Serotype 10A in Case Patients with Invasive Pneumococcal Disease: A Pilot Study of PCR-Based Serotyping in New Jersey

Abstract

In 2008, the New Jersey Department of Health (NJDOH) identified a 21.1% increase in reported invasive pneumococcal disease (IPD). In 2009, NJDOH piloted nucleic acid-based serotyping to characterize serotypes causing IPD. From April through September, NJDOH received specimens from 149 of 302 (49%) case patients meeting our case definition. An uncommon serotype, 10A, accounted for 25.2% of IPD overall and was identified in 12 counties, but it was associated with one county (rate ratio = 5.4, 95% confidence interval [CI] 2.1, 11.8). NJDOH subsequently conducted a case-control study to assess the presentation of and clinical risk factors for 10A IPD. Case patients with 10A IPD were more likely to have had immunosuppression, asthma, and multiple chronic medical conditions than control subjects had (odds ratio [OR] = 2.6, 95% CI 1.1, 6.3; OR=4.7, 95% CI 1.7, 13.2; and OR=2.3, 95% CI 1.0, 5.2, respectively). State-based pneumococcal serotype testing identified an uncommon serotype in New Jersey. Continued pneumococcal serotype surveillance might help the NJDOH identify and respond to future serotype-specific increases.

Streptococcus pneumoniae (S. pneumoniae) is a leading cause of bacterial pneumonia, bacteremia, and meningitis. In 2008, the New Jersey Department of Health (NJDOH) observed a 21.1% increase in reported invasive pneumococcal disease (IPD) compared with IPD cases reported in 2007 (95% confidence interval [CI] 8.6, 35.2, 8.33 cases/100,000 population in 2008 vs. 6.88 cases/100,000 population in 2007), including increases in four counties. In response, in 2009, NJDOH began a pilot project to characterize the serotypes causing IPD and to determine if serotypes not included in current vaccines contributed to these increases. When our investigation identified an unusual serotype, 10A, NJDOH conducted a case-control study to assess the presentation of and clinical risk factors for 10A IPD.

METHODS

In New Jersey, hospitals and local health departments report cases of IPD to NJDOH using the electronic Communicable Disease Reporting and Surveillance System (CDRSS). We defined a case as S. pneumoniae isolated from a normally sterile site (e.g., blood or cerebrospinal fluid) in a New Jersey resident.

During the period April 1–September 30, 2009, we requested that clinical laboratories submit all S. pneumoniae isolates from patients with IPD to the state public health laboratory. We matched electronic CDRSS case reports containing available demographic and clinical information to serotyped specimens for analysis.

We serotyped isolates using a modified version of the multiplex polymerase chain reaction (PCR) protocol developed and validated by Pai et al.¹ following personal communication with the principal investigator (Personal communication, Bernard Beall, Chief, Streptococcus Laboratory, Respiratory Disease Branch, Division of Bacterial Diseases, National Center for Immunization and Respiratory Diseases, Centers for Disease Control and Prevention [CDC], April 2009). We verified the multiplex PCR protocol using 66 previously serotyped S. pneumoniae isolates obtained from the Streptococcus Laboratory of the Respiratory Diseases Branch of the Division of Bacterial Diseases at CDC. We used conventional serotyping by immunoprecipitation reaction (Denka Seiken Co., Ltd., Tokyo, Japan) to confirm the multiplex PCR on 30 of the first 40 specimens received; all serotypes were confirmed.

We subsequently conducted a case-control study to identify risk factors for IPD due to an unusual serotype, 10A. We defined a case as IPD caused by serotype 10A; control subjects were patients with IPD caused by any other serotype. We excluded people aged ≤5 years because only one child in this age group had infection with serotype 10A.

We abstracted medical records using a standardized data abstraction form with the reviewer blinded to pneumococcal serotype. Following chart abstraction, we classified patients as having an underlying condition if a medical indication for the pneumococcal polysaccharide vaccine (PPSV23) was present, based on Advisory Committee on Immunization Practices (ACIP) recommendations in place in 2008 and current recommendations.^2,3 We used Clinical and Laboratory Standards Institute (CLSI) breakpoints to determine antibiotic susceptibilities from minimal inhibitory concentration (MIC) or disk diffusion zone diameter results reported by clinical laboratories;⁴ if these results were unavailable, antibiotic susceptibility interpretations were accepted from laboratories confirming the use of 2009 CLSI breakpoints (for all antibiotics except penicillin, which was excluded due to systematic error in interpretation).

Analysis

Rates and rate ratios of IPD were calculated using population estimates from the U.S. Census Bureau⁵ and Stata^® version 10.⁶ We used chi-square and Fisher's exact tests to compare proportions and calculated odds ratios (ORs) and 95% CIs for all demographic and clinical variables, using SAS^® version 9.2.⁷

OUTCOMES

Serotype analysis

In 2009, NJDOH received report of 791 cases of IPD, a rate of 9.03 cases per 100,000 population (95% CI 8.42, 9.69). During the study period, we received specimens from 149 of 302 (49.3%) reports of IPD that met our study definition. Among the 149 available S. pneumoniae isolates, six were nonviable, leaving 143 isolates for analysis. Case patients included in our analyses were similar in age (median 57.9 vs. 59.5 years) and gender (53.9% vs. 50.3% female, p=0.54) to those who were excluded, but were less likely to be black (16.1% vs. 25.5%, p=0.13).

Among our 143 S. pneumoniae isolates, serotype 19A was the most commonly identified (n=41, 28.7%), followed by serotype 10A (n=36, 25.2%), 6C (n=13, 9.1%), 7F (n=12, 8.4%), and 3 (n=10, 7.0%) (Table 1). There were only two cases with serotypes included in the seven-valent conjugate vaccine (PCV7) in use during the study period (data not shown). Among people aged ≥5 years (n=129), serotype 10A was the most commonly identified serotype, accounting for 35 (27.1%) IPD cases (data not shown).

Table 1.

Frequency of the most common serotypes and their antibiotic sensitivity from a case-control study: New Jersey, April–September 2009

^aNumber susceptible to antibiotic/total with available data

PCV13 = 13-valent pneumococcal conjugate vaccine

PPSV23 = pneumococcal polysaccharide vaccine

Serotype 10A infection occurred in 12 of 21 (57.1%) New Jersey counties. The rate of serotype 10A for one county that reported an overall rate increase in 2008 was 2.92/100,000 population, 5.4 (95% CI 2.1, 11.8) times the average rate in all counties where 10A was detected (0.54/100,000 population) and 3.8 (95% CI 1.2, 13.3) times the rate of the county reporting the second highest 10A frequency (0.77/100,000 population). No other counties had statistically different rates compared with the average rate in all counties where 10A was detected.

Antibiotic susceptibility

Individual antibiograms from serotype 10A specimens demonstrated substantial variability without a dominant pattern. Serotype 10A demonstrated decreased susceptibility to erythromycin (57.9%), tetracycline (70.6%), and trimethoprim/sulfamethoxazole (78.6%) (Table 1). One case of penicillin resistance was identified in a case of 10A IPD with meningitis.

Case-control study

We reviewed the medical records of 127 of 129 (98.4%) patients aged ≥5 years for whom we had received isolates. The demographic and clinical characteristics of serotype 10A IPD compared with other serotypes are displayed in Table 2. Of the 33 serotype 10A IPD cases, 18 cases (55%) occurred among people aged 15–64 years, 13 cases (39%) occurred among people aged ≥65 years, and two cases (6%) occurred among children aged 5–14 years, with no statistically significant difference in age category between case patients and control subjects. Pneumonia was the most common presentation between both groups (serotype 10A, 88%; other serotypes, 79%; p=0.25). No difference occurred between IPD caused by serotype 10A and that caused by other serotypes in the frequency of death, intubation, intensive care unit admission, or sepsis, with an overall mortality of 16%.

Table 2.

Clinical characteristics of serotype 10A vs. all other IPD: New Jersey, April–September 2009

^aPercentages do not sum to 100% because syndromes were not mutually exclusive.

^bOther diagnoses included septic arthritis (n=3), endocarditis (n=2), sinusitis (n=1), laryngitis (n=1), peritonitis (n=1), and facial cellulitis (n=1).

^cOne case patient and two control subjects were excluded because of incomplete severity data.

^dBased on ACIP guidelines in effect before 2009

^eBased on 2010 ACIP guidelines

^fExcluding asthma

^gSmoking and asthma were not qualifying underlying conditions on the basis of ACIP recommendations in effect before 2009.

^hVaccination status was known for 34 of 41 high-risk adults aged 19–64 years, with at least one underlying condition.

ⁱVaccination status was known for 53 of 59 adults aged ≥65 years.

IPD = invasive pneumococcal disease

OR = odds ratio

CI = confidence interval

NC = not calculated

ICU = intensive care unit

ACIP = Advisory Committee on Immunization Practices

HIV = human immunodeficiency virus

PPV23 = pneumococcal polysaccharide vaccine

Among all 127 IPD cases, 69% had an underlying medical condition, on the basis of ACIP recommendations for vaccination in 2008, with no significant difference between case patients and control subjects; however, case patients were more likely to have had immunosuppression (OR=2.6, 95% CI 1.1, 6.3), asthma (OR=4.7, 95% CI 1.7, 13.2), or two or more underlying conditions (OR=2.3, 95% CI 1.0, 5.2). The latter association was stronger when smoking and asthma were included, based on the 2009 ACIP guidelines (OR=3.0, 95% CI 1.2, 7.3). Only 50% of cases with high-risk conditions had documented vaccination, with no difference between case patients with 10A and control subjects.

LESSONS LEARNED

We used the existing infrastructure for surveillance of invasive S. pneumoniae in New Jersey to conduct state-based serotype surveillance using nucleic acid amplification. Our surveillance revealed that an unusual serotype, 10A, represented 25% of IPD cases in New Jersey during our study period, with the majority of cases occurring among people aged ≥15 years. Although serotype 10A IPD was observed in a broad geographic distribution, the rate of IPD due to serotype 10A was significantly higher in one New Jersey county reporting a rate increase in 2008.

Data from New Jersey differed from national data from Active Bacterial Core surveillance (ABCs) on the proportion of serotype 10A, which represented only 1.5% of cases during 2009 (Personal communication, Ruth Link-Gelles, CDC, June 2011). Nearby ABCs sites, including New York State, Connecticut, and Maryland, did not report higher frequencies of serotype 10A in 2009 (Personal communication, Matthew Moore, CDC, 2011). The New Jersey serotype distribution otherwise reflected national trends, with few cases of IPD in PCV7 strains, and a predominance of 19A and serotypes included in the 13-valent conjugate vaccine (PCV13), licensed in 2010.

Serotype-specific outbreaks of pneumococcal disease have been commonly identified among closed populations (e.g., military recruits and medical facilities) and less frequently within communities.^8–10 While it is possible that the high frequency of serotype 10A in our sample represents an outbreak associated with the county, it is difficult to attribute this finding to an outbreak without further strain characterization. An alternate possibility is that residents of this county have a higher frequency of risk factors for serotype 10A IPD, placing them at higher risk for infection.

In our case-control study, patients with asthma, immunosuppression, and/or multiple chronic conditions appear to be more likely to have IPD due to serotype 10A, consistent with a low invasive disease potential among healthy people, as previously suggested.¹¹ Serotype 10A is included in PPSV23, but only 50% of case patients with high-risk conditions had documented PPSV23 vaccination; in a national survey, only 18.5% of high-risk adults were vaccinated.¹² The NJDOH continues to provide education and public health messaging to increase PPSV23 vaccination rates in these vulnerable adults.

Limitations

Although subject to limitations of a pilot study, including short duration and small sample size, our study illustrates the potential value of PCR-based pneumococcal serotyping at the state level to identify serotype-specific outbreaks or regional serotype trends. Unfortunately, serotype surveillance in New Jersey has not continued due to a lack of ongoing funding, as this study was supported by a one-time laboratory grant. Since the time of our study, the rate of IPD in New Jersey has declined, presumably as a result of uptake in PCV13. Further serotype surveillance would be needed to evaluate the current rate of serotype 10A IPD.

CONCLUSION

State-based pneumococcal serotype testing identified IPD due to an uncommon serotype in New Jersey. Continued pneumococcal serotype surveillance might help the NJDOH identify and respond to future serotype-specific increases.