OBJECTIVE

We know that diabetes predisposes to common infections, such as cellulitis and pneumonia. However, the correlation between the level of glycemic control and the rate of infection is unknown.

RESEARCH DESIGN AND METHODS

We examined the association between glycemic control in patients with diabetes and the incidence of infection in the entire population of patients with diabetes in a large HMO. During the study period, we first selected an HbA1c test for each patient and then searched for an infection diagnosis in the 60 days that followed the test. A multivariate logistic regression analysis was performed to determine the independent effect of HbA1c on the likelihood of being diagnosed with an infection. We were able to control for many confounders, such as other chronic illness, time since the diagnosis of diabetes, and use of steroids before the infection.

RESULTS

We identified 407 cases of cellulitis. Multivariate logistic regressions for cellulitis showed a 1.4-fold increased risk among patients with HbA1c >7.5% (58 mmol/mol). Factors such as obesity, Parkinson’s disease, peripheral vascular disease, and prior treatment with prednisone predisposed to cellulitis. There was an increase of 12% in the odds of cellulitis for every 1% (11 mmol/mol) elevation in HbA1c (odds ratio [OR] 1.12; CI 1.05–1.19). A similar analysis showed a trend toward an increased risk of pneumonia in patients with HbA1c >7.5% (58 mmol/mol) (OR 1.1; CI 0.9–1.4).

CONCLUSIONS

Poor glycemic control was associated in this study with the development of cellulitis. The study also suggests that exposure to oral prednisolone increases the risk of cellulitis, pneumonia, and upper respiratory infection.

Diabetes is a leading cause of morbidity and mortality (1). The estimated prevalence of diabetes among adults age ≥18 years was 12.2% in 2015 in the U.S., and it was the seventh leading cause of death (1). The estimated prevalence of diabetes among adults age ≥21 years was 9.7% in 2015 in Israel (2).

It has been demonstrated that good glycemic control decreases the prevalence of both microvascular and macrovascular complications (3). In addition to such complications, patients with diabetes have an increased risk of infection. Some rare infections are pathognomonic or more common among those with diabetes, such as emphysematous pyelonephritis, malignant otitis externa, Fournier’s gangrene, and mucormycosis (4). However, common infections, such as pneumonia and cellulitis, have a greater impact in individuals with diabetes. This was demonstrated by a retrospective cohort study that compared patients with diabetes with matched patients without diabetes and showed that the risk ratio for infection was 1.21, the risk of infection-related hospitalization was 2.17, and the risk of infection-related death was 1.92 (5).

Poor glycemic control has been associated with poor outcomes, such as microvascular complications. However, the relation between glycemic control and the risk of infection is unclear. There have been several studies with conflicting results (4). A population-based case-control study showed that poor long-term glycemic control increased the risk of hospitalization with pneumonia for patients with diabetes compared with patients without a diagnosis of diabetes. The relative risk (RR) for those with HbA1c <7% (53 mmol/mol) was 1.22, whereas the RR for those with HbA1c ≥9% (75 mmol/mol) was 1.60 (6). A different study that compared the risk of recurrent urinary tract infection (UTI) among women with and without diabetes found that glycemic control did not influence the risk (7). Another study that compared the incidence of UTI among patients with and without diabetes found that patients with poor glycemic control (HbA1c >8% [64 mmol/mol]) were at higher risk of UTI than those with fair control (RR 12.4) (8).

The aim of this study was to examine the association between glycemic control in patients with type 2 diabetes and the incidence of infection.

Study Design

This historical prospective cohort study examined the association between glycemic control in patients with diabetes and the incidence of eight infections using the Meuhedet HMO database.

The Meuhedet HMO serves 1.2 million patients in Israel. Patients’ medical data are stored in a comprehensive database that serves different health service providers when treating Meuhedet patients. According to the Israeli National Health Insurance Act legislated in 1994, all citizens must be registered with one of the four HMOs that are obliged to insure every citizen wishing to join, irrespective of age, sex, physical condition, or any other criterion. Citizens can move freely among the HMOs, which operate in every town and region. According to Israeli Health Ministry statistics, the Meuhedet population is a little younger than those of the other HMOs, with every part of the Israeli population being well represented in its database.

The following eight infections were selected for analysis: cellulitis, cholecystitis, diverticulitis, upper respiratory tract infection (URTI), UTI, influenza, sinusitis, herpes zoster, and pneumonia. The occurrence of an infection was identified using a community-based diagnosis, from the visit diagnosis field in the patient’s electronic medical record. The physician fills in this field before completing an office visit, and the diagnosis is taken from a list of ICD-9 diagnoses. All the different terms that appear in the Meuhedet database to describe the selected infections were used. A unique infection was defined as an infection that was recorded in the visit diagnosis field with no other infection diagnosis in the previous 14 days.

Glycemic control was assessed using the HbA1c test. The first HbA1c test during the study period was selected for each patient, and a search was then performed for an infection diagnosis in the 60 days after the test.

Population

The Meuhedet database uses a diabetes register to identify all patients with diabetes. We identified all patients who were born in 1964 or earlier and registered with a diagnosis of diabetes (n = 56,626). The exclusion criteria were: having an oncologic disease (n = 10,439), receiving antineoplastic or immunosuppressive medication other than corticosteroids (n = 3,490), undergoing dialysis (n = 549), hemoglobin level <9 g/dL (n = 1,026), and glucose-6-phosphate dehydrogenase deficiency (n = 197). Patients who died during the study period were also excluded (n = 1,566). A total of 13,846 patients were excluded using these criteria; the rest were included in the study cohort only if they had at least one HbA1c test during the study period and fulfilled one of the following criteria: had at least two HbA1c tests with a value >6.4% (46 mmol/mol) during the study period and received at least one medication for diabetes during the study period. The final cohort included 33,637 patients (Fig. 1).

Figure 1

Flow diagram of study inclusion and exclusion criteria. DM, diabetes mellitus; G6PD, glucose-6-phosphate dehydrogenase.

Figure 1

Flow diagram of study inclusion and exclusion criteria. DM, diabetes mellitus; G6PD, glucose-6-phosphate dehydrogenase.

Close modal

Data Collection

The study period was from October 2014 to September 2017. We collected the following demographic information: age, sex, and socioeconomic index. Comorbid conditions were obtained from the electronic medical record field for chronic disease diagnosis or visit diagnosis. Information on the following diagnoses was collated: ischemic heart disease (IHD), peripheral vascular disease (PVD), cerebrovascular accident, congestive heart failure (CHF), asthma, chronic obstructive pulmonary disease (COPD), Parkinson’s disease, and dementia. Chronic kidney disease (CKD) was defined as an estimated glomerular filtration rate of <60 mL/min/1.73 m2. The MDRD study equation was used for glomerular filtration rate calculation. Additional information obtained was date of diabetes diagnosis, BMI, and use of diabetes medication. The type and dose of systemic steroids, either orally or per injection, and inhaled steroids 90 days before and during the study period were recorded. The data were collected during November 2018.

Data Analysis

First, we compared the HbA1c test results that were followed by an infection with those that were not. A multivariate logistic regression analysis was then performed to determine the independent effect of HbA1c on the likelihood of being identified with an infection, after adjustment for all confounders (odds ratio [OR] and 95% CI). Data analyses were conducted using SPSS (version 25). A probability level of <0.05 was considered significant in all analyses. This study was approved by the institutional review board of the Meuhedet (number 01-02-10-17).

Description of Patients With Diabetes

Of 33,637 patients in the study cohort, 47% (15,857 patients) were women. The median age at the beginning of the study was 65.9 years (interquartile range 59–71.9); 72% (24,259 patients) were treated with glucose-lowering medications other than insulin, 21% (7,170 patients) were treated with insulin as a single medication or in combination with other drugs, and 7% were not treated with any diabetes medication. Patient characteristics appear in Table 1.

Table 1

Characteristics of the study cohort and univariate analysis

All (n = 33,637)With cellulitis (n = 407)With pneumonia (n = 441)With URTI (n = 3,161)With UTI (n = 1,004)
Age, years      
P  NS <0.01 <0.001 <0.001 
 50–59 9,733 (28.9) 120 (29.5) 99 (22.4) 997 (31.5) 221 (22) 
 60–64 6,632 (19.7) 72 (17.7) 78 (17.7) 684 (21.6) 168 (16.7) 
 65–69 6,940 (20.6) 86 (21.1) 92 (20.9) 664 (20.4) 196 (19.5) 
 70+ 10,332 (30.7) 129 (31.7) 172 (39.0) 836 (26.4) 419 (41.7) 
Sex      
P  NS NS <0.001 <0.001 
 Female 15,857 (47.1) 182 (44.7) 215 (48.8) 1,665 (52.7) 757 (75.4) 
 Male 17,780 (52.9) 225 (55.3) 226 (51.2) 1,496 (47.3) 247 (24.6) 
Diabetes medications      
P  <0.001 <0.01 <0.01 0.14 
 Noninsulin only 24,259 (72.1) 265 (65.1) 291 (66) 2,258 (71.4) 730 (72.7) 
 Insulin with/without other medications 7,170 (21.3) 121 (29.7) 121 (27.4) 773 (23.2) 223 (22.2) 
 None 2,211 (6.6) 21 (5.2) 29 (6.6) 170 (5.4) 51 (5.1) 
Comorbidity (multiple variables)      
 IHD 10,682 (31.8) 146 (35.9) 192 (43.5)* 1,030 (32.6) 326 (32.5) 
 PVD 4,086 (12.1) 121 (29.7)* 67 (15.2)* 338 (12.3) 155 (15.4)* 
 Cerebrovascular accident 5,100 (15.2) 68 (16.7) 85 (19.3)* 507 (16) 187 (18.6)* 
 CHF 2,706 (8.0) 59 (14.5)* 85 (19.3)* 253 (8) 91 (9.1) 
 Asthma 6,384 (19.0) 86 (21.1) 157 (35.6)* 870 (27.5)* 262 (26.1)* 
 COPD 5,889 (17.5) 77 (18.9) 162 (36.7)* 720 (22.8)* 204 (20.3)* 
 Parkinson’s disease 830 (2.5) 19 (4.7)* 13 (2.9) 60 (1.9)* 49 (4.9)* 
 Dementia 1,350 (4.0) 24 (5.9) 28 (6.3)* 93 (2.9)* 72 (7.2)* 
 CKD& 4,256 (12.7) 72 (17.8)* 90 (20.5)* 369 (11.7) 152 (15.2)* 
BMI, kg/m2      
P  <0.001 0.02 <0.001 <0.01 
 <25 3,417 (10.2) 30 (7.9) 43 (10.1) 272 (9) 88 (9.3) 
 25–29 10,946 (32.5) 98 (25.7) 124 (29.2) 1,012 (33.4) 288 (30.3) 
 30–34 10,342 (30.7) 108 (28.3) 140 (33) 990 (32.7) 329 (34.6) 
 35+ 6,884 (20.5) 146 (38.2) 117 (27.6) 752 (24.9) 245 (25.8) 
Years since diabetes diagnosis      
P  0.02 0.29 0.55 0.23 
 <5 13,528 (40.2) 144 (35.6) 167 (38.7) 1,290 (41.7) 387 (39.4) 
 5+ 19,287 (57.3) 260 (64.4) 264 (61.3) 1,801 (58.3) 596 (60.6) 
Socioeconomic index§      
P  0.03 <0.01 <0.01 0.01 
 1–5 5,778 (17.2) 90 (22.5) 101 (23.1) 621 (19.9) 210 (21.3) 
 6–10 13,697 (40.7) 167 (41.8) 163 (37.3) 1,288 (41.3) 400 (40.5) 
 11–15 11,815 (35.1) 126 (31.5) 156 (35.7) 1,057 (33.9) 322 (32.6) 
 16–20 1,816 (5.4) 17 (4.3) 17 (3.9) 154 (4.9) 55 (5.6) 
All (n = 33,637)With cellulitis (n = 407)With pneumonia (n = 441)With URTI (n = 3,161)With UTI (n = 1,004)
Age, years      
P  NS <0.01 <0.001 <0.001 
 50–59 9,733 (28.9) 120 (29.5) 99 (22.4) 997 (31.5) 221 (22) 
 60–64 6,632 (19.7) 72 (17.7) 78 (17.7) 684 (21.6) 168 (16.7) 
 65–69 6,940 (20.6) 86 (21.1) 92 (20.9) 664 (20.4) 196 (19.5) 
 70+ 10,332 (30.7) 129 (31.7) 172 (39.0) 836 (26.4) 419 (41.7) 
Sex      
P  NS NS <0.001 <0.001 
 Female 15,857 (47.1) 182 (44.7) 215 (48.8) 1,665 (52.7) 757 (75.4) 
 Male 17,780 (52.9) 225 (55.3) 226 (51.2) 1,496 (47.3) 247 (24.6) 
Diabetes medications      
P  <0.001 <0.01 <0.01 0.14 
 Noninsulin only 24,259 (72.1) 265 (65.1) 291 (66) 2,258 (71.4) 730 (72.7) 
 Insulin with/without other medications 7,170 (21.3) 121 (29.7) 121 (27.4) 773 (23.2) 223 (22.2) 
 None 2,211 (6.6) 21 (5.2) 29 (6.6) 170 (5.4) 51 (5.1) 
Comorbidity (multiple variables)      
 IHD 10,682 (31.8) 146 (35.9) 192 (43.5)* 1,030 (32.6) 326 (32.5) 
 PVD 4,086 (12.1) 121 (29.7)* 67 (15.2)* 338 (12.3) 155 (15.4)* 
 Cerebrovascular accident 5,100 (15.2) 68 (16.7) 85 (19.3)* 507 (16) 187 (18.6)* 
 CHF 2,706 (8.0) 59 (14.5)* 85 (19.3)* 253 (8) 91 (9.1) 
 Asthma 6,384 (19.0) 86 (21.1) 157 (35.6)* 870 (27.5)* 262 (26.1)* 
 COPD 5,889 (17.5) 77 (18.9) 162 (36.7)* 720 (22.8)* 204 (20.3)* 
 Parkinson’s disease 830 (2.5) 19 (4.7)* 13 (2.9) 60 (1.9)* 49 (4.9)* 
 Dementia 1,350 (4.0) 24 (5.9) 28 (6.3)* 93 (2.9)* 72 (7.2)* 
 CKD& 4,256 (12.7) 72 (17.8)* 90 (20.5)* 369 (11.7) 152 (15.2)* 
BMI, kg/m2      
P  <0.001 0.02 <0.001 <0.01 
 <25 3,417 (10.2) 30 (7.9) 43 (10.1) 272 (9) 88 (9.3) 
 25–29 10,946 (32.5) 98 (25.7) 124 (29.2) 1,012 (33.4) 288 (30.3) 
 30–34 10,342 (30.7) 108 (28.3) 140 (33) 990 (32.7) 329 (34.6) 
 35+ 6,884 (20.5) 146 (38.2) 117 (27.6) 752 (24.9) 245 (25.8) 
Years since diabetes diagnosis      
P  0.02 0.29 0.55 0.23 
 <5 13,528 (40.2) 144 (35.6) 167 (38.7) 1,290 (41.7) 387 (39.4) 
 5+ 19,287 (57.3) 260 (64.4) 264 (61.3) 1,801 (58.3) 596 (60.6) 
Socioeconomic index§      
P  0.03 <0.01 <0.01 0.01 
 1–5 5,778 (17.2) 90 (22.5) 101 (23.1) 621 (19.9) 210 (21.3) 
 6–10 13,697 (40.7) 167 (41.8) 163 (37.3) 1,288 (41.3) 400 (40.5) 
 11–15 11,815 (35.1) 126 (31.5) 156 (35.7) 1,057 (33.9) 322 (32.6) 
 16–20 1,816 (5.4) 17 (4.3) 17 (3.9) 154 (4.9) 55 (5.6) 

Data are n (%).

*

P < 0.05.

&

Data are missing for 193 patients.

Data are missing for 2,048 patients.

Data are missing for 822 patients.

§

Data are missing for 549 patients.

Infection and HbA1c Measurement

During a period of 60 days after the first HbA1c test, 4,806 patients had at least one infection. Among these infections, there were 3,161 cases of URTI, 1,004 cases of UTI, 441 cases of pneumonia, and 407 cases of cellulitis. Other infections were rare (<100 cases).

The mean HbA1c test result with all subsequent infections was 7.2% (55 mmol/mol); the mean HbA1c test results with subsequent cellulitis and pneumonia infections were 7.6% (60 mmol/mol; P < 0.001) and 7.2% (55 mmol/mol), respectively; the mean HbA1c test results were significantly lower for patients diagnosed with URTI and UTI, at 7.2% (55 mmol/mol) and 7.1% (54 mmol/mol), respectively.

Cellulitis

In univariate analysis, there was a statistically significant association between the incidence of cellulitis and HbA1c >7.5% (58 mmol/mol; χ2 [1] = 14.99; P < 0.001). A multivariate logistic regression, including all the variables associated with the incidence of cellulitis, showed a 1.4-fold increased risk of cellulitis among patients with HbA1c >7.5% (58 mmol/mol; OR 1.4; CI 1.1–1.7). Odds were increased for patients with BMI ≥35 kg/m2 (OR 2.6; CI 1.7–3.9), Parkinson’s disease (OR 2.0; CI 1.3–3.3), PVD (OR 3.0; CI 2.4–3.8), or prior treatment with prednisone (OR 2.3; CI 1.3–4.0) (Table 2). A multivariate logistic regression with HbA1c level as a continuous variable showed an increase of 12% in the odds of cellulitis for every 1% (11 mmol/mol) elevation in HbA1c (OR 1.12; CI 1.1–1.2) (Table 3).

Table 2

Multivariate logistic regression for cellulitis

ORP95% CI for OR
HbA1c >7.5% 1.38 <0.01 1.11–1.71 
Female vs. male 0.83 0.08 0.67–1.02 
Age, years    
 <60  0.67  
 60–64 0.85 0.30 0.63–1.15 
 65–69 0.86 0.33 0.64–1.16 
 70+ 0.89 0.39 0.67–1.17 
BMI, kg/m2    
 <25  <0.001  
 25–29 1.02 0.95 0.67–1.54 
 30–34 1.17 0.45 0.78–1.77 
 35+ 2.59 <0.001 1.73–3.87 
Parkinson’s vs. no Parkinson’s disease 2.03 <0.01 1.26–3.29 
PVD vs. no PVD 3.00 <0.001 2.38–3.78 
Prior treatment with prednisone vs. no treatment 2.27 <0.01 1.29–4.00 
ORP95% CI for OR
HbA1c >7.5% 1.38 <0.01 1.11–1.71 
Female vs. male 0.83 0.08 0.67–1.02 
Age, years    
 <60  0.67  
 60–64 0.85 0.30 0.63–1.15 
 65–69 0.86 0.33 0.64–1.16 
 70+ 0.89 0.39 0.67–1.17 
BMI, kg/m2    
 <25  <0.001  
 25–29 1.02 0.95 0.67–1.54 
 30–34 1.17 0.45 0.78–1.77 
 35+ 2.59 <0.001 1.73–3.87 
Parkinson’s vs. no Parkinson’s disease 2.03 <0.01 1.26–3.29 
PVD vs. no PVD 3.00 <0.001 2.38–3.78 
Prior treatment with prednisone vs. no treatment 2.27 <0.01 1.29–4.00 
Table 3

Multivariate logistic regression for cellulitis with HbA1c analyzed as continuous variable

ORP95% CI for OR
HbA1c 1.12 <0.001 1.05–1.19 
Female vs. male 0.83 0.08 0.67–1.02 
Age, years    
 <60  0.77  
 60–64 0.87 0.35 0.64–1.17 
 65–69 0.88 0.40 0.65–1.19 
 70+ 0.91 0.52 0.69–1.21 
BMI, kg/m2    
 <25  <0.001  
 25–29 1.02 0.92 0.68–1.55 
 30–34 1.19 0.41 0.79–1.79 
 35+ 2.62 <0.001 1.75–3.91 
Parkinson’s vs. no Parkinson’s disease 2.06 <0.01 1.27–3.34 
PVD vs. no PVD 2.98 <0.001 2.37–3.75 
Prior treatment with prednisone vs. no treatment 2.28 <0.01 1.29–4.02 
ORP95% CI for OR
HbA1c 1.12 <0.001 1.05–1.19 
Female vs. male 0.83 0.08 0.67–1.02 
Age, years    
 <60  0.77  
 60–64 0.87 0.35 0.64–1.17 
 65–69 0.88 0.40 0.65–1.19 
 70+ 0.91 0.52 0.69–1.21 
BMI, kg/m2    
 <25  <0.001  
 25–29 1.02 0.92 0.68–1.55 
 30–34 1.19 0.41 0.79–1.79 
 35+ 2.62 <0.001 1.75–3.91 
Parkinson’s vs. no Parkinson’s disease 2.06 <0.01 1.27–3.34 
PVD vs. no PVD 2.98 <0.001 2.37–3.75 
Prior treatment with prednisone vs. no treatment 2.28 <0.01 1.29–4.02 

Pneumonia

In univariate analysis, HbA1c measurements were not significantly higher among patients with pneumonia infection compared with patients with no infection. A multivariate logistic regression, including variables associated with the incidence of pneumonia, showed a trend toward an increased risk of pneumonia in patients with HbA1c >7.5% (58 mmol/mol; OR 1.1; CI 0.9–1.4). Significantly increased odds were shown for patients with asthma (OR 1.6; CI 1.2–2.0), COPD (OR 1.9; CI 1.5–2.4), IHD (OR 1.3; CI 1.0–1.6), CHF (OR 1.7; CI 1.3–2.2), CKD (OR 1.4; CI 1.1–1.8), prior treatment with prednisone (OR 1.9; CI 1.2–3.1), or prior treatment with inhaled steroids (OR 2; CI 1.4–3.0). Significantly decreased odds were shown for patients with higher socioeconomic status (SES) compared with patients with SES of 1–5: SES 6–10, OR of 0.6 (CI 0.5–0.8); SES 11–15, OR of 0.7 (CI 0.5–0.9); and SES 16–20, OR of 0.5 (CI 0.3–0.9).

URTI

In univariate analysis, there was a statistically significant association between the incidence of URTI and HbA1c <7.5% (58 mmol/mol; χ2 [1] = 3.88; P = 0.049). A multivariate logistic regression, including all the variables associated with the incidence of URTI, showed a 0.8-fold reduced risk of URTI among patients with HbA1c >7.5% (58 mmol/mol; OR 0.8; CI 0.8–0.9). Odds for patients in age-group 65–70 were decreased compared with those in age-group <60 (OR 0.8; CI 0.8–0.9), and odds for patients in age-group 70+ were decreased compared with those in age-group <60 (OR 0.8; CI 0.7–0.8). The odds for patients with dementia were decreased compared to those without dementia (OR 0.8; CI 0.6–0.99). Odds were increased for women (OR 1.3; CI 1.2–1.4) and patients with asthma (OR 1.5; CI 1.3–1.6), COPD (OR 1.2; CI 1.1–1.4), prior treatment with prednisone (OR 1.5; CI 1.2–2.0), or prior treatment with inhaled steroids (OR 1.6; CI 1.3–1.9). Odds were also increased for patients who were treated with glucose-lowering medications other than insulin (OR 1.5; CI 1.2–1.8) or insulin as a single medication or in combination with other drugs (OR 1.2; CI 1.0–1.5).

UTI

In univariate analysis, there was a statistically significant association between the incidence of UTI and HbA1c <7.5% (58 mmol/mol; χ2 [1] = 6.98; P = 0.008). However, this association was not significant in a multivariate logistic regression (OR 0.9; CI 0.8–1.1). Odds were increased for women (OR 3.5; CI 3–4.1), patients in the age group 70+ years compared with patients age <60 years (OR 1.6; CI 1.3–1.9), and patients with PVD (OR 1.4; CI 1.1–1.7), asthma (OR 1.3; CI 1.1–1.5), or Parkinson’s disease (OR 1.5; CI 1.1–2.1). Odds were decreased for patients with higher SES compared with patients with SES 1–5: SES 6–10, OR of 0.7 (CI 0.6–0.9), and SES 11–15, OR of 0.7 (CI 0.6–0.8).

Prior Treatment With Prednisone or Inhaled Steroids

Prior treatment with prednisone at minimum dose of 5 mg was recorded in 506 patients, 33% (167 patients) of whom were treated with prednisone only once during the study period; 60% (299 patients) had a diagnosis of either asthma or COPD. At least once during the study period, 713 patients were treated with inhaled steroids; only 78% (554 patients) had a diagnosis of either asthma or COPD.

Prior treatment with prednisone was significantly associated with cellulitis, pneumonia, and URTI but not with UTI, whereas prior treatment with inhaled steroids was significantly associated with pneumonia and URTI.

We found that HbA1c >7.5% (58 mmol/mol) was associated with a 1.4-fold increased risk of cellulitis. There was a 1.12-fold increased risk of cellulitis for every 1% (11 mmol/mol) elevation in HbA1c. There was a trend toward higher HbA1c level and higher incidence of pneumonia, a finding that may have become significant with a larger sample and is supported by other studies (4,6). Consistent with many other studies, we did not find a positive association between poor glycemic control and the incidence of UTI (811); for patients with UTI, HbA1c was slightly lower, but no association was found in multivariate analysis. Interestingly, we found an association between URTI and HbA1c <7.5% (58 mmol/mol). This may be explained by the fact that patients who frequently visit their physician tend to have better medication adherence and glycemic control (12). Patients who visit less often may simply self-medicate for URTI.

The large number of patients in our cohort allowed for the analysis of confounders that were not examined in previous studies, including comorbidities and prior steroid treatment. Another interesting finding was a 2.0- and 3.0-fold increased risk of cellulitis among patients with Parkinson’s disease and PVD, respectively. The latter may be related to insufficient blood flow and skin breakdown (13), but a different explanation should be sought for the association with Parkinson’s disease. Information regarding active foot disease or ulceration was unavailable. As expected, an association was found between pneumonia and asthma and COPD (14,15). An association was also found between IHD, CHF, and CKD (16).

We found that treatment with prednisone within 90 days before the infection was associated with increased risk of cellulitis, pneumonia, and URTI, but we did not find a similar association with other systemic steroids. A similar positive association was found between inhaled steroids taken within 90 days before the infection and pneumonia and URTI. Although this association is known from previous studies (17), it may indicate the severity of a patient’s chronic lung disease, so the steroid use may not be related directly to an increased risk of pneumonia. We were unable to ascertain the indication for treatment with steroids, nor the exact dose taken by each patient; however, 59% of them had a diagnosis of asthma or COPD, and 33% were treated with prednisone only once during the study period. The study design excluded patients with a diagnosis of an oncologic disease as well as patients who received antineoplastic or immunosuppressive medications.

This study was designed to assess the relationship of diabetic control and infection in a temporal association. We used a large community-based cohort that includes all patients with diabetes who registered with the Meuhedet HMO, which insures 13% of the Israeli population. This large cohort allowed a multivariate analysis of confounders that were not included in previous studies. This study design allowed a good estimation of a patient’s glycemic control at the time of the infection.

A limitation of this study is that it used a real-world database that relies on clinical diagnoses by many different physicians. As an observational study, it did not allow adjustment for unknown confounders. Another limitation is that purchase of steroid medication was used as a surrogate marker for steroid use, which may not reflect actual steroid use.

This study adds to the existing data suggesting that poor glycemic control is associated with cellulitis and is the first to assess the additional risk as a variable of glycemic control analyzed as a continuous variable.

Our results have possible clinical implications. Cellulitis was a common infection in our cohort, with 72.6 cases per 1,000 patient-years. Therefore, we consider a 1.4-fold increase to be clinically significant. Compared with insulin treatment, newer diabetes medications allow the targeting of a lower HbA1c level with a reduced risk of hypoglycemia. A lower HbA1c level may reduce the risk of cellulitis, especially for patients with Parkinson’s disease or PVD. Within the limitations of our study, there may be an increased risk of infection with steroid use for those with comorbid diabetes.

Future research should examine the possible efficacy of treating chronic cellulitis with tighter glycemic control if there are no contraindications to additional medications. The association of steroid treatment and the incidence of infection in populations with diabetes should be further investigated.

This article is featured in a podcast available at https://www.diabetesjournals.org/content/diabetes-core-update-podcasts.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. G.Z. was the main researcher and contributed to literature review, study design, data collection, and manuscript writing. F.H.S. contributed to study design, data analysis, and manuscript writing. A.D.H. contributed to literature review, study design, and manuscript writing. A.D.H. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

1.
National Center for Chronic Disease Prevention and Health Promotion
.
National Diabetes Statistic Report, Estimates of Diabetes and Its Burden in the United States
. Accessed 2 March 2019. Available from https://www.cdc.gov/diabetes/pdfs/data/statistics/national-diabetes-statistics-report.pdf
2.
Israeli Ministry of Health
.
National Program for Quality Indicators in Community Healthcare 2014-2016
. Accessed 2 March 2019. Available from https://www.israelhealthindicators.org/publications
3.
Stratton
IM
,
Adler
AI
,
Neil
HAW
, et al
.
Association of glycaemia with macrovascular and microvascular complications of type 2 diabetes (UKPDS 35): prospective observational study
.
BMJ
2000
;
321
:
405
412
4.
Pearson-Stuttard
J
,
Blundell
S
,
Harris
T
,
Cook
DG
,
Critchley
J
.
Diabetes and infection: assessing the association with glycaemic control in population-based studies
.
Lancet Diabetes Endocrinol
2016
;
4
:
148
158
5.
Shah
BR
,
Hux
JE
.
Quantifying the risk of infectious diseases for people with diabetes
.
Diabetes Care
2003
;
26
:
510
513
6.
Kornum
JB
,
Thomsen
RW
,
Riis
A
,
Lervang
HH
,
Schønheyder
HC
,
Sørensen
HT
.
Diabetes, glycemic control, and risk of hospitalization with pneumonia: a population-based case-control study
.
Diabetes Care
2008
;
31
:
1541
1545
7.
Gorter
KJ
,
Hak
E
,
Zuithoff
NP
,
Hoepelman
AI
,
Rutten
GE
.
Risk of recurrent acute lower urinary tract infections and prescription pattern of antibiotics in women with and without diabetes in primary care
.
Fam Pract
2010
;
27
:
379
385
8.
Hirji
I
,
Guo
Z
,
Andersson
SW
,
Hammar
N
,
Gomez-Caminero
A
.
Incidence of urinary tract infection among patients with type 2 diabetes in the UK General Practice Research Database (GPRD)
.
J Diabetes Complications
2012
;
26
:
513
516
9.
Bartelink
ML
,
Hoek
L
,
Freriks
JP
,
Rutten
GE
.
Infections in patients with type 2 diabetes in general practice
.
Diabetes Res Clin Pract
1998
;
40
:
15
19
10.
Boyko
EJ
,
Fihn
SD
,
Scholes
D
,
Abraham
L
,
Monsey
B
.
Risk of urinary tract infection and asymptomatic bacteriuria among diabetic and nondiabetic postmenopausal women
.
Am J Epidemiol
2005
;
161
:
557
564
11.
Hammar
N
,
Farahmand
B
,
Gran
M
,
Joelson
S
,
Andersson
SW
.
Incidence of urinary tract infection in patients with type 2 diabetes. Experience from adverse event reporting in clinical trials
.
Pharmacoepidemiol Drug Saf
2010
;
19
:
1287
1292
12.
Dobbins
JM
,
Elliott
SW
,
Cordier
T
, et al
.
Primary care provider encounter cadence and HbA1c control in older patients with diabetes
.
Am J Prev Med
2019
;
57
:
e95
e101
13.
Dryden
M
,
Baguneid
M
,
Eckmann
C
, et al
.
Pathophysiology and burden of infection in patients with diabetes mellitus and peripheral vascular disease: focus on skin and soft-tissue infections
.
Clin Microbiol Infect
2015
;
21
(
Suppl. 2
):
S27
S32
14.
Rivero-Calle
I
,
Cebey-López
M
,
Pardo-Seco
J
, et al
.
Lifestyle and comorbid conditions as risk factors for community-acquired pneumonia in outpatient adults (NEUMO-ES-RISK project)
.
BMJ Open Respir Res
2019
;
6
:
e000359
15.
Zaidi
SR
,
Blakey
JD
.
Why are people with asthma susceptible to pneumonia? A review of factors related to upper airway bacteria
.
Respirology
2019
;
24
:
423
430
16.
Advisory Committee on Immunisation Practices - CDC
.
Prevention of pneumococcal disease: recommendations of the Advisory Committee on Immunization Practices (ACIP), 1997
. Accessed 21 June 2019. Available from https://www.cdc.gov/mmwr/preview/mmwrhtml/00047135.htm
17.
Morjaria
JB
,
Rigby
A
,
Morice
AH
.
Inhaled corticosteroid use and the risk of pneumonia and COPD exacerbations in the UPLIFT study
.
Lung
2017
;
195
:
281
288
Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at https://www.diabetesjournals.org/content/license.