Streptococcus pneumoniae Penicillin-Nonsusceptible Strains in Invasive Infections in Sao Paulo, Brazil

Abstract

Streptococcus pneumoniae infections are an important cause of morbidity and mortality in developing countries. Although penicillin is the antibiotic of choice for the treatment of most infections, strains with resistance to this and other antibiotics are increasing in many parts of the world.1, 2 Data on prevalence of penicillin-nonsusceptible pneumococci (PNS-P) in South America are scant. In Brazil a prevalence of 17.5% was previously reported but no detailed information regarding its importance in pediatrics was provided.3 The present study was conducted in Sao Paulo, Brazil, in two teaching hospitals (Santa Casa de São Paulo and Instituto de Infectologia Emilio Ribas) to investigate the prevalence of resistance to penicillin (PEN), trimethoprim-sulfamethoxazole (TMP-SMX), chloramphenicol and ceftriazone (CT) in isolates from blood, cerebrospinal fluid and pleural effusion among children with invasive infections. The study also sought to identify the most prevalent pneumococcal serotypes among children with invasive infections. Methods. The Santa Casa is a referral hospital in the central area of Sao Paulo with 150 pediatric beds. An average of 300 patients are admitted every month. The Instituto de Infectologia Emilio Ribas is a hospital for infectious diseases with 50 pediatric beds and is a referral center for meningeal infections in Sao Paulo. From July, 1989, to July, 1993, all isolates of Streptococcus pneumoniae obtained from cerebrospinal fluid, blood or pleural effusion from patients <12 years old were routinely investigated for penicillin resistance. Isolates were identified as S. pneumoniae on the basis of alphahemolysis on blood agar and susceptibility to ethylhydrocuprein (optochin). All strains were screened for PEN resistance with the disk diffusion test using a 1-μg oxacillin disk, according to the procedures recommended by the National Committee for Clinical Laboratory Standards.4 The inhibition zone <20 mm was considered resistant. Disk diffusion testing according to the National Committee for Clinical Laboratory Standards procedure for TMP/SMX 2.5/23.75 μg, and chloramphenicol 30 μg were performed for every isolate. Inhibition zones >19 mm to TMP/SMX and >21 mm to chloramphenicol were defined as susceptible. For TMP/SMX inhibition zones between 16 and 18 mm were defined as intermediate susceptibility. The MIC for PEN was determined for each isolate using the broth dilution assay recommended by the National Committee for Clinical Laboratory Standards, using cation-adjusted Mueller-Hinton broth with 3% lysed horse blood.4S. pneumoniae isolates were defined as PEN-susceptible in MIC were <0.1 μg/ml, and PNS-P when MIC were ≥0.1 μg/ml. PNS-P were further subdivided into intermediate (0.1 μg/ml ≤ MIC ≤ 1.0 μg/ml) and highly resistant strains (MIC ≥ 2.0 μg/ml). PNS-P isolates were evaluated for CT susceptibility using the E test manufactured by AB Biodisk, Sweden, and defined as resistant to CT if MIC were ≤2.0 μg/ml. Serotyping was performed in the WHO pneumococcal reference laboratory in Pennsylvania on the basis of capsular swelling with type-specific antiserum (quellung reaction). Results. Patients. One hundred one pneumococcal isolates were obtained between July, 1989, and July, 1993; 60 were from cerebrospinal fluid, 35 from pleural effusion and 6 from blood. The diagnoses and age distribution are shown in Table 1. Resistant strains. The prevalence of PNS-P in our study was 9.8%. One strain was highly resistant (MIC = 4.0 μg/ml) and 9 were intermediately resistant strains (0.1 μg/ml ≤ MIC ≤ 1.0 μg/ml). Of the 10 PNS-P isolates 9 were obtained from the 58 children younger than 2 years of age whereas only one was obtained from the 43 children older than 2 years. Sixty-three (62.7%) strains were nonsusceptible to TMP/SMX (53 resistant and 10 intermediate) and two were nonsusceptible to chloramphenicol. One PNS-P was resistant to CT (MIC = 2.0 μg/ml). The other 9 strains tested against CT were susceptible (0.064 μg/ml < MIC < 1.0 μg/ml). Serotypes. A total of 86 strains were serotyped. Among these 72 (82.7%) were serotypes, contained or antigenically related, in the 23-valent pneumococcal polysaccharide vaccine. Table 2 shows the most prevalent serotypes obtained from children younger than 2 years, from patients infected by PNS-P strains and from patients with a fatal outcome. Outcome. There were 15 fatal pneumococcal invasive infections, 14 caused by meningitis and 1 by bacteremia, giving an overall case fatality rate of 14.8%. Of the 31 children <2 years of age with meningitis, 12 (38.7%) died, and 2 (6.6%) of the 30 children older than 2 years of age with the same diagnosis died. We were not able to correlate fatal outcome with infections caused by PNS-P strains. Two of our patients with meningitis caused by a PNS-P strain had a favorable outcome. Two children with meningitis caused by a PNS-P strain had a poor outcome with severe neurologic damage. One was infected by a CT-resistant and PEN-intermediately resistant strain (MIC = 2.0 μg/ml for CT and 0.125 μg/ml for PEN). This child was initially managed with CT (100 mg/kg/day) and dexamethasone (0.6 mg/kg/day) on admission. The culture remained positive after 48 h and the therapeutic schedule was changed to a combination of CT and vancomycin. The other child was infected by a fully PEN-resistant strain (MIC = 4.0 μg/ml). The children admitted because of pneumonitis had a favorable outcome and were treated with ampicillin or an oxacillin and chloramphenicol combination. Discussion. Strains of PNS-P have spread rapidly worldwide; however, the patterns of resistance differ from one world region to another.2 In this study we have shown a prevalence of PNS-P strains of 9.8% among invasive pneumococci isolates in Sao Paulo. All of the PNS-P strains exhibited intermediate resistance with the exception of one strain that was highly resistant (MIC = 4.0 μg/ml). Most of the nonsusceptible strains were recovered from children younger than 2 years of age. A recent survey carried out in Brazil, including children and adults, found that 17.4% of strains recovered from blood, respiratory tract secretions, ocular secretions and others, between 1988 to 1992, were resistant to penicillin.3 Another survey in Uruguay reported a resistance prevalence of 6%.5 The increase of the prevalence of PNS-P strains may be associated with reduction of susceptibility to other betalactam and non-beta-lactam drugs. Two strains were nonsusceptible to chloramphenicol and 62.4% were nonsusceptible to TMP/SMX. A high prevalence of TMP/SMX resistance has also been observed in Asia, Europe and the US and may reflect the widespread use of this drug.2, 6 This finding is a matter of concern as this drug is recommended in outpatient management of respiratory infections in developing countries. Patients with pneumonitis caused by PNS-P were successfully treated with ampicillin or the combination of oxacillin and chloramphenicol. Our data suggest that penicillin G or ampicillin treatments would probably be effective for pneumococcal pneumonitis caused by strains for which the MIC of penicillin ranged from 0.12 to 1 μg/ml as has been previously reported.7 In Brazil most medical centers use the ampicillinchloramphenicol combination for initial meningitis therapeutics in patients younger than 5 years of age. This approach has been substituted in developed countries by third generation cephalosporins because of the increasing bacterial resistance and the poor bactericidal activity of chloramphenicol against pneumococci.8 However, there have been an increasing number of reports in which cephalosporin treatment failed.9 Combination therapy in which third generation cephalosporin and vancomycin are used has been proposed for the treatment of meningitis caused by PNS-P.9 In Brazil mortality from pneumococcal invasive infection is high. The development of effective conjugate pneumococcal vaccines including serotypes most prevalent in our country will raise hope of reducing mortality and morbidity caused by those infections. Acknowledgments. We thank Robert Austrian, M.D. (WHO Collaborating Center for Reference and Research on Pneumococci, Department of Research Medicine, University of Pennsylvania School of Medicine, Philadelphia, PA) for serotyping our pneumococcal strains. Eitan N. Berezin, M.D.; Eduardo S. Carvalho, M.D.; Silvana Casagrande, M.Sc.; Maria Cristina Brandileone, M.Sc.; Igor M. Mimica, M.D.; Calil K. Farhat, M.D. Department of Pediatrics and Microbiology and Santa Casa de São Paulo (ENB, IMM) Instituto de Infectologia Emilio Ribas Universidade Federal de São Paulo (ESC, CKF) Instituto Adolfo Lutz (SC, MCB) Sao Paulo, Brazil