Despite a well-established and highly accepted pertussis vaccine program, reported pertussis has been increasing in the US since the 1990s. In 2010 and 2012, national pertussis cases exceeded levels not observed since the 1950s.1 Several factors are likely contributing to these increases, including improved provider awareness and reporting, advances in pertussis diagnostics and, more recently, waning immunity from acellular pertussis vaccines.2,3
Since the mid 1990s, an array of pertussis diagnostic test options has become available for routine diagnosis, including culture, direct fluorescent antibody testing (DFA), polymerase chain reaction (PCR) and serology. Diagnostic testing practices heavily impact pertussis surveillance. Case notification is based on standardized surveillance definitions that define acceptable diagnostic tests, which may or may not fully align with current testing practices. Pertussis surveillance data have become increasingly important to those studying emerging epidemiologic trends, particularly in the wake of recent surges in national reporting. To date, there has been no comprehensive analysis of US pertussis diagnostic testing trends. Using nationally reported surveillance data, we describe changes in pertussis diagnostic testing over time in the US and discuss the potential implications of these changes.
MATERIALS AND METHODS
Data for our analysis were obtained from the National Notifiable Diseases Surveillance System (NNDSS), a passive surveillance system for notifiable diseases reported in the US. Pertussis is notifiable in all 50 states and the District of Columbia. Cases are reported to state health departments by medical providers, diagnostic laboratories and local health departments; states then report confirmed, probable and unknown cases to Centers for Disease Control and Prevention (CDC) using the Council of State and Territorial Epidemiologists (CSTE)/CDC pertussis case definition. The earliest CSTE case definition for pertussis was established in 1990 and required isolation of Bordetella pertussis and clinical compatibility (≥2 weeks of cough plus 1 classic sign or symptom: paroxysms, whoop or posttussive vomiting) to classify a case as confirmed.4 Clinical compatibility in the absence of another diagnosis was and continues to be sufficient to classify a case as probable, and clinically compatible patients epidemiologically linked to laboratory-confirmed cases are classified as confirmed. By 1995, the case definition had been modified to classify as confirmed all cases from which B. pertussis was isolated, with acute cough of any duration.5 In 1997, PCR was added to the case definition as an accepted laboratory test for diagnosis, requiring clinical compatibility for confirmation of the case.6,7 To date, neither serology nor DFA tests have been accepted for inclusion in the CSTE pertussis case definition. Massachusetts, however, has accepted and reported serology results for case confirmation since 1987; a practice CDC informally recognizes.8
Several types of case-data are reported to NNDSS, including demographic information, clinical symptoms, antibiotic treatment, vaccination history, epidemiologic data (contact history) and laboratory test results. Four types of pertussis diagnostic test results are collected as part of NNDSS: bacterial culture, serologic tests (acute and convalescent), PCR and DFA. Diagnostic test results are coded qualitatively as positive, negative, indeterminate, pending, not done, unknown or parapertussis. Information on test kit names, manufacturers and submitting laboratories is not collected through NNDSS. There are inconsistencies among states in what types of information are reported to CDC, which are the result of variations in state-specific electronic reporting capacity; however, most send the majority of the data elements described above. During 2007 to 2012, laboratory testing results were unavailable from 5 states: Alaska, Delaware, Illinois, Missouri and New Jersey.
All pertussis cases reported to NNDSS during 1990 to 2012 with either known or unknown laboratory testing results were included in this analysis. A pertussis laboratory result for culture, PCR, serology and DFA testing coded as positive, negative or indeterminate was considered indicative of a completed test. We included all reported laboratory test results, not just those that were positive, allowing for assessment of preferential testing practices for diagnosing pertussis.
Trends in reported diagnostic test results were analyzed by test type. Test types were then stratified by age group and state to look for age-related and geographic patterns in testing. To assess temporal changes in age-specific laboratory testing, data were analyzed at 3 points in our study period: 1990, 2001 and 2012. We also assessed whether individual or multiple diagnostic test types were employed for diagnosis of pertussis and whether this changed over time. Finally, cases were stratified by diagnostic testing methods, epidemiologic linkage and CSTE case classification to assess variations in case classification practices.
During 1990 to 2012, 291,290 confirmed, probable and unknown cases of pertussis were reported to CDC through NNDSS. Of those, 186,766 (64.1%) included at least 1 known pertussis laboratory test result. The percentage of cases with unknown laboratory testing results varied ranging from 19.7% in 1991 to 48.4% in 2005 (mean, 33.3%). In 2012, approximately 32% of cases had unknown or missing laboratory test results (Fig. 1).
Between 52% and 80% of all cases reported to CDC during 1990 to 1997 were tested using DFA, culture or both. The use of DFA peaked in 1991 when 70% of cases included a DFA result. By 2012, only 1.4% of cases included a DFA result. Reporting of cases with a completed culture test was greatest during 1993 (60%), but the percentage gradually decreased to 6.2% during 2012 (Fig. 1).
The first reports of PCR testing sent to CDC were received in 1995 and included 20 cases from 7 states. During 2012, 42 states reported at least 1 completed PCR test, and the percentage of all reported cases diagnosed using PCR, either alone or in combination with another test, was 62% (Fig. 1). Of cases reported with a completed diagnostic test, 91% included a PCR result.
Reports of serologic testing varied from 1995, when the first reports of completed serology tests were reported to CDC, to 2012. The percentage of all cases with completed serology testing ranged from 1.6% in 1995 to more than 20% in 2006. As of 2012, less than 5% of all reported cases included a completed serology test. During 1996 to 2004, Massachusetts contributed an annual average of 71% of all serology testing reported to CDC. In 2005, however, serologic test results were reported from an increasing number of states, and although Massachusetts continued to report more serologic testing proportionately than any other single state, they no longer represented the majority of all serology test results reported to CDC (Fig. 2). During 1995 to 2012, the total number of states submitting serology test results to CDC ranged from 16 in 1995 to 37 in 2005 and 2006. In 2012, 34 states reported serology test results.
During 2007 to 2012, 37 states and the District of Columbia reported more cases tested using PCR alone than any other single test type or combination of laboratory tests. Four states reported sole use of other diagnostic tests more frequently than PCR: serology was more prevalent in Arizona, Massachusetts and Utah; and culture was reported for a majority of cases in Georgia. Five states reported the use of multiple test types more often than the use of a single testing method. Of those, Maryland, North Carolina, Rhode Island and Vermont reported combined use of culture and PCR most frequently. In contrast, most cases in Hawaii were tested with a combination of culture and DFA (Table 1).
The use of multiple diagnostic tests for the diagnosis of a single pertussis case was common throughout the 1990s. In particular, DFA and culture were frequently used in tandem and comprised more than 54% of all cases with known test results reported to CDC during 1990 to 1995. During 1990 to 1996, an average of 53% of all cases with known laboratory testing included more than 1 test type (range, 44.8%–60.1%; Fig. 3). As the use of PCR increased (either alone or in combination with other test types), the overall frequency of multiple diagnostic tests declined. The combination of culture and PCR outpaced the use of culture and DFA beginning in 2004 and continues to be the most frequently reported combination of pertussis diagnostics. In 2012, the use of multiple diagnostic tests was reported in only 8% of cases with known diagnostic test results, whereas more than 84% of cases reported with a known laboratory result were tested using PCR exclusively. Serology is commonly used alone for diagnosing pertussis. In 2012, 93% of all cases with a positive serology result had no other laboratory test result present. When serology was reported in combination with another test type, the most frequent combination was with PCR.
Table, Supplemental Digital Content 1, https://links.lww.com/INF/C287, shows reports of laboratory-tested pertussis case-patients by age group and reported test type during years 1990, 2001 and 2012. In 1990, culture and DFA (used either alone or in combination) were the only test types reported among all age groups. During 2001, the combined use of culture and DFA continued to be the most commonly reported testing paradigm among infants, children and adolescents 11–14 years of age and adults ≥20 years of age (range, 34.7%–46.8%). The use of PCR alone was common among infants and children aged less than 15 years (range, 16.1%–21.0%). In 2012, between 82.4% and 88.4% of laboratory-diagnosed cases aged ≤19 years were tested solely using PCR. The use of culture or DFA was infrequent regardless of age group. Exclusive use of serology was most prevalent among adults ≥20 years of age (see Table, Supplemental Digital Content 1, https://links.lww.com/INF/C287).
Of all cases reported to NNDSS during 2000 to 2012, approximately 72% were classified as confirmed and 28% as probable; <1% of cases were unclassified. Of confirmed cases, 65% had a corresponding laboratory result. Of confirmed cases with no laboratory result reported, 39% were epidemiologically linked to a laboratory-confirmed case of pertussis; the remaining cases were either misclassified or had inadequate information. Trends in case classification among those with known laboratory results were similar. Of those 2012 cases diagnosed solely by a positive serology test, only 19% were reported as confirmed; 78% were from Massachusetts, the only state permitted to confirm pertussis cases using serologic assays.
Because the majority of laboratory results reported through NNDSS are positive, when restricted to only positive laboratory results, similar trends were observed throughout our analysis. Inclusion of negative, indeterminate and parapertussis results did not impact the outcome of any of our analyses.
Culture is being replaced by faster molecular methods of disease detection for diagnosis of many pathogens, and the availability of PCR for diagnosis of pertussis has been expanded to nearly all commercial and public health laboratories.9–12 Inclusion of PCR in the CSTE/CDC pertussis case definition in 1997 likely contributed to its increased acceptance as a routine diagnostic test in the US, and the decline in the use of multiple testing types for pertussis diagnosis corresponds directly to this increase. Although age-specific testing practices were noted, the use of PCR increased, whereas DFA and culture decreased among all age groups, mirroring the overall trends observed nationally.
Although Massachusetts continues to contribute a substantial proportion of serology results reported nationally, a majority of states now report pertussis serology results to NNDSS. Similar to PCR, the availability of commercial pertussis serologic assays has increased substantially. In addition, clinicians may favor serology over other pertussis diagnostics because obtaining blood specimens is often preferred over nasopharyngeal (NP) specimens and providers often do not stock appropriate NP specimen collection and transport supplies in their offices. The overall burden of pertussis observed in Massachusetts progressively declined during 2006 to 2011, which is in contrast to other states, many of which experienced peaks in disease reporting during this time period, especially 2010.1 This likely influenced the proportionate increase in serologic reporting among other states, excluding Massachusetts.
Pertussis diagnostics, surveillance and our understanding of pertussis epidemiology are inextricably linked. As a result, case definitions must be assessed in the context of changing diagnostic methods. At present, the CSTE/CDC case definition for pertussis includes both clinical and laboratory components. Although culture is sufficient to confirm pertussis among persons with a cough of any duration, a positive PCR result must also meet the clinical case definition. Public health professionals are considering making PCR confirmatory without requiring an individual to meet the clinical case definition and including serology as an acceptable method of case identification.13–15 Revisions have already been made to the CSTE pertussis case definition for 2014, which allows clinically compatible, PCR-positive infants with cough of any duration to be classified as probable, and discussions are currently underway to expand this modification to case-patients of all ages.16
Increasing the use of PCR has provided some benefits for patients, clinicians and public health. From a clinical standpoint, PCR provides faster diagnosis and treatment. The significant increase in diagnostic sensitivity provided by PCR compared with culture has resulted in improved case ascertainment and identification of pertussis clusters and has allowed for faster response to emerging pertussis outbreaks. However, PCR can have compromised specificity depending on the assay used and specimen collection practices. Several pertussis pseudooutbreaks have resulted from misdiagnosis based on PCR testing because of cross contamination or inadequately interpreted results.17,18 Many diagnostic laboratories, including large commercial laboratories, initially used a single target PCR assay unable to distinguish B. pertussis from some other Bordetella species; however, more laboratories are now incorporating an additional parapertussis-specific molecular target, increasing the diagnostic specificity.19–22 Thus, when used appropriately to test symptomatic patients, PCR is a powerful diagnostic tool for detecting pertussis.23,24
Similar to PCR, serology could substantially increase the identification of pertussis infections. However, the use of inadequately validated serologic assays with low specificity can result in substantial increases in misdiagnoses and costly public health response, as recently observed in Arizona.25 However, in the context of a reliable, standardized assay, serology has the potential to advance our understanding of the true burden of pertussis among older children, adolescents and adults, who often present later in the course of illness when other testing methods are less sensitive. CDC and Food and Drug Administration have developed and analytically validated an anti-pertussis toxin IgG serologic assay for pertussis; 20 county and state public health laboratories have been trained to use this assay and 2 (Minnesota and Wisconsin) have implemented the assay for routine use.26 Unfortunately, commercially available pertussis serology assays have not yet been adequately validated for clinical use and many have not been calibrated to a universally accepted reference standard. Despite these challenges, many other countries accept pertussis serologic assays for diagnosis and reporting, including Australia, where serology is used routinely.27–29 CDC is evaluating serologic assays currently available in the US and the data generated will help guide their routine use in the US.
Although the relative speed of diagnosis and sensitivity afforded by PCR and serology is appealing, the specificity culture provides remains invaluable. Pseudooutbreaks of pertussis have been identified partly through the use of culture confirmation.17 Culture also provides isolates of B. pertussis, which are required for monitoring antimicrobial susceptibility and assessing molecular changes occurring within the bacterium. In recent years, isolates have played a significant role in understanding the genetic make-up of circulating B. pertussis strains which informs the interpretation of epidemiologic trends, vaccine impact and the development of future pertussis vaccines.30–32 Thus, it is important that public health, clinical and commercial laboratories maintain culture capacity to both confirm the presence of pertussis and ensure continued access to B. pertussis isolates.
The results presented in our analysis are based on passive notifiable disease reporting, which inherently underrepresents the true burden of pertussis. Pertussis surveillance data are influenced by a number of factors. Provider awareness and diagnostic testing preferences for pertussis directly impact case identification, and differences in case investigation practices among states directly affect case reporting. Thus, our findings describe diagnostic trends associated with reported pertussis, which may differ with trends in pertussis diagnostic practices observed in clinical practice. Given these considerations, it is difficult to determine to what degree changes in diagnostic testing are responsible for the increases in pertussis reporting that have occurred in the last 30 years.
Despite the increasing reliance on culture-independent diagnostics for pertussis, it is important for public health, clinical and commercial laboratories to maintain culture capacity to monitor changes in pertussis at the molecular level.32 Epidemiologic trends must be considered in the context of changing diagnostic tests, and modifications to surveillance case definitions should be considered to better reflect current testing practices, while taking into account the accuracy of available diagnostic tests.
1. National Notifiable Diseases Surveillance System. . Division of Health Informatics and Surveillance, Center for Surveillance, Epidemiology and Laboratory Services, Office of Public Health Scientific Services, Centers for Disease Control and Prevention, U.S. Department of Health and Human Services 1990–2012 Atlanta, GA
2. Klein NP, Bartlett J, Rowhani-Rahbar A, et al. Waning protection after fifth dose of acellular pertussis vaccine in children. N Engl J Med. 2012;367:1012–1019
3. Misegades LK, Winter K, Harriman K, et al. Association of childhood pertussis with receipt of 5 doses of pertussis vaccine by time since last vaccine dose, California, 2010. JAMA. 2012;308:2126–2132
4. Centers for Disease Control and Prevention. . Case definitions for public health surveillance. MMWR. 1990;39(RR-13)
5. Council of State and Territorial Epidemiologists (CSTE). . CSTE Position Statement 1994-11-f: Pertussis. 1994 Minneapolis, MN
6. Centers for Disease Control and Prevention. . Case definitions for infectious conditions under public health surveillance. MMWR. 1997;46(RR-10)
7. Council of State and Territorial Epidemiologists (CSTE). . CSTE Position Statement 97-ID-09: public health surveillance, control and prevention of pertussis. CSTE National Meeting 1997 NY Saratoga Springs
8. Marchant CD, Loughlin AM, Lett SM, et al. Pertussis in Massachusetts, 1981–1991: incidence, serologic diagnosis, and vaccine effectiveness. J Infect Dis. 1994;169:1297–1305
9. Mancini N, Carletti S, Ghidoli N, et al. The era of molecular and other non-culture-based methods in diagnosis of sepsis. Clin Microbiol Rev. 2010;23:235–251
10. Wolk DM, Dunne WM Jr. New technologies in clinical microbiology. J Clin Microbiol. 2011;49:S62–67
11. Cronquist AB, Mody RK, Atkinson R, et al. Impacts of culture-independent diagnostic practices on public health surveillance for bacterial enteric pathogens. Clin Infect Dis. 2012;54 Suppl 5:S432–S439
12. Massay SC.. A survey of state public health laboratories for pertussis diagnostics. Lab Medicine. 2007;38:169–171
13. Cherry JD, Grimprel E, Guiso N, et al. Defining pertussis epidemiology: clinical, microbiologic and serologic perspectives. Pediatr Infect Dis J. 2005;24(5 Suppl):S25–S34
14. Shakib JH, Wyman L, Gesteland PH, et al. Should the pertussis case definition for public health reporting be refined? J Public Health Manag Pract. 2009;15:479–484
15. Cherry JD, Tan T, von Konig CHW, et al. Clinical definitions of pertussis: summary of a Global Pertussis Initiative roundtable meeting, February 2011. Clin Infect Dis. 2012;54:1756–1764
16. Council of State and Territorial Epidemiologists (CSTE). . CSTE Position Statement 13-ID-15: revision of the pertussis surveillance case definition to more accurately capture the burden of disease among infants <1 year of age. CSTE National Meeting 2013 Pasadena, CA
17. CDC. . Outbreaks of respiratory illness mistakenly attributed to pertussis—New Hampshire, Massachusetts, and Tennessee, 2004–2006. MMWR. 2007;56:837–842
18. Mandal S, Tatti KM, Woods-Stout D, et al. Pertussis pseudo-outbreak linked to specimens contaminated by Bordetella pertussis
DNA from clinic surfaces. Pediatrics. 2012;129:e424–e430
19. Reischl U, Lehn N, Sanden GN, Loeffelholz MJ.. Real-time PCR assay targeting IS81 of Bordetella pertussis
and molecular basis for detecting Bordetella holmesii
. J Clin Microbiol. 2001;39:1963–1966
20. Register KB, Sanden GN.. Prevalence and sequence variants of IS481 in Bordetella bronchiseptica: implications for IS481-based detection of Bordetella pertussis
. J Clin Microbiol. 2006;44:4577–4583
21. Rodgers L, Martin SW, Cohn A, et al. Epidemiologic and laboratory features of a large outbreak of pertussis-like illnesses associated with cocirculating Bordetella holmesii and Bordetella pertussis—
Ohio, 2010–2011. Clin Infect Dis. 2013;56:322–331
22. Burgos-Rivera B, Lee AD, Bowden KE, et al. Evaluation of level of agreement in Bordetella
species identification in three U.S. laboratories during a period of increased pertussis. J Clin Microbiol. 2015;53:1842–1847
23. Tatti KM, Sparks KN, Boney KO, et al. Novel multitarget real-time PCR assay for rapid detection of Bordetella
species in clinical specimens. J Clin Microbiol. 2011;49:4059–4066
25. Acosta A, Yasmin S, Ejigiri G, et al. Increasing use of serology in pertussis diagnosis, Arizona [abstract 1663]. 2013 Pasadena, CA Council of State and Territorial Epidemiologists annual meeting
26. Menzies SL, Kadwad V, Pawloski LC, et al.Pertussis Assay Working Group. Development and analytical validation of an immunoassay for quantifying serum anti-pertussis toxin antibodies resulting from Bordetella pertussis
infection. Clin Vaccine Immunol. 2009;16:1781–1788
27. Quinn HE, McIntyre PB.. Pertussis epidemiology in Australia over the decade 1995–2005 – trends by region and age group. CDI. 2007;31:105–116
28. Tondella ML, Carlone GM, Messonnier N, et al. International Bordetella pertussis assay standardization and harmonization meeting report. 2007 Atlanta, Georgia, United States Centers for Disease Control and Prevention
29. Guiso N, Berbers G, Fry NK, et al.EU Pertstrain group. What to do and what not to do in serological diagnosis of pertussis: recommendations from EU reference laboratories. Eur J Clin Microbiol Infect Dis. 2011;30:307–312
30. Schmidtke AJ, Boney KO, Martin SW, Skoff TH, Tondella ML, Tatti KM.. Population diversity among Bordetella pertussis
isolates, United States, 1935–2009. Emerg Infect Dis. 2012;18:1248–1255
31. Cassiday PK, Skoff TH, Faulkner AE, Connelly L, Jawahir S, Tondella ML.. Temporal changes in the predominance of pulsed-field gel electrophoresis profiles of Bordetella pertussis isolates from the United States, 2000–2012 [abstract P-12]. 2013 Dublin, Ireland
32. Pawloski LC, Queenan AM, Cassiday PK, et al. Prevalence and molecular characterization of pertactin-deficient Bordetella pertussis
in the United States. Clin Vaccine Immunol. 2014;21:119–125