Risk Factors and Mortality Rates Associated With Invasive Group B Streptococcus Infections Among Patients in the US Veterans Health Administration | Infectious Diseases | JAMA Network Open | JAMA Network
[Skip to Navigation]
Sign In
Figure 1.  Kaplan-Meier Curves of Time to Death Following an Invasive Group B Streptococcus Infection by Infectious Syndrome
Kaplan-Meier Curves of Time to Death Following an Invasive Group B Streptococcus Infection by Infectious Syndrome
Figure 2.  Incidence of Invasive Group B Streptococcus Infections Among Patients in the Veterans Health Administration, 2008-2017
Incidence of Invasive Group B Streptococcus Infections Among Patients in the Veterans Health Administration, 2008-2017

Active Veterans Affairs health care users included those who had an outpatient primary care, specialty care clinic visit, or a hospital admission in the year evaluated. Invasive infections were defined as a positive culture from a normally sterile site.

Figure 3.  Incidence of Invasive Group B Streptococcus (GBS) Infections Among Patients in the Veterans Health Administration, Assessing Interactions Among Body Mass Index (BMI), Hemoglobin A1c (HbA1c), and Age
Incidence of Invasive Group B Streptococcus (GBS) Infections Among Patients in the Veterans Health Administration, Assessing Interactions Among Body Mass Index (BMI), Hemoglobin A1c (HbA1c), and Age

BMI calculated as weight in kilograms divided by height in meters squared. To convert HbA1c to proportion of total hemoglobin, multiply by 0.01.

Table 1.  Characteristics of Patients in the Veterans Health Administration With Invasive Group B Streptococcus Infections, by Age
Characteristics of Patients in the Veterans Health Administration With Invasive Group B Streptococcus Infections, by Age
Table 2.  Characteristics and Outcomes of Patients in the Veterans Health Administration With Invasive Group B Streptococcus Infections by Clinical Syndrome
Characteristics and Outcomes of Patients in the Veterans Health Administration With Invasive Group B Streptococcus Infections by Clinical Syndrome
1.
Skoff  TH, Farley  MM, Petit  S,  et al.  Increasing burden of invasive group B Streptococcal disease in nonpregnant adults, 1990-2007.  Clin Infect Dis. 2009;49(1):85-92. doi:10.1086/599369PubMedGoogle ScholarCrossref
2.
Francois Watkins  LK, McGee  L, Schrag  SJ,  et al.  Epidemiology of invasive group B Streptococcal infections among nonpregnant adults in the United States, 2008-2016.  JAMA Intern Med. 2019;179(4):479-488. doi:10.1001/jamainternmed.2018.7269PubMedGoogle ScholarCrossref
3.
Farley  MM, Harvey  RC, Stull  T,  et al.  A population-based assessment of invasive disease due to group B Streptococcus in nonpregnant adults.  N Engl J Med. 1993;328(25):1807-1811. doi:10.1056/NEJM199306243282503PubMedGoogle ScholarCrossref
4.
Pitts  SI, Maruthur  NM, Langley  GE,  et al.  Obesity, diabetes, and the risk of invasive group B Streptococcal disease in nonpregnant adults in the United States.  Open Forum Infect Dis. 2018;5(6):ofy030. doi:10.1093/ofid/ofy030PubMedGoogle Scholar
5.
US Centers for Disease Control and Prevention. Active Bacterial Core surveillance. https://www.cdc.gov/abcs/reports-findings/surv-reports.html. Accessed July 11, 2019.
6.
Quan  H, Sundararajan  V, Halfon  P,  et al.  Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data.  Med Care. 2005;43(11):1130-1139. doi:10.1097/01.mlr.0000182534.19832.83PubMedGoogle ScholarCrossref
7.
Tazi  A, Morand  PC, Réglier-Poupet  H,  et al.  Invasive group B Streptococcal infections in adults, France (2007-2010).  Clin Microbiol Infect. 2011;17(10):1587-1589. doi:10.1111/j.1469-0691.2011.03628.xPubMedGoogle ScholarCrossref
8.
Morozumi  M, Wajima  T, Takata  M, Iwata  S, Ubukata  K.  Molecular characteristics of group B Streptococci isolated from adults with invasive infections in Japan.  J Clin Microbiol. 2016;54(11):2695-2700. doi:10.1128/JCM.01183-16PubMedGoogle ScholarCrossref
9.
Camuset  G, Picot  S, Jaubert  J,  et al.  Invasive group B Streptococcal disease in non-pregnant adults, Réunion Island, 2011.  Int J Infect Dis. 2015;35:46-50. doi:10.1016/j.ijid.2015.04.006PubMedGoogle ScholarCrossref
10.
Harrison  LH, Ali  A, Dwyer  DM,  et al.  Relapsing invasive group B Streptococcal infection in adults.  Ann Intern Med. 1995;123(6):421-427. doi:10.7326/0003-4819-123-6-199509150-00004PubMedGoogle ScholarCrossref
11.
Huang  K-F, Hung  M-H, Lin  Y-S,  et al.  Independent predictors of mortality for necrotizing fasciitis: a retrospective analysis in a single institution.  J Trauma. 2011;71(2):467-473. doi:10.1097/TA.0b013e318220d7faPubMedGoogle ScholarCrossref
12.
Hadeed  GJ, Smith  J, O’Keeffe  T,  et al.  Early surgical intervention and its impact on patients presenting with necrotizing soft tissue infections: a single academic center experience.  J Emerg Trauma Shock. 2016;9(1):22-27. doi:10.4103/0974-2700.173868PubMedGoogle ScholarCrossref
13.
Thuny  F, Giorgi  R, Habachi  R,  et al.  Excess mortality and morbidity in patients surviving infective endocarditis.  Am Heart J. 2012;164(1):94-101. doi:10.1016/j.ahj.2012.04.003PubMedGoogle ScholarCrossref
14.
Ballard  MS, Schønheyder  HC, Knudsen  JD,  et al; International Bacteremia Surveillance Collaborative.  The changing epidemiology of group B Streptococcus bloodstream infection: a multi-national population-based assessment.  Infect Dis (Lond). 2016;48(5):386-391. doi:10.3109/23744235.2015.1131330PubMedGoogle ScholarCrossref
15.
Lambertsen  L, Ekelund  K, Skovsted  IC, Liboriussen  A, Slotved  H-C.  Characterisation of invasive group B Streptococci from adults in Denmark, 1999 to 2004.  Eur J Clin Microbiol Infect Dis. 2010;29(9):1071-1077. doi:10.1007/s10096-010-0941-zPubMedGoogle ScholarCrossref
16.
Bergseng  H, Rygg  M, Bevanger  L, Bergh  K.  Invasive group B Streptococcus (GBS) disease in Norway, 1996-2006.  Eur J Clin Microbiol Infect Dis. 2008;27(12):1193-1199. doi:10.1007/s10096-008-0565-8PubMedGoogle ScholarCrossref
17.
Lamagni  TL, Keshishian  C, Efstratiou  A,  et al.  Emerging trends in the epidemiology of invasive group B Streptococcal disease in England and Wales, 1991-2010.  Clin Infect Dis. 2013;57(5):682-688. doi:10.1093/cid/cit337PubMedGoogle ScholarCrossref
18.
Abat  C, Chaudet  H, Raoult  D, Colson  P.  Increasing trend of invasive group B Streptococcal infections, Marseille, France.  Clin Infect Dis. 2014;58(5):750-751. doi:10.1093/cid/cit777PubMedGoogle ScholarCrossref
19.
Agha  Z, Lofgren  RP, VanRuiswyk  JV, Layde  PM.  Are patients at Veterans Affairs medical centers sicker? a comparative analysis of health status and medical resource use.  Arch Intern Med. 2000;160(21):3252-3257. doi:10.1001/archinte.160.21.3252PubMedGoogle ScholarCrossref
20.
Dursa  EK, Barth  SK, Bossarte  RM, Schneiderman  AI.  Demographic, military, and health characteristics of VA health care users and nonusers who served in or during Operation Enduring Freedom or Operation Iraqi Freedom, 2009-2011.  Public Health Rep. 2016;131(6):839-843. doi:10.1177/0033354916676279PubMedGoogle ScholarCrossref
21.
Keyhani  S, Ross  JS, Hebert  P, Dellenbaugh  C, Penrod  JD, Siu  AL.  Use of preventive care by elderly male veterans receiving care through the Veterans Health Administration, Medicare fee-for-service, and Medicare HMO plans.  Am J Public Health. 2007;97(12):2179-2185. doi:10.2105/AJPH.2007.114934PubMedGoogle ScholarCrossref
22.
Landrum  MB, Keating  NL, Lamont  EB,  et al.  Survival of older patients with cancer in the Veterans Health Administration versus fee-for-service Medicare.  J Clin Oncol. 2012;30(10):1072-1079. doi:10.1200/JCO.2011.35.6758PubMedGoogle ScholarCrossref
23.
Kothari  NJ, Morin  CA, Glennen  A,  et al.  Invasive group B Streptococcal disease in the elderly, Minnesota, USA, 2003-2007.  Emerg Infect Dis. 2009;15(8):1279-1281. doi:10.3201/eid1508.081381PubMedGoogle ScholarCrossref
24.
Jones  BE, Jones  MM, Huttner  B,  et al.  Trends in antibiotic use and nosocomial pathogens in hospitalized veterans with pneumonia at 128 medical centers, 2006-2010.  Clin Infect Dis. 2015;61(9):1403-1410. doi:10.1093/cid/civ629PubMedGoogle ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Original Investigation
    Infectious Diseases
    December 27, 2019

    Risk Factors and Mortality Rates Associated With Invasive Group B Streptococcus Infections Among Patients in the US Veterans Health Administration

    Author Affiliations
    • 1Geriatric Research Education and Clinical Center, VA Northeast Ohio Healthcare System, Cleveland
    • 2Specialty Care Center of Innovation, VA Northeast Ohio Healthcare System, Cleveland
    • 3Division of Infectious Diseases and HIV Medicine, Department of Medicine, Case Western Reserve University School of Medicine, Cleveland, Ohio
    • 4Department of Population and Quantitative Health Sciences, Case Western Reserve University School of Medicine, Cleveland, Ohio
    • 5Cleveland Institute for Computational Biology, Case Western Reserve University School of Medicine, Cleveland, Ohio
    JAMA Netw Open. 2019;2(12):e1918324. doi:10.1001/jamanetworkopen.2019.18324
    Key Points español 中文 (chinese)

    Question  What are the risk factors and mortality rates associated with different types of invasive group B Streptococcus (GBS) infections?

    Findings  In this cohort study of 5175 veterans with 5497 cases of invasive GBS infections, all-cause 30-day mortality was highest among patients with peritonitis (28%) and pneumonia or empyema (17%) and lowest among those with osteomyelitis (1%) or joint infection (3%). The incidence of invasive GBS infections among patients with poorly controlled diabetes was 4-fold higher than among patients with well-controlled diabetes.

    Meaning  Poor long-term glycemic control was associated with increased risk of invasive GBS infections.

    Abstract

    Importance  The incidence of invasive infections caused by group B Streptococcus (GBS) continues to increase in the United States. Although diabetes is a key risk factor for invasive GBS, the influence of long-term glycemic control is not well characterized; other risk factors and mortality rates associated with specific types of invasive GBS infections are unknown.

    Objective  To investigate risk factors and mortality rates associated with specific invasive GBS infectious syndromes.

    Design, Setting, and Participants  This cohort study used US Veterans Health Administration data to assess active users of the Veterans Affairs health care system between January 1, 2008, and December 31, 2017. Data analysis was conducted from April 2018 to August 2019.

    Exposures  Invasive GBS infections.

    Main Outcomes and Measures  The specific types of infectious syndromes and comorbid conditions among patients with an invasive GBS infection were evaluated.

    Results  Between 2008 and 2017, 5175 patients in the Veterans Affairs health care system experienced 5497 invasive GBS infections (5027 [97.1%] men, 3737 [72.2%] white, and 4545 [87.8%] non-Latino); all-cause 30-day mortality was 8.7% (451 of 5175). The most frequently observed infections were osteomyelitis (1171 [21.3%]), bacteremia without focus (1009 [18.4%]), skin or soft-tissue infections (919 [16.7%]), and pneumonia or empyema (694 [12.6%]). All-cause 30-day mortality following the index infection for each patient was highest among patients with peritonitis (38 of 138 [27.5%]) and pneumonia or empyema (116 of 664 [17.5%]) and lowest among those with osteomyelitis (15 of 1075 [1.4%]) or joint infection (17 of 501 [3.4%]). The most common comorbid conditions among patients with invasive GBS infections were diabetes (3364 [65.0%]), obesity (2669 [51.6%]), and chronic heart conditions (1633 [31.6%]). From 2008 to 2017, the incidence of invasive GBS infections increased from 9.23 to 11.67 cases per 100 000 person-years (P = .049). Stratification by body mass index showed the highest incidence of infections among patients at the extremes of body mass index (body mass index <18.5, 25.1 cases per 100 000 person-years; body mass index ≥40, 31.0 cases per 100 000 person-years). The incidence among patients with diabetes and poor long-term glycemic control (ie, hemoglobin A1c ≥9.5%) was 4-fold greater than among patients with diabetes and good glycemic control (ie, hemoglobin A1c <7.5%) (78.3 cases vs 19.0 cases per 100 000 person-years; unadjusted incident rate ratio, 4.1; 95% CI, 3.7-4.4; P < .001).

    Conclusions and Relevance  In this cohort study, comorbid conditions and 30-day mortality varied among types of invasive GBS infections. Obesity and poor long-term glycemic control were associated with invasive GBS infections. Efforts to reverse the trend of an increased incidence of invasive GBS infections should continue to support reducing obesity and focus on improving glycemic control.

    Introduction

    The incidence of invasive group B Streptococcus (GBS) infections among nonpregnant adults in the United States tripled between 1990 and 2016.1,2 In the last 3 decades, analyses of population-based data from the US Centers for Disease Control and Prevention Active Bacterial Core surveillance (ABCs) program has consistently identified older age and diabetes mellitus as risk factors for invasive GBS infections, with more recent data indicating an association with obesity as well.1-4 These findings raise additional questions about risk factors and mortality rates associated with specific types of GBS infections. Furthermore, the association of long-term glycemic control among individuals with diabetes with the risk of developing an invasive GBS infection has not been characterized. We postulated that underlying comorbid conditions and outcomes varied according to the type of invasive GBS infection and that the risk of invasive GBS infections was higher among individuals with poor long-term glycemic control and those with extreme obesity. To explore these questions, we used microbiological, clinical, and administrative databases from the US Veterans Health Administration (VHA) to investigate risk factors and mortality associated with different types of invasive GBS infectious syndromes among users of the Veterans Affairs (VA) health care system.

    Methods
    Study Design and Data Sources

    The institutional review board at the VA Northeast Ohio Healthcare System approved the study protocol and granted a waiver of informed consent because the research was no more than minimal risk and the waiver would not adversely affect the participants’ rights and welfare. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.

    We conducted a cohort study of patients treated at the VA health care system from January 1, 2008, through December 31, 2017, using the VA Informatics and Computing Infrastructure to access the VHA Corporate Data Warehouse (CDW). Data were extracted from the CDW on April 20, 2018. Cultures growing GBS and the anatomic site of those cultures were identified by searching microbiology tables in the CDW. The following data were also obtained from the CDW: patient demographic characteristics, including height and weight, laboratory values, and, using codes from the International Classification of Diseases, Ninth Revision (ICD-9) and ICD-10, infectious diagnoses and comorbid conditions. Data indicating date of death and Medicare enrollment among VA health care users were extracted from the from the VHA Vital Status File on June 25, 2019.

    Case Ascertainment and Clinical Characteristics

    Inclusion criteria were VA health care users with at least 1 case of invasive GBS infection, defined as a culture from a normally sterile site that grew GBS (ie, blood, bone, or cerebrospinal, synovial, pleural, and peritoneal fluids).5 Exclusion criteria were patients who were not active users of the VA health care system in the year of the positive culture. Active use was defined as a hospital admission or an outpatient primary or specialty care clinic visit in that year. Among patients with multiple cultures consistent with an invasive GBS infection, a culture was considered a separate case if it occurred at least 30 days after the incident case. Separate cases that occurred in the same patient were considered recurrences.

    Only incident cases were included when analyzing age, sex, self-reported race/ethnicity, comorbid conditions, all-cause mortality at 30 days and 1 year, the number of recurrent cases, Charlson Comorbidity Index (CCI) score, body mass index (BMI, calculated as weight in kilograms divided by height in meters squared), percentage of glycated hemoglobin or hemoglobin A1c (HbA1c), and type of infectious syndrome. We determined the CCI score using ICD codes.6 To calculate BMI, we used the first height and weight measurements within the same calendar year as the infection; if these were not available, we used the first BMI from a previous calendar year, using a last-observed, carry-forward logic for missing values. For HbA1c percentage among patients with an ICD code for diabetes, we similarly used the first value within the same calendar year as the infection; if the HbA1c percentage was not present during the year, it was deemed a missing value.4

    To determine the types of invasive GBS infection, we considered the culture site in conjunction with the ICD codes associated with each case. Culture sites informed a preliminary classification of cases according to the following hierarchy: cerebrospinal fluid, heart-related, pleural fluid, synovial fluid, bursal fluid, bone, intra-abdominal fluid, and blood. We used ICD codes associated with each case to confirm the specific invasive GBS infection syndromes (eTable 1 in the Supplement). We considered outpatient ICD codes entered 7 days before to 30 days after a GBS culture to be associated with the given culture. Similarly, we considered inpatient ICD codes entered for an admission that included a GBS culture or began within 30 days of a GBS culture to be associated with that culture. We also assessed if the cases were polymicrobial, defined as isolation of bacteria different from GBS in the same sample that grew GBS. Specifically, we recorded the presence of Staphylococcus aureus, Pseudomonas aeruginosa, or other bacteria.

    Cases of invasive GBS infection in which the anatomical culture site was consistent with the clinical syndrome were considered concordant (eg, cultures from bone or peritoneal fluid were concordant with an ICD code for osteomyelitis or intra-abdominal infection, respectively). For cases in which only blood cultures were positive for GBS, we searched ICD codes to identify associated infectious syndromes. If the cultures and ICD codes did not indicate an infectious syndrome other than bacteremia, the case was designated bacteremia without focus. We classified 2 types of invasive GBS infection cases as discordant: (1) cases identified by a GBS culture but without any ICD code indicating an infectious syndrome and (2) cases in which the site of the GBS culture was not consistent with the infectious syndrome indicated by an associated ICD code (eg, a patient with a bone culture growing GBS and an ICD code for an unrelated infection, such as pneumonia).

    Statistical Analysis

    To determine rates of invasive GBS infection among VA patients, we considered all cases (ie, concordant, discordant, incident, and recurrent) and calculated person-years at risk from the cohort of active VA health care users. For each calendar year, each patient who received inpatient or outpatient clinical care in the VA contributed 1 person-year at risk if they were alive at the end of the year; if they died in the given year, they contributed the fraction of the year during which they were alive. The determination of active users for the denominator was made for each year of the study. Overall and age-stratified incidence rates with 95% CIs were calculated using Poisson methods. We performed similar calculations including only VA health care users who were not enrolled in Medicare and compared them with estimates of incidence for different age strata from ABCs data.5 Additionally, incidence rates were calculated and compared across strata of BMI including less than 18.5 (ie, underweight) and 40 and greater (ie, extreme obesity). Similarly, incidence rates were calculated for patients without diabetes and for patients with diabetes with different levels of long-term glycemic control according to HbA1c percentage, as follows: less than 7.5% (ie, well controlled), 7.5% to 9.4%, and 9.5% or higher (ie, poorly controlled). We used Poisson regression models to estimate the incidence rate ratios of invasive GBS infection in different age and BMI strata as well as among patients with diabetes and, among patients with diabetes, in different long-term glycemic control strata. Univariate and multivariable models were estimated using the stratum of age, BMI, or glycemic control with the lowest crude incident rate as the reference level. Crude incidence rates across combinations of age, BMI, and HbA1c percentage groups were graphically displayed. Kaplan-Meier survival curves for the year following incident infection were estimated by type of invasive GBS infection; the survival curves of the most common syndromes were compared using a log-rank test. The Mann-Kendall test was used to assess trends for incidence rates over time, with a 2-sided P < .05 considered statistically significant. All statistical analyses were performed using R version 3.5.1 (R Project for Statistical Computing) including functions from the epitools, trend, ICD, and survival packages.

    Results
    Clinical Characteristics

    Between 2008 and 2017, GBS caused 5497 invasive infections among 5175 VA health care users. Most patients were male (5027 [97.1%]), white individuals (3737 [72.2%]), and non-Latino individuals (4545 [87.8%]) (Table 1). The most common invasive GBS infection was osteomyelitis (1171 [21.3%]), followed by bacteremia without focus (1009 [18.4%]), skin or soft-tissue infections (919 [16.7%]), and pneumonia or empyema (694 [12.6%]), which together accounted for nearly 70% of incident cases of invasive GBS infection. An additional 418 patients (7.6%) had a culture indicating an invasive GBS infection without a concordant ICD code. Compared with patients with a concordant index case, discordant cases had significantly higher rates of mortality (73 [17.5%] vs 378 [7.9%]; P < .001) and malignant neoplasm (114 [27.3%] vs 1034 [21.7%]; P = .005) and lower rates of diabetes (215 [51.4%] vs 3149 [66.1%]; P < .001) and obesity (171 [40.9%] vs 2498 [52.4%]; P < .001).

    Further assessment of patients with invasive GBS infections yielded insights into differences in underlying conditions and outcomes associated with each syndrome (Table 2). The most common comorbid conditions among patients with invasive GBS infections were diabetes (3364 [65.0%]), obesity (2669 [51.6%]), and chronic heart conditions (1633 [31.6%]). Those with osteomyelitis were most likely to experience recurrent infection (77 of 1075 [7.2%]) and had the highest prevalence of diabetes (928 [86.3%]) and peripheral vascular disease (342 [31.8%]). Among cases of osteomyelitis, 308 (28.7%) had ICD codes specifying involvement of the foot and/or ankle (data not shown). Patients with pneumonia or empyema were older, with a mean (SD) age exceeding 70 (11.8) years, were more likely to have 4 or more comorbid medical conditions (226 of 664 [34.0%]), and had higher prevalence of chronic heart conditions (271 [40.8%]), pulmonary conditions (267 [40.2%]), cancer (197 [29.7%]), and cerebrovascular disease (153 [23.0%]). Patients with peritonitis often had underlying chronic liver disease (64 of 138 [46.4%]).

    Of 5175 cases of invasive GBS infection identified among VA users, 1400 (27.1%) were polymicrobial. Staphylococcus aureus and P aeruginosa were isolated in 566 (40.4%) and 70 (5.0%) polymicrobial cases, respectively. The proportion of polymicrobial cases varied by type of invasive GBS infection, ranging from 58.4% (628 of 1075) in cases of osteomyelitis to 15% or less in cases of meningitis (10 of 78 [12.8%]), endocarditis (54 of 392 [13.8%]), skin or soft-tissue infections (110 of 847 [13.0%]), and joint infection (76 of 501 [15.2%]).

    Mortality by Syndrome

    All-cause mortality at 30 days was 8.7% (451 of 5175) among patients with any type of invasive GBS infection; mortality was lowest among those with osteomyelitis (15 of 1075 [1.4%]) or joint infections (17 of 501 [3.4%]) and highest among patients with peritonitis (38 of 138 [27.5%]), followed by pneumonia or empyema (116 of 664 [17.5%]), bacteremia without focus (125 of 968 [12.9%]), and meningitis (10 of 78 [12.8%]). The overall 1-year mortality rate following any type of invasive GBS infection was 22.9% (1185 deaths among 5175 index cases). Survival curves of the of 4 most common invasive GBS infections showed decreased survival for those with pneumonia or empyema and bacteremia without focus, particularly in the first 90 days after infection (90 day mortality, 25.3% [168 of 664] and 19.1% [185 of 968], respectively). Mortality at 90 days was 3.3% (36 of 1075) among patients with osteomyelitis and 6.7% (58 of 847) among patients with skin and soft-tissue infections (Figure 1).

    Incidence Rates

    Between 2008 and 2017, the incidence of invasive GBS infections among VA health care users increased from 9.23 cases per 100 000 person-years to 11.67 cases per 100 000 person-years (P = .049). The incidence of osteomyelitis increased from 1.2 cases per 100 000 person years in 2008 to 3.3 cases per 100 000 person-years in 2017 (P < .001). The incidence of infection did not appear to increase with age, with similar incidence rates observed among all age groups older than 50 years (Figure 2). However, after excluding VA health care users who were enrolled in Medicare, the incidence of invasive GBS infection increased according to age, in a pattern similar to that of estimates for the general US population2 (eFigure 1 in the Supplement).

    Stratification by BMI showed the highest incidence of invasive GBS infections among VA health care users at the extremes of BMI (BMI <18.5, 25.1 cases per 100 000 person-years; BMI ≥40, 31.0 cases per 100 000 person-years), while those with a BMI between 18.5 and 39.9 all had a similar, lower incidence (ie, 7.6-11.0 cases per 100 000 person-years) (Figure 3A; eTable 2 in the Supplement). Patients with BMI less than 18.5, who accounted for 2.1% (116 of 5497) of events, had a greater prevalence of chronic pulmonary conditions (47 of 113 [41.6%]) and paralysis (20 of 113 [17.7%]) and a lower prevalence of diabetes (38 of 113 [33.6%]) and heart conditions (26 of 113 [23.0%]) than patients with a BMI of 18.5 or greater (data not shown).

    After aggregating cases and person-years across 10 years, the incidence of invasive GBS infections among VA health care users with diabetes and a HbA1c level of 7.5% or lower was nearly 4-fold greater than that among patients without diabetes (19.0 [95% CI, 18.0-20.0] cases per 100 000 person-years vs 4.9 [95% CI, 4.7-5.2] cases per 100 000 person-years; incidence rate ratio, 3.9; 95% CI, 3.6-4.1; P < .001) (Figure 3B; eTable 2 in the Supplement). The incidence of invasive GBS infections increased among patients with diabetes with poor glycemic control. The incidence among those with an HbA1c level of 7.5% to 9.4% was 33.2 (95% CI, 31.2-35.3) cases per 100 000 person-years, which increased to 78.3 (95% CI, 73.4-83.4) cases per 100 000 person-years among those with an HbA1c level of 9.5% and higher (P < .001 for both rates compared with those with HbA1c ≤7.5%). Compared with patients with well-controlled diabetes, those with an HbA1c level of 9.5% or higher had an unadjusted incident rate ratio of 4.1 (95% CI, 3.4-4.4; P < .001).

    Given that the rates of invasive GBS infections varied by age, BMI, and HbA1c percentage, we calculated and plotted the rates across combinations of those variables to visualize their interactions (Figure 3C-E). Patients who were underweight (ie, BMI <18.5) were excluded. Among patients without diabetes, the rate of invasive GBS ranged from 0.9 (95% CI, 0.6-1.3) cases per 100 000 patient-years among patients aged 18 to 49 years with obesity to 29.0 (95% CI, 18.4-43.5) cases per 100 000 patient-years among patients aged 75 years or older with extreme obesity. Among patients with an HbA1c level of at least 9.5%, the rates across age and BMI combinations ranged from 34.8 (95% CI, 20.6-54.9) cases per 100 000 person-years among patients aged 75 years or older with obesity to 104.9 (95% CI, 87.7-124.3) cases per 100 000 person-years among patients aged 50 to 74 years with extreme obesity. In every combination of age and BMI, the rate among patients with well-controlled diabetes was between the rate among patients with poorly controlled diabetes and patients without diabetes, ranging from 4.8 (95% CI, 2.1-9.4) cases per 100 000 person-years among patients aged 18 to 49 years with obesity to 42.2 (95% CI, 28.3-60.6) cases per 100 000 patient-years among patients aged 75 years or older with extreme obesity. Aggregating across all age and BMI groups, the rate among patients with well-controlled diabetes was significantly different from patients without diabetes and patients with poorly controlled diabetes (incidence rate ratio compared with patients without diabetes, 3.9; 95% CI, 3.6-4.1; P < .001; incidence rate ratio compared with patients with poorly controlled diabetes, 0.24; 95% CI, 0.22-0.26; P < .001). A similar pattern of increased incidence of invasive GBS infections among patients with higher HbA1c percentage levels for the most common invasive GBS syndromes is detailed in eFigure 2 in the Supplement. A multivariable Poisson regression model estimated far higher incidence rate ratios for poor glycemic control (ie, HbA1c ≥9.5%) than for advanced age or obesity (HbA1c ≥9.5%: incidence rate ratio, 13.30; 95% CI, 12.26-14.41; aged ≥85 years: incidence rate ratio, 6.99; 95% CI, 5.10-9.87; BMI ≥40: incidence rate ratio, 2.37; 95% CI, 2.17-2.59) (eTable 2 in the Supplement).

    Discussion

    In this large cohort of patients from the US VHA, the incidence of invasive GBS infection was associated with a diagnosis of diabetes and the degree of long-term glycemic control among those with diabetes. Patients with HbA1c levels of 9.5% or higher had a 4-fold increase in the risk of invasive GBS compared with patients with HbA1c levels less than 7.5%. Invasive GBS infection was also associated with BMI, with increased incidence occurring at the extremes of BMI (ie, <18.5 and ≥40). We uncovered differences in the underlying clinical characteristics and outcomes of different types of invasive GBS infection; patients who developed pneumonia or empyema were more likely to have cancer and chronic diseases of the heart, lungs, or kidneys and had a 30-day mortality of approximately 15%, comparable with that among patients with bacteremia without focus and meningitis. In contrast, patients with osteomyelitis (the most frequently observed infection in our cohort) were more likely to have diabetes and peripheral vascular disease, with a 30-day mortality rate of 1%, similar to that among patients with skin or soft-tissue infection.

    In the United States and other countries, the most common syndromes of invasive GBS infections are bacteremia without focus and skin or soft-tissue infections.2,7,8 In VHA patients, the proportion of osteomyelitis among cases of invasive GBS infections was greater than in the general US population; furthermore, the incidence of osteomyelitis more than doubled between 2008 and 2017, from 1.2 to 3.3 cases per 100 000 person-years. Based on the analysis of ICD codes, at least 28% of those cases involved the foot and/or ankle and a high proportion (86%) of these patients had diabetes. Therefore, it is likely that many of these were diabetic foot infections with osteomyelitis, an important syndrome in the epidemiology of invasive GBS infections.9 In the VHA cohort, osteomyelitis and skin or soft-tissue infections were associated with recurrent infection, similar to a previous report of relapsing invasive GBS infections in adults.10

    Invasive GBS infection is a burden for veterans, with some types of infection associated with substantial mortality. In VHA patients, the 30-day all-cause mortality rate among patients with invasive GBS infection was 8.7%. Although a different measure, this is comparable with the 6.5% case fatality of invasive GBS infection estimated by the ABCs program in the nonpregnant adult US population.2 Surprisingly, both necrotizing fasciitis and endocarditis involving GBS carried a lower risk of 30-day mortality compared with what is typically reported for these infections.11-13 In contrast, the greatest risk of mortality occurred among those with peritonitis and pneumonia or empyema. These syndromes were common in patients of advanced age and chronic diseases of the liver, heart, and lungs, contributing to their mortality. In discordant cases (where the diagnosis of infection was not captured in the electronic health record of patients with invasive GBS infection), there was a higher rate of mortality and cancer, suggesting that invasive GBS infection may have been only a single aspect of a complex and/or prolonged medical course or that patients were transferred to a non-VA hospital.

    The increased risk of invasive GBS infection among patients with diabetes is well recognized in population-based surveillance studies in the United States.1-4 According to the ABCs program, diabetes is present in 53.4% of patients with invasive GBS infection, and diabetes confers the highest adjusted relative risk in young people without obesity.4 In the VHA cohort, where 66% of patients with invasive GBS infection had diabetes, the incidence of invasive GBS infection increased with rising HbA1c percentages, a measure not available in population-based surveillance data. The increased incidence of invasive GBS infections among those with poorly controlled diabetes was consistent across age groups and patients who had normal weight, overweight, and obesity.

    Data from ABCs also indicated that obesity is independently associated with an increased risk of invasive GBS infection. Specifically, Pitts et al4 reported a 5-fold higher risk in patients with a BMI of 40 or greater compared with patients with overweight. They also detected an interaction between diabetes and BMI in which the relative risk of infection for those with a BMI of 40 or greater was higher among patients without diabetes; patients with a BMI of less than 18.5 were excluded because of small numbers.4 Among VHA patients, there was a higher incidence of invasive GBS infection in patients with a BMI of 40 or greater as well as in patients with a BMI less than 18.5. Our multivariable modeling, which included long-term glycemic control among patients with diabetes, similarly indicated that extreme obesity was independently associated with an increased risk of invasive GBS infection, with an adjusted relative risk of 2.4 compared with individuals who had overweight.

    Older adults with chronic medical conditions are known to be vulnerable to invasive GBS infection.1,2,14-18 Among VA health care users aged 18 to 64 years, rates of invasive GBS infection were similar to estimates for the US population; however, the rate of invasive GBS infections did not increase further in VA health care users aged 65 years or older. A potential explanation is that VA health care users aged 65 years or older who were enrolled in Medicare had access to health care services outside the VHA, and cases of invasive GBS infection in those patients were not captured by the VHA. When considering invasive GBS infections only among VA health care users who were not enrolled in Medicare, the rates of invasive GBS infections were similar to rates for the US population.5 Because VA health care users typically have a greater burden of chronic health conditions than their non-VA counterparts, it is unlikely that differences in baseline health status accounted for lower rates of invasive GBS infection in VA health care users aged 65 years or older.19,20 Indeed, there were comparable rates of diabetes, obesity, and other chronic medical conditions among VHA cases and the cases summarized in the ABCs reporting. While there is not yet a vaccine or protective measures specific to GBS infections for nonpregnant adults, VA health care may include preventative services, continuity of care, and earlier care of chronic illnesses that could affect GBS infection.21,22

    Limitations

    This study has limitations. First, VA health care users are predominantly white and non-Latino men and have a different sociodemographic and health status than the rest of the US population,19 which limits the generalizability of these results. Specifically, the high burden of chronic medical conditions among VA health care users may overestimate the risk that these conditions pose for developing infections.20 However, findings from the VHA cohort are consistent with population-based studies that identified comorbid conditions as risk factors for invasive GBS infections.1-4,14,23 Second, individuals who were pregnant or post partum at the time of infection were not specifically excluded; this likely had negligible impact given that the study population was 97% men and most were aged 65 years or older. Third, we relied on definitions based on ICD codes as well as site of GBS culture to determine the type of invasive GBS infection. In particular, pneumonia, a common diagnosis among hospitalized patients, may be overrepresented. Our GBS culture data were obtained from microbiology tables within the VHA CDW, which are carefully curated. Additionally, our definitions, tailored to GBS infections, were similar to those previously described24 and did not impose categories for every GBS culture as indicated by the number of discordant or unclassified cases. Furthermore, we reviewed data from 200 patients (approximately 4% of cases) to verify the GBS culture data, infectious syndrome, and other details (eg, age, comorbid conditions, mortality). Fourth, individuals with an incident case occurring before the study period and those with a recurrent case occurring after the study period may affect the measured invasive GBS infection rate. The 10-year study period as well as the overall low rate of recurrent infection indicated these influences were minimal. Fifth, neither the serotypes nor antibiotic susceptibilities of GBS isolates were reported because these are not routinely determined by microbiology laboratories serving VA medical centers.

    Conclusions

    This analysis of the epidemiology of invasive GBS infection among patients in the US VHA demonstrated that underlying characteristics and associated mortality vary among different syndromes and underscored that poor long-term glycemic control in patients with diabetes (determined by an elevated HbA1c percentage) was an important and potentially modifiable risk factor. Efforts to reverse the increasing rates of invasive GBS infections in adults should continue to address diabetes, obesity, and other chronic medical conditions. Among those with diabetes, improving glycemic control may help mitigate the risk of invasive GBS infection.

    Back to top
    Article Information

    Accepted for Publication: November 1, 2019.

    Published: December 27, 2019. doi:10.1001/jamanetworkopen.2019.18324

    Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2019 Jump RLP et al. JAMA Network Open.

    Corresponding Author: Robin L. P. Jump, MD, PhD, Geriatric Research Education and Clinical Center 111O(W), VA Northeast Ohio Healthcare System, 10701 East Blvd, Cleveland, OH 44106 (robinjump@gmail.com).

    Author Contributions: Drs Jump and Perez had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

    Concept and design: Jump, Perez.

    Acquisition, analysis, or interpretation of data: All authors.

    Drafting of the manuscript: Jump, Wilson, Banks, Perez.

    Critical revision of the manuscript for important intellectual content: Jump, Baechle, Briggs, Song, Zappernick, Perez.

    Statistical analysis: Wilson.

    Obtained funding: Jump, Perez.

    Administrative, technical, or material support: Baechle, Briggs, Banks, Song, Zappernick.

    Supervision: Jump, Perez.

    Conflict of Interest Disclosures: Dr Jump reported receiving grants from Accelerate Diagnostics and serving on the advisory boards of Pfizer and Merck outside the submitted work. Dr Perez reported receiving grants from Merck and Accelerate Diagnostics outside the submitted work. No other disclosures were reported.

    Funding/Support: Drs Jump and Perez were supported by Pfizer, and this work was supported by funds and facilities provided by the Cleveland Geriatric Research Education and Clinical Center and the Specialty Care Center of Innovation at the VA Northeast Ohio Healthcare System.

    Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

    Disclaimer: The findings and conclusions in this document are those of the authors, who are responsible for its content, and do not necessarily represent the views of the US Department of Veterans Affairs or of the US government.

    References
    1.
    Skoff  TH, Farley  MM, Petit  S,  et al.  Increasing burden of invasive group B Streptococcal disease in nonpregnant adults, 1990-2007.  Clin Infect Dis. 2009;49(1):85-92. doi:10.1086/599369PubMedGoogle ScholarCrossref
    2.
    Francois Watkins  LK, McGee  L, Schrag  SJ,  et al.  Epidemiology of invasive group B Streptococcal infections among nonpregnant adults in the United States, 2008-2016.  JAMA Intern Med. 2019;179(4):479-488. doi:10.1001/jamainternmed.2018.7269PubMedGoogle ScholarCrossref
    3.
    Farley  MM, Harvey  RC, Stull  T,  et al.  A population-based assessment of invasive disease due to group B Streptococcus in nonpregnant adults.  N Engl J Med. 1993;328(25):1807-1811. doi:10.1056/NEJM199306243282503PubMedGoogle ScholarCrossref
    4.
    Pitts  SI, Maruthur  NM, Langley  GE,  et al.  Obesity, diabetes, and the risk of invasive group B Streptococcal disease in nonpregnant adults in the United States.  Open Forum Infect Dis. 2018;5(6):ofy030. doi:10.1093/ofid/ofy030PubMedGoogle Scholar
    5.
    US Centers for Disease Control and Prevention. Active Bacterial Core surveillance. https://www.cdc.gov/abcs/reports-findings/surv-reports.html. Accessed July 11, 2019.
    6.
    Quan  H, Sundararajan  V, Halfon  P,  et al.  Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data.  Med Care. 2005;43(11):1130-1139. doi:10.1097/01.mlr.0000182534.19832.83PubMedGoogle ScholarCrossref
    7.
    Tazi  A, Morand  PC, Réglier-Poupet  H,  et al.  Invasive group B Streptococcal infections in adults, France (2007-2010).  Clin Microbiol Infect. 2011;17(10):1587-1589. doi:10.1111/j.1469-0691.2011.03628.xPubMedGoogle ScholarCrossref
    8.
    Morozumi  M, Wajima  T, Takata  M, Iwata  S, Ubukata  K.  Molecular characteristics of group B Streptococci isolated from adults with invasive infections in Japan.  J Clin Microbiol. 2016;54(11):2695-2700. doi:10.1128/JCM.01183-16PubMedGoogle ScholarCrossref
    9.
    Camuset  G, Picot  S, Jaubert  J,  et al.  Invasive group B Streptococcal disease in non-pregnant adults, Réunion Island, 2011.  Int J Infect Dis. 2015;35:46-50. doi:10.1016/j.ijid.2015.04.006PubMedGoogle ScholarCrossref
    10.
    Harrison  LH, Ali  A, Dwyer  DM,  et al.  Relapsing invasive group B Streptococcal infection in adults.  Ann Intern Med. 1995;123(6):421-427. doi:10.7326/0003-4819-123-6-199509150-00004PubMedGoogle ScholarCrossref
    11.
    Huang  K-F, Hung  M-H, Lin  Y-S,  et al.  Independent predictors of mortality for necrotizing fasciitis: a retrospective analysis in a single institution.  J Trauma. 2011;71(2):467-473. doi:10.1097/TA.0b013e318220d7faPubMedGoogle ScholarCrossref
    12.
    Hadeed  GJ, Smith  J, O’Keeffe  T,  et al.  Early surgical intervention and its impact on patients presenting with necrotizing soft tissue infections: a single academic center experience.  J Emerg Trauma Shock. 2016;9(1):22-27. doi:10.4103/0974-2700.173868PubMedGoogle ScholarCrossref
    13.
    Thuny  F, Giorgi  R, Habachi  R,  et al.  Excess mortality and morbidity in patients surviving infective endocarditis.  Am Heart J. 2012;164(1):94-101. doi:10.1016/j.ahj.2012.04.003PubMedGoogle ScholarCrossref
    14.
    Ballard  MS, Schønheyder  HC, Knudsen  JD,  et al; International Bacteremia Surveillance Collaborative.  The changing epidemiology of group B Streptococcus bloodstream infection: a multi-national population-based assessment.  Infect Dis (Lond). 2016;48(5):386-391. doi:10.3109/23744235.2015.1131330PubMedGoogle ScholarCrossref
    15.
    Lambertsen  L, Ekelund  K, Skovsted  IC, Liboriussen  A, Slotved  H-C.  Characterisation of invasive group B Streptococci from adults in Denmark, 1999 to 2004.  Eur J Clin Microbiol Infect Dis. 2010;29(9):1071-1077. doi:10.1007/s10096-010-0941-zPubMedGoogle ScholarCrossref
    16.
    Bergseng  H, Rygg  M, Bevanger  L, Bergh  K.  Invasive group B Streptococcus (GBS) disease in Norway, 1996-2006.  Eur J Clin Microbiol Infect Dis. 2008;27(12):1193-1199. doi:10.1007/s10096-008-0565-8PubMedGoogle ScholarCrossref
    17.
    Lamagni  TL, Keshishian  C, Efstratiou  A,  et al.  Emerging trends in the epidemiology of invasive group B Streptococcal disease in England and Wales, 1991-2010.  Clin Infect Dis. 2013;57(5):682-688. doi:10.1093/cid/cit337PubMedGoogle ScholarCrossref
    18.
    Abat  C, Chaudet  H, Raoult  D, Colson  P.  Increasing trend of invasive group B Streptococcal infections, Marseille, France.  Clin Infect Dis. 2014;58(5):750-751. doi:10.1093/cid/cit777PubMedGoogle ScholarCrossref
    19.
    Agha  Z, Lofgren  RP, VanRuiswyk  JV, Layde  PM.  Are patients at Veterans Affairs medical centers sicker? a comparative analysis of health status and medical resource use.  Arch Intern Med. 2000;160(21):3252-3257. doi:10.1001/archinte.160.21.3252PubMedGoogle ScholarCrossref
    20.
    Dursa  EK, Barth  SK, Bossarte  RM, Schneiderman  AI.  Demographic, military, and health characteristics of VA health care users and nonusers who served in or during Operation Enduring Freedom or Operation Iraqi Freedom, 2009-2011.  Public Health Rep. 2016;131(6):839-843. doi:10.1177/0033354916676279PubMedGoogle ScholarCrossref
    21.
    Keyhani  S, Ross  JS, Hebert  P, Dellenbaugh  C, Penrod  JD, Siu  AL.  Use of preventive care by elderly male veterans receiving care through the Veterans Health Administration, Medicare fee-for-service, and Medicare HMO plans.  Am J Public Health. 2007;97(12):2179-2185. doi:10.2105/AJPH.2007.114934PubMedGoogle ScholarCrossref
    22.
    Landrum  MB, Keating  NL, Lamont  EB,  et al.  Survival of older patients with cancer in the Veterans Health Administration versus fee-for-service Medicare.  J Clin Oncol. 2012;30(10):1072-1079. doi:10.1200/JCO.2011.35.6758PubMedGoogle ScholarCrossref
    23.
    Kothari  NJ, Morin  CA, Glennen  A,  et al.  Invasive group B Streptococcal disease in the elderly, Minnesota, USA, 2003-2007.  Emerg Infect Dis. 2009;15(8):1279-1281. doi:10.3201/eid1508.081381PubMedGoogle ScholarCrossref
    24.
    Jones  BE, Jones  MM, Huttner  B,  et al.  Trends in antibiotic use and nosocomial pathogens in hospitalized veterans with pneumonia at 128 medical centers, 2006-2010.  Clin Infect Dis. 2015;61(9):1403-1410. doi:10.1093/cid/civ629PubMedGoogle ScholarCrossref
    ×