In this nationwide cohort of VA patients, we analyzed over 1250 cases of osteomyelitis caused by GBS. Of these, 40% were identified using an expanded case definition that included a diagnosis code for osteomyelitis and isolation of GBS from cultures of anatomically relevant non-sterile sites. These cases would not have been captured by the case definitions conventionally used to conduct population-based surveillance of invasive GBS disease, including GBS osteomyelitis. By incorporating cases identified through our expanded osteomyelitis criteria that included GBS recovered from non-sterile sites, we demonstrated that the disease burden caused by GBS osteomyelitis appears to be greater than previously reported. recognized.
As noted by the US CDC as well as in our previous assessment of a cohort of US veterans, osteomyelitis is one of the most common invasive infections caused by GBS in non-pregnant adults. [3, 5]. A study of 3 large counties in California found that between 1995 and 2021, the rate of GBS osteomyelitis increased from 0.7 to 2.4 cases per 100,000 population . These studies all used well-established definitions of invasive disease, requiring either positive blood cultures or blood cultures to diagnose osteomyelitis. In clinical practice, however, the diagnosis of osteomyelitis may require serological, imaging, and microbiological investigations, all of which must be interpreted in the context of each patient. For this study, we applied less stringent microbiological criteria, combined with diagnostic and procedural (diagnostic imaging) codes available in clinical and administrative databases, to allow recognition of probable GBS osteomyelitis. These are aligned with the criteria used by Kremers et al.. in their description of the epidemiology of osteomyelitis in Olmsted County, Minnesota, in which they characterized cases as possible, probable, and certain osteomyelitis based on the availability of pathological and microbiological data as well as the quality of the sample taken . The inclusion of an expanded definition considering cultures from non-sterile sites, considered here as probable GBS osteomyelitis, nearly doubled the number of cases in our cohort, with little clinical difference between those considered to have GBS. Definite and probable GBS osteomyelitis. Even if we only consider cases of monomicrobial GBS osteomyelitis, where the pathogenic role of GBS is less debatable, we obtain 20% more cases using the expanded definition. A population-based study that assessed hospitalizations due to invasive and noninvasive GBS infections in Louisville, Ky., also confirmed osteomyelitis to be a common presentation of GBS infection and found disease rates non-invasive 3 to 4 times higher than those for invasive disease; this study was based on clinical and microbiological criteria different from those applied here . These observations, together with our results, suggest that the strict case definitions used to assess invasive GBS disease in population surveillance studies may miss a notable proportion of GBS osteomyelitis cases.
GBS osteomyelitis cases identified using the strict invasive disease case definitions (classified as definitive GBS osteomyelitis) had similar demographic and clinical characteristics to cases identified with GBS using our expanded criteria (probable GBS osteomyelitis). As in previous epidemiological studies of invasive GBS infection, obesity and especially diabetes mellitus were prevalent in this cohort of patients with GBS osteomyelitis. [4, 16]. Diabetes mellitus that was not well controlled (HbA1c >7.5%) occurred more frequently in people with definitive GBS osteomyelitis, consistent with previous reports that describe elevated HbA1c as an important risk factor for invasive GBS disease (5). A high proportion of cases also had peripheral vascular disease. Taken together, these results suggest that diabetic foot infections contributed to a large proportion of GBS osteomyelitis cases in this cohort.
The serious consequences of GBS osteomyelitis in this cohort of American veterans deserve comment. Short-term mortality, even for cases of invasive GBS infection, appeared low. This may be a function of the sensitivity of GBS to most empiric treatment regimens in patients with suspected osteomyelitis. It should be noted, however, that one in 10 patients with GBS osteomyelitis dies within one year, which may reflect the burden of comorbidities in this patient group, and mortality was similar in patients with GBS. definite versus probable GBS osteomyelitis. In contrast, among lower limb GBS osteomyelitis cases, those considered to have definitive osteomyelitis were more likely to undergo amputation. This difference may result from selection bias, as patients who underwent amputation were also more likely to have a bone sample available for culture and isolation of GBS, thus fulfilling the criteria for a invasive disease.
The large proportion of cases considered to have probable GBS osteomyelitis who had polymicrobial infection may also reflect differences in the types of specimens taken, which included non-sterile sites like wounds and ulcers. For these samples, the microbiology laboratory may choose to report GBS only if it grows in significant amounts, while any presence of S. aureus and P. aeruginosa is usually reported. If so, GBS can probably be considered a pathogen, giving credence to polymicrobial osteomyelitis. Alternatively, the microbiology laboratory may report any presence of GBS independently of the presence of other pathogens, in which case the pathogenic role of GBS is more difficult to discern. Unfortunately, in this study, we were unable to assess the protocols and practices of the microbiology laboratories that processed these samples.
Cultures taken from patients with established osteomyelitis were significantly less likely to have other organisms, including S. aureus and P. aeruginosa. Nevertheless, these pathogens have been found in up to a quarter of cases defined by isolation of GBS from a sterile site (that is to say, bones and blood). In this scenario, it is difficult to establish which is the predominant pathogen, as they may exhibit synergistic interactions that enhance their colonization, virulence, and persistence. . Mortality and lower limb amputation rates in patients whose cultures were monomicrobial for GBS were similar to those in the entire cohort. These results suggest that GBS has as much pathogenic potential as other bacterial species commonly recognized in association with osteomyelitis, such as S. aureus. Like most antibiotic regimens that treat S. aureus osteomyelitis are effective against GBS, the risk of inadvertently selecting an inappropriate treatment for GBS is low .
Our study has several important limitations. First, the results of this study may not be generalizable to other populations due to the different sociodemographic and health patterns that exist in the VA population. Specifically, VA patients are predominantly non-Latin white males with a high chronic disease burden. . Second, the definitions used in this study were based on clinical and administrative data, including ICD codes, CPT codes for amputation and imaging studies, and GBS culture site identification; ambiguity in labeling and coding errors may have led to misclassification of some cases. Third, we did not review the results of imaging studies to confirm the diagnosis of osteomyelitis because radiological interpretations are not readily available through the CDW, which may overestimate the number of cases of probable GBS osteomyelitis. Fourth, there is a risk of misclassification of GBS cases identified from abscesses arising in a sterile site or from exposed bones that are no longer sterile, and cases with discrepancies between blood and bone cultures. Finally, practices for isolating and reporting GBS in clinical specimens, particularly in polymicrobial cultures from non-sterile sites, may not be uniform across different microbiology laboratories and may have changed over the study period. . Likewise, this retrospective study did not report serotypes or antibiotic susceptibilities of GBS isolates because these are not routinely determined in microbiology laboratories serving VA medical centers.