Detection of Penicillin Nonsusceptible Streptococcus agalactiae by Laboratories That Participate in the College of American Pathologist’s Proficiency Testing Program

Report

Streptococcus agalactiae (i.e., group B Streptococcus) is a major pathogen among neonates, pregnant women (1, 2) and nonpregnant adults with comorbidities such as diabetes or obesity (3). Susceptibility to the penicillins remains nearly universal, and beta-lactams are the primary agents for both intrapartum prophylaxis and treatment of S. agalactiae infection in all ages and for intrapartum prophylaxis. Rare isolates of S. agalactiae with measured MICs at or above the Clinical and Laboratory Standards Institute (CLSI) susceptible breakpoint of ≤0.12 μg/mL (penicillin) and ≤0.25 μg/mL (ampicillin) have been identified (4–6). Reference broth microdilution MIC is the gold standard method for evaluating antimicrobial susceptibility against S. agalactiae, but clinical laboratories rarely use this method as it is labor-intensive (7). The ability of clinical laboratories to detect penicillin/ampicillin nonsusceptibility among S. agalactiae is largely unknown as, to our knowledge, no broad assessment of clinical laboratories’ ability to detect a penicillin nonsusceptible isolate has been conducted.

An isolate of S. agalactiae recovered from an 85-year-old woman with community onset pneumonia in Japan was sent to laboratories that participate in the “D” Bacteriology Proficiency Testing Survey in 2021 by the College of American Pathologists (CAP). The isolate was sent as a simulated vaginal-rectal swab with an accompanying clinical vignette of a 27-year-old pregnant woman who was allergic to penicillin. Laboratories were asked to identify the principal pathogen and perform antimicrobial susceptibility testing (AST). The S. agalactiae isolate had been determined to have a penicillin MIC of 0.25 μg/mL by reference broth microdilution performed according to CLSI standards (5) by 2 laboratories prior to sending, and confirmed by a third laboratory after survey results were received and analyzed. The isolate was Sequence Type 1, as determined by whole genome sequencing, and harbored the Penicillin Binding Protein (PBP)2x gene mutation Q557E, which is known to be associated with decreased susceptibility to the beta-lactams (8). Given the absence of intermediate or resistant breakpoints for penicillin and S. agalactiae (5), laboratories were instructed to report isolates that were nonsusceptible as resistant for the purposes of the survey. The isolate was resistant to erythromycin, clindamycin, and levofloxacin.

Seven hundred and thirty-seven laboratories reported penicillin susceptibility testing results in the survey. Commercial automated systems were the most common test method (73.1%) (Table 1) and disk diffusion accounted for 18.7% of tests performed by participating laboratories. Only 10.4% of all laboratories correctly reported the isolate as nonsusceptible (resistant) to penicillin. Resistance was found by 42.9% of Sensititre users (ThermoFisher), 36.1% of gradient diffusion users, and 33.3% of Vitek users (bioMérieux). Interestingly, CLSI standardized methods performed by the clinical laboratories, including agar dilution (n = 1 laboratory, 0% identified resistance), disk diffusion (n = 138 laboratories, 2.9% identified resistance), and broth microdilution (n = 2 laboratories, 0% identified resistance) did not correctly identify elevated penicillin MICs in this isolate.

TABLE 1.

Penicillin susceptibility results for laboratories participating in the D Survey, 2021, for S. agalactiae

Test method	N (%) laboratories	% of laboratories reporting as
		Sa	Ib	Rc	No result
Disk diffusion	138 (18.7%)	94.9	1.4	2.9	0.7
Agar dilution	1 (0.1%)	100	0.0	0.0	0.0
Broth microdilution	2 (0.3%)	100	0.0	0.0	0.0
Gradient diffusion	36- (4.9%)	55.6	5.6	36.1	2.8
Phoenix (BD, Sparks MD)	89 (12.1%)	100	0.0	0.0	0.0
Microscan (Beckman Coulter, Sacramento CA)	216 (29.3%)	94.9	0.0	4.2	0.9
Sensititre (ThermoFisher, Lenexa KS)	7 (0.9%)	57.1	0.0	42.9	0.0
Vitek (bioMerieux, Durham, NC)	9 (1.2%)	55.6	0.0	33.3	11.1
Vitek2 (bioMerieux, Durham, NC)	218 (29.6%)	68.8	2.3	19.3	9.6
Other / Not Specified	21 (2.8%)	85.7	0.0	14.3	0.0
TOTAL	737 (100.0%)	84.8	1.2	10.4	3.5

^aS, susceptible.

^bI, intermediate.

^cR, resistant.

Other beta-lactams (ampicillin, cefotaxime, and ceftriaxone) were susceptible for this isolate, with MICs of 0.25 μg/mL, as measured by broth microdilution. Six hundred and forty-five laboratories reported ampicillin results, all of which were susceptible except 1 laboratory (Vitek2 user) that reported the isolate as intermediate, despite the lack of intermediate category for ampicillin and S. agalactiae. In addition, 390 laboratories reported results for cefotaxime and 693 for ceftriaxone. Two laboratories reported intermediate (despite no intermediate category) and 1 laboratory resistant results for cefotaxime (all Vitek2 users). Two laboratories reported resistant results to ceftriaxone (1 disk diffusion user and 1 unspecified MIC method).

Penicillin nonsusceptibility among S. agalactiae remains rare in most geographic areas. Roughly 6% of clinical isolates in Japan show reduced susceptibility to penicillin by CLSI breakpoints (9), a phenotype that is associated with amino acid substitutions in PBP 2X (10, 11). In addition, penicillin nonsusceptible S. agalactiae have been reported globally at much lower rates. A recent evaluation of the U.S. Centers for Disease Control and Prevention Active Bacterial Core surveillance in 8 states between 2015 and 2017 showed 0.87% (n = 54) of invasive S. agalactiae were penicillin nonsusceptible, with point mutations in the pbp2x gene (12). Similarly, a follow up study showed between 0 to 0.4% of isolates were beta-lactam nonsusceptible by participating US State, with an increasing proportion of strains with first-step pbp2x mutations associated with reduced susceptibility (6). Most penicillin nonsusceptible S. agalactiae isolates reported in the literature are recovered from elderly adults, although the serotype III/ST109, penicillin nonsusceptible lineage of S. agalactiae has been emerging among young infants with invasive infections in Mozambique (13). These isolates are typically resistant to macrolides and fluoroquinolones (as was the challenge isolate in this survey), leaving few treatment options (14).

The absence of a resistant breakpoint for penicillin relates to the rare occurrence of isolates with MICs >0.12 μg/mL at the time clinical breakpoints were established by CLSI. In contrast, EUCAST has defined resistance for S. agalactiae as >0.25 mg/L (indications other than meningitis) and >0.125 mg/L (meningitis) (www.eucast.org). Little clinical data are available to evaluate the impact of penicillin MICs >0.12 μg/mL, although pharmacokinetics alone suggest concern for the ability to achieve sufficient exposure of penicillin to such isolates in the central nervous system (such as neonatal meningitis), as penicillin penetration across the blood-brain barrier is restricted, even in the presence of inflammation (15).

The MICs for penicillin nonsusceptible S. agalactiae are near the susceptible breakpoint, typically 0.25 to 0.5 μg/mL, meaning that misclassification of these isolates is not infrequent due to the inherent imprecision of AST, even by reference methods. A study conducted in 2013 showed 13 of 28 (46.4%) penicillin nonsusceptible S. agalactiae tested repeatedly susceptible by the Vitek2 (bioMérieux). Similarly, a recent study evaluated penicillin disk diffusion (using a 10U penicillin disk) and found high rates of very major errors (100% by current CLSI nonsusceptible breakpoint of <24 mm). In contrast, use of 1 μg oxacillin, 30 μg ceftizoxime, or 30 μg ceftibuten disk as a surrogate for penicillin were associated with 95.5%, 92.2%, and 94.1% categorical agreement with reference agar dilution MIC, respectively (16). Others have suggested use of a 128 μg/mL ceftibuten agar screen for the detection of penicillin nonsusceptible S. agalactiae, although this method is associated with specificity of only 81.6% (17).

This report documents the poor ability of laboratories that participate in the CAP D Survey to detect penicillin nonsusceptibility in S. agalactiae. Strengths of this study include a broad cross-section of test methods and U.S. and international laboratories. We cannot exclude the possibility that some laboratories edit all S. agalactiae penicillin results to susceptible, regardless of MIC results, as CLSI guidance indicates that testing does not need to be routinely performed because of the extreme rare incidence of nonsusceptible isolates (5) at this time. Similarly, we cannot exclude clerical errors in reporting that may have occurred when laboratories filled out proficiency test report forms. Laboratories should be aware that penicillin nonsusceptibility to S. agalactiae is increasing in some regions and should send nonsusceptible isolates to a reference laboratory for confirmation of nonsusceptible results. In addition, CLSI could consider reevaluation of disk diffusion breakpoints for penicillin, as only 2.9% of laboratories using disk diffusion correctly identified nonsusceptibility in this isolate. Finally, laboratories should partner with patient-facing clinical colleagues to be vigilant for possible penicillin treatment failures of invasive S. agalactiae infections, even if the isolate is reported as penicillin-susceptible. Continued monitoring of penicillin nonsusceptibility through genomic testing of isolates collected in surveillance programs are clearly critical to detecting increases in the emergence of penicillin nonsusceptibility given the challenges with phenotypic susceptibility testing for S. agalactiae at this date.