two enemas could be administered at a 48 hour interval. This procedure was only allowed twice during the study. If the child produced liquid stools for more than 1 day or more than two or three stools per day depending on age, the dose could be decreased by one pair of sachets per day to a minimum of one pair of sachets every other day and possibly to transitory interruption. Outcomes were assessed at day 42 and day 84 after initiating treatment. There were no clinically relevant differences between the two treatment groups at baseline for clinical or biological parameters.

At day 42 the median stool frequency (interquartile range) was not significantly different for babies (aged 6 to 12 months) taking PEG 4000 compared to babies taking lactulose. However, for toddlers (aged 13 months to 3 years) taking PEG 4000 (n=51) the stool frequency increased significantly more than for toddlers taking lactulose (n=45) (8 stools per week, interquartile range 6 to 10 versus 6, interquartile range 5 to 7; P = 0.013). At day 84 there were no significant differences in either babies or toddlers for both treatment groups regarding stool frequency. At day 42 significantly more children taking lactulose (14 out of 41, 34%) reported a higher frequency of hard stools compared to children taking PEG 4000 (4 out of 46, 9%; P = 0.003). This remained the case at day 84 (for PEG 4000 3 out of 47, 6% versus lactulose 11 out of 40, 28%; P = 0.008).

At day 42 significantly more children taking lactulose (19 out of 44, 43%) reported using enemas compared to children taking PEG 4000 (14 out of 48, 30%). This remained the case at day 84 (for PEG 4000 8 out of 48, 17% versus lactulose 17 out of 42, 41%; P = 0.012). Faecal impaction was diagnosed in significantly more patients taking lactulose compared to children taking PEG 4000 (for PEG 4000 1 out of 51, 2% versus lactulose 6 out of 45, 13%; P = 0.049). There were no significant differences in the doses used for both medications in either babies or toddlers (babies taking PEG 1 sachet/day, median interquartile range 0.9 to 1 versus babies taking lactulose 1 sachet/day, 1 to 1.3 and toddlers taking PEG 1 sachet/day, 1 to 3 versus toddlers taking lactulose 1.1 sachet/day, 0.9 to 1.5). Treatment was stopped in one child in the lactulose group because of lack of efficacy.

A double-blind RCT conducted in the Netherlands⁶⁷ (2004) [EL=1+] compared the clinical efficacy and safety of PEG 3350 plus electrolytes (PEG 3350+E [Transipeg®, Mundipharma Medical Company]) and lactulose in paediatric idiopathic constipation. The RCT included 91 children aged 6 months to 15 years with constipation (49 male). During the run-in phase (1 week before treatment) no laxatives were allowed and at the end all patients received one enema daily for 3 days. Children age 6 years or under received 60 ml of Klyx (sodium dioctylsulfosuccinate and sorbitol) and children over 6 years received 120 ml Klyx. During the initial phase children were randomised to receive either PEG 3350+E or lactulose for 8 weeks (children aged 6 months to 6 years (inclusive) received 1 sachet (2.95 g) of PEG 3350+E per day or 1 sachet (6 g) of lactulose per day and children older than 6 years received 2 sachets (5.9 g) of PEG 3350+E per day or 2 sachets (12 g) of lactulose per day). Overall treatment success was defined as three or more bowel movement per week and one encopresis episode or less every 2 weeks.

After 8 weeks there were no significant differences regarding both defecation frequency per week and encopresis frequency per week for children taking PEG 3350+E compared with children taking lactulose. Success percentages were significantly greater for children taking PEG 3350+E compared to children taking lactulose (56, 95% CI 39 to 70 versus 29, 95% CI 16 to 44; P = 0.02). Significantly more sachets a day were taken by children on lactulose compared to children on PEG 3350+E (mean 2.4, SD 0.4 versus mean 1.99, SD 0.3; P = 0.03). No serious or significant side effects were recorded. Significantly more adverse effects (abdominal pain, pain at defecation and straining at defecation) were seen in patients taking lactulose compared to PEG 3350+E (P < 0.05). There were no significant differences between the two groups regarding: bloating, diarrhoea, flatulence, nausea, hard stool consistency and vomiting.

Significantly more children complained of bad palatability of PEG 3350+E compared to lactulose and this caused the premature withdrawal of one patient. There were no significant differences at baseline between the two groups with respect to: age, sex, defecation frequency, encopresis, large amounts of stool and faecal impaction. Nine children dropped out of the study: four children in the PEG 3350+E group and five in the lactulose group. Two children in each group were lost to follow-up. Overall treatment success was independent of age (under 6 years and 6 years or over) and use of laxatives for more than 1 year prior to the start of the study. In children treated for less than 1 year a significant difference in success was found between those treated with PEG 3350 and those treated with lactulose (63% versus 1% respectively; P = 0.02).

An open label RCT (crossover) conducted in the USA⁶⁸ (2002) [EL=1-] compared the efficacy of PEG 3350 and lactulose in the treatment of chronic constipation in children. Forty-four children aged 2 to 16 years (mean 7.8 ± 3.7) referred for evaluation of constipation were included. Children were randomised to receive either PEG 3350 without electrolytes 10 g/m2/day orally for 2 weeks (mean weight adjusted dose: 0.3 g/kg/day, range 0.2 to 0.5) or lactulose 1.3 g/kg/day orally for 2 weeks. There was no washout period in between the two medications. Outcome measures were stool frequency, stool form, ease of passage, effectiveness (global assessment, as reported by parent or guardian) and laxative preference (based on efficacy, ease of administration and side effects). The mean number of bowel movements, the stool form (mean sum of scores) and the easy of stool passage (mean sum of scores) were not significantly different in children taking PEG 3350 compared to children taking lactulose. PEG 3350 was significantly more effective than lactulose (PEG 3350 (n=37): 84% effective versus lactulose (n=37): 46% effective; P = 0.002). Seventy-three percent of patients said they preferred PEG 3350 compared to 27% who said they preferred lactulose. Seven patients withdrew during the first 2 week treatment period due to lack of efficacy of the assigned intervention: six of these patients were taking lactulose at the time of withdrawal.

Osmotic laxatives versus faecal softeners

An open label RCT conducted in Iran⁷² (2007) [EL=1-] compared the clinical efficacy and safety of liquid paraffin and lactulose in the treatment of idiopathic childhood constipation. The study included 247 children (127 male) with chronic idiopathic constipation aged 2 to 12 years (mean age 4.1 years ± 2.1). All children received one or two enemas daily for 2 days to clear any rectal impaction (30 ml/10 kg body weight of paraffin oil). Children were randomised to receive either liquid paraffin orally (1 to 2 ml/kg twice daily) for 8 weeks or lactulose orally (1 to 2 ml/kg twice daily) for 8 weeks. For determination of the best dose for each child, parents were asked to increase the volume of each drug by 25% every 3 days as required to yield one or two firm to loose stools. Outcomes were measured during the first 4 weeks and during the last 4 weeks of treatment.

Stool frequency per week during the first and last 4 weeks of treatment increased significantly more in children taking liquid paraffin (n=127) compared to children taking lactulose (n=120) (first 4 weeks 12.1 ± 3.2 versus 9.2 ± 2.1, P < 0.001 and in the last 4 weeks 13.1 ± 2.3 versus 8.1 ± 3.1; P < 0.001). Encopresis frequency per week during the first 4 weeks of treatment decreased significantly more in children taking liquid paraffin compared to children taking lactulose (first 4 weeks 1 ± 4.3 versus 2 ± 4.6; P = 0.07). During the last 4 weeks no child on liquid paraffin experienced encopresis compared to a frequency of 3 ± 4.1 in children taking lactulose; P < 0.001. Success rate was significantly larger during the first 4 weeks and at the end of 8 weeks of treatment in children taking liquid paraffin compared to children taking lactulose (first 4 weeks: 90% versus 52%; P < 0.001), (end of 8 weeks: 85% versus 29%; P < 0.001). The final mean effective dose was significantly larger in children taking lactulose compared to children taking liquid paraffin (2.08 ml/kg/day ± 0.21 versus 1.72 ml/kg/day ± 0.13; P < 0.001).

An open label RCT conducted in Turkey⁷³ (2005) [EL=1-] determined and compared the efficacy, safety and optimal dose of liquid paraffin and lactulose in children with chronic idiopathic constipation. The study included 40 children aged 2 to 12 years (22 male, mean age 3.7 years ± 2.7) referred for evaluation of constipation with evidence of faecal impaction. Children were randomised to receive either liquid paraffin or lactulose for 8 weeks. The medication was administered orally as a suspension at 1 ml/kg, twice a day for each drug. For determination of the best dose for each child, parents were asked to increase or decrease the volume of each drug by 25% every 3 days as required, to yield two firm to loose stools per day. The maximum dose used throughout the study was 3 ml/kg/day for each drug. Outcomes were measured at 4 and 8 weeks after initiation of treatment. Stool frequency and stool consistency were recorded by the parents in daily diary forms (stool consistency scoring: 1, hard; 2, firm; 3, loose stools).

The mean stool consistency during the first 4 weeks of treatment improved significantly more for children taking lactulose (n=20) compared to children taking liquid paraffin (n=20) (1.71 ± 0.5 versus 2.17 ± 0.5; P < 0.01). There were no significant differences in stool consistency when comparing both groups during the last 4 weeks of treatment. The stool frequency per week increased significantly more in children taking liquid paraffin compared to children taking lactulose, both during the first and the last 4 weeks of treatment (first 4 weeks: mean 13.3 ± 4.2 versus 10.2 ± 4.4; P < 0.05), (last 4 weeks: mean 16.1 ± 2.2 versus 12.3 ± 6.6; P < 0.05). The optimal dose of drugs was not significantly different for children taking liquid paraffin compared to children taking lactulose (mean 1.88 ml/kg/day ± 0.27 versus 2.08 ml/kg/day ± 0.27). These data were reported in a table and it was assumed that this represented the whole study period.

Data reported in text for the last 4 weeks of treatment established the optimal dose for liquid paraffin was 1.72 ± 0.18 and for lactulose it was 1.82 ± 0.57. Compliance rate during the first 4 weeks of treatment was not significantly different when comparing the two groups. At the end of 8 weeks significantly more children complied with taking liquid paraffin than with taking lactulose (n=90% versus n=60%; P = 0.02).

Stimulant laxatives versus faecal softeners

A single blind RCT conducted in the USA⁷⁴ (1982) [EL=1-] compared the efficacy of mineral oil and standardised senna concentrate (Senokot®, Reckitt Benckiser Healthcare (UK) Limited) in the treatment of idiopathic constipation in children. The RCT included 37 children aged 3 to 12 years treated for chronic idiopathic constipation in a specialist clinic. Children received a 5-day course of oral bisacodyl (most patients) and daily enema for 3 to 5 days in addition (a minority). Children were randomised into two groups. Group 1 (n=19) received mineral oil orally twice a day in doses sufficient to induce loose stools and leakage of oil per rectum. After the first week of treatment, the dose was reduced until the leakage ceased. This dose (range 1.5 to 5.0 ml/kg/day) was maintained for minimum of 3 months. The second group (n=18) received senna (tablet or syrup) in doses sufficient to induce at least 1 bowel movement daily during the first 2 weeks of treatment. This dose was maintained for 3 months. Tapering was accomplished by changing from daily to every other day and then every third day of medication. Treatment lasted approximately 6 months. Children in the mineral oil group were followed up for an average of 10.1 months while children in the senna group were followed up for an average of 10.5 months.

At 1 month the percentage of patients experiencing daily bowel movement was not significantly different when comparing the two groups. At 3 months all children on mineral oil were experiencing daily bowel movements compared to 72% of children on senna (P < 0.05). At the final follow-up significantly more children on mineral oil were experiencing daily bowel movements compared to children on senna (mineral oil: 89% versus senna 50%; P < 0.05). At all three assessment points daily soiling decreased significantly more in children taking mineral oil compared to children taking senna (at 1 month: mineral oil 11% of patients versus senna 39%, P < 0.05; at 3 months: mineral oil 11% of patients versus senna 50%, P < 0.05; at final follow-up: mineral oil 6% of patients versus senna 44%, P < 0.05). Sixty-eight per cent of children on mineral oil were reliably compliant with medication during the first 3 months of treatment compared to 78% of children on senna. At the latest follow-up 55% of children on mineral oil successfully discontinued regular medication compared to 22% of children on senna. An additional 33% of children discontinued senna because of unacceptable symptom control. Forty-five percent of children in each group remained on regular medication. There were significantly more episodes of recurrence or treatment of symptoms per month in children taking senna compared to children taking mineral oil (senna mean 0.34 ± 0.36 versus mineral oil 0.09 ± 0.08; P < 0.01). There were no significant baseline differences between the two groups regarding mean age, median age at onset of symptoms, percentage of patients who had received prior treatment for constipation, gender ratio, faecal soiling, overt retentive behaviour, enuresis, ‘difficult’ toilet training and primary failure of toilet training. One patient on mineral oil was lost to follow-up after the 3-month visit and not considered in the results. There was no attrition or loss to follow-up in the senna group.

Laxatives versus placebo

A double-blind RCT (cross over, multicentre) conducted in the UK⁷⁵ (2008) [EL=1+] assessed the efficacy and safety of PEG 3350+E for the treatment of chronic idiopathic constipation in children. The study included 51 children aged 24 months to 11 years (29 girls) with chronic constipation for at least 3 months. Children were randomised to receive PEG 3350+E (6.9 g powder/sachet) or placebo (6.9 g powder/sachet) for 2 weeks with a 2-week washout period in between. The dosing regimen for PEG 3350+E and placebo for children aged 2 to 6 years was: 1 sachet/day on days 1 and 2, 2 sachets/day taken together on days 3 and 4), 3 sachets/day (2 morning, 1 evening) on days 5 and 6 and 4 sachets/day (2 morning, 2 evening) on days 7 and 8. For children aged 7 to 11 years the dosing regimen was: 2 sachets/day taken together on days 1 and 2, 2 sachets/day taken together on days 3 and 4), 5 sachets/day (2 in the morning, 3 in the evening) on days 5 and 6, and 6 sachets/day (3 in the morning, 3 in the evening) on days 7 and 8. For both groups if diarrhoea was present, doses were decreased by two sachets or parents were instructed to miss a day of medication. If there were loose stools doses were decreased by one sachet.

Children on PEG 3350+E experienced significantly more complete defecations per week compared to children on placebo, both for the intention to treat (ITT) population and the per protocol (PP) population. Results for the ITT population were PEG+E (n=47): mean 3.12, SD 2.050, range 0.00 to 8.87 versus placebo (n=48): mean 1.45, SD1.202, range 0.00 to 3.73; treatment difference 1.64; P < 0.001, 95% CI 0.99 to 2.28. Results for the PP population were PEG+E (n=36): mean 3.63, SD 1.980, range 0.00 to 8.87 versus placebo (n=36): mean 1.63, SD 1.229, range 0.00 to 3.73; treatment difference 1.96; P < 0.001, 95%CI 1.19 to 2.72. Data do not include the washout period. Children on PEG 3350+E (ITT population) experienced significantly more defecations in general compared to children on placebo (PEG+E (n=47): mean 5.68, SD 2.771 versus placebo (n=47): mean 4.10, SD 2.503; treatment difference 1.58; P = 0.003, 95% CI 0.55 to 2.60). Children on PEG 3350+E (ITT population) experienced significantly less pain on defecation compared to children on placebo (PEG+E (n=47): mean 0.49, SD 0.727 versus placebo (n=47): mean 0.77, SD 0.863; treatment difference -0.28; P = 0.041, 95% CI −0.52 to −0.01). Children on PEG 3350+E (ITT population) experienced significantly less straining on defecation compared to children on placebo (PEG+E (n=47): mean 0.72, SD 0.789 versus placebo (n=47): mean 1.37, SD 1.041; treatment difference -0.65; P = 0.001, 95%CI -0.97 to -0.33). The stool consistency improved significantly more in children on PEG 3350+E compared to children on placebo (PEG+E (n=47): 1.73, SD 0.497 versus placebo (n=47): 2.21, SD 0.556; treatment difference -0.48; P = 0.001, 95% CI -0.68 to -0.27). The percentage of hard stools decreased significantly more in children on PEG 3350+E compared to children on placebo (PEG+E (n=47): 14.64, SD 26.041 versus placebo (n=47): 38.19, SD 39.508; treatment difference -23.55; P < 0.001). There were no significant differences between children on PEG 3350+E and children on placebo regarding abdominal pain on defecation and faecal incontinence. The mean effective dose of PEG 3350+E was 0.6 g/kg/day in children aged 2 to 6 years and 0.7 g/kg/day in children aged 7 to 11 years.

One double-blind RCT (multicentre) conducted in the USA⁷⁶ (2008) [EL=1+] established the efficacy and best starting dose of PEG 3350 in the short-term treatment of children with idiopathic constipation. The study included 103 children aged 4 to 16 years (69 boys, mean age 8.5 years ± 3) with chronic constipation. Patients taking other laxatives were only included if they had less than three bowel movements per week while taking the laxative. All children received behavioural treatment consisting of instructions to sit on the toilet for 10 minutes twice after meals, positive reinforcement using age-appropriate printed calendars and special stickers for days without episodes of faecal incontinence and others with bowel movements. Children were randomly assigned in blinded fashion in a 1:1:1:1 ratio within each participant site into four groups. Group 1 received PEG 3350 without electrolytes at 0.2 g/kg/day single dose (maximum 8.5 g per day), group 2 received PEG 3350 without electrolytes at 0.4 g/kg/day single dose (maximum 17 g per day), group 3 received PEG 3350 without electrolytes at 0.8 g/kg/day single dose (maximum 34 g per day) and the last group received a placebo. Treatment lasted 3 weeks. Assessments were conducted at 7 and 14 days after medication started. Response to treatment was defined as 3 or more bowel movements (BM) during the second week of treatment. Patients were considered failures and withdrawn from study if they had no BM for 7 days or developed faecal impaction at any point; however intention to treat analysis was performed.

There were no significant differences in baseline characteristics between the four groups. The percentage of children who responded to treatment was significantly higher when comparing each and all treatment groups with placebo (group 1 (n=26): 77%, group 2 (n=27): 74%, group 3 (n=26): 73%, placebo (n=24): 42%; P < 0.04 each group versus placebo; P = 0.026 all treatment groups versus placebo). There were no significant differences between treatment groups regarding this outcome. There were no significant predictors of success by controlling for age, duration of constipation, prior laxative use, presence of stool in rectum, gender and presence of faecal incontinence at baseline. There was a significant increase in the final number of bowel movements in the different treatment groups compared to placebo (overall difference between treatment groups and placebo P = 0.017; P = 0.015 dose-response trend). Note that figures for the after treatment were reported in a graph from which it is difficult to extract the data.

There was no significant difference in weekly number of faecal incontinence episodes among the four groups. Stool consistency became softer in all treatment groups compared with placebo and comparing all treatment groups with each other. Changes in stool consistency measured as number of faecal incontinence episodes were: group 1: before 2.8 ± 0.8, after 2.1 ± 0.7; group 2: before 2.6 ± 0.9, after 1.7 ± 0.6; group 3: before 2.9 ± 0.7, after 1.5 ± 0.7; placebo: before 3.0 ± 0.8, after 2.4 ± 0.9; P < 0.003 each group versus placebo; P < 0.003 test for trend; P < 0.003 overall difference amongst treatment groups. Straining decreased in all treatment groups compared with placebo, particularly for those in group 2 and group 3 (mean straining scores: group 1: before 2.3 ± 1.1, after 1.4 ± 0.9; group 2: before 1.9 ± 1.2, after 1.0 ± 1.0; group 3: before 2.0 ± 1.0 after 0.9 ± 0.6; placebo: before 2.7 ± 1.2, after 1.5 ± 1.2; P < 0.003 each group versus placebo; P < 0.003 test for trend; P < 0.003 overall difference between treatment groups. There were no significant difference amongst groups regarding incidence and severity of adverse effects (group 1: 9 out of 26 [34.6%], group 2: 16 out of 27 [59.3%], group 3: 17 out of 26 [65.4%], placebo: 14 out of 24 [58.3%]).

There were no differences in the type of non-gastrointestinal (GI) related events, the most common being headaches. There was a higher incidence of GI related events in patients receiving PEG versus placebo. As the dose of PEG increased, it also increased incidence of flatulence, abdominal pain, nausea and diarrhoea. There were no electrolyte abnormalities or differences in laboratory values among groups. Treatment failure was similar in all treatment groups but lower than the placebo (number of children who failed in group 1 was 6 out of 26 [4 BM frequency criteria, 2 with stool impaction]; in group 2 it was 7 out of 27 [3 BM frequency criteria, 4 with stool impaction]; in group 3 it was 7 out of 26 [6 BM frequency criteria, 1 with stool impaction] and in placebo (n=24) it was 14 [all related to BM frequency criteria]). Fourteen patients did not complete the 2-week treatment: eight because of treatment failure (five with impaction [two Group 1, three Group 2] and three with more than 7 days without a BM [two Group 1, one Group 3]), three because of adverse events, there was one withdrawal (lack of response (placebo)) and two cases of non-compliance (one Group 2, one Group 3). Three serious adverse events occurred requiring hospitalisation (two cases of impaction and one case of exacerbation of bipolar or depression).

Laxatives versus other interventions

Laxatives versus laxatives

Osmotic laxatives versus faecal softeners

Osmotic laxatives versus stimulant laxatives

Stimulant laxatives versus faecal softeners

Bulk forming laxatives

No evidence was found for the clinical effectiveness of bulk forming laxatives for ongoing treatment and/or maintenance in children with chronic idiopathic constipation.

Laxatives versus placebo

Laxatives versus other interventions

Adverse effects up to 6 months of treatment

Adverse effects at between 6 and 12 months of treatment

Adverse effects at/after 12 months of treatment

Bulk forming laxatives

No evidence was found of adverse effects of the medium- to long-term use of bulk forming laxatives.

Tables 5.1, 5.2 and 5.3 summarise the results of these studies.

Table 5.1

Adverse effects up to 6 months of treatment.

Table 5.2

Adverse effects between 6 and 12 months of treatment.

Table 5.3

Adverse effects at or after 12 months of treatment.

Infant formulae

One double-blind RCT (crossover) conducted in The Netherlands⁹⁰ (2007) [EL=1+] tested the hypothesis that Nutrilon Omneo (new formula, NF) would have a positive effect on stool characteristics in constipated children. The study included 38 otherwise healthy, term infants with constipation aged 3 to 20 weeks who received at least two bottles of milk-based formula per day (19 boys, median age 1.7 months). Infants were randomised to receive either NF or standard formula (SF) in period 1 and crossed over after 3 weeks to treatment period 2. Each treatment period lasted 3 weeks. Feeding patterns were not described. NF composition (per 100 ml) differed from the SF in that its protein content was higher, 100% of it was based on whey protein hydrolysate (no casein, no intact whey protein) and it contained a mixture of prebiotic oligosaccharides (galacto-oligosaccharide [GOS] and fructo-oligosaccharide [lcFOS]), a higher concentration of sn-2 palmitic acid and a lower lactose content.

Defecation frequency, improvement of hard to soft stools and number of children experiencing no painful defecation were not significantly different between the two treatment groups after period 1. After the crossover, painful defecation and defecation frequency were not significantly different between the periods on NF and SF. Seventeen percent (n=4) of infants had soft stools when receiving NF but hard stools with SF, compared to no infant with soft stools when receiving SF and no infant with hard stools when receiving NF (P = 0.046).Throughout the study there were no serious adverse effects in either group. Both formulae were well tolerated.

Only 24 children (63%) completed the crossover study. In period 1, three patients on SF dropped out: two patients stopped because of severe constipation and one patient switched to hypoallergenic feeding, because of suspected cows' milk protein allergy. Parents of one patient decided that they did not want to cross over because she was free of symptoms and they started openly with NF instead. Three patients dropped out after switching to NF: two patients stopped after less than one week because of recurrence of constipation symptoms and one patient was lost to follow-up. Seven patients dropped out after switching to SF: six patients stopped after one week because of recurrence of constipation symptoms and one patient was lost to follow-up. Data analysis was based on the group of 35 patients that completed period 1 and the subgroup analysis of 24 patients who completed the crossover. There were no significant differences in baseline characteristics between the two groups.

One prospective case series conducted in Italy⁹¹ (2003) [EL=3] investigated whether a new infant formula commercially available in Italy was useful as a dietary option in infants with minor feeding problems. The study included 604 formula-fed healthy term infants aged up to 3 months seen by paediatrician because of colic and/or constipation and/or regurgitation (age at entry of the total sample was 1.35 months ± 0.77, gender not reported). Of these, 232 infants were diagnosed with constipation, defined as a stool frequency of less than one stool a day. During 14 days all infants received a new formula (NF)^{†††††}. The feeding volume was based on a feeding on demand procedure. The feeding frequency was decided by the parents and not influenced by the study protocol.

The study found that 147 infants (63.4%) reported an increase in the number of stools per day during the study period compared to baseline, with a significant average increase of 0.42 (95% CI 0.55 to 0.27; P < 0.005). The average increase between day 1 and day 7 was 0.41 (95% CI 0.51 to 0.23; P < 0.05) and between day 7 and day 14 it was 0.04 (NS). There was no improvement of symptoms in 85 infants (26.6%). Mean parent evaluation of formula (on a score of 1 to 10) was 7.9 ± 1.8. A total of 550 parents (91%) gave a positive judgement (score 6 to 10). A total study population of 932 infants were enrolled and 604 (65%) completed the study protocol. A total of 358 infants were excluded from the study: 154 completed only the first step and did not return for the visit on day 14 while 131 infants were excluded because of incomplete data. Seventy-three infants required medication during the first week of study and were therefore excluded. The proportion of these infants who had constipation was not reported in the paper. It should be noted that stool consistency was not assessed in the study.

A prospective case series conducted in Spain⁹² (2008) [EL=3] assessed the prevalence of mild gastrointestinal disorders (MGDs) in infants fed with artificial milk formulas in paediatric practice and evaluated the effectiveness and satisfaction with dietetic treatment, specifically elaborated formulas belonging to the Novalac® line of products (United Pharmaceuticals). The study included 3487 infants with MGDs and fed with artificial milk formulae (52.2% boys, aged 1 to 17 weeks). Of these, 604 infants had constipation. For 30 days constipated children received Novalac Anti-Constipation®, a formula with an adapted concentration of magnesium and lactose. No other details regarding feeding volume or frequency were provided.

In total, 91.6% of cases of constipation resolved within 7 days, but this was not clearly defined in the paper. The number of daily stools increased significantly at the end of the study when compared to baseline (baseline: mean 0.6 ± 0.7 versus at 30 days: mean 1.7 ± 0.8). The percentage of children having normal stools increased significantly at the end of the study when compared to baseline (baseline: 33.40% versus at 30 days: 95.60%). The percentage of children presenting with pain or discomfort on defecation was significantly reduced at the end of the study when compared to baseline (baseline: 90.0% versus at 30 days: 10.4%). The percentage of children needing external help at defecation was significantly reduced at the end of the study when compared to baseline (baseline: 76.1% versus at 30 days: 8.8%). Ninety percent of parents reported being satisfied with the treatment. Adverse events (for all formulae, no subgroup analysis) were reported in 3.9% infants of the total population. Effectiveness was evaluated among 1441 infants (total population) who completed follow-up. Premature study termination due to adverse events occurred in 2.7% of cases, parent decision in 6.9%, loss to follow-up in 1.64%, protocol violations in 2.46% and non-specified reasons in 16.62%.

One open label RCT conducted in Taiwan⁹³ (2007) [EL=1-] evaluated a commercialised formula, Novalac-IT® (Intestinal Transit), against a ‘strengthened regular formula’, the traditional approach in infants with digestive problems in Taiwan. The study included 93 children aged 2 to 6 months referred to the paediatric gastroenterology clinic at a medical centre with constipation for 2 weeks or more and fed exclusively with formula milk (47 boys, mean age 3.8 months ± 1.7). Children were randomised to receive either a magnesium-enriched infant formula, Novalac-IT or a 20% strengthened infant formula for 2 months. Children were assessed at 2 weeks, 1 month and 2 months.

Outcomes measured were remission, improvement or failure according to a severity scoring system based on stool consistency, frequency and volume of stools and difficulties in defecation (1 to 3 mild constipation; 4 to 6 moderate; 7 or 8 severe). Asymptomatic children were considered in remission, a decrease in severity of 4 or more was considered a good response and a decrease in severity of 1 to 3 a fair response. If the score did not change or increased it was considered treatment failure. The severity scoring system comprised the following variables:

• stool consistency (hard stool 0, no hard stool 1, hard and long form 2)
• difficulties with defecation (no difficulties 0, irritability 1, crying 2)
• frequency of defecation (more than 3 times per week 0, 1 to 3 times per week 1, less than once per week 2)
• stool weight (more than 35 g/kg/week 1, 20 to 35 g/kg/week 2, less than 20 g/kg/week 3).

The number of children who improved was not significantly different in the two groups at 2 weeks. At 1 month significantly more children on Novalac-IT had improved compared to children on the strengthened formula (39 out of 47 [83%] versus 23 out of 46 [50%]; P = 0.002). At 2 months significantly more children on Novalac-IT had improved compared to children on the strengthened formula (42 [89%] versus 25 [54%]; P < 0.001). The number of children free of symptoms at 2 weeks was not significantly different between the treatment groups. However, both at 1 month and at 2 months, significantly more children on Novalac-IT were free of symptoms compared to children on the strengthened formula (at 1 month: Novalac-IT: 28 out of 47 [60%] versus strengthened formula: 16 out of 46 [35%]; P = 0.029; at 2 months: Novalac-IT: 35 out of 47 [75%] versus strengthened formula: 18 out of 46 [39%]; P < 0.001). There were no significant differences in the baseline characteristics (clinical or demographic) between the two groups. It should be noted that participation in the trial was proposed before a more complete diagnostic workup for cows' milk protein allergy, Hirschsprung's disease and others was conducted.

One open label RCT conducted in Italy⁹⁴ (2005) [EL=1-] evaluated the efficacy on digestive problems of a formula based on palmitic acid predominantly esterified at the β-position, oligosaccharides (GOS and FOS) with a prebiotic activity, partially hydrolysed protein, low lactose content and higher density. The study included 95 formula-fed healthy term infants aged 4 months or less with constipation, defined as a stool frequency of less than one stool a day (64.2% with hard stools) (50 boys, age at study entry in the intervention group was 1.55 months ± 0.88 and in the control group was 1.28 months ± 0.66). Children were randomised to receive either the new formula (NF) (Omneo®/Conformil®, Numico) or a standard formula (SF) for 14 days. The feeding volume was based on a feeding on demand procedure. Feeding frequency was decided by the parents and not influenced by the study protocol.

The stool frequency increased significantly more in children receiving NF compared to children receiving SF, both on day 7 and on day 14 (number/day, mean ± SD) (day 7: NF group (n=55): 1.79 ± 0.96 versus SF group (40): 1.31 ± 0.89; difference: 0.48 (95% CI: 0.09 to 0.87); P = 0.02]); (day 14: NF group (n=55): 2.04 ± 1.04 versus SF group (40): 1.64 ± 0.99, difference: 0.40 (95% CI: -0.03 to 0.83); P = 0.07). The stool frequency (number/day, mean ± SD) also increased significantly more in children receiving NF compared to children receiving SF, after adjusting for gender, age at entry, maternal instruction, parity, birth weight, number of feedings per day and stool frequency at entry (mean adjusted difference in stool frequency between the two groups for days 0 to 7 was 0.60 [CI 95% 0.19 to 1.01; P = 0.004] and for days 0 to 14 was 0.53 [95% CI 0.11 to 0.90; P = 0.015]). Post-treatment outcomes for stool consistency were not reported. There were no significant differences in the baseline characteristics between the two groups. No dropouts or children lost to follow-up were reported.

Increasing fibre

One double-blind RCT conducted in the Netherlands⁹⁵ (2008) [EL=1+] assessed the clinical efficacy and safety of a dietary fibre mixture and compared it with lactulose in the treatment of childhood constipation. The study included 135 children referred to hospital outpatient clinic for idiopathic constipation. Children were randomised to receive either a yogurt drink (125 ml) with 10 g of mixed dietary fibre (fibre mixture contained per 100 ml of solution: 3.0 g transgalacto-oligosacharides, 3.0 g inulin, 1.6 g soy fibre, 0.33 g resistant starch) or a yogurt drink containing lactulose (10 g/125 ml, Duphalac Lactulose®, Solvay Healthcare Limited). Forty-two children received yogurt with the fibre mix (20 boys, median age 5.5 years, range 1 to 12 years), whereas 55 children received the yogurt containing lactulose (23 boys, median age 5.0 years, range 1 to 12 years). Both products were taken at breakfast and when two or more bottles were needed they were also taken at lunch. The daily amount of fibre/fluid intake depended on the patient's body weight. If persistent diarrhoea was reported, the original dose was reduced by 50%. If clinical parameters compared to baseline did not improve 3 weeks after the start of intervention period, step-up medication (polyethylene glycol (PEG) 3350) was given per protocol. There was an intervention period lasting 8 weeks and a weaning period lasting 4 weeks when doses where reduced.

Defecation frequency per week and number of patients with one or more faecal incontinence episodes per week at 8 weeks was not significantly different between the two groups. Stool consistency (mean) was significantly softer in the lactulose group compared to the fibre group, both at 3 and at 8 weeks (at 3 weeks: fibre (n=42) 3.5, lactulose (n=55) 4.5; P < 0.01 and at 8 weeks: fibre 3.6, lactulose 4.0; P = 0.01). The number of patients using step-up medication at 3 weeks was significantly smaller in the group taking fibre than in the group taking lactulose (fibre: 13, lactulose: 7; P = 0.028) but there were no significant differences regarding this outcome at 8 and at 12 weeks. No serious or significant side effects were recorded. In the fibre group one child experienced dose-related persistent diarrhoea compared to two children in the lactulose group. No significant differences were found in baseline characteristics between the two groups. Thirty-three patients left the study: 22 in the fibre group after 1 to 56 days (median 7) and 11 in the lactulose group after 1 to 51 days (median 8) (P = 0.020). All those patients refused to drink the yogurt. Three patients were lost to follow-up: one on fibre and two on lactulose. Despite the high drop-out rate (24.4%) intention-to-treat analysis was not performed.

One double-blind RCT (crossover) conducted in the USA and Italy⁹⁶ (2004) [EL=1+] evaluated whether fibre supplementation with glucomannan is beneficial in the treatment of children with idiopathic constipation. The study included 31 otherwise healthy children ( ) older than 4 years who had chronic idiopathic constipation for 6 months or longer with or without encopresis (16 boys, age range 4.5 to 11.7 years, mean age 7.1 years ± 2.0). Disimpaction was carried out with one or two phosphate enemas if rectal impaction felt during rectal examination. Fifty-eight percent of patients continued with their pre-evaluation laxative during the whole study period. Children were randomised to receive Glucomannan B (one capsule containing glucomannan, a polysaccharide of d-glucose and d-mannose, equal to 450 mg of alimentary fibre) or Glucomannan A (one capsule containing maltodextrins as placebo). After 4 weeks children were switched to the other treatment for another 4 weeks, with no washout period in between. Both glucomannan and placebo were given at a dose of 100 mg/kg body weight daily (maximal 5 g per day), rounded to the nearest 500 mg because each capsule contained 500 mg. Each capsule was either: opened and sprinkled on food and given with 50 ml of fluid per capsule; given as a solution, whereby the content of each 500 mg capsule was mixed with 50 ml of fluid of the child's choice; or swallowed as a capsule with 50 ml of fluid for each capsule. In addition, parents were instructed to have the child sit on the toilet four times daily after meals and to keep a stool diary.

No enemas were given during each treatment period unless rectal disimpaction felt during rectal examination at assessment visits. Successful treatment was rated by physician and defined as 3 or more bowel movements per week and 1 or no soiling episode in the last 3 weeks with no abdominal pain. Parents' global assessments related to whether they believed that the child was better during the first or second treatment period.

Stool consistency and frequency of soiling episodes per week were not significantly different when comparing the fibre treatment period with the placebo period. However, significantly more children on placebo reported having less than 3 bowel movements per week compared to children on fibre (placebo (n=31): 52% versus fibre (n= 31): 19%; P < 0.05). Significantly more physicians rated the fibre treatment as ‘successful’ when compared to placebo (45% versus 13%; P < 0.05). Significantly more parents in the fibre period rated their children as ‘improved’ when compared to parents in the placebo period (68% versus 13%; P < 0.05). Successful treatment (physician rating) and improvement (parent rating) were independent of low or acceptable fibre intake (P > 0.6). Significantly more children who were taking laxatives at enrolment were treated successfully with fibre than with placebo (P < 0.01). Children with constipation only were significantly more likely to be treated successfully with fibre than those with constipation and encopresis (69% versus 28%; P < 0.04). No significant side effects, such as new onset of abdominal pain, bloating, abdominal distension, excessive gas, diarrhoea or anaphylactic symptoms, were reported. No significant differences in baseline characteristics between the two groups were observed.

Forty-six children were originally recruited. Thirteen children did not attend their appointment: seven children randomized to placebo first and six children randomized to fibre first. Two constipated girls only completed the first 4 weeks of the study: one received placebo and one received fibre and both recovered from chronic constipation and abdominal pain during the first four weeks of treatment and did not return for the 8-week visit. Data from the 13 children who entered the study and were randomised but did not come for follow-up and the two children who did not complete the study were excluded from the analysis. Initial data from these 15 children were not significantly different from the data of the 31 children who completed the study, except soiling frequency per week which was significantly less (4.0 ± 1.4; P < 0.001). Data analysis thus includes 31 children with idiopathic constipation with or without encopresis. Despite the high attrition rate (28%) intention-to-treat analysis was not performed.

One double-blind RCT (pilot study) conducted in Spain⁹⁷ (2006) [EL=1+] evaluated the effect of a palatable cocoa husk supplement that is rich in fibre on intestinal transit time and other indices of constipation in children with idiopathic chronic constipation. The study included 56 children aged 3 to 10 years (22 boys, mean age 6.3 years ± 2.2) referred to paediatric gastroenterology outpatients' clinic with chronic idiopathic constipation, defined in accordance with Rome II diagnostic criteria. Children were randomised to receive either a cocoa husk supplement rich in dietary fibre (one sachet (5.2 g): 4 g cocoa husk plus 1 g betafructosans) or placebo (one sachet (5.2 g): glucose, cocoa flavouring and excipients) during 4 weeks. The fibre supplement of cocoa husk contained 53.2 g of fibre (39.6 g of total fibre and 13.6 g of betafructosans) per 100 g of product. Insoluble fibre represented 37.2% and soluble fibre represented 2.4% of the total fibre. Cellulose and uronic acids were the main type of insoluble fibre and soluble fibre, respectively. In addition both groups received the same standardised toilet training procedures during the study period. Doses for both products in children aged 3 to 6 years were one sachet before lunch and one sachet before dinner and in children aged 7 to 10 years it was two sachets before lunch and dinner. Parents were instructed to dissolve the content of the sachets in 200 ml of whole milk before ingestion.

The number of bowel movements per week (mean) did not differ significantly between the two treatment groups. The percentage of children reporting hard stool consistency decreased significantly more in children taking the cocoa husk supplement when compared to children taking placebo (cocoa husk group: 41.7 versus placebo group: 75.0; P = 0.017). Significantly more children on the cocoa husk group reported a subjective improvement in stool consistency compared to children on placebo (cocoa husk group (n=24), improvement 14, no improvement 10 versus placebo group (n=24), improvement 6, no improvement 18; P = 0.039). Subjective improvement in pain on defecation was not significantly different between the two groups. No significant adverse effects, such as a new onset of abdominal pain, bloating, abdominal distension, excessive gas, diarrhoea or anaphylactic symptoms, were reported during the 4 week period with either treatment. There were no significant differences in baseline characteristics between the two groups. Eight children withdrew from the study before its completion (five children discontinued study because of the difficulty of the protocol and three were excluded because of the presence of positive antigliadin and antiendomysium antibodies). Data refer only to 48 participants who completed the study. Intention to treat analysis was not performed.

One prospective case series conducted in Italy⁹⁸ (2000) [EL=3] evaluated the efficacy of glucomannan as a treatment for chronic constipation in children with severe neurological damage. The study included 20 children with severe neurological damage and constipation of at least 12 months duration (14 boys, mean age 5.7 ± 4.2 years). In most patients evacuation was not possible without enema. Children were fed by mouth with semi-liquid diet including formula and puréed food. All children received treatment for disimpaction with enemas for 2 or 3 days (not clear what medication was used). After that children were randomised to receive either glucomannan at a dose of 100 mg/kg two times a day or placebo at the same dose, for 12 weeks. Both glucomannan and placebo consisted of a 500 mg capsule which was given orally mixed with 100 ml of water. An arbitrary scoring system was used for assessment of symptoms:

• stool consistency: 1 pellets; 2 hard; 3 soft; 4 loose; 5 liquid
• presence of painful defecation: 1 often; 2 occasionally; 3 none.

None of the outcomes changed significantly at any of the study periods for the placebo group when compared to baseline. The number of stools per week significantly increased in the glucomannan group at all assessment points when compared to baseline (at 4 weeks: mean 4.0 ± 1.3; at 8 weeks: 3.3 ± 1.0; at 12 weeks: 3.8 ± 0.9; P < 0.001 for all). Stool consistency significantly improved in the glucomannan group at all assessment points when compared to baseline (mean score at 4 weeks: 2.4 ± 0.5; at 8 weeks: 2.8 ± 0.7; at 12 weeks: 2.7 ± 0.7; P < 0.001 for all). Painful defecation improved significantly only at the 12 week assessment for the glucomannan group compared to baseline (mean score at 12 weeks: 1.9 ± 1.2; P < 0.01). Laxative use was significantly reduced in the glucomannan group at the 4 and 12 week assessments (mean number of laxatives per week at 4 weeks: 0.3 ± 0.8; at 12 weeks: 0.3 ± 0.5; P < 0.01). There were no significant differences in baseline characteristics between the two groups. One patient receiving glucomannan withdrew from the study after three weeks of treatment because of concomitant increase in seizure frequency associated with blood level of phenobarbital below the therapeutic range.

One prospective case series (pilot study) conducted in Hong Kong⁹⁹ (2000) [EL=3] evaluated the fibre intake of severe developmentally disabled children living in a residential institution along with the possibility of reducing the use of laxatives by increasing their fibre intake. The study included 20 severely developmentally disabled children (age range 3 to 17 years) with idiopathic constipation who were able to take oral feeding and medically stable. All children received fibre supplementation with wheat bran (All Bran®, Kellogg's) added in breakfast. During stage 1 (20 days), 15 g was added to each serving of breakfast (total fibre intake 17 g). Following stage 1 there was a period of 10 days where children received their normal diet without any supplementation. During stage 2 (6 weeks) 19 g was added to each serving of breakfast (total fibre intake 21 g). Baseline fibre intake was around 2 g/day.

The number of laxatives per week decreased significantly at the end of stage 1 when compared to baseline (baseline: 1.22, SD 0.36 versus end of stage 1: 0.9, SD 0.75; P < 0.05) and at the end of stage 2 when compared to baseline (baseline: 1.22, SD 0.36 versus end of stage 2: 0.7, SD 0.40; P < 0.01) but there were no significant differences when comparing end of stage 1 and end of stage 2. Outcomes for bowel movements were not reported in the paper.

An open label non-RCT conducted in the USA¹⁰⁰ (1955) [EL= 1-] evaluated the effectiveness of a palatable mixture containing prune and fig concentrate and non-diastatic malt syrup neutralised with potassium carbonate for the treatment of idiopathic constipation in infants and children. The study included 200 infants and children aged 3 months to 8 years with idiopathic constipation. One group had a prune and fig concentrate (Prune-Malt®, Benson-Nuen Laboratories Inc) added to their diet for three weeks and the control group received no intervention. The prune and fig concentrate was given to infants aged 3 weeks to 1 year as 2 tablespoonfuls daily added to milk or juice. Children aged 1 to 4 years received 3 tablespoonfuls daily added to milk or food and children aged 4 to 8 years received 4 tablespoonfuls daily added to milk or food. No changes were made to their usual diet and no drugs were given. No definitions or scoring system were given for: ‘improvement’, ‘no improvement’, ‘return to normality’, ‘good’ ‘fair’ and ‘poor’.

Twenty-eight children who received the prune and fig concentrate returned to normality compared to 16 children in the control group. Fifty-one children who received the prune and fig concentrate improved compared to 25 children in the control group. Only 21 children who received the prune and fig concentrate did not improve compared to 59 children in the control group. In total 132 parents rated the treatment as good, 47 as acceptable and 21 as poor (P values not reported in the study). No comparison was made between baseline characteristics of the two groups, although the author stated that wherever possible, cases of equal severity and ages were equally divided between the two groups. No attrition or loss to follow-up was reported.

Probiotics

A double blind RCT conducted in Taiwan⁷⁹ (2007) [EL=1+] investigated the effect of probiotics (lactobacillus casei rhamnosus, Lcr35) alone in the treatment of chronic constipation in children and to compare the effect with magnesium oxide (MgO) and placebo, respectively. The study included 45 children (23 male) under 10 years with chronic constipation. Children were randomised into three groups to receive during 4 weeks: MgO 50 mg/kg/day, twice a day; Lcr35 8 × 10⁸ colony forming units (CFUs) per day (antiobiophilus 250 mg, two capsules, twice a day); or placebo (starch in content). Lactulose use (1 ml/kg/day) was allowed when there was no stool passage noted for 3 days. Glycerine enema was used only when there was no defecation for more than 5 days or when abdominal pain was suffered due to stool impaction.

Defecation frequency significantly increased in children taking MgO and probiotic compared to placebo (MgO (n=18): mean 0.55 times/day ± 0.13; probiotic (n=18): 0.57 times/day ± 0.17; placebo (n=9): 0.37 times/day ± 0.10; P = 0.006 [placebo versus probiotic]; P = 0.01 [MgO versus placebo]). However, there were no significant differences between children taking probiotic and children taking MgO regarding this outcome. The percentage of children having hard stools was significantly lower in children taking MgO and in those taking probiotic compared to placebo (MgO: 23.5% ± 7.9; probiotic: 22.4% ± 14.7; placebo: 75.5% ± 6.1; P = 0.02 [placebo versus probiotic]; P = 0.03 [MgO versus placebo]) but there were no significant differences between children taking probiotic and children taking MgO regarding this outcome.

Children taking placebo had to make use of glycerine enemas significantly more often than children taking either MgO or placebo (MgO: 1.3 times ± 1.9, probiotic: 1.6 times ± 1.9, placebo: 4.0 times ± 2.1; P = 0.04 [placebo versus probiotic]; P = 0.03 [MgO versus placebo]) but there were no significant differences between children taking probiotic and children taking MgO regarding this outcome. There were no significant differences regarding use of lactulose and faecal soiling amongst the three groups. Significantly more patients were successfully treated with MgO or probiotic compared to placebo (MgO 72.2%, probiotic: 77.8%, placebo: 11.1%; P = 0.01 [placebo versus probiotic], P = 0.01 [MgO versus placebo]). However, there were no significant differences between children taking probiotic and children taking MgO regarding this outcome.

No adverse effects were noted in the probiotic and placebo groups and only one patient in the MgO group suffered from mild diarrhoea. There were no significant differences at baseline amongst the three groups regarding: gender, age of enrolment, age of onset of constipation, duration of constipation, previous treatment, defecation period, stool consistency, abdominal pain, faecal soiling, bleeding during defecation, use of enema and taking fruits or vegetables daily. Four patients discontinued medication during the study period: two in the MgO group, one in the probiotic group and one in the placebo group. Two patients suffered from acute gastroenteritis (not clear whether as a consequence of the study medication) and two patients were lost to follow-up.

A triple-blind RCT conducted in Poland¹⁰¹ (2005) [EL=1+] assessed the effectiveness of lactobacillus rhamnosus GG (LGG) as an adjunct to lactulose in the treatment of constipation in children. The study included 84 children aged 2 to 16 years with idiopathic constipation defined as less than 3 bowel movements per week for at least 12 weeks (ages: lactulose plus LGG group 79 months ± 47, lactulose plus placebo group 65 months ± 36; gender not reported). All children received treatment for disimpaction with phosphate and saline enema before study treatment started. Children were then randomised to receive during 12 weeks either lactulose 70%, 1 ml/kg/day (in two divided doses) plus 10⁹ CFUs of LGG or lactulose 70%, 1 ml/kg/day (in two divided doses) plus placebo. From weeks 13 to 24 patients were instructed to continue the use of lactulose or other laxatives as needed. Treatment success was defined as 3 or more spontaneous bowel movements per week with no episodes of faecal soiling.

Treatment success at 12 and 24 weeks was not significantly different between the two treatment groups. The average number of spontaneous bowel movements per week, episodes of faecal soiling per week and straining frequency per week were not significantly different when comparing both treatment groups at 4, 8 and 12 weeks. The percentage of patients using laxatives at 24 weeks was not significantly different between the two groups. LGG was well tolerated. The number of patients experiencing side effects was not significantly different between the two groups and the side-effects profile of LGG was similar to that of placebo: three patients in the LGG group versus five patients in the placebo group developed abdominal pain.

One patient in the LGG group developed vomiting and one in the placebo group experienced headache. There were no significant differences in baseline characteristics between the two groups. Five children in the LGG group discontinued the intervention (four because of clinical improvement, one developed abdominal pain) versus three patients in placebo group who discontinued the study without receiving any intervention (two refused to participate and one because of another reason, not provided). Outcomes for stool consistency were not reported in the paper.

One prospective case series (pilot study) conducted in The Netherlands¹⁰² (2007) [EL=3] determined the therapeutic effect of a combination of probiotic strains, containing the bifidobacteria B. bifidum, B. infantis and B. longum and the lactobacilli L. casei, L. plantarum and L. rhamnosus, on childhood constipation. The study included 20 children aged 4 to 16 years referred to outpatient clinic with idiopathic constipation, as defined by Rome III criteria (10 boys, median age 8 years). All children received treatment for disimpaction using rectal enema (Klyx, sodium-dioctylsulfosuccinate and sorbitol) once daily for 3 days. For the following 4 weeks children received a daily probiotics mixture of 4×10⁹ CFUs containing bifidobacteria B. bifidum, B. infantis and B. longum and lactobacilli L. casei, L. plantarum and L. rhamnosus. During the treatment period children were instructed to start toilet training. Toilet training consisted of sitting on the toilet three times per day for 5 minutes after each meal with the intention of trying to defecate. Use of laxatives was not allowed during treatment period.

The frequency of bowel movements (BMs) per week in the total sample did not change significantly at weeks 2 and 4 when compared to baseline. The frequency of BMs per week in 12 children presenting with more than 3 BMs per week at baseline increased significantly at weeks 2 and 4 when compared to baseline (baseline: median 1.0, range 0.0 to 2.0; week 2: median 3.0, range 0.0 to 7.0), P = 0.01; week 4: median 3.0, range 0.0 to 10.0; P = 0.009). The number of children reporting hard stools did not change significantly at week 2 and week 4 compared to baseline. At week 4, hard stools appeared in five children who had also had hard stools at baseline. One child with normal stools at baseline reported hard stools only at the end of the study. Two of the seven children who presented with hard stools reported normal stools at the end of the study. The number of faecal incontinence episodes per week decreased significantly at both week 2 and week 4 when compared to baseline (baseline: median 4.0, range 0.0 to 35.0; week 2: median 1.5, range 0.0 to 14.0; week 4: median 0.3, range 0.0 to 7.0; P = 0.007 and P = 0.001 respectively). There were no side effects, such as vomiting, bloating and increased flatulence, during the study period. No attrition or loss to follow-up was reported.

Excluding cows' and goats' milk^††††

A double-blind crossover RCT conducted in Italy¹⁰³ (1998) [EL=1+] compared the effects of cows' milk and soy milk in children with chronic constipation. Sixty-five consecutive children diagnosed with chronic idiopathic constipation underwent an observation period during weeks 1 and 2 when all medications were stopped. During weeks 3 and 4, one group (n=33) was randomly assigned to receive cows' milk and unrestricted diet and the other (n=32) had cows' milk and its derivatives excluded from their diet and received soy milk instead. During week 5 there was a ‘washout’ period for both groups with unrestricted diet and intake of soy or cows' milk and its derivatives. During weeks 6 and 7 patients were switched to the other type of milk. After the two study periods children with a response to the cows' milk free diet were given the soy milk diet for another month and then underwent a 2 week double-blind challenge with cows' milk at hospital. Children with eight or more bowel movements during a treatment period were considered to have a response.

Children were randomly assigned to receive cows' milk or a placebo containing soy milk. If no clinical reactions were observed within 12 hours, patients were discharged and the challenge continued at home. A qualitative faecal score was defined as 1 (mushy or liquid stool), 2 (soft faeces and no pain in passing stools) or 3 (hard faeces and difficulty and pain on passing stools). Patients were followed up for a mean period of 10 months (range 3 to 20).

During the observation period (n=65) the number of bowel movements was a median of 4 (25th to 75th percentile: 3 to 5) and the qualitative faecal score (QFS) was 3 for all 65 patients. During the two study periods neither the number of bowel movements, nor the qualitative faecal score changed significantly for the cows' milk group (n=65) compared to the observation period. For the group who had a response to the soy milk diet (n=44) the number of bowel movements increased significantly (median: 10, 25th to 75th percentile, 4 to 12) and 44 patients stopped having pain or difficulty passing stools (QFS 1 n=2; QFS 2 n=42; QFS 3 n=21) (P < 0.001 for all variables). During the challenge with cows' milk (n=44) no patients in the placebo group (soy milk) showed any clinical reactions. Patients in the cows' milk group did not have any acute reaction, but in all of them constipation associated with hard stools and discomfort on defecation reappeared after 5 to 10 days on the diet. The cows' milk-free diet was therefore recommenced, with a consequent normalisation of bowel movements in all patients.

Neither the number of bowel movements nor the qualitative faecal score were specifically measured during the challenge period. During the follow-up period none of the children with response had constipation. Cows' milk was reintroduced into the diets of 15 children after 8 to 12 months of the cows' milk-free diet and in all cases constipation returned within 5 to 10 days. Children with no response to soy milk diet were treated with high doses of laxatives, with subsequent improvement in stool frequency. In all cases symptoms returned once treatment with laxatives was stopped.

There were significant baseline differences in the groups of children with and those without a response. Anal fissures with erythema or oedema were more common among those with a response (40 of 44 patients versus 9 of 21, P < 0.001). Furthermore, at diagnosis, symptoms of suspected intolerance to cows' milk were more common in children with a response (11 of 44 patients versus 1 of 21; P = 0.05): recurrent bronchospasm in four patients, rhinitis in four and dermatitis in three. Six patients were withdrawn from the study during the cows' milk study period (on days 9 to 12) because of the reappearance of constipation and other related disorders. For children withdrawn from study during the cows' milk study period the number of bowel movements per period was prorated. Intention to treat analysis was used. Patients included in this study were highly selected and this might have led to overestimation of the frequency of cows' milk intolerance as a cause of constipation. Paediatricians who referred the patients may have pre-selected them as being likely to have a food intolerance since the study centre specialised in the treatment of food allergies. The inclusion of patients with no response to laxatives may have also contributed to this issue. It should be noted that the two types of milk taste different from one another, thus undermining the degree of blinding achievable.

A small prospective case series and embedded randomised controlled challenge conducted in Italy¹⁰⁴ (2006) [EL=3] evaluated the histology and manometry characteristics of patients with food intolerance-related constipation. Thirty-six children (age range 9 months to 10 years) with chronic constipation underwent a cows' milk-free diet for 4 weeks, following a 2-week observation period where all medications were stopped. After 12 weeks all patients cured on the cows' milk free diet or oligoantigenic diet (n=17) underwent a 2-week double-blind placebo-controlled challenge with cows' milk at the hospital. Patients were randomised to receive either cows' milk or ass's milk as placebo. If no clinical reactions (not specified which ones) occurred after 12 hours, patients were discharged and the challenge continued at home with bottles coded A or B. The challenge was stopped when a clinical reaction occurred. Outcome measures were number of bowel movements per week, appearance of stools and child's degree of difficulty in passing stools. The last two measures were combined in a QFS. A score of 1 was given if mushy or liquid stools, 2 if soft faeces and no pain in passing stools and 3 if hard stools and difficulty and pain on passing stools.

During the observation period both for patients further diagnosed with food intolerance (n=17; 14 to cows' milk only, 3 with multiple food intolerance) and for patients with constipation unrelated to food intolerance (n=19) the number of bowel movements per week and the QFS were the same (number of bowel movements: median 1.5, 25th to 75th percentile 1–2; qualitative faecal score: 1 n=0, 2 n=0, 3 n=36). During the elimination diet period the number of bowel movements per week in patients with food intolerance (n=17) significantly increased (median 5, P < 0.01, 25th to 75th percentile 3–7) and no children presented with hard stools or difficulty and pain on passing stools (QFS 1 n=1, QFS 2 n=16, QFS 3 n=0; P < 0.01 for the three values). For patients with constipation unrelated to food intolerance (n=19) both the number of bowel movements per week and the QFS remained the same as during the observation period and were significantly different from the results obtained in the group with food allergy (P < 0.01).

During the cows' milk challenge period cows' milk readministration caused the reappearance of constipation in all cases, very often associated with painful defecation, within 5 days after the commencement of the challenge (median 2 days, range 1–5 days). These symptoms disappeared on returning to the cows' milk-free diet or oligoantigenic diet in the three patients with multiple food intolerance. Patients with chronic constipation caused by food intolerance showed at baseline a higher frequency of a personal history of previous food intolerance (P < 0.01) and concomitant signs of food intolerance (bronchospasm four cases, dermatitis two cases; P = 0.05) than patients with constipation unrelated to food intolerance.

A second small prospective case series and embedded randomised controlled challenge conducted in Italy¹⁰⁵ (2005) [EL=3] evaluated the histologic data in patients with food intolerance-related constipation. Fifty-two infants and children with chronic constipation unresponsive to previous treatments underwent a 2-week observation period where all medications were stopped and at the end of the second week they were given a clean-out with a single dose of PEG 4000 (0.75 g/kg). For the next 4 weeks cows' milk and all its derivatives were excluded from the diet of all patients. Patients unresponsive to a cows' milk-free diet were placed on an oligoantigenic diet for 4 weeks (also excluding cows' milk). After 12 weeks all patients cured on cows' milk free or oligoantigenic diet underwent a 2-week, double-blind, placebo-controlled challenge with cows' milk at hospital. Patients were randomised to receive either cows' milk or ass's milk as placebo. If no clinical reactions (not specified) occurred after 12 hours, patients were discharged and the challenge continued at home with bottles coded A or B. The challenge was stopped when a clinical reaction occurred. Outcome measures were number of bowels movements per week and the QFS. Both were recorded by parents during the observation period and the elimination diet period. The qualitative faecal score was defined as 1 (mushy or liquid stool), 2 (soft faeces and no pain in passing stools) and 3 (hard faeces and difficulty and pain on passing stools). Children with eight or more bowel movements during a treatment period were considered to have a response. Normalised stools habits were defined as: bowel frequency of at least five evacuations per week with the elimination of soft stools without pain.

During the observation period both patients with food intolerance (n=30) and patients with constipation unrelated to food intolerance (n=22) had a median of 1.5 bowel movements per week (25th to 75th percentile 1–2) and all 52 patients a QFS of 3. During the elimination diet period the number of bowel movements per week increased significantly for patients with food intolerance (median 5, 25th to 75th percentile 4–7; P < 0.001) and no children presented with hard stools or difficulty and pain on passing stools (QFS 1 n=2, QFS 2 n=28, QFS 3 n=0; P < 0.01 for the three values). For patients with constipation unrelated to food intolerance both bowel movements per week and QFS remained the same as during the observation period. For all children cows' milk readministration caused the reappearance of constipation within 5 days after commencing the challenge (median 2 days, range 1–5 days).

Patients with chronic constipation caused by food intolerance showed at baseline a higher frequency of a personal history of previous food intolerance (P = 0.02) and concomitant signs of food intolerance (bronchospasm five cases, rhinitis four cases, dermatitis two cases) than patients with constipation unrelated to food intolerance (P = 0.03). No difference was observed between the 24 patients with cows' milk intolerance and the six patients with multiple food intolerance for outcome measures considered (number of bowel movements and qQFS), either at baseline or on elimination diet. However, in comparison with patients intolerant to cows' milk alone, patients suffering from multiple food intolerance were older (P = 0.04) and had a higher frequency of family history of atopic disease (P = 0.03). It should be noted that the high frequency of chronic constipation owing to food intolerance found in this study was likely due to a selection bias, as mainly food-intolerant patients are treated at the centre where the study was conducted.

Another small prospective case series conducted in Italy¹⁰⁶ (1995) [EL=3] aimed to investigate the possible relation between constipation and cows' milk protein (CMP) allergy (CMPA). The study sample comprised 27 infants considered to have idiopathic constipation. During the first 7 days all patients were being fed the same diet as at the time of diagnosis: various forms of commercial formula derived from cows' milk or whole cows' milk and its derivatives. For the next month all patients started a CMP-free diet. Three patients aged younger than 12 months were fed a formula containing soy protein and the others received soy milk or ass's milk (eight cases) and all cows' milk derivatives were excluded. After a month all patients whose symptoms abated underwent a cows' milk challenge. Cows' milk was given for a maximum of 10 days; then these patients started again an exclusion diet for 1 month and then a second cows' milk challenge was performed. Outcome measures were number of stools per day and QFS. The QFS was defined as in the studies described above.

During the first month of the CMP-free diet there was a significant improvement in symptoms in 21 patients: the frequency of stools significantly increased, faeces were soft and none of the infants had any discomfort when passing stools (mean number of stools per day on unrestricted diet (a): 0.24 ± 0.10; on first CMP-free diet (b): 1.04 ± 0.120; QFS on unrestricted diet (a): 2.85 ± 0.05; on CMP-free diet (b): 1.90 ± 0.08). During the first challenge constipation returned within 48 hours after the reintroduction of cows' milk, passing stools became painful and in seven patients with abdominal pain, ingestion of cows' milk was discontinued on day 4 (mean number of stools per day on first CMP challenge (c): 0.31 ± 0.14; QFS on CMP challenge (c): 2.75 ± 0.11).

During the second period of CMP-free diet the stools became normal again in the 21 patients and the symptoms accompanying constipation disappeared (mean number of stools per day on second CMP-free diet (d): 1.05 ± 0.11; significance: (b) and (d) versus (a) and (c), P < 0.0005) (QFS for second CMP-free diet: 1.85 ± 0.10; P < 0.001). During the second challenge symptoms reappeared within 24 to 48 hours: all 21 patients had painful passage of stools and for this reason the challenge was suspended on the third day.

Six patients did not improve on the first CMP-free diet period (mean number of stools per day on unrestricted diet: 0.18 ± 0.12; on first CMP-free diet: 0.20 ± 0.13) and their difficulty in passing stools did not change (QFS: control: 3; first CMP-free diet: 3). These patients were subsequently treated with lactulose and only a partial regression in symptoms was observed. They were permanently given an unrestricted diet, except for one infant who had episodes of recurrent bronchospasm related to the ingestion of cows' milk.

Patients were followed up monthly for a mean period of 18 months (range 10 to 30 months). Reintroduction of cows' milk was cautiously attempted in 16 children 6 to 9 months after the diagnosis of CMP allergy-dependant constipation. In eight children CMP did not cause the onset of any problems and it was reintroduced on a permanent basis; in eight patients CMP led to the reappearance of constipation within 2 to 3 days after introduction, and these infants were still following CMP-free diet at the time the paper was written. No harmful reactions with either soy milk or ass's milk were reported. It is important to note that significant differences at baseline were found between patients who were cured with the CMP-free diet and those whose condition did not improve with this diet. Patients who were cured with the CMP-free diet were more likely to have a history of CMP allergy or symptoms of CMP allergy (atopic dermatitis or recurrent episodes of bronchospasm) at the time they entered the study than those whose condition did not improve with this diet (15 out of 21 versus 1 out of 6; chi square= 3.75; P < 0.05).

Increasing fluid intake

One open label RCT conducted in the USA¹⁰⁷ (1998) [EL=1-] aimed to determine whether or not increasing fluid intake by either excess water intake or excess hyperosmolar liquid intake would significantly alter the course of simple constipation in children. The study included 90 prepubertal children with moderate to severe idiopathic constipation (31 boys [47.46%], mean age 7.5 years, age range 2.5 to 12.5 years). Children were randomised into two intervention groups and one control group. During 2 weeks one intervention group was instructed to increase water intake by 50% on the basis of the total measured oral liquid intake during the baseline week. The second group received supplemental liquid in the form of hyperosmolar liquids: Kool-Aid, juice, soda pop or other liquids known to contain more than 600 mOsm/l. The control group received no intervention.

Neither increasing water intake nor increasing hyperosmolar liquid intake significantly increased stool frequency or improved stool consistency or difficulty with stool passage within groups when comparisons were made with previous weeks, or between the three groups during the same week (analysis of variance). A second round of analysis excluded all subjects who failed to comply with at least 75% of assigned intervention and this did not change the study outcomes. No comparison was made of baseline characteristics between the three groups. The study originally included 108 children but only 90 completed the entire study as assigned. Eighteen children failed to comply with 75% of the intervention.

Increasing physical activity

One open non-randomised controlled trial conducted in Israel¹⁰⁸ (2009) [EL=1-] assessed the effect that stepping while standing had on constipation in children with severe cerebral palsy (CP). The trial included 22 children (aged 3.5 to 10 years) with a diagnosis of spastic quadriplegic CP with gross motor function classification system (GMFCS) level 4 or 5. All children were unable to stand and walk with a traditional walker or rollator because of insufficient upper extremity control, would attempt to step when supported in a standing position and had flexion contractures of less than 30° in the hips and the knees. Eleven children began a trial of the David Hart Walker (HW) orthosis in addition to their physical therapy sessions (six males, mean age 6.1 years ± 2.1) and 11 children who were matched for age and gender with the study group (six males, mean age 6.7 years ± 1.6) underwent a program with a standing frame (SF) as part of their physical therapy session. At entry the proportion of constipation in both groups was equal (6 out of 11 [54.5%]). After 6 months the study (HW) group had significantly reduced their level of constipation (1 out of 11 [9.1%]) and the control (SF) group had no change in constipation (6 out of 11 [54.5%]) (P = 0.02). It should be noted that the sample size was very small and that the paediatric evaluation of disability inventory (PEDI) was higher at baseline in the study group compared to the control group (indicating better self care, mobility and social function). There was no attrition or loss to follow up in either group.

Dietary modifications

There is no evidence for the clinical effectiveness of dried or fresh fruits, fruit juices, vegetables, cereals, fructo-oligosaccarides, omega 3 fish oils or excluding goats' milk from the diet for ongoing treatment or maintenance in children with chronic idiopathic constipation.

Increasing fibre

One double-blind RCT [EL=1+] showed that there were no significant differences between a yogurt drink with mixed dietary fibre (transgalacto-oligosacharides, inulin, soy fibre and resistant starch) and a yogurt drink containing lactulose at increasing defecation frequency per week and decreasing the number of patients with 1 or more faecal incontinence episodes per week. The study also showed that the stool consistency was significantly softer in the lactulose group compared to the fibre group. The number of patients using step-up medication at 3 weeks was significantly smaller in the group taking fibre than in the group taking lactulose but there were not significant differences regarding this outcome at 8 and at 12 weeks.

One double-blind RCT (pilot study) [EL=1+] showed that a cocoa husk supplement rich in dietary fibre (cocoa husk plus betafructosans) was more effective than placebo at decreasing the number of children reporting hard stool consistency and increasing the number of children reporting a subjective improvement in stool consistency. The study also showed that there were no significant differences between the cocoa husk supplement and placebo at subjectively improving pain on defecation and increasing the number of bowel movements per week.

One prospective case series [EL=3] showed that fibre supplementation with wheat bran was effective at decreasing the number of laxatives used per week

One open label non-RCT [EL=1-] showed that a palatable mixture containing prune and fig concentrate and non-diastatic malt syrup neutralised with potassium carbonate was effective at improving constipation. One hundred and thirty two parents rated the treatment as good, 47 as acceptable and 21 as poor.

Supplements

There is no evidence for the clinical effectiveness of supplements containing partially hydrolysed guar gum, iron or pectin for ongoing treatment or maintenance in children with chronic idiopathic constipation

One double-blind RCT (crossover) [EL=1+] showed that glucomannan (a polysaccharide of d-glucose and d-mannose, equal to 450 mg of alimentary fibre) was more effective than placebo at successfully treating constipation as per physician rating and improving children's symptoms as per parent rating. Successful treatment (physician rating) and improvement (parent rating) were independent of amount of fibre intake from the treatment. Significantly more children who were also taking laxatives were treated successfully with glucomannan than with placebo. Children with constipation only were significantly more likely to be treated successfully with glucomannan compared with children with constipation and encopresis

One prospective case series [EL=3] showed that glucomannan was effective at significantly increasing the number of stools per week, improving the stool consistency and painful defecation and reducing laxative use.

Probiotics

One double blind RCT [EL=1+] showed that there were no significant differences between probiotic (lactobacillus casei rhamnosus, Lcr35) and magnesium oxide (MgO) at increasing daily defecation frequency and decreasing the percentage of children having hard stools and both were more effective than placebo at increasing daily defecation frequency and decreasing the percentage of children having hard stools. There were no significant differences between the three treatments at decreasing faecal soiling. Children taking placebo had to make use of glycerine enema significantly more often than children taking either MgO or probiotic (lactobacillus casei rhamnosus, Lcr35) but there were no significant differences between children taking probiotic (lactobacillus casei rhamnosus, Lcr35) and children taking MgO regarding this outcome. There were no significant differences between the three groups regarding the need to use lactulose. Significantly more patients were successfully treated with MgO or probiotic (lactobacillus casei rhamnosus, Lcr35) compared to placebo but there were no significant differences between children taking probiotic and children taking MgO regarding this outcome.

One triple blind RCT [EL=1+] showed that there were no significant differences between probiotic (lactobacillus rhamnosus GG) plus lactulose and placebo plus lactulose at increasing the average number of spontaneous bowel movements per week and decreasing the episodes of faecal soiling per week, the straining frequency per week and the number of patients using laxatives.

One prospective case series [EL=3] showed that a probiotics mixture (bifidobacteria B. bifidus, B. infantis and B. longum plus lactobacilli L. casei, L. plantarum and L. rhamnosus) was effective at significantly decreasing the number of faecal incontinence episodes per week only in children presenting with less than 3 bowel movements per week at baseline. The study also showed that the probiotics mixture was not effective at improving stool consistency.

Infant formulae

One double-blind RCT (crossover) [EL=1+] showed that there were no significant differences between Nutrilon Omneo (new formula, NF), a formula with higher protein content, 100% of it based on whey protein hydrolysate (no casein, no intact whey protein), a mixture of prebiotic oligosaccharides (GOS and lcFOS), a higher concentration of sn-2 palmitic acid and a lower lactose content and a standard formula (SF) at reducing painful defecation and increasing defecation frequency. The study also showed that NF was significantly more effective than SF at improving the stool consistency.

One open label RCT [EL=1-] showed that Novalac-IT, a magnesium-enriched infant formula, was significantly more effective than a 20% strengthened regular infant formula at improving stool consistency, increasing stool frequency and reducing difficulties in defecation.

One open label RCT showed that a new formula (NF) (Omneo / Conformil) based on palmitic acid predominantly esterified at the β-position, oligosaccharides (GOS and FOS) with a prebiotic activity, partially hydrolysed protein, low lactose content and higher density was significantly more effective than a standard formula at increasing stool frequency.

One prospective case series showed that a new formula (NF)^††† was effective at increasing stool frequency.

One prospective case series [EL=3] showed that Novalac Anti-Constipation, a formula with an adapted concentration of magnesium and lactose, was effective at significantly increasing the number of daily stools and the number of children having normal stools, as well as at reducing the number of children presenting with pain or discomfort on defecation and the number of children needing external help at defecation.

Soy milk

One double-blind RCT (crossover) [EL=1+] showed that excluding cows' milk and its derivatives from the diet and giving soy milk instead was more effective than giving an unrestricted diet including cows' milk and its derivatives at significantly increasing the number of bowel movements, improving stool consistency and reducing the pain or difficulty on passing stools in children with chronic constipation and suspected food allergies, but was not effective in children in whom food allergies were not suspected at baseline.

Excluding cows' milk protein from the diet

One double-blind RCT (cross over) [EL=1+] showed that excluding cows' milk and its derivatives from the diet was more effective than giving an unrestricted diet including cows' milk and its derivatives at significantly increasing the number of bowel movements, improving stool consistency and reducing the pain or difficulty on passing stools in children with chronic idiopathic constipation and suspected food allergies, but was not effective in children in whom food allergies were not suspected at baseline.

Three small case series and embedded randomised controlled challenges [EL=3] showed that a cows' milk-free diet was effective at increasing the number of bowel movements, improving stool consistency and reducing the pain or difficulty on passing stools in children with chronic constipation and food intolerance, but was not effective in children with constipation unrelated to food intolerance

Increasing fluid intake

One open label RCT [EL=1-] showed that increasing liquid intake by either excess water intake or excess hyperosmolar liquid intake did not have significant impact on stool frequency, stool consistency or difficulty with stool passage in constipated children when compared to controls who did not increase their fluid intake.

Increasing physical activity

One open non-randomised controlled trial [EL=1-] showed that a device which allows children with severe cerebral palsy to step while standing was more effective than passive standing in improving symptoms of constipation.

Increasing physical activity

Despite the fact that there is no good quality evidence for the effectiveness of increasing physical activity to improve constipation, it is the opinion of the GDG that exercise should be encouraged. It is a common clinical observation that a lack of physical activity can be a contributing factor in constipation. While recognising that physical activity is not in itself a treatment for constipation, the GDG felt that it was important to encourage children to be physically active, as it may decrease the likelihood that they will develop constipation again once an episode has been medically treated, bearing in mind what is achievable and appropriate for the individual child. It has been recommended by the Department of Health⁸⁸ that children should do at least 60 minutes of moderate intensity physical activity per day as part of a healthy lifestyle.

Fibre-rich foods

No evidence was found to suggest that increasing fibre-rich foods, such as fruits, vegetables and cereals, is effective in treating or managing constipation. The GDG felt that encouraging children to eat more fibre, when they are already having a healthy balanced diet with sufficient fibre, could be detrimental. A high fibre intake in this case could exacerbate symptoms and potentially increase soiling. It is the opinion of the GDG that children should be advised to eat a healthy diet, including fibre containing foods, as outlined by the Paediatric Group of the British Dietetic Association in ‘Food for the Growing Years’ and ‘Food for the School Years’.^109,110

Fibre supplements

The evidence for using fibre supplements, such as prune and fig concentrate, cocoa husks and glucomannan, in the treatment of constipation is very limited. It is the view of the GDG that this evidence is not enough to recommend these products in the treatment or ongoing management of idiopathic constipation.

Probiotics

The GDG felt it was not possible to recommend specific probiotics at this stage as there is little evidence (only small trials, admittedly well conducted): the three studies refer to three different probiotics and in one case the probiotic was given in addition to lactulose. Additionally, some probiotics are not available commercially and the commercially available probiotics do not always say what their active ingredient is.

Infant formulas

The GDG examined four studies, each on a different infant formula, none of which are used in the UK. The GDG believes that there is not enough evidence to suggest that any of the formulas are clinically effective in the treatment or ongoing management of constipation.

The GDG believes that the current common practice of switching from one infant formula to another to alleviate constipation may be detrimental. It takes time to trial infants with different feeds and this often delays treatment with laxatives.

Excluding cows' milk

Although there is some evidence for excluding cows' milk from the diet to improve constipation, the opinion of the GDG is that the studies are of a poor quality and the selection of participants was biased. In the studies which were reviewed, both soy and ass's milk were used in the placebo group. Recommendations in the UK are that children with suspected cows' milk protein intolerance should be given feeds based on extensively hydrolysed proteins. ^††89 Soy and ass's milk are inappropriate alternatives to cows' milk and should be avoided due to a risk of allergenic cross-reactivity.

Replacing goats' milk

No evidence was found on replacing cows' milk with goats' milk in the diet to improve constipation in children. The recommendation from the Department of Health^† is that goats' milk is not suitable to be used as an infant feed because of its high renal solute load, inadequate vitamin and mineral content and doubtful microbiological safety.⁸⁹ Infant formulas based on goats' milk are not available in the UK. In addition, goats' milk protein can be as sensitising as cows' milk protein and is therefore not recommended when a cows' milk protein allergy is suspected.

Increasing fluid intake

The GDG found little evidence for the effectiveness of increasing fluid intake in children with chronic constipation. Despite this, it is the GDG's view that increasing fluid intake to recommended levels is essential. Without sufficient fluid intake, the use of osmotic laxatives will lead to dehydration which can itself contribute to constipation.

Conventional treatment alone versus conventional treatment plus biofeedback

Meta-analysis of four RCTs comparing conventional treatment alone versus conventional treatment plus biofeedback showed that treatment success was not significantly different between the two treatment groups either in the medium term (figure 5.2) or in the long term (figure 5.3).

Figure 5.2

Conventional treatment alone versus conventional treatment plus biofeedback: treatment success at medium term (6 months).

Figure 5.3

Conventional treatment alone versus conventional treatment plus biofeedback: treatment success at long term (12 months).

One parallel RCT conducted in the USA¹¹⁵ (1990) [EL=1+] determined whether outcome in chronically constipated and encopretic children with abnormal defecation dynamics could be improved with biofeedback training. The study included 43 children with chronic constipation and encopresis and abnormal defecation dynamics (33 boys, mean age 8.9 years, age range 5 to 16 years). Children were randomised to receive for 6 months conventional treatment alone (CT) (n=19) or conventional treatment plus biofeedback (BF) (n=22). CT consisted of use of laxatives, increase of dietary fibre and scheduled toileting. Disimpaction was carried out using enemas (type and dose not reported). For the maintenance phase children received magnesium oxide (milk of magnesia, MOM) at approximately 2 ml/kg body weight/day to induce at least one bowel movement daily and prevent faecal retention. Doses were decreased gradually to maintain daily bowel movement and prevent faecal retention and soiling. Children in the BF group received the same CT plus up to six sessions of biofeedback therapy 5 to 9 days apart. One session included approximately 30 to 35 defecation trials and lasted approximately 45 minutes. Patients in both groups were instructed to discontinue laxative therapy at 6 months ± 0.5 after initiation of therapy. Outcome measured was recovery rate at 7 and 12 months after initiation of treatment. Patients were considered to have recovered if they had 3 or more bowel movements per week and 2 or less soiling episodes per month while not receiving laxatives for 4 weeks.

At 7 months significantly more children in the BF group recovered compared to the CT group (BF (n=22): 12 [55%] versus CT (n=19): 1 [5%]; P < 0.001). Recovery rates did not differ between boys and girls in general and within the biofeedback group in particular. Prior unsuccessful treatment was not related to treatment outcome in either group. Patients with an initial abdominal faecal mass (severe constipation) were significantly more likely to recover with BF training than with CT alone (46% versus 0%, P < 0.02). At 12 months significantly more children in the BF group recovered compared to the CT group (BF: 11 [50%] versus CT: 3 [16%]; P < 0.05).

A boy aged 14 years in the BF group had a relapse. He had severe faecal impaction with enormous abdominal distension initially. Faecal impaction recurred 4 months after successful discontinuation of MOM. At the time the study was written he had no soiling but required intermittent treatment for constipation. One boy in the CT was lost to follow-up 1 month after treatment began. At that visit he was taking MOM and his soiling had resolved. One boy was lost to follow-up in the BF group after the first biofeedback session. Baseline characteristics were not significantly different between the two groups apart from gender: there were more girls in the BF group than in the CT group (41% versus 5%, P < 0.02). During initial evaluation severe constipation (an abdominal faecal mass present) was significantly more frequent in girls than in boys (90% versus 48%, P < 0.03). It was not completely clear who measured outcomes and how and whether questionnaires were piloted. Intention to treat analysis was not performed.

A parallel RCT conducted in the Netherlands¹¹⁶ (1996) [EL=1+] evaluated the effect of biofeedback training and conventional treatment on defecation dynamics and outcome in chronically constipated children. The study included 192 children with paediatric constipation (126 boys, median age 8 years). Patients were randomised to receive conventional laxative treatment alone (CT) (n=94) or conventional laxative treatment and biofeedback (BF) (n=98). Patients on CT received five outpatient visits lasting approximately 30 minutes during which laxative treatment and information from a diary containing defecation frequency and encopresis and/or soiling episodes were discussed. High fibre diet was advised but additional fibre supplements were not prescribed. Patients were instructed to try to defecate on the toilet for 5 minutes immediately after each meal. During the first 3 days patients were to use daily enemas (120 ml sodium-dioctylsulfosuccinate, 1 mg sorbitol, 250 mg per ml, Klyx) at home. If on day 3 enemas still resulted in large amounts of stool, they were continued for a maximum of 7 days. After the initial 3-day enema treatment, patients started oral laxatives with lactitol betagalactoside sorbitol (Importal®, Novartis) (one sachet of 5 g/10 kg body weight/day divided into two doses). Enemas were given whenever spontaneous defecation was delayed for more than 3 days.

Motivation was enhanced by praise and small gifts. Children in the BF group received five outpatient visits, including the same conventional treatment as described above, in combination with five biofeedback training sessions. As far as possible, both groups received equal attention. The treatment period lasted 6 weeks. Treatment was considered successful if the patients achieved 3 or more bowel movements per week and less than 2 soiling or encopresis episodes per month while not receiving laxatives for 4 weeks. Patients were assessed after the last visit of the intervention period at 6 weeks, then at 6 months, 1 year and 1½ years.

Treatment success was reported as number of children cured, and was not significantly different between the two groups at any of the assessment points (at 6 weeks, CT: 31 out of 94 [33%] versus CT+BF: 31 out of 98 [32%]; at 6 months, CT: 48 out of 93 [52%] versus CT+BF: 44 out of 94 [47%]; at 1 year, CT: 54 out of 92 [59%] versus CT+BF: 46 out of 92 [50%]; at 1½ year: CT: 52 out of 92 [57%] versus CT+BF: 44 out of 92 [48%]).

At baseline, patients were comparable for gender, age, frequency of gastrointestinal complaints and urinary problems. During the intervention period, three patients in the CT group refused manometry at the end of the treatment period: one patient was successfully treated and the parents refused permission for manometry; one patient was unsuccessfully treated and refused manometry; and one patient was lost to follow-up after two visits. Two patients of the BF group discontinued treatment: one patient aged 5 years did not cooperate and another patient discontinued treatment because his parents could not afford the cost of transport. At 6 months, five patients were lost (four patients in the CT+BF and one patient in the CT group), and at 1 year eight patients were lost to follow-up (two in the CT+BF and one in the CT group). Patients lost to follow-up were withdrawn from further analysis.

A parallel RCT conducted in Australia¹¹⁷ (1998) [EL=1+] determined whether surface electromyographic (EMG) biofeedback training produced sustained faecal continence in medical treatment resistant and/or treatment dependent children with anismus. The study included 29 children aged 4 years or more (24 boys, age range 4.8 to 14.9 years). Children were randomised to receive electromyographic biofeedback training and conventional medical treatment (BF) or conventional medical treatment alone (CT).

Up to four sessions of biofeedback were conducted at weekly intervals for each patient, each session consisting of approximately 30 to 35 defecation attempts. The aim was to achieve 10 relaxations of the external anal sphincter without visual feedback in two successive sessions. If this occurred in fewer than four sessions then biofeedback was discontinued. At completion of training, children were followed at monthly intervals by a single paediatrician, who gave verbal reinforcement of the skills learned during training.

CT alone comprised laxative therapy, behaviour modification and dietary advice. Laxative therapy occurred in two phases. The initial disimpaction phase comprised 3-day cycles of 5 ml Microlax enemas (sodium citrate) on day 1, one 5 mg bisacodyl tablet after school and 1 in the evening of day 2. Up to four cycles (12 days) were undertaken. Further cycles were prescribed if there was later evidence of stool re-accumulation. During the maintenance phase different laxatives were administered: liquid paraffin 5 to 30 ml once or twice a day, senna granules and or bisacodyl tablets. Medication use was decreased to a level consistent with maintenance of continence as monitored by bowel diary. Standard paediatric behaviour modification consisted of clarification during a joint parent-child interview of the postulates underlying physiological basis for encopresis.

The bowel training programme used positive reinforcement for successful defecation in the toilet and additional reinforcement for each 24 hours without soiling. Reinforcement consisted of parental praise and use of star chart diary (fitness training card) to indicate soiling free days. A regular sitting programme of 5 to 10 minutes toilet time within 30 minutes of each meal was basis of the programme. Dietary advice, general counselling and support were provided by a paediatrician. Psychiatric assessment or treatment was initiated when indicated clinically. It was unclear how long the CT lasted for.

Treatment success was assessed at 6 months after initiation of therapy. Full remission was defined as no medication and no soiling for at least 4 weeks; full remission on medication was defined as on medication and no soiling for at least 4 weeks; Partial remission defined as soiling no more than once a week, regardless of medication used. The use of medication was attempted by all those not in full remission, not only those who were worse or not improved. The remainder were those who were soiling more than once a week, regardless of medication use. Improvement was defined as progression by at least one level from baseline status, but without achieving full remission.

There were no significant differences between both treatment groups regarding the number of children who achieved full remission (BFT+CT (n=14): 2 [14%] versus CT (n=15): 2 [13%]; 95% CI on difference −24% to 26%). There were no significant differences between both treatment groups after combining the number of children who achieved full remission and the number of children who improved (BFT+CT: 2 (14%) versus CT: 4 (27%); P = 0.7, 95% CI on difference −46% to 23%). Three out of 14 patients in the BFT group completed the training in three sessions and the remainder underwent four sessions. Only one patient was unable to demonstrate relaxation of the external anal sphincter with attempted defecation. Only one patient (same as previous) was unable to defecate the biofeedback balloon by the time of their final session. All patients complied well with the instructions and procedures involved in the training. Two patients complained of transient discomfort when the biofeedback apparatus was inserted. No other adverse effects were seen or reported. At baseline there were slightly more subjects with primary encopresis in the biofeedback group than in the control group. No attrition or loss to follow-up was reported. It should be noted that no definition of constipation was given and also the study included a very small number of children.

Laxatives versus laxatives plus behavioural intervention versus laxatives plus behavioural intervention plus biofeedback

A parallel RCT conducted in the USA¹¹⁸ (2002) [EL=1+] compared short- and long-term effectiveness of three additive treatment protocols in children experiencing chronic encopresis. The study included 87 children aged 5 to 15 years who had experienced encopresis for a minimum of 6 months, defined as at least weekly episodes of faecal soiling for at least 6 months (72 boys, mean age at time of enrolment 8.6 ± 2.0 years, age range 5 to 13 years). Children were randomised to receive intensive medical therapy (IMT), intensive medical therapy plus enhanced toilet training (ETT) or intensive medical therapy plus enhanced toilet training and anal sphincter biofeedback (BF).

In the IMT group one of two paediatric gastroenterologists directed the treatment: colonic disimpaction with a series of enemas followed by sufficient laxative therapy to produce at least one soft stool each day without associated pain. Laxatives prescribed were magnesium oxide (milk of magnesia, MOM) and/or senna. Laxative dosages were adjusted regularly to produce one to three soft bowel movements daily. An enema or suppository was administered if the child had not produced a bowel movement during a 48-hour period. No specific dietary recommendations or manipulations were undertaken. Families received instructions and a brochure detailing the treatment protocol and the need for children to attend the toilet at least twice daily, preferably after breakfast and supper.

Children in the ETT group received similar enema and laxative therapy, with a clinical psychologist adjusting the laxative dose. The only difference from the previous therapy was that laxative therapy was decreased gradually when children demonstrated a stable bowel frequency with no soiling episodes. As long as the child had daily bowel movements of normal size for a week, the laxative dose was decreased by one quarter. This process was continued until laxative therapy was discontinued. If the child did not pass daily bowel movements of normal size, the laxative dose was increased. Parents and child were instructed on the psychophysiology of constipation and encopresis, and on how responding to early rectal distension cues along with regular toileting was critical to avoid reimpaction and to establish regular bowel habits. Various incentive programs were established, depending on the developmental age and the motivation of the child. Target behaviours were spontaneous trips to the toilet and clean pants. Toilet training was ‘enhanced’ because instructions were given on the role of paradoxical constriction of the external anal sphincter, and because appropriate defecation straining was modelled. The therapist sat on a portable toilet and demonstrated how to relax the legs and feet, how to take in a deep breath and hold it while sitting up straight, and how to push down with the held breath and pull in from the lower abdomen to propel out a stool. The child then replicated this while sitting on a portable toilet. The child received ‘hand feedback’ by placing one hand on the abdomen just below the navel to feel the abdomen move out when the breath was pushed down, and placing the second hand just below the first to feel inward movement with contraction of the rectus abdominous. Parents were instructed to prompt these behaviours at home. Additionally, 8 to 12 minutes of ‘toilet time’ was scheduled daily, beginning 15 to 30 minutes after the same two meals. During these times, children were instructed to practice tensing and relaxing the external anal sphincter for the first 4 minutes, with the objective of localising control of, and fatiguing, the external anal sphincter, and to mechanically stimulate the rectum. To desensitise children to toilet sitting, the second 4 minutes were spent ‘having fun’ while being read to or playing games. During the final 4 minutes, the child was to strain and attempt to have a bowel movement while relaxing his or her legs and feet. This routine toilet sitting was discontinued 2 weeks after the last scheduled treatment session.

The third group received the same instructions given to the other two groups and simultaneously received surface electromyographic biofeedback training. The same two psychologists who worked with the ETT group also worked with the BF group. It was unclear how long each of the treatments lasted. Data concerning toileting habits were collected for 14 consecutive days, before and after the initial outpatient visit, and again at 3 months, 6 months and 12 months after initiation of therapy. Treatment was considered successful if the child experienced no episodes of faecal soiling during the 2-week assessment 12 months after initiation of therapy.

There were no significant differences between the three groups at any time regarding mean soiling frequency (at 3 months: IMT 0.54, SD 0.68 versus ETT 0.22, SD 0.21 versus BF 0.34, SD 0.51; at 6 months: IMT 0.44, SD 0.52 versus ETT 0.38, SD 0.45 versus BF 0.20, SD 0.26 and at 12 months: IMT 0.33, SD 0.48 versus ETT 0.36, SD 0.53 versus BF 0.27, SD 0.37). At 3 months, 6 months and 12 months, the number of children who responded in the ETT group was significantly greater than in either the IMT or the BF group (at 3 months: IMT 45% versus ETT 85% versus BF 61%; at 6 months: IMT 41% versus ETT 74% versus BF 58%; and at 12 months: IMT 41% versus ETT 78% versus BF 61%; P < 0.05). These results were very stable over time (P < 0.001). With all three regimens, the response to treatment during the first 2 weeks of therapy strongly correlated with response to treatment at 3, 6 and 12 months (r > 0.90, P < 0.0001 in all cases). Of those children who had significant improvement after 2 weeks of therapy, 86 continued to improve at 3 months, 83 at 6 months and 81 at 12 months.

There were no significant differences between the three groups in the number of children cured at 12 months (IMT: 10 out of 29 [34.5%] versus ETT: 12 out of 27 [44.4%] versus BF: 11 out of 31 [35.5%]). There were no significant differences between the three groups at any time regarding the number of bowel movements passed in the toilet each day (mean at 3 months: IMT 1.44, SD 0.57 versus ETT 1.21, SD 0.49 versus BF 1.25, SD 0.64; at 6 months: IMT 1.36, SD 0.61 versus ETT 1.31, SD 0.63 versus BF 1.12, SD 0.60 and at 12 months: IMT 1.30, SD 0.61 versus ETT 1.01, SD 0.51 versus BF 1.16, SD 0.67). There were no significant differences between the three groups at any time regarding self-initiated toileting each day (mean at 3 months: IMT 1.53 times/day, SD 0.77 versus ETT 1.62 times/day, SD 0.82 versus BF 1.40 times/day, SD 0.71; at 6 months: IMT 1.49 times/day, SD 0.60 versus ETT 1.67 times/day, SD 0.95 versus BF 1.34 times/day, SD 0.72 and at 12 months: IMT 1.40 times/day, SD 0.76 versus ETT 1.31 times/day, SD 0.83 versus BF 1.31 times/day, SD 0.69). There were no significant differences between the three groups regarding laxative use at 12 months (IMT: 17 out of 29 [58.6%] versus ETT: 9 out of 27 [33.3%] versus BF: 17 out of 31 [54.8%]). There were no significant differences in baseline clinical or demographic characteristics between the three groups. It should be noted that no definition of constipation was given and no sample size calculation was performed.

One parallel RCT conducted in Croatia¹¹⁹ (2002) [EL=1+] assessed the success of biofeedback method versus conventional method in the treatment of chronic idiopathic constipation in childhood over a 12-week period and followed up the effect of biofeedback treatment on defecation dynamics and other anorectal manometric parameters in 49 children aged over 5 years (27 male) with chronic idiopathic constipation. Children were randomised to receive conventional treatment alone (CT, n=24) or conventional treatment plus biofeedback (BF, n=25). Conventional treatment consisted of oral administration of lactulose (Portalak®, Belupo) (240 mg/day or 10 ml syrup) with dose titration for the patient to have at least three stools per week. When spontaneous defecation failed to occur for more than 3 days in spite of appropriate therapy an enema was used. In addition, a fibre-rich diet and attempting defecation after meals were advised.

Biofeedback was conducted using a pressure technique. The child and the parents were instructed on how to perform Kegel exercises at home. Exercises included alternating 10-second contraction and relaxation of the sphincter and pubo-rectal muscle, performed five times a day in 20 cycles. Treatment lasted for 12 weeks. Treatment was considered successful if a frequency of 3 or more stools per week and less than 2 episodes of soiling or encopresis per month were achieved without laxatives. Therapeutic success was evaluated by the use of questionnaires distributed on weekly visits.

The number of children cured was significantly higher in the BF group compared to the CT group (BF: 21 out of 15 [84%] versus CT: 15 out of 24 [62.5%], P < 0.05). All children completed treatment. There were no significant differences in baseline characteristics between the two groups. It should be noted that the study included a small number of children and no sample size calculation was performed. There were insufficient details reported on who measured the outcomes and how they were measured.

One retrospective cohort study conducted in the USA¹²⁰ (1995) [EL=2+] evaluated whether patients who received biofeedback treatment continued with improved outcome compared with patients who received conventional treatment alone. The study included 129 (97 boys) aged 5 to 18 years with chronic constipation and encopresis (1 or more soiling episode per week). One group received conventional treatment plus biofeedback (BF) and the other group received conventional treatment alone (CT).

At least two and up to six weekly training biofeedback sessions were given. Each session included approximately 30 to 35 defecation trials and lasted approximately 45 to 60 minutes. The number of training sessions given depended on how soon the child learned to relax the external sphincter. Sessions stopped after ten relaxations of the external sphincter could be accomplished without visual feedback in each of two successive training sessions. CT comprised the use of laxatives, increase of dietary fibre and scheduled toileting (child instructed to defecate for 5 minutes after each meal and after returning from school for the initial months, and try to defecate at least daily once they could recognise the urge to defecate). Disimpaction was carried out with enemas (type and dose not reported). For maintenance magnesium oxide (milk of magnesia, MOM) was administered at approximately 2 ml/kg body weight daily to induce at least one bowel movement daily and prevent faecal retention. Doses were decreased gradually to maintain daily bowel movement and to prevent faecal retention and soiling. Occasionally mineral oil or senna were used instead of MOM. It was unclear how long the CT lasted for.

The follow-up period for the CT group was 4.2 years ± 2.5 and for the BF group it was 4.1 years ± 2.4. The mean age of the CT group initially was 9.1 years ± 3.3 and at follow-up 13.4 years ± 3.3; of the BF group initially it was 10.4 years ± 3.2 and at follow-up 14.5 years ± 3.3. Patients were considered to have recovered if they had 3 or more bowel movements per week and 2 or fewer soiling episodes per month while off laxatives for at least 1 month.

There were no significant differences between groups in any of the outcomes measured (mean stool frequency per week: BF (n=63) 5 ± 3 versus CT (n=66) 6 ± 3; percentage of children soiling: BF 35% versus CT 24%; mean soiling frequency per week: BF 1 ± 2 versus CT 1 ± 2; recovery rate: BF 28 children [44%] versus CT 41 children [62%]; and laxative use: BF 25% children versus CT 18% children).

Of 64 patients who originally received biofeedback one patient did not return after the first unsuccessful biofeedback session and was lost to follow-up. The 63 patients included in the biofeedback group were combined from two studies (as clinical characteristics of both groups were similar): 21 patients from one RCT (included already in this review, see Loening-Baucke, 1990) and 42 patients who had not recovered after at least 6 months of conventional treatment. Twenty-three patients had been originally included in the RCT but one boy was lost to follow-up after the first biofeedback session and a second patient received a central nervous system shunt during the follow-up period and was excluded from the analysis. Baseline characteristics were comparable between both groups except for the presence of an abdominal faecal mass (BF: 60 children versus CT: 41 children; P < 0.05). Age and follow-up age were not related to outcome in either group. The length of follow-up was significantly related to recovery for the biofeedback group (P < 0.02) and for all patients (P < 0.01) but showed no relationship for the conventionally treated group.

Conventional treatment alone versus conventional treatment plus behavioural intervention

A parallel RCT conducted in the Netherlands¹²¹ (2008) [EL=1+] evaluated the clinical effectiveness of behavioural therapy with laxatives compared with conventional treatment in treating functional constipation in childhood. The study included 134 children (76 boys) with functional constipation aged 4 to 18 years referred to a gastrointestinal outpatient clinic. Children were randomised to receive conventional treatment alone (CT, n=67) or laxatives and behavioural therapy (BT, n=67). All children received treatment for disimpaction with daily Klyx enemas (sodium-dioctylsulfosuccinate and sorbitol, 60 ml/day for children aged 6 years or under; 120 ml/day for children aged over 6 years) for 3 consecutive days before starting treatment. During the maintenance phase children received PEG 3350, 1 sachet (10 g) per day, and if treatment was considered to have insufficient effect the dose was increased by one sachet. If spontaneous defecation was delayed for more than 3 days, parents were advised to give an enema or bisacodyl suppository of 5 mg. In the BT group it was preferred to give oral bisacodyl tablets of 5 mg instead of rectal laxatives. During BT, paediatric psychologists adjusted the laxative dose and consulted the paediatric gastroenterologist when necessary. In both treatment groups patients kept a bowel diary.

The protocolised BT was developed by paediatric psychologists of the authors' hospital. The protocol consisted of two age-related modules: a module for children aged 4 to 8 years and a module for children aged 8 years and over. The learning process for the child and the parents was based on five sequential steps (know, dare, can, will and do). This approach was derived from a multidisciplinary BT to treat children with defecation disorders. For all involved psychologists, a detailed manual for both age-related modules was available to ensure a standard delivery of therapy. Visits lasted approximately 45 minutes.

Conventional treatment was conducted by paediatric gastroenterologists. Visits lasted approximately 20 to 30 minutes when laxative treatment and bowel diary were discussed. Patients and their parents received education to explain that symptoms are not harmful and are common in children with functional constipation and that a positive, non-accusatory approach is essential. Children were instructed not to withhold stools when they felt the urge to defecate. Motivation was enhanced by praise and small gifts from the paediatric gastroenterologists.

For both t groups a total of 12 visits were scheduled during 22 weeks with similar intervals between treatment sessions. Children were assessed at the last visit (post-treatment time point) and 6 months after the 22-week treatment ended (follow-up). The time between baseline assessment and follow-up was approximately 1 year. Treatment was considered successful if patients achieved a defecation frequency of 3 or more times per week and a faecal incontinence frequency of once every 2 weeks or less, irrespective of laxative use. A secondary outcome measured was stool withholding behaviour.

Compared with the BT group, defecation frequency in the CT group was significantly increased (incidence rate ratio (IRR) = 0.75, 95% CI 0.59 to 0.96; P = 0.021). This effect was mainly caused by a difference between interventions at post-treatment (CT: mean 7.2, 95% CI 6.1 to 8.5 versus BT: mean 5.4, 95% CI 4.3 to 6.7) and not at follow-up (CT: 6.6, 95% CI 5.0 to 8.8 versus BT: 5.3, 95% CI 4.4 to 6.3). There was no statistically significant difference between both treatment groups regarding faecal incontinence per week (post-treatment CT: mean 2.1, 95% CI 0.8 to 5.8 versus BT: 5.0, 95% CI 2.1 to 12.0; follow-up CT: mean 6.4, 95% CI 3.5 to 11.7 versus BT: 8.6, 95% CI 4.0 to 18.3; IRR=2.36, 95% CI 0.77 to 7.31; P = 0.135). At post-treatment, success rate was higher in the CT group than in the BT group (CT 62.3%, 95% CI 51.1 to 76.1 versus BT: 51.5%, 95% CI 39.7 to 66.9). However, no statistically significant difference between treatments was found (IRR=0.83, 95% CI 0.60 to 1.14; P = 0.249). At follow-up, the number of children successfully treated declined in both groups but again the difference was not statistically significant (CT: 57.3%, 95% CI 46.6 to 70.4 versus BT: 42.3%, 95% CI 31.8 to 56.4; IRR=0.74, 95% CI 0.52 to 1.05; P = 0.095). There were no significant differences between both treatment groups in the proportion of children who exhibited stool withholding behaviour at follow-up.

It should be noted that during treatment 2 out of 64 (3.1%) in the CT group and 9 out of 65 (13.8%) in the BT group discontinued the intervention (P = 0.054). At follow-up, four patients dropped out in CT. There was one loss of contact and three children were referred for BT directly after CT, making them unsuitable for follow-up measurements. Questionnaires were not returned by three patients in both intervention arms at post-treatment and by nine patients (CT six, BT three) at follow-up. Except for painful defecation (65.0% CT versus 43.1% BT, P = 0.014), there were no significant differences between the two groups in baseline sociodemographic factors or for clinical characteristics. An intention-to-treat analysis was conducted. Because of withdrawal before treatment start, attrition during the study, failure to fill out questionnaires or research procedure violations, missing data occurred. Imputation of missing values was used to make intent-to-treat analyses feasible.

Behavioural intervention plus laxatives versus laxatives only

A small parallel RCT (multicentre) conducted in the USA¹²² (2003) [EL=1+] examined the utility and effectiveness of an internet-based version of enhanced toilet training. The study included 24 children aged 6 to 12 years, soiling at least once a week, who had no medical diagnosis other than constipation that could explain their faecal incontinence (19 boys, mean age 8.46 years, SD 1.81). Children were randomised to receive the web intervention (n=12, 10 boys) or no intervention (n=12, 9 boys). All children were instructed to start with a basic regimen of one square of senna (Ex-Lax®, Novartis) twice a day.

The intervention was a web-based programme for the treatment of paediatric encopresis (U-CAN-POOP-TOO). This was a child-focused programme which targeted primarily at children aged 5 to 10 years but designed to be used by the child and the parents together. The program comprised three core modules which took 60 to 90 minutes to complete, with all users instructed to review them during the first week. The modules were: ‘The body’ (anatomy, physiology and pathophysiology of digestion), ‘How to poop’ (behavioural techniques for treatment of encopresis) and ‘Medication’ (clean-out and laxative treatment). New modules were assigned each week based on a follow-up assessment completed by the user about their child's status. Not all modules were necessarily used by all users: only those modules identified as relevant were assigned and reviewed. However, all modules could be viewed by all users. Follow-up comprised 17 to 20 questions, depending on the week. The system contained a total of 22 modules, each taking 5 to 10 minutes to review. Exposure to the program lasted for 3 weeks after which an assessment was conducted.

The number of faecal accidents per week decreased significantly more in the web group compared with the group with no web intervention (web group: mean 0.50, SD 0.85 versus no web: mean 8.27, SD 13.83). The number of bowel movements passed in the toilet per week increased significantly more in the web group compared to the no-web group (change from pre- to post-assessment:152% versus - 16%; P = 0.001). Using the bathroom without prompts also increased significantly more in the web group compared to the no-web group (change from pre- to post-assessment: 109% versus -37%; P = 0.021). Using the bathroom with prompts was not significantly different between the two groups. There were no significant differences in baseline characteristics between the two groups: age, gender, race, stage of bowel movement training, length of current laxative regimen or any of the outcomes measured. No dropouts or children lost to follow-up were reported. It should be noted that the study included a very small number of children.

Laxatives plus behaviour modification versus laxatives plus behaviour modification plus psychotherapy

A quasi-RCT conducted in the UK¹¹³ (1986) [EL=1+] reported the authors' experience with children who presented with faecal soiling, with or without constipation, who were treated by incentive-based behavioural modification, with or without psychotherapy, and consider factors that might predict the outcome for a non-intensive approach and in particular to draw attention to social background as a prognostic indicator. The study included 47 children who presented with faecal soiling, with or without constipation (26 boys, age not reported). For all children in cases where constipation was severe with large faecal masses they were initially admitted to the ward. They were then continued on whatever laxative they had been on before referral. Where no laxative had previously been used the child was offered a dose of lactulose twice daily (amount not reported). If there was no accumulation of faeces no laxatives were prescribed. No other laxatives were used in this study and in general their use was minimised, with the parents encouraged to stop the treatment with laxatives as soon as a regular bowel habit was established. In none of the children were suppositories used at any time. All the children were encouraged to take a high residue diet and in particular were asked to take bran with their breakfast cereal.

Children were randomised to receive behaviour modification (BM, n=26) only or behaviour modification plus psychotherapy (BM+Psy, n=21). BM was carried out by a paediatrician. All children were placed on a star chart regimen and offered varying coloured stars for ‘sitting on the toilet’ and ‘remaining unsoiled for a full day’. In some cases stars were awarded to encourage children who were reluctant to take bran in their diet. A contract was negotiated between the child and the parent (usually the father) for an award to be made at the discretion of the paediatrician. The child was to understand that the giving of the award would depend on response to treatment. ‘Demystification’, alleviation of guilt and use of explanatory diagrams were used. Children were seen at intervals of 6 weeks by the paediatrician for between 3 months and 1 year and were subjected to shows of affection and interest, which included careful and serious inspection of the charts. Failure to keep a star chart on two successive visits resulted in firm statement of displeasure. Two further failures led to the stopping of treatment and discharge with the option of psychiatric referral. Discharge of cured patients was at discretion of the parents.

Children in the BM+Psy group received the same BM as previously described. In psychotherapy children were seen by the child psychiatrist at roughly monthly intervals for between 2 and 12 months. At each appointment the mother (and also the father in four cases) was seen for 15 to 30 minutes to explore her feelings in respect of the child's bowel problem and its effect on the family and on her own relationship with the child. Whenever possible the mother's own history was explored and other emotional problems discussed where relevant, such as expressions of grief, anger or depression. The child was seen for 15 to 30 minutes for play, including picture drawing, games and sharing of their own toys and belongings. Their feelings concerning their problem were also explored. The behavioural star chart was also often brought and reviewed and the child praised and encouraged according to progress. The mother and child were seen together, sometimes early in treatment, sometimes later, depending on their relationship and success with management of the problems, to assess overall progress.

One year after initiating treatment success was assessed. Children were considered cured if they had at least five normal stools each week without soiling and only occasional use of laxatives (less than once a week). Children were considered improved if they had at least three stools each week and soiling less than once a week. Non-responders were children who had less than three stools each week or soiling more than once a week. These children were considered as failing to improve, despite the fact that in most cases there was less soiling than at the beginning of treatment. Treatment success did not differ between the groups. It is not possible to report the figures here, as they were only analysed by the authors according to compliance with treatment and with the children's social class, but not according to treatment groups. Four children left the study and 13 failed to keep adequate 'star charts'. Two children were subsequently found to be cured. It should be noted that no definition of constipation was given. Additionally the study included a small number of children and no sample size calculation was performed.

Systemic/family therapy: externalising versus behavioural approach

A retrospective audit conducted in the UK¹²³ (1998) [EL=3] aimed to assess the effectiveness of externalising treatment (EXT) compared to other traditional treatments (OTH) in children with soiling problems. The audit included 108 children treated for soiling problems (45 aged 3 to 5 years, 63 aged over 6 years) and their families. Referrals included ‘faecal soiling’, ‘encopresis’, ‘psychological soiling’, ‘failed toileting’, ‘constipation with overflow’ and ‘deliberate soiling’. It should be noted that some children were clearly diagnosed in the referral letter as ‘constipated’ or ‘not constipated’, but in some referral letters it was not stated whether the referring doctor had checked for constipation. Families who received EXT (n=54) were only included in the study if the treatment approach included: externalising the poo from the first interview with the child and family; developing a narrative with the child and family where they could see themselves as capable, skilful and determined ‘to teach the poo a lesson’, ‘outwit the poo’ or ‘defeat the poo’; not using rewards, interpretation, confrontation or paradoxical interventions as therapeutic manoeuvres; and attempting to see the whole family at least once. Other treatments (OTH) (n=54) included a mixed group of traditional treatments with predominantly (but not only) a behavioural approach in a family systems context. There were no elements of externalising in any OTH sessions. The treatment given depended only on the current approach of the therapist who received the referral.

Treatment lasted an average of 7.8 months for the EXT group and 6.6 months for the OTH group. At a minimum of 6 months' follow-up (mean 23 months), all parents (including those who left the study) were sent a questionnaire and asked whether there had been any further soiling incidents since they were last seen and the frequency of these incidents in the past month. Where children had returned for paediatric consultation, the frequency of soiling stated in paediatric notes was recorded even if parents did not reply to the audit. GPs were also asked whether they were aware of any further soiling after treatment had ended.

Significantly more children who received EXT stopped soiling or improved compared to children who received OTH, however this outcome was assessed (from notes: EXT 42 out of 47 versus OTH 30 out of 40, P = 0.02; from GP follow-up: EXT 29 out of 37 versus OTH 24 out of 42, P = 0.045; from parent follow-up: EXT 24 out of 38 versus OTH 13 out of 35, P = 0.026). Significantly more parents assessed EXT as helpful compared to OTH (number of parents: 24 versus 10; P = 0.0001). Externalising proved to be superior for boys, for children aged 6 years or over, for those with frequent soiling at the outset, for those with over 2 years' continuous soiling and those diagnosed as constipated on referral. The average number of appointments was not significantly different between the groups. There were no significant differences between the groups on baseline variables. It was unclear exactly how many children left the study or were lost to follow-up.