Outpatient

A number of studies have been performed on the rate of infection among patients with diabetes in the primary care and other outpatient settings. In a Canadian cohort of 1,779 patients with DM matched to 11,066 without DM the patients with DM had an increased risk of infection (adjusted odds ratio 1.21, adjusted for confounding variables). with skin and soft tissue infections having the strongest association with having DM. Interestingly, DM was not found in this study to be associated with head and neck, musculoskeletal, or viral infections (3). Another large Canadian study including more than a million individuals matched those with DM to those without, and assessed all physician and hospital claims for infectious disease. It found that almost half of all individuals with DM had a claim for an infectious disease within a cohort year compared with 38% of those without DM. The risk ratio was skewed most towards those with DM for upper respiratory tract infections, cystitis, and pneumonia (4).

Inpatient

In one study, having diabetes led to a 2-fold increased risk for hospitalization when presenting with an infection to the emergency room, and up to 12% of inpatient admissions in patients with diabetes were the consequence of an infection (5). A South Korean showed that those with diabetes had a significantly greater risk of infection-related ICU admission and death when hospitalized with infections of skin or soft tissue, central nervous system infection, or bone and joints (6). The previously-mentioned Canadian retrospective cohort study showed that, while the overall risk ratio for infection in those with diabetes versus without was 1.21, this number rose to 2.17 and 1.92 when considering infection which led to hospitalization and death, respectively (4). These and other studies (7,8) have revealed that not only is diabetes associated with (and causative of) an increased risk of infection, but also with higher rates of hospitalization, ICU stays, and death related to these infections. Of note, many of these patients with DM have other comorbidities which may not be able to be fully controlled for in these epidemiologic studies demonstrating higher estimates of the risk of infection with DM.

One important entity to consider in the inpatient arena is sepsis. The studies on sepsis are not consistent —though some show worse outcomes from DM, others have suggested either no effect or even a protective effect from DM (9). Data for the latter come from studies assessing acute respiratory failure and respiratory distress syndrome in the ICU, and it may be that the blunted immune response that we see in some patients with DM are responsible for the findings (i.e., reduced inflammation and injury related to impaired neutrophil function as described in section on Innate Immunity) (10). More studies are needed to better understand in what specific clinical contexts DM results in higher risk.

Innate Immunity

Adaptive Immunity

The adaptive immune system is activated in response to specific pathogens and involves the immunologic memory for those pathogens. There are two components of adaptive immunity, the humoral and cellular, which carry out the major purposes of generating an antibody and cellular (involving B and T cells) response to a specific antigen. The adaptive humoral immunity is involved in antibody production. This appears to be preserved in those with DM on the basis of overall appropriate response to various vaccines (29,30). However, there may be some dysregulation of the adaptive cellular immunity. There is a depletion of memory CD4+ cells that has been noted prior to the development of type 1 DM (31). Furthermore, a dysfunction of and an impaired response of these cells to Streptococcus pneumoniae has been described (32). However, the dysfunction seen in DM with innate immunity is better understood than that the dysfunction in adaptive immunity (33).

Cytokine Signaling Defects

Cytokines play a vital role in the signaling cascades which underpin the immune system, allowing for full activation of both innate and adaptive immune responses. Multiple points of dysregulation have been identified in DM. With stimulation, multiple of these cytokines have been shown to be secreted at lower levels than would be typical with stimulation. In vitro studies done on monocytes isolated from patients without DM showed suppression of IL-1β, IL-2, IL-6, and IL-10 secretion in the presence of hyperglycemia (34-36). In diabetic mice, there was noted to be immune dysregulation and also inflammation which was related to IL-22 deficiency reversed with provision of IL-22 (37). There is also evidence for impaired interferon gamma (IFN-γ) and TNF alpha (TNF-α) production from T cells in the setting of methylglyoxal, a compound which is increased in those with DM (38). The end result of these deficits is that there is attenuation of the phagocytic and cellular immune response.

Respiratory Infections

Pneumonia is a frequently-seen infection in those with DM. In a large Danish population-based case-control study of 34,239 patients, the relative risk for hospitalization from community-acquired pneumonia was 1.26 compared with patients without DM. Furthermore, the risk appeared to be correlated with level of glycemic control with relative risk (RR) for those with HbA1c <7% being 1.22, versus a RR of 1.6 when HbA1c was ≥9% (43). A Portuguese study similarly showed DM prevalence was higher in those with pneumonia and that outcomes were worse with a longer hospital stay and significantly higher mortality in patients with DM versus those without (15.2% vs 13.5%) (44). These trends were also seen in another Danish study which showed mortality was greater in those with type 2 DM compared with other patients at both 30 and 90 days after the initial pneumonia episode (45). The presentation of pneumonia is potentially different in those with DM as typically bacterial pneumonia presents with a purulent cough and pleuritic chest pain, symptoms which are less commonly seen in those with DM. The hypothesis is that the lowered immune defense results in a decreased inflammatory response and symptoms. Notably, altered consciousness is more common on presentation with pneumonia in those with DM. The causative agent of pneumonia is similar in those with and without DM with the most common being Streptococcus pneumoniae (46). However, there is an over-representation of organisms such as Staphylococcus aureus and Klebsiella pneumoniae related to skin colonization and more frequent aspiration in those with DM. Management is using combination therapy with amoxicillin/clavulanate or cephalosporin and a macrolide or doxycycline versus monotherapy with respiratory fluoroquinolone (47).

Diabetes also represents an important risk factor for contracting tuberculosis (TB). The odds of developing tuberculosis appear to be higher in those with DM compared to those without with the odds ratio ranging from 2.44-8.33 in various studies. Furthermore, severity of DM appears to be correlated with greater risk of contracting TB based on studies comparing the incidence of TB in those with insulin-dependent versus non-insulin dependent DM. Risk of treatment failure despite good adherence to medication regimen and also death from tuberculosis all appear to be increased (48). There are side effects of medications targeted at tuberculosis with particular relevance in DM. Isoniazid can cause peripheral neuropathy that could be mistaken for diabetic neuropathy, and pyridoxine should be administered to ameliorate this risk. Rifampicin has the ability to cause hyperglycemia. Rifampicin also is a powerful inducer of the cytochrome P450 system leading to increased clearance of multiple DM agents (i.e., sulfonylurea, pioglitazone, meglitinides). Hence while the regimen to treat TB seen in patients with DM is the same as the regimen in those without, special attention must be paid to these DM specific issues.

Skin and Soft Tissue Infections

There is a significantly increased risk of skin and soft tissue infections (SSTI) in DM. Using a large administrative claims database (HealthCore Integrated Research Database), Suaya et al were able to demonstrate complications from SSTI were five times higher, and hospitalization four times higher, in patients with DM than those without DM in their cohort (49). The most common agent of infection is Staphylococcus aureus (S aureus). The foot represents the most common site of infection as well which has been expertly covered in another chapter (50).

For cellulitis, the diagnosis is typically clinical, as opposed to an abscess which often is cultured for organism identification and resistance profiling. The decision for antibiotics in cellulitis and with abscess is often empiric, and in those with DM it is imperative that the choice cover Staphylococcus species. Potential first line therapy for outpatient oral antibiotics includes doxycycline, clindamycin, trimethoprim-sulfamethoxazole, and cephalexin. If there is admission due to SIRS criteria being present or a suspected complication, then the recommendations change to IV predominant choices including vancomycin, linezolid, and ceftaroline (51).

Necrotizing fasciitis is a life-threatening condition involving the subcutaneous fat and deep fascial layers. A retrospective report of 59 necrotizing fasciitis cases at a single center revealed that 11 of the cases had DM, and another study showed that 51% of 84 patients in the cohort with necrotizing fasciitis had DM (52,53). Though most commonly a polymicrobial infection, Klebsiella pneumoniae is also commonly seen as a single isolate (53). Prognosis appears to be poorer in those with DM, with a higher rate of limb loss than that seen in those without DM. Broad spectrum antibiotics are utilized, related to the frequent polymicrobial nature of the infection. However, ischemia compromises appropriate antibiotic concentration at the site, therefore the management is primarily surgical involving a combination of debridement, necrosectomy, and fasciotomy frequently (54).

Fournier gangrene represents a particularly serious SSTI, defined as a necrotizing skin infection of the scrotum and penis or vulva. Typically, patients are between the age of 50-60 years and DM represents a serious risk factor for development. Usually, the infection will begin in the perianal or retroperitoneal region and then spreads to the genitalia or as a urinary tract infection which then also moves towards the genitalia. There will be necrosis and crepitus which is an indication of involvement of the underlying skin and soft tissue. The etiology is usually a mix of aerobic and anaerobic bacteria which can commonly include S aureus and Pseudomonas (51). Surgical debridement is typically necessary. This condition has become of particular relevance with the release of a warning from the US Food and Drug Administration in August 2018 that Fournier’s represented a rare but serious complication of sodium-glucose cotransporter 2 inhibitors (SGLT2i), a newer but increasingly utilized class of DM medications (55). There have been 55 cases of Fournier’s reported to the FDA between March 2013 and January 2019 (56). However, it is important to note that while this number appears to be higher than for other anti-hyperglycemic agents, that there has not been a clear establishment of causality here. In fact, other studies including a meta-analysis of randomized controlled trials involving SGLT2i and a “real-world” study using IBM MarketScan were unable to confirm an increased risk of Fournier’s Gangrene for those using SGLT2i versus those who did not (57,58).

Sternal wound infections after surgery are also known to occur more frequently in those with DM. After coronary artery bypass, presence of DM is one the strongest predictors for a deep sternal wound infection (odds ratio 2.6) (59). Improved glycemic control in the post-operative period has been shown to be able to significantly reduce the rate of sternal wound infection (60-62). Furnary et al in their seminal study were able to demonstrate in a prospective manner that use of an IV insulin infusion to keep glucose <200 mg/dL resulted in a 66% reduction in risk of sternal infection compared with nonrandomized historical controls (60).

Gastrointestinal Infections

The presence of DM has been known to worsen viral hepatitis. The outcome in chronic hepatitis C infection is worse in those with DM as compared to those without, corresponding to a significantly increased risk for cirrhosis and a reduced response to antiviral therapy in those with DM and hepatitis C (63). Emphysematous cholecystitis is a rare progression of acute cholecystitis, defined as presence of gas in the gallbladder wall for which DM serves as a major risk factor. The prevalence of DM among those with emphysematous cholecystitis has been described as high as 50% in the literature (64). Sonography is able to detect the presence of gas within the gallbladder wall and an abdominal radiograph will show a curvilinear lucency around the gallbladder. CT can also detect the condition with nearly 100% sensitivity. The two most common causative agents are Clostridium perfringens and Escherichia coli. The treatment is often surgical with prompt cholecystectomy although for a mild case antibiotic therapy can be initiated but without improvement within 3-4 days then the recommendation is still for cholecystectomy (65).

Genitourinary Infections

Urinary tract infections (UTI) are significantly more common in patients with DM with a large UK study showing an incidence of UTI of 46.9 per 1000 person-years among those who had type 2 DM versus 29.9 for those without DM (66). The most common pathogens in those with DM are Escherichia coli and the Enterobacteriaceae (Klebsiella, Proteus, Enterobacter, etc.). There is an increased risk of drug-resistant organisms being present. In making the diagnosis, beyond the typical clinical symptoms of dysuria and increased frequency, it is important to note that a urine culture really should be obtained in all individuals with DM prior to therapy (67). Outcomes are worse in those with DM, being associated with increased relapse and reinfection (at 7.1% and 15.9% respectively in those with DM versus 2% and 4.1% in those without) (68). Asymptomatic bacteriuria should not be treated even in those with DM. In those with symptoms, trimethoprim-sulfamethoxazole and ciprofloxacin are good choices but the general consensus is to follow local infection and resistance trends and tailor antibiotic therapy towards those organisms (67). From a lower UTI, there is also more risk of progression to pyelonephritis (both emphysematous and nonemphysematous) which tends to be more severe requiring hospitalization and are often bilateral in those with DM (69,70).

Head and Neck Infections

Consistent with the findings in many types of infections, head and neck infections are more common and appear to be more severe in those with DM. An assessment of 185 patients at the National Taiwan University Hospital with deep neck infections found that those with DM had a significantly higher rate of abscess formation than those without (89.3% vs 71.3%) and that surgical drainage was required more frequently as well (86% vs 65.2%) (71). One concerning entity commonly associated with DM is necrotizing external otitis. This term references an infection which has spread to the temporal and adjacent bones at the base of the skull. The causative agent is often Pseudomonas aeruginosa. The increased frequency of this in those with DM appears to be related in part to the vascular compromise seen in DM combined with a pseudomonal vasculitis (72). A certain level of suspicion for this condition needs to be present when there is external otitis in a patient at risk of necrotizing progression. Imaging is often needed to confirm with CT used. Systemic antibiotics are a requirement along with local treatment of the canal (i.e., cleaning, antimicrobial topicals, etc.). The antibiotic chosen needs to have antipseudomonal action such as the fluoroquinolones with the understanding that poor vascularization of the area often means higher doses are required (72).

Fungal Infections

Infection with Candida species is common in those with DM (73). Skin and soft tissue along with mucosa can be commonly impacted. Assessment of mouth swabs from patients with and without DM revealed that there was a higher frequency of Candida infection in the patients with DM (74). This appears to be related to decreased salivary pH and salivary flow which promoted colonization with the yeast. But there is also the possibility that in DM there is increased ability of Candida to bind to its receptor (75). Onychomycosis, a fungal infection of the nails, is also very common with DM with some studies suggesting up to 1/3 of individuals with DM are impacted. Diagnosis is made based on a positive fungal culture for a dermatophyte or microscopy showing fungus prior to initiation of therapy. Oral agents tend to be the most efficacious but topical lacquers are used as well (76).

Beyond the skin and mucosa, there is a high rate of genitourinary infections with Candida species. Among fungal infections of the urinary tract system, the vast majority involve Candida species. For both outpatient and inpatient urinary tract infections where Candida species was the causative agent, DM is present as a comorbid condition in 29% and 39% of cases respectively. Initial work-up would be similar to that of any other for urinary tract infection including urinalysis and culture. If there is a concern for Candida then that should be communicated with the reporting lab as there can be slow growth on certain cultures. It is important to remember that symptomatic urinary tract infections with Candida are rare. However, if truly a symptomatic infection is felt to be present, then fluconazole is typically the first line therapy of choice because of its ability to accumulate in high concentrations within the urine. If there is progression to pyelonephritis then speciation and sensitivities will be required given the presence of resistant strains and often dual agent therapy is required (77).

Mycotic genital infections are driven by Candida species. in both men and women (78). A very large percentage of women will experience symptomatic vulvovaginal candidiasis within their lifetime, and DM represents a risk factor for development of infection, with worsening glycemic control predisposing to an even higher risk (73). This appears to be related to known defects in immune cell function (PMNs and macrophages) and the impact of glucose in the urine, leading to worsening virulence in addition to improved adherence of yeast. Clinically, women present with pruritis and discomfort. Diagnosis is best made with microscopy and can be suspected in those with a negative amine (“whiff”) test and normal pH. Treatment is typically with a short course of intravaginal imidazole or triazole as opposed to oral fluconazole (78). For male patients, a similar finding of increased risk of balanitis (i.e., inflammation of the glans penis and prepuce) with DM and with increasing HbA1c is noted (73,79). This condition also occurs almost exclusively in males who have not been circumcised. Diagnosis is best made based on clinical presentation (burning and itching of penis worse after intercourse) in combination with a subpreputial culture. Imidazole or triazole creams represent the mainstay of therapy (78).

It is worth mentioning that SGLT2i as a class are known to be associated with genital infections, particularly mycotic genital infections caused by Candida species. Using information from two US based health insurance databases, Dave et al were able to demonstrate in a retrospective cohort study that use of an SGLT2i compared with dipeptidyl peptidase 4 inhibitors and glucagon-like peptide 1 receptor agonists led to an approximately three-fold increase in risk of genital infection (80). This might be mitigated in part by improved hygiene (i.e., rinsing the genital area with water after every episode of urination and before going to bed) (81). Interestingly, the association between use of SGLT2i and genital infections does not appear to extend to UTIs (82). This may be related to increased volume and flow through the urinary tract preventing excess bacterial load from developing (83). As SGLT2i are increasingly a mainstay of therapy for patients with type 2 DM, this association with mycotic genital infections needs to be carefully considered.

Mucormycosis is a rare but life-threatening fungal infection caused by a mucormycetes, a group of molds (commonly the mucor and rhizppus species). A metanalysis of cases of mucormycosis showed that DM was the commonest underlying condition present in 40% of cases (84). Frequently the DM associated is type 2 and uncontrolled. It carries with it a high fatality rate which ranges from 32-57%. The most common site of infection (66% of cases in a large 929 case series) is the sinus-cavity leading to rhinocerebral infection which can also impact the orbits. In these cases, there are often symptoms of sinus congestion/inflammation, fever, facial swelling along with ophthalmoplegia, cellulitis, and cranial nerve palsies. Necrotic eschars in the nasal cavity and on the hard palate are classic findings and also indicate a rapidly progressive infection. The respiratory tract is the second most commonly affected location (16%) with endobronchial lesions commonly found in those with DM. The subsequent invasion of the vasculature can lead to more distant infection. Finally, the third most common site is the skin (10%). Cutaneous mucormycosis presents with erythematous or ulcerative necrotic lesions which can also lead to osteomyelitis (85,86). Tissue biopsy is a requirement for diagnosis while various imaging modalities can help provide further evidence of an infection. Debulking of the infection through surgery is necessary. Amphotericin B and isavuconazole have been utilized as adjunctive therapies (85,87).

COVID-19 and DM

Becoming increasingly clear is the interaction between DM and SARS-CoV-2 (causative agent of coronavirus disease 2019 (COVID-19)). Multiple risk factors for contracting and having a more severe course of COVID-19 have been identified, including advanced age and male gender, but both type 1 and type 2 DM are now known to be important risk factors for morbidity and mortality with the disease (88). An assessment of patients who contracted COVID-19 and were tested within the Vanderbilt University Medical system demonstrated that there was a significantly increased risk for hospitalization in those with DM compared to without DM (odds ratio 3.36 for type 2 and 3.9 for type 1) and also for more severe disease course (odds ratio 3.42 for type 2 and 3.35 for type 1) (89). From a cohort of patients in England, there is evidence that poorly-controlled DM as compared with well-controlled DM (HbA1c 6.5-7% versus ≥10%) results in significantly increased mortality in both type 1 and type 2 DM (hazard ratio 2.23 and 1.61 respectively) (90). Increased mortality has also been seen in another English cohort (83). A more comprehensive review on this topic is offered by Lim et al, where the multiple points at which COVID-19 and DM interact – including the impact of glucotoxicity on the lungs, increased thromboembolic risk, worsened oxidative stress. and inappropriately high levels of cytokine production leading to organ damage – are outlined (91).

Glycemic Control and Diabetes Therapies

There is good evidence that glycemic control is correlated with infection. A study of 69,318 patients with type 2 DM in Denmark revealed an association between increased risk for community- and hospital-treated infection in those with higher HbA1c ≥10.5% compared with HbA1c 5.5-<6.4% (92). Similarly, in a large English cohort there was an increasing risk of infection in parallel with HbA1c for patients with both type 1 and type 2 (2). In a Taiwanese study looking at outcomes from a community-based health screening program, the authors found that fasting plasma glucose >200 mg/dL and DM was associated with the highest risk of infection and also a 3-fold higher risk of death than those without DM (93). Looking at an older population, the risk of certain infections was significantly higher in those with poor glycemic control HbA1c >8.5% compared with good glycemic control (relative risk infections ranging from 1.28-2.38) (94). Intervening to lower glucose appears to mitigate the risks. Zerr et al assessed incidence of sternal wound infection in patients with and without DM before and after implementation of a postoperative continuous IV insulin protocol to keep blood glucose <200 mg/dL. They found that lower glucose in the first 2 days postoperatively was associated with a decrease in deep wound infection from 2.4% to 1.5% (62).

Insulin, in both translational and clinical studies, has been suggested to have a protective effect against infection risk in those with DM (Table 3). A large surgical ICU trial assessing tight (80-110 mg/dL) versus conventional (treatment with insulin only if glucose >215 mg/dL) glycemic control using IV insulin found a lower mortality with tight glycemic control, and the greatest reduction in mortality was seen in those with sepsis leading to multi-organ dysfunction. In those treated with IV insulin, there was a significant reduction in the risk of developing sepsis (46%) (95). While these data were later brought into question by the findings of the NICE-SUGAR trial (96) which demonstrated increased mortality with intensive glycemic control using IV insulin, other studies have suggested that there is improvement in rates of infection with use of insulin (60,97) particularly in the post-cardiac surgery setting for sternal wound infections. Furthermore, the increased risk of mortality in the intensive glycemic control arm in NICE-SUGAR is felt to be related to an excess of hypoglycemia seen in the cohort compared with normal care. If improved glycemic control can be achieved without causing a significant increase in hypoglycemia, a different outcome may be able to be achieved (98). Translational studies have been able to show in vivo that T cells which lack insulin receptor expression are unable to proliferate and produce cytokines properly and that insulin enables the cells to take up nutrient which supports their function (99). As mentioned earlier chemotaxis of PMNs is impaired in patients with DM and it has been noted that provision of glucose and insulin can restore these to baseline (15).

While data for immune function improvement and infection using other antihyperglycemic medications are relatively sparse, there are some suggestions that therapies beyond insulin can have an immune-restorative effect (Table 3). Metformin has demonstrated an ability to increase the number and action of CD8+ tumor-infiltrating lymphocytes, resulting in improved production of cytokines IL-2, TNFα, and IFNγ (100). In a mice model with an absence of the TNF receptor-associated factor 6 (TRAF6), there is a relative inability to generate memory T cells that is related to defects in fatty acid metabolism. When TRAF6-deficient antigen specific effector T cells and TRAF6-deficient mice were exposed to metformin, there was a restoration of the production of memory T cells (101). While a number of other diabetes therapies have well-established anti-inflammatory effects (i.e., peroxisome proliferator-activated receptor-γ agonists, glucagon-like peptide-1 receptor agonists, sodium-glucose cotransporter 2 inhibitors) data are lacking as to their specific impact on immune function with regards to infection risk apart from their reduction of hyperglycemia (102-104).

Immunomodulating Therapies

There are data showing that the use of granulocyte-colony stimulating factor (G-CSF), which induces differentiation and release of PMN from marrow, is able to assist with healing in foot infections (105). While a subsequent meta-analysis did not show that the use of G-CSF was able to significantly impact the resolution or healing of wounds, there was a reduction in risk of needing amputation or surgical intervention (106).

While not a “medication”, physical activity has long been known to be associated with improvement in the immune system which are known to extend as well to those with DM (107). In diabetic rats, exercise was able to improve the neutrophil and lymphocyte count significantly (108).