LACK OF UTILITY OF SEROTYPING MULTIPLE COLONIES FOR DETECTION OF SIMULTANEOUS NASOPHARYNGEAL CARRIAGE OF DIFFERENT PNEUMOCOCCAL SEROTYPES

Abstract

Streptococcus pneumoniae is commonly carried asymptomatically in the nasopharynx of young children, beginning as early as the first week of life. More than 95% of children are colonized during the first 2 years of life and can sequentially lose and acquire as many as six different serotypes during that time. 1 Although data are limited, the frequency of simultaneous carriage of multiple serotypes has been reported to be as high as 7.5% in Australian Aboriginal children 2 and 29.5% in children colonized with pneumococci in Papua New Guinea. 3 For several years our laboratories have been involved in studies examining the impact of vaccination on pneumococcal carriage, as well as the use of nasopharyngeal isolates as surrogates for monitoring trends in antibiotic resistance. 4–11 In these studies three to five colonies from each nasopharyngeal specimen were randomly selected for serotyping. We have recently reviewed these data to determine the frequency of detection of simultaneous carriage of more than one serotype and the usefulness of serotyping multiple colonies. Methods. Nasopharyngeal specimens were collected with calcium alginate swabs from children 1 to 60 months of age in South Africa (RSA) (Calgiswab type1; Harwood Products) or cotton-tipped swabs in Israel inoculated directly onto either blood agar plates supplemented with 5 μg/ml gentamicin (RSA) or placed in modified Stuart’s transport medium (Medical Wire & Equipment) before inoculation on blood agar (Israel). After incubation at 35–37°C for 24 to 48 h, three to five colonies with morphologic characteristics suggestive of S. pneumoniae were isolated, and identification confirmed by optochin sensitivity and bile solubility (RSA) or slide agglutination (Israel; Phadebact; Pharmacia Diagnostics, Uppsala, Sweden). Serogrouping/serotyping was done on overnight cultures of the individual colonies by the quellung method using antisera purchased from the Statens Seruminstitut, Copenhagen, Denmark. Results. The results of serogrouping/typing multiple colonies are shown in Table 1. In South Africa 2 different serogroups/types were found in only 22 of the 1653 (1.3%) specimens examined. No differences in the number of multiple serogroups/types were seen when up to 5 colonies were examined. In Israel serogrouping/typing of 3 individual colonies from each of 246 nasopharyngeal specimens yielded only 6 (2.4%) with more than 1 serogroup/type (Table 1).Table 1: Frequency of simultaneous carriage of multiple pneumococcal serogroups/typesData from a study in New Guinea in which 50 colonies from each of 10 children were serotyped suggest that when more than one pneumococcal serotype is simultaneously carried, one type is usually present in greater numbers. 3 In New Guinea the proportion of the dominant serotype ranged from 73 to 96%. Using the simple binomial formula q∧n = (1 −P) where q = 1 − concentration of organisms, ∧ = exponentiation operator, P = probability of finding one or more colonies and n = number of colonies serotyped, the number of colonies that would need to be examined and the probability of detecting the less common serotype for various proportions of different pneumococcal serotypes were calculated (Table 2). If both serotypes were present in equal numbers, examining five colonies would be sufficient to detect both serotypes ∼97% of the time. If, as in New Guinea, the less common serotype represents only 4 to 27% of the total pneumococcal population, 11 to 59 colonies from every specimen would need to be serotyped to have a 95% probability of picking up the second pneumococcus. Similarly the likelihood of detecting the less common serotype by examining 5 or 10 colonies is 76 and 94% if the less common serotype comprises only one-fourth of the population and 23 and 40% if only 5% of the population is the lesser serotype.Table 2: Likelihood of detecting carriage of more than one pneumococcal serotypeDiscussion. It is likely that the number of multiply colonized children in our studies is much higher than we have observed by serotyping three to five colonies per specimen. How then can a more accurate estimate of the prevalence of colonization with more than one pneumococcal serotype be obtained? The obvious choice would be to examine more colonies per specimen. The quelling method, the current standard for pneumococcal serotyping, is labor-intensive and expensive. Serotyping one colony can take as long as 30 min and cost up to $100.00. For many studies of nasopharyngeal carriage, sample sizes of ≥100 pneumococcal carriers are required to achieve statistical significance. Serotyping 1000 to 6000 colonies to detect strains that are present in small numbers is impractical. Gundel and Okura 12 found that carriage of multiple serotypes could be detected by intraperitoneal injection of mice with saliva. Once the dominant serotype of pneumococcus in the specimen was identified, additional aliquots of the same saliva were mixed with anticapsular serotype-specific antibodies and then reinjected into the mice. 12, 13 This procedure was repeated until no new serotypes were isolated. This study identified simultaneous carriage of multiple serotypes in 70 of 95 children. Although the method used by Gundel is highly sensitive, the animal inoculations and the large amounts of serotype-specific antisera required make this technique even more costly and time-consuming than the agar plate method in current use. Clearly a more practical, reasonably inexpensive yet sensitive method for isolation and screening of multiple pneumococcal serotypes needs to be developed. A slide agglutination method using specific antibody-adsorbed latex particles 14 identified carriage of multiple serotypes in 10% of Gambian patients. 15 Although this technique is simple and inexpensive and allows serotyping of a broth suspension of multiple colonies from a single agar plate (a “sweep”), no direct comparison of this technique with the quelling method has been done to determine the rate of false positive reactions. Another possible approach would be a quantitative enzyme-linked immunosorbent assay; however, this method would be very expensive given the amount of serogroup/type-specific antisera required. The best answer may be a quantitative PCR assay to detect pneumococcal capsule polysaccharide genes. Each of the 90 serogroup/types of pneumococci produces a distinct capsule, synthesized by a cassette of genes coding for a complex pathway of monosaccharide synthesis, activation, transport and polymerization. The nucleotide sequences of several loci encoding capsule types, including type 19F 16 and type 3, 17 have been described. Differences between these loci could identify DNA sequences that could be used in a PCR assay to distinguish mixtures of pneumococci of different capsular types at very low concentrations. Obviously the sequences of at least the most prevalent serogroups/types would need to be characterized before such an assay could be used in place of serotyping by the quelling method. The need for development of a more sensitive assay to detect simultaneous carriage of multiple pneumococcal serotypes is urgent. Although vaccination with pneumococcal conjugate vaccines has been shown to reduce nasopharyngeal carriage of vaccine serotypes, there appears to be a concomitant increase in the carriage of nonvaccine serotypes. 4, 5 Two explanations are possible 18 : (1) that the removal of the vaccine serotypes allows “unmasking” of serotypes that were present but were difficult to detect because of their lower concentration relative to the vaccine serotypes; (2) that vaccine serotypes are being “replaced” by nonvaccine serotypes, i.e. that the proportion of children who carry nonvaccine serotypes actually increases with the decrease in prevalence of vaccine serotypes. If this is associated with an increase in the incidence of pneumococcal disease caused by these types, the overall impact of the pneumococcal conjugate vaccine would be decreased. Because of the limitations in the number of serotypes that will be included in a pneumococcal conjugate vaccine, other strategies such as the development of pneumococcal vaccines with common protein antigens should continue to be pursued. Limited data available from a large pneumococcal conjugate vaccine trial suggest no increase in invasive disease caused by nonvaccine serotypes, 19 but ongoing monitoring is important. Distinguishing between serotype unmasking and replacement in the current vaccine studies is difficult without a more sensitive method for detecting carriage of multiple serotypes. Until such a test is developed, we would not recommend serotyping of multiple colonies because the yield of different pneumococcal serotypes is unlikely to outweigh the cost and the effort expended and could actually result in erroneous conclusions as to the role of unmasking vs. serotype replacement after vaccination. Acknowledgments. We thank Brian Plikaytis for statistical assistance and Benjamin Schwartz for helpful comments.