A Ten-Year Case-Control Study of Passive Smoke Exposure as a Risk Factor for Pertussis in Children

Mark A Schmidt, PhD, MPH; Samantha K Kurosky, MS; John P Mullooly, PhD;
Colleen Chun, MD; Sheila Weinmann, PhD

Perm J 2015 Summer; 19(3):59-63 [Full Citation]



Objectives: This retrospective study was initiated in febrile neutropenic inpatients with solid tumors in 4 community hospitals, to discover how the Multinational Association for Supportive Care in Cancer (MASCC) risk index score (RIS) of 21 or greater correlated with complications occurring in 198 episodes: whether it could help determine which patients not to admit, the savings of not admitting patients without complications, and whether an algorithm could facilitate management of those not admitted.
Methods: Febrile neutropenic episodes in patients with solid tumors were identified electronically between October 1, 2008, and November 15, 2010. Electronic charts were reviewed manually for inclusion criteria and data extraction. Episodes were stratified by an MASCC RIS below 21 or 21 or greater. Complications were correlated with the index.
Results: Inclusion criteria were met in 198 episodes. Sensitivity, specificity, and positive and negative predictive values of the MASCC RIS vs complications were 94%, 29.6%, 57.7%, and 82.9%, respectively. In episodes with an RIS 21 or greater, 42.3% had complications, misclassifying to low risk 69 episodes with complications. "Unable to eat" correlated with complications in 84% of episodes. In 3 patients stratified to no complication, a complication developed 24 hours after admission.
Conclusions: An MASCC RIS of 21 or greater could not be used as a criterion for "no complication/do not admit." Inability to eat should be an admission criterion. Savings of approximately $1 million per 100 uncomplicated admissions could be realized if appropriate criteria for nonadmission could be devised. An algorithm to facilitate outpatient management is suggested.


Pertussis, or whooping cough, is a respiratory tract infection and a major public health burden because of its high infectivity and severe manifestations among infants. Pertussis is caused by the bacterial species, Bordetella pertussis, which is transmitted from person to person via aerosolized droplets. Rates of childhood immunization against pertussis in the US are high, with more than 83% of children younger than age 3 years having received 4 doses of pertussis antigen-containing vaccine.1 However, children too young to be fully immunized, older underimmunized children, and immunized teenagers and adults with waned immunity are at risk of pertussis infection acquired from symptomatically infected individuals. Whereas adolescents and adults may have a relatively mild illness, pertussis in infants is particularly severe. According to the Centers for Disease Control and Prevention in 2011, infants younger than age 2 months who were too young to be immunized accounted for 57% of all infant hospitalizations and 85% of all infant deaths due to pertussis in the US.2

Passive cigarette smoke exposure (PSE) in the household as a possible risk factor for pertussis infection among children has not been well characterized in the literature. Our prior work investigating the impact of PSE on pneumococcal infection3 prompted us to conduct a case-control study within the population of a large managed care organization to compare household smoking exposure histories recorded in the electronic medical record (EMR) of pediatric pertussis cases with those of matched controls.


We conducted our study among the member population of the Kaiser Permanente Northwest (KPNW) Health Plan, using a similar protocol and the same control group described for our investigation of PSE as a risk factor for invasive pneumococcal disease.3 We selected pertussis cases occurring from January 1, 1996, through December 31, 2005, among KPNW members from birth through age 12 years. Potential cases were identified from either of the following: 1) a microbiology laboratory database as having pertussis confirmed by culture, direct fluorescent-antibody testing, or polymerase chain reaction or 2) the EMR as having a pertussis-related code from the International Classification of Diseases, Ninth Revision (ICD-9; 033, 033.0, 033.8, 033.9, and 484.3). When ICD-9 codes were used for case identification, we manually reviewed the EMR to exclude those without documented laboratory confirmation of pertussis infection.

Where possible, we used our control population from the study of invasive pneumococcal disease. Among this group, we determined which potential controls were closest in age to our cases on their "reference date," defined as the collection date of the pertussis-positive specimen from the case, and matched 1 control to each pertussis case, by age group (0-2, 3-5, and 6-12 years) and type of Health Plan account. We matched controls and cases on the basis of the reference date to avoid confounding the relationship between smoking exposure and pertussis occurrence by secular changes in pertussis incidence and capture of smoking history within the EMR. Because the study design involved collecting PSE information from the EMR of family members if available to us, we matched group members based on Health Plan account (child plus family members or child alone) to equalize our access to the family member's EMR between cases and their matched controls. For 11 pertussis cases without a suitable control from this control population, we randomly selected 1 control for each case from the KPNW member population, using the same criteria.

We reviewed the health records of all cases and new controls, plus the records of their family members where available, to collect PSE history, using standardized data collection instruments. We categorized each case and control as definitely exposed, probably exposed, probably not exposed, definitely not exposed, or unknown (Table 1). For all cases and controls, we collected information about receipt of pertussis antigen-containing vaccines before the collection date of the pertussis-positive specimen (cases) or equivalent reference date (controls) through electronic abstraction of the Vaccine Safety Datalink vaccine dataset.4,5 We defined vaccine exposure as "vaccinated" if the subject had received at least one pertussis antigen-containing vaccine before the reference date or "unvaccinated" if they had not. We determined whether cases and controls were up to date with the receipt of pertussis antigen-containing vaccines, as appropriate for age, using Advisory Committee on Immunization Practices recommendations.6

We constructed conditional logistic regression models to investigate the relationship between PSE and pertussis. Our main model included only the matched case-control pairs for whom PSE history was categorized for both members of the pair. Because of the small number of study participants, we considered participants to be "exposed" if they were categorized as "definitely exposed" or "probably exposed." We considered participants to be "unexposed" if they were categorized as "definitely not exposed" or "probably not exposed." Variables evaluated for confounding in multivariable models included sex, race, history of being breastfed, daycare attendance, and receipt of pertussis antigen-containing vaccine before the reference date. The latter was retained in the final multivariable model.

To account for any possible selection bias related to the exclusion of case-control pairs in which at least one of the members had an unknown PSE history, we conducted a sensitivity analysis, including all case-control pairs in two separate models. For Model 1, we compared the combination of unexposed and unknown pairs with exposed pairs; for Model 2, we compared the combination of unexposed pairs with exposed and unknown pairs. We included vaccine history in both models to adjust for potential confounding.

This protocol was reviewed and approved by the KPNW institutional review board.



We identified 65 laboratory-confirmed pertussis cases meeting our criteria, 33 of which occurred in infants younger than 1 year of age. Positive test results for pertussis included 53 cases confirmed by culture, 5 by direct fluorescent-antibody testing, 4 by culture and direct fluorescent-antibody testing, and 3 by polymerase chain reaction. Table 2 describes demographic and exposure information of pertussis cases and matched controls. Cases and controls were similar with respect to sex (p = 0.73) and white vs nonwhite race (p = 0.57). A similar proportion of cases (68%; n = 44) and controls (78%; n = 51) had received a pertussis antigen-containing vaccine (p = 0.24). Among those vaccinated, a similar proportion of cases (77%; n = 34) and controls (40; 78%) were up to date with pertussis vaccination for their age (p = 0.89). Because cases and controls were similar by up-to-date vaccination status and because our small sample size would potentially limit our ability to consider additional variables in our main model, we decided to consider only whether our subjects were vaccinated or unvaccinated for the remainder of the analysis. A higher proportion of cases than controls was categorized as exposed to PSE (37% vs 27%), but the control group had a higher proportion of subjects with missing data on PSE (18% in controls vs 8% in cases).

For 15 (23%) of the case-control pairs, 1 or both members had unknown PSE, and these pairs were excluded from our main-effects analysis. Among the remaining 50 case-control pairs, 28 (56%) had concordant case-control exposure histories (7 in which both members were exposed and 21 in which both members were unexposed). In our conditional logistic regression model, which calculated the odds ratio (OR) using data from the 22 discordant pairs, we found no statistically significant relationship between pertussis and PSE in the household (OR = 1.2; 95% confidence interval [CI] = 0.5-2.8; Table 3). Our results remained similar after adjusting for receipt of pertussis antigen-containing vaccine (adjusted OR = 1.2; 95% CI = 0.5-2.9). We included the 15 case-control pairs with at least 1 member with unknown PSE in our sensitivity analyses. In Model 1, we placed those with missing PSE data in the unexposed category; in Model 2, we placed them in the exposed category. Thus, the range of ORs that could have resulted from this study if there had been no missing data was 1.1 to 1.7.




We did not detect an association between laboratory-confirmed pertussis and PSE in the household in this analysis; however, this remains an important research question to study. In an in vitro model, cigarette smoke exposure was associated with an increased number of B pertussis organisms binding to buccal epithelial cells from nonsmokers compared with unexposed cells.7 Among cigarette smokers, studies have described decreased mucociliary clearance,8 decreased levels of circulating immunoglobulins, decreased natural killer cell activity, depressed neutrophil chemotaxis and phagocytic activity, and decreased release of pro-inflammatory cytokines.9 Cigarette smoking also is associated with increased permeability of the respiratory epithelium.10 Further research is needed to investigate whether similar immune alterations occur in children exposed to cigarette smoke.

Among children, PSE has been associated with an increased risk of invasive bacterial infections from Streptococcus pneumoniae,11,12 Neisseria meningitidis,13-18 Haemophilus influenzae type b,18,19 pulmonary Mycobacterium tuberculosis infection,20 and viral infections, including respiratory syncytial virus.21 In addition, smokers with pertussis infection are more infectious than nonsmokers because of greater severity and duration of cough.22 This is especially important, as household members are the estimated source for 60% to 83% of infant cases of pertussis, and adults have been responsible for introducing pertussis into the household in 26% of household outbreaks and for 42% of all household secondary cases.23-26

Our main study strengths were the use of information from EMR and laboratory records, reducing exposure misclassification and outcome misclassification, and a study period that extended over 10 years. Our main limitation was the small sample size of the study population because of a low incidence of pertussis during the study period.27 Our sample size was further restricted by incomplete household PSE history for some study participants, although we accounted for this through our sensitivity analysis and this limitation did not alter our findings.


Pertussis is now considered a resurgent disease; its incidence has increased overall during the past 3 decades, with more notable increases and multiple outbreaks since the mid- to late 2000s.28 In 2012, Oregon had an outbreak of pertussis and recorded more than 900 cases (23.3 cases per 100,000)—its highest occurrence since 1953.29 The incidence rate was highest among infants (253/100,000 persons), followed by children aged 10 to 14 years (104/100,000), aged 1 to 4 years (81/100,000), and aged 5 to 9 years (67/100,000). Twenty-six infants had severe disease requiring hospitalization. If PSE is linked to increased risk or severity of pertussis in childhood, this would allow targeted education for parents about eliminating household PSE and would assist clinicians in assessing the child's risk for pertussis infection. Improved documentation of PSE in the EMR would assist future studies of the relationship between PSE and pertussis risk in children.

Disclosure Statement

The author(s) have no conflicts of interest to disclose.

Financial support: Centers for Disease Control and Prevention (through agreement 200-2002-00732 with America's Health Insurance Plans [AHIP]). The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the funding agency. The findings and conclusions do not necessarily represent the views or policies of the Department of Health and Human Services or AHIP, nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government.


Kathleen Louden, ELS, of Louden Health Communications provided editorial assistance.

1.    Elam-Evans LD, Yankey D, Singleton JA, Kolasa M; Centers for Disease Control and Prevention. National, state, and selected local area vaccination coverage among children aged 19-35 months—United States, 2013. MMWR Morb Mortal Wkly Rep 2014 Aug 29;63(34):741-8.
    2.    Terranella A, Asay GR, Messonnier ML, Clark TA, Liang JL. Pregnancy dose Tdap and postpartum cocooning to prevent infant pertussis: a decision analysis. Pediatrics 2013 Jun;131(6):e1748-56. DOI: https://doi.org/10.1542/peds.2012-3144.
    3.    Chun CS, Weinmann S, Riedlinger K, Mullooly JP. Passive cigarette smoke exposure and other risk factors for invasive pneumococcal disease in children: a case-control study. Perm J
2015 Winter;19(1):38-43. DOI: https://doi.org/10.7812/TPP/14-010.
    4.    Chen RT, Glasser JW, Rhodes PH, et al. Vaccine Safety Datalink project: a new tool for improving vaccine safety monitoring in the United States. The Vaccine Safety Datalink Team. Pediatrics 1997 Jun;99(6):765-73. DOI: https://doi.org/10.1542/peds.99.6.765.
    5.    Chen RT, DeStefano F, Davis RL, et al. The Vaccine Safety Datalink: immunization research in health maintenance organizations in the USA. Bull World Health Organ 2000;78(2):186-94.
    6.    Diphtheria, tetanus, and pertussis: recommendations for vaccine use and other preventive measures. Recommendations of the Immunization Practices Advisory committee (ACIP). MMWR Recomm Rep 1991 Aug 8;40(RR-10):1-28.
    7.    El Ahmer OR, Essery SD, Saadi AT, et al. The effect of cigarette smoke on adherence of respiratory pathogens to buccal epithelial cells. FEMS Immunol Med Microbiol 1999 Jan;23(1):27-36. DOI: https://doi.org/10.1111/j.1574-695X.1999.tb01713.x.
    8.    Sherman CB. The health consequences of cigarette smoking. Pulmonary diseases. Med Clin North Am 1992 Mar;76(2):355-75.
    9.    Arcavi L, Benowitz NL. Cigarette smoking and infection. Arch Intern Med 2004 Nov 8;
164(20):2206-16. DOI: https://doi.org/10.1001/archinte.164.20.2206.
    10.    Jones JG, Minty BD, Lawler P, Hulands G, Crawley JC, Veall N. Increased alveolar epithelial permeability in cigarette smokers. Lancet 1980 Jan 12;1(8159):66-8. DOI: https://doi.org/10.1016/S0140-6736(80)90493-6.
    11.    Gessner BD, Ussery XT, Parkinson AJ, Breiman RF. Risk factors for invasive disease caused by Streptococcus pneumoniae among Alaska native children younger than two years of age. Pediatr Infect Dis J 1995 Feb;14(2):123-8.

DOI: https://doi.org/10.1097/00006454-199502000-00008.
    12.    O'Dempsey TJ, McArdle TF, Morris J, et al. A study of risk factors for pneumococcal disease among children in a rural area of West Africa. Int J Epidemiol 1996 Aug;25(4):885-93. DOI: https://doi.org/10.1093/ije/25.4.885.
    13.    Fischer M, Hedberg K, Cardosi P, et al. Tobacco smoke as a risk factor for meningococcal disease. Pediatr Infect Dis J 1997 Oct;16(10):979-83. DOI: https://doi.org/10.1097/00006454-199710000-00015.
    14.    Haneberg B, Tønjum T, Rodahl K, Gedde-Dahl TW. Factors preceding the onset of meningococcal disease, with special emphasis on passive smoking, symptoms of ill health. NIPH Ann 1983 Dec;6(2):169-73.
    15.    Kriz P, Bobak M, Kriz B. Parental smoking, socioeconomic factors, and risk of invasive meningococcal disease in children: a population based case-control study. Arch Dis Child 2000 Aug;83(2):117-21. DOI: https://doi.org/10.1136/adc.83.2.117.
    16.    Lee CC, Middaugh NA, Howie SR, Ezzati M. Association of secondhand smoke exposure with pediatric invasive bacterial disease and bacterial carriage: a systematic review and meta-analysis. PLoS Med 2010 Dec 7;7(12):e1000374. DOI: https://doi.org/10.1371/journal.pmed.1000374.
    17.    Murray RL, Britton J, Leonardi-Bee J. Second hand smoke exposure and the risk of invasive meningococcal disease in children: systematic review and meta-analysis. BMC Public Health 2012 Dec 10;12:1062. DOI: https://doi.org/10.1186/1471-2458-12-1062.
    18.    Pereiró I, Díez-Domingo J, Segarra L, Ballester A, Albert A, Morant A. Risk factors for invasive disease among children in Spain. J Infect 2004 May;48(4):320-9. DOI: https://doi.org/10.1016/j.jinf.2003.10.015.
    19.    Arnold C, Makintube S, Istre GR. Day care attendance and other risk factors for invasive Haemophilus influenzae type b disease. Am J Epidemiol 1993 Sep 1;138(5):333-40.
    20.    Altet MN, Alcaide J, Plans P, et al. Passive smoking and risk of pulmonary tuberculosis in children immediately following infection. A case-control study. Tuber Lung Dis 1996 Dec;77(6):537-44. DOI: https://doi.org/10.1016/S0962-8479(96)90052-0.
    21.    DiFranza JR, Masaquel A, Barrett AM, Colosia AD, Mahadevia PJ. Systematic literature review assessing tobacco smoke exposure as a risk factor for serious respiratory syncytial virus disease among infants and young children. BMC Pediatr 2012 Jun 21;12:81. DOI: https://doi.org/10.1186/1471-2431-12-81.
    22.    De Serres G, Shadmani R, Duval B, et al. Morbidity of pertussis in adolescents and adults. J Infect Dis 2000 Jul;182(1):174-9. DOI: https://doi.org/10.1086/315648.
    23.    Baptista PN, Magalhães VS, Rodrigues LC. The role of adults in household outbreaks of pertussis. Int J Infect Dis 2010 Feb;14(2):e111-4. DOI: https://doi.org/10.1016/j.ijid.2009.03.026.
    24.    de Greeff SC, Mooi FR, Westerhof A, et al. Pertussis disease burden in the household: how to protect young infants. Clin Infect Dis 2010 May 15;50(10):1339-45. DOI: https://doi.org/10.1086/652281.
    25.    Bisgard KM, Pascual FB, Ehresmann KR, et al. Infant pertussis: who was the source? Pediatr Infect Dis J 2004 Nov;23(11):985-9. DOI: https://doi.org/10.1097/01.inf.0000145263.37198.2b.
    26.    Wendelboe AM, Njamkepo E, Bourillon A, et al; Infant Pertussis Study Group. Transmission of Bordetella pertussis to young infants. Pediatr Infect Dis J 2007 Apr;26(4):293-9. DOI: https://doi.org/10.1097/01.inf.0000258699.64164.6d.
    27.    Bancroft J, Byster L. Selected reportable communicable disease summary: 2008-2009—Pertussis [Internet]. Salem, OR: Oregon Department of Human Services; 2010 Dec [cited 2015 May 13]. Available from: https://public.health.oregon.gov/DiseasesConditions/CommunicableDisease/DiseaseSurveillanceData/AnnualReports/arpt0809/Documents/arpt0809.pdf.
    28.    Clark TA. Changing pertussis epidemiology: everything old is new again. J Infect Dis 2014 Apr 1;209(7):978-81. DOI: https://doi.org/10.1093/infdis/jiu001.
    29.    Selected reportable communicable disease summary: 2012 state of Oregon [Internet]. Portland, OR: Oregon Health Authority; 2013 Sep [cited 2014 Apr 3]. Available from: http://public.health.oregon.gov/DiseasesConditions/CommunicableDisease/DiseaseSurveillanceData/AnnualReports/arpt2012/Pages/index.aspx.


Click here to join the eTOC list or text ETOC to 22828. You will receive an email notice with the Table of Contents of The Permanente Journal.


2 million page views of TPJ articles in PubMed from a broad international readership.


Indexed in MEDLINE, PubMed Central, EMBASE, EBSCO Academic Search Complete, and CrossRef.




ISSN 1552-5775 Copyright © 2021 thepermanentejournal.org

All Rights Reserved