Antimicrobial Susceptibility Patterns among Viridans Group Streptococcal Isolates from Infective Endocarditis Patients from 1971 to 1986 and 1994 to 2002

Abstract

To determine whether changes in antimicrobial resistance have occurred among viridans group streptococci, we retrospectively examined 50 viridans group streptococcal isolates recovered from patients with infective endocarditis over 3 decades. Resistance rates (percent resistant isolates 1971 to 1986 and 1994 to 2002) were as follows: levofloxacin, 0 and 9; penicillin and clindamycin, 0 and 4; and erythromycin and azithromycin, 11 and 26, respectively.

Viridans group streptococci cause 20 to 40% of breakthrough bacteremia episodes in neutropenic patients (2, 4, 9, 14, 21) and 30 to 40% of native valve infective endocarditis cases (1). In hematology patients, fluoroquinolones are used as antibacterial prophylaxis agents during the neutropenic period after hematopoietic stem cell transplantation or chemotherapy alone (24). Many institutions have reported episodes of breakthrough bacteremia with fluoroquinolone-resistant viridans group streptococci during fluoroquinolone prophylaxis (5, 10). Previously, our group reported a 16% rate of breakthrough bacteremia with fluoroquinolone-resistant viridans group streptococci during levofloxacin prophylaxis in hematology patients undergoing autologous stem cell transplantation (24). In recent years, several case reports of infective endocarditis due to penicillin-resistant viridans group streptococci have been published (15-18, 26). In light of these reports, we retrospectively examined antimicrobial susceptibility patterns, including fluoroquinolone susceptibility, in a collection of viridans group streptococcal isolates spanning 3 decades, to see if changes in antimicrobial resistance had occurred.

Since 1971, selected blood isolates from bacterial endocarditis patients have been collected and stored at −70°C in the Infectious Diseases Research Laboratory as part of the Mayo Clinic Endocarditis Registry. These organisms have clinical relevance from the antimicrobial susceptibility perspective because they are of proven pathogenic potential in immunocompetent hosts. Twenty-eight isolates collected from the years 1971 to 1986 and 24 isolates collected from the years 1994 to 2002 were studied. These two periods represent times before and after the U.S. Food and Drug Administration approval of ciprofloxacin in 1987. MICs were determined by broth microdilution in cation-adjusted Mueller-Hinton broth supplemented with 2.5% lysed horse blood and interpreted in accordance with National Committee for Clinical Laboratory Standards (NCCLS) guidelines (19, 20). Gemella morbillorum ATCC 27825 was used as a quality control strain. The concentration ranges tested were 0.125 to 128 μg/ml (in doubling dilutions) for penicillin (Sigma-Aldrich Co., St. Louis, Mo.), erythromycin (USP, Rockville, Md.), azithromycin (Pfizer, Groton, Conn.), clindamycin (Sigma-Aldrich Co.), vancomycin (Sigma-Aldrich Co.), levofloxacin (Pharmaceutical Research Institute, Spring House, Pa.), gatifloxacin (Bristol-Myers Squibb, Plainsboro, N.J.), ciprofloxacin (USP), moxifloxacin (Bayer Corporation Pharmaceutical Division, West Haven, Conn.), and garenoxacin (BMS-284756; Bristol-Myers Squibb). The NCCLS has not published susceptibility breakpoints for the last four quinolones. Due to the initial identification date of many of the isolates, they were biochemically reidentified and classified into one of five viridans streptococcal species groups-mitis, anginosus, mutans, salivarius, or sanguinis—according to criteria set forth by Facklam (11). For isolates where identification was unclear following biochemical testing, 16S rRNA sequence analysis was performed (24); one isolate from 1971 to 1989 was identified as Gemella species, and another isolate from 1994 to 2002 was identified as Streptococcus gallolyticus subspecies gallolyticus. Among the remaining isolates from 1971 to 1986 and from 1994 to 2002, there were seven and eight of the S. mitis species group, two and seven of the S. anginosus species group, eight and five of the S. mutans species group, three and two of the S. salivarius species group, and seven and one of the S. sanguinis species group, respectively. The two non-viridans group streptococcal isolates were excluded from the analysis.

In isolates from 1994 to 2002 high rates of nonsusceptibility were measured to erythromycin (26%), azithromycin (26%), levofloxacin (22%), and penicillin (13%) (Table 1). The lowest rates of nonsusceptibility (for the same time period) were those to vancomycin (9%) and clindamycin (4%).

TABLE 1.

MIC₉₀s (in micrograms per milliliter) for and susceptibility patterns of viridans group streptococci for the time periods 1971 to 1986 and 1994 to 2002^a

Antimicrobic	1971-1986 (n = 27)			1994-2002 (n = 23)
	MIC90 (range)	% not S	% R	MIC90 (range)	% not S	% R	Pb
Penicillinc	0.125 (0.125-0.125)	0	0	0.5 (0.125-4)	13	4	0.09
Clindamycind	0.125 (0.125-0.25)	0	0	0.125 (0.125-128)	4	4	0.46
Erythromycind	4 (0.125-8)	11	11	16 (0.125-128)	26	26	0.27
Azithromycine	8 (0.125-32)	11	11	16 (0.125-128)	26	26	0.27
Vancomycinf	1 (0.5-2)	4	4	2 (0.5-2)	9	9	0.59
Levofloxacing	4 (0.125-4)	11	0	8 (0.25-32)	22	9	0.44
Moxifloxacin	0.25 (0.125-2)	NA	NA	0.25 (0.125-4)	NA	NA
Gatifloxacin	0.5 (0.125-0.5)	NA	NA	0.5 (0.125-8)	NA	NA
Ciprofloxacin	8 (0.5-16)	NA	NA	8 (0.5-64)	NA	NA
Garenoxacin	0.25 (0.125-1)	NA	NA	0.25 (0.125-1)	NA	NA

^aAbbreviations: S, susceptible; I, intermediate; R, resistant; NA, not applicable.

^bFisher's exact test for percent not susceptible, 1971 to 1986 compared to 1994 to 2002 period.

^cNCCLS breakpoints: S, ≤0.12 μg/ml; I, 0.25-2 μg/ml; R, ≥4 μg/ml.

^dNCCLS breakpoints: S, ≤0.25 μg/ml; I, 0.5 μg/ml; R, ≥1 μg/ml.

^eNCCLS breakpoints: S, ≤0.5 μg/ml; I, 1 μg/ml; R, ≥2 μg/ml.

^fNCCLS breakpoints: S, ≤1 μg/ml.

^gNCCLS breakpoints: S, ≤2 μg/ml; I, 4 μg/ml; R, ≥8 μg/ml.

The newer fluoroquinolones, moxifloxacin and gatifloxacin, demonstrated better in vitro activity than did levofloxacin and ciprofloxacin, a finding shared by other studies (6, 7, 12, 13). Nevertheless, elevated MICs of levofloxacin or ciprofloxacin were associated with elevated MICs of the newer fluoroquinolones. There was no statistically significant difference (by Fisher's exact test) in levofloxacin nonsusceptibility rates and MIC₉₀s (MICs at which 90% of the isolates tested are inhibited) between the two time periods (Table 1).

Among 1994 to 2002 isolates, higher MIC₉₀s were found for ciprofloxacin (8 μg/ml) and levofloxacin (8 μg/ml) compared with other studies, wherein MIC₉₀s ranged from 2 to ≥4 μg/ml for ciprofloxacin and 1 to 2 μg/ml for levofloxacin (6, 7, 12). In the same studies, MIC₉₀s of gatifloxacin (0.25 to 0.5 μg/ml), moxifloxacin (0.25 μg/ml), and garenoxacin (0.06 μg/ml) were comparable to ours (Table 1) (6, 7, 12). In contrast, the level of penicillin nonsusceptibility among the 1994 to 2002 isolates (13%) was lower than the 28 to 56% level reported in recent surveillance blood culture studies (6-8, 12, 13, 22, 28). Three of our isolates (two of S. mitis and one of the S. sanguinis species group) from 1994 to 2002 were penicillin nonsusceptible. The level of erythromycin (26%) and clindamycin (4%) nonsusceptibility among the 1994 to 2002 isolates in our study was comparable to rates of 40 to 44% (6-8, 12, 13, 22, 28) and 4.2 to 10% (6, 7, 12, 13), respectively, in these same studies.

For the three vancomycin-nonsusceptible viridans group streptococcal isolates in the present report, the MICs were 2 μg/ml, a one-dilution difference from susceptibility (MIC ≤ 1 μg/ml) and within the margin of error of the assay (20). There have been previous reports of viridans group streptococci for which vancomycin MICs were slightly elevated (3, 23, 25, 27).

There are limitations to our study. First, the small sample size did not provide the power to detect statistically significant differences, if present, in susceptibility between the two time periods. Second, the viridans group streptococcal isolates were from a tertiary care center and may not be reflective of endocarditis isolates seen at community hospitals. Third, since the patient population from which isolates are recovered may influence antimicrobial susceptibility patterns, isolates from endocarditis patients may not reflect the fluoroquinolone susceptibility pattern of isolates from other populations. Diekema et al., for example, found a trend of reduced susceptibility to ciprofloxacin (34 versus 46%) in patients with a diagnosis of cancer versus those without such a diagnosis (P = 0.07) (7). Fourth, the antimicrobial use history of these patients was unknown, a factor which may have influenced antimicrobial susceptibility. Fifth, not all viridans group streptococcal strains were archived from endocarditis patients diagnosed at our institution during the time periods of the study, resulting in potential selection bias. The levels of nonsusceptibility reported must therefore be interpreted in the context of these limitations.

The strength of the data presented lies in the uniqueness of this strain collection. Most published studies do not report antimicrobial susceptibility patterns of viridans group streptococci based on clinical diagnosis. To our knowledge, there are no published antimicrobial susceptibility pattern surveys of viridans group streptococci that include strains collected from both recent and past endocardial infections. Larger prospective surveillance studies are needed to monitor antimicrobial resistance in viridans group streptococci from defined patient populations such as neutropenic hematology and endocarditis patients. Increasing levels of antimicrobial resistance could impact the rate of breakthrough bacteremia with viridans group streptococci in neutropenic patients receiving fluoroquinolone prophylaxis and may also influence antimicrobial prevention and management of infective endocarditis.