In the prevaccine era pertussis was a major cause of mortality among children in the United States with an average of >160 000 cases and 5000 deaths reported annually in the 1920s and 1930s. Although reported pertussis cases fell substantially after introduction of whole cell pertussis vaccines in the 1940s, cases continued at low levels despite high rates of coverage. After a nadir of 1016 reported cases in 1976, the incidence of reported pertussis progressively increased. 1 Analyses for the 1980s 2 and the first half of the 1990s 1 highlighted a rapid increase in adolescents and adults; both a real increase in pertussis incidence and improved ascertainment and reporting were thought to have contributed to the increase. 1, 2 A recent analysis found a major increase in the incidence of pertussis among infants <4 months of age since 1980. 3 We reviewed pertussis deaths reported to CDC in the 1990s to characterize the clinical course of the illness and factors potentially associated with a fatal outcome.
MATERIALS AND METHODS
Sources of data
We identified patients who were reported to have had pertussis with a fatal outcome in the 1990s by reviewing reports to the CDC through both the National Notifiable Diseases Surveillance System (NNDSS) and the Supplementary Pertussis Surveillance System (SPSS) from state and local health departments. Surveillance personnel in all states were asked to verify cases and to identify previously unreported deaths attributed to pertussis. To determine characteristics of the fatal cases, we reviewed pertussis case reports and available medical records (hospital discharge summaries, death certificates, and autopsy reports). We compared these data to data on fatal cases of pertussis reported to CDC in the 1980s. 2
A death attributed to pertussis was defined as confirmed if: (1) Bordetella pertussis was isolated from a clinical specimen by culture; or (2) the case was clinically compatible with pertussis (paroxysmal cough, apnea or acute respiratory decompensation in individuals with chronic respiratory disease) and had supporting laboratory or epidemiologic evidence. The supporting evidence accepted included at least one of the following: a positive direct fluorescent antibody test result; a positive PCR test result or a history of exposure to a culture-positive or PCR-positive patient (epidemiologically linked); or autopsy findings consistent with pertussis.
Pertussis deaths that occurred from January 1, 1990, through December 31, 1999, were included. Clinical diagnoses, such as pneumonia, apnea and encephalopathy, were accepted as recorded on the case report form or in the medical record.
An infant was defined as a child <12 months of age.
Evaluation of completeness of reporting
Capture-recapture analyses are used to estimate the total number of cases of a disease in a population. 4 The estimate results from an analysis of the proportion of overlapping and nonoverlapping cases reported to two (or more) independent reporting systems. To estimate the total number of pertussis deaths in the United States, we performed a capture-recapture analysis using the Chapman method. 5 We compared fatal cases reported to CDC, with pertussis deaths identified through death certificate files for the years 1990 through 1999 maintained by the National Center for Health Statistics (NCHS). For NCHS deaths data were examined on deaths for which pertussis (ICD-9-CM International Classification of Diseases, 9th revision, Clinical Modification code 339) was listed on multiple cause of death tapes. To identify patients reported to both systems (overlapping cases) and cases reported to one but not both systems, the following matching criteria were applied: year and month of death; state and county of residence; state and county of death; age; and sex. To correspond with the age variable in the NCHS dataset, the age data in the CDC database was adjusted from age at onset to age at death using data from the case reports. We calculated a goodness-of-fit-based confidence interval 6 for the capture-recapture estimate of the total number of pertussis deaths. Completeness of reporting was calculated as the ratio of the number of reported deaths to the estimated total number of pertussis deaths.
Pertussis mortality rates
We estimated the average annual pertussis mortality rate by dividing the number of pertussis deaths reported to the CDC in the 1990s by the sum of mid-year postcensal estimates of the US population of infants during 1990 through 1999. 7 We performed Poisson regression analyses 8 using SAS software 9 to: (1) estimate rate ratios for comparison of rates by race group, ethnicity group and calendar year; (2) calculate likelihood ratio-based confidence intervals for the pertussis mortality rates and the estimated rate ratios; and (3) model mortality rates by year as a continuous variable.
One hundred three deaths attributed to pertussis were identified in the 1990s; the average annual pertussis mortality rate was 0.04 death per million population. Laboratory confirmation of pertussis was obtained by culture in 65 cases, by direct fluorescent antibody test in 29 cases, by PCR in 6 cases, by an epidemiologic link to a culture- or PCR-positive case in 2 cases and by histopathologic findings only in 1 case. Autopsy was performed in 52 of 93 cases with available information. The deaths were evenly divided among male (51 deaths) and female (52 deaths) cases.
Of the 103 patients who died 93 (91%) were infants, including 84 (82%) who were <4 months of age (Table 1). The average annual pertussis mortality rate among infants was 2.40 deaths per million. In the 1980s 61 infant deaths (1.67 per million infants) were reported. 2 Compared with the 1980s, the number of infant deaths increased by 52% and the infant pertussis mortality rate increased by 44%; the case-fatality ratio (CFR) was 0.5% in both decades. Within the 1990s we estimated that the mortality rate increased by 9.0% per year [95% confidence interval (95% CI) 1.5 to 17.2] (Table 2).
Reported infant pertussis deaths in the 1990s were concentrated during the winter months. Of 93 fatal infant cases, 43 (46%) had onset of pertussis from December through March (Fig. 1); during these months the CFR was 0.9% compared with a CFR of 0.4% in other months (P < 0.01).
Race and ethnicity, vaccination status
Of 97 (78%) fatal cases with available information on race, 76 were identified as white, and 33 (34%) of 98 individuals with ethnicity data were Hispanic, including 31 (35%) of 89 infants (Table 2). For cases with information on race group, the overall mortality rate was 2.27 (95% CI 1.83 to 2.78) deaths per million infants. The mortality rate for Native Americans was significantly higher than that for whites, with an estimated rate ratio (RR) of 4.33 (95% CI 1.32 to 10.4). For pertussis deaths with information on ethnicity, the overall mortality rate was 2.30 (95% CI 1.85 to 2.81) deaths per million infants. Hispanics had a higher mortality rate than non-Hispanics (RR 2.65; 95% CI 1.69 to 4.07).
Collection of data on race and ethnicity for pertussis cases reported to the NNDSS and SPSS began in 1995, and the data were largely incomplete from most states. Data from California were complete for 1631 of 1739 (93.8%) cases from 1995 to 1999. In California the case-fatality ratio for pertussis cases among Hispanic infants was not significantly different from the rate among non-Hispanic infants (6 deaths among 908 cases, 0.7%vs. 4 deaths among 723 cases, 0.6%).
Among the 93 reported infant cases, the number of doses of pertussis toxoid-containing vaccine received was 0 dose for 83 infants (including 3 of 4 infants ≥6 months of age), 1 dose for 7 infants, 2 doses for 1 infant and 3 doses for 1 infant. The vaccination history was unknown for 1 infant. Five of the 8 children ≥1 year of age had received ≤3 doses (Table 3); none of the 5 children, including 3 with severe but stable neurologic conditions (cerebral palsy, 2; spinal-muscular atrophy, 1) had a contraindication to pertussis vaccination. Vaccination histories were unknown for the 2 adult cases. Nine of the 10 noninfant cases had severe preexisting health conditions.
Thirty-nine (51%) of 76 fatal infant cases with data available were born at <37 weeks gestational age, including 22 (29%) who were born at <35 weeks gestational age. Among the subgroup of infants with onset of pertussis at 2 to 11 months of age, 11 (48%) of 23 with data available were born at <35 weeks gestational age. Eleven fatal cases were among infants from twin or triplet births including 7 born at a gestational age of <37 weeks of age.
Laboratory results and coinfection
Elevated white blood cell counts (WBC) were frequent. At least 1 value for a WBC was available for 73 patients; differential counts were available for 53 patients. The median WBC was 62 000, and the median percentage of lymphocytes was 49%.
The results of testing for respiratory syncytial virus (RSV) were available for 55 of 93 infants. Eight of 55 (15%) infants were positive for RSV; 2 additional infants had RSV infections confirmed within the 3 weeks preceding presentation with pertussis. All 10 cases with acute or recent RSV infections had onset of pertussis between December and March. Eight infants (including 1 with RSV infection) had confirmation of other viral infections (influenza, 3 cases; parainfluenza, 2 cases; cytomegalovirus, 2 cases; adenovirus, 1 case).
Eight infants were documented to have additional bacterial pathogens in cultures from blood or autopsy lung tissue within the first seven days of hospitalization (S treptococcus pneumoniae, 5 cases; Haemophilus influenzae, 1 case; Escherichia coli, 1 case; Pseudomonas aeruginosa, 1 case.). All infants whose courses were complicated by infection with S.pneumoniae or H. influenzae had onset of pertussis between November and February.
All fatal cases developed apnea, severe respiratory difficulty or both. Four infants were declared dead in prehospital settings. Of the 99 patients admitted to hospitals, 89 were placed on mechanical ventilation during their hospitalization; 3 patients with severe underlying medical problems were not mechanically ventilated, and for 7 patients, available records are insufficient to confirm mechanical ventilation. Apnea was reported in 68 (73%) of 93 infants before or soon after admission (Table 4). Eighty-five (86%) of 99 patients had infiltrates on chest radiograph at admission (Table 4); the infiltrates were frequently described as “streaky,” “interstitial” or “atelectatic-like.”
Pulmonary hypertension was diagnosed in 38 (41%) of fatal infant cases with supportive findings on echocardiogram (31 cases), autopsy (3 cases) and clinical examination (2 cases) or by hemodynamic monitoring (1 case); for 1 case medical records were insufficient to establish the basis for the diagnosis. The echocardiograms of 7 other infants did not show pulmonary hypertension, 27 infants did not have an echocardiogram and available records were inadequate to establish whether echocardiograms were performed for 21 infants. From 1990 through 1994, 13 (32%) of 41 infants had documented pulmonary hypertension compared with 23 (44%) of 52 cases between 1995 and 1999. Pulmonary hypertension was recognized more often among full-term infants than among infants born prematurely [54% (20 of 37) vs. 28% (11 of 39), P = 0.04)]. Pulmonary hypertension was frequently severe, rapidly progressive, and refractory to nitric oxide therapy. Extracorporeal membrane oxygenation (ECMO) was used in 18 patients, including 14 patients with documented pulmonary hypertension.
Among infants the median duration of symptoms before hospitalization was 6 days (interquartile range, 3 to 10 days); median survival once hospitalized was 3 days (interquartile range, 2 to 11 days) for infants who did not receive ECMO and 21 days (interquartile range 8 to 34 days) for infants who received ECMO.
Source of infection
Report of previous exposure to a person with cough illness was documented for 46 (49%) infants; sources of exposure included parents (18 cases), both parents and siblings (12 cases), siblings (10 cases), parents and other household members (3 cases) and other individuals in the household (3 cases).
Completeness of reporting
During 1990 through 1999, 103 pertussis deaths were reported to CDC and 82 pertussis deaths were recorded on death certificates sent to NCHS. A total of 53 cases were reported to both institutions. Based on these matches, the capture-recapture analysis estimated that a total of 159 pertussis deaths (95% CI 145 to 181) may have occurred in the United States during 1990 through 1999. The estimated completeness of reporting of pertussis deaths to the CDC was 65%, to the NCHS was 52%, and to the 2 systems combined was 83% (Table 5). There was a modest increase in the estimated completeness of reporting to the CDC between the periods 1990 through 1994 (61%) and 1995 through 1999 (70%).
In the United States infants, especially very young infants, account for most cases of severe pertussis and its associated complications and mortality. 10 Most infants <4 months of age are too young to be directly protected against pertussis by vaccination because at least two to three doses of pertussis toxoid-containing vaccine are required to provide effective protection. The current US infant vaccination schedule recommends three doses of pertussis toxoid-containing vaccine at 2, 4 and 6 months of age.
In the 1990s the incidence of reported pertussis among infants in the US increased by 49% compared with the 1980s, and the increased incidence was almost exclusively among infants <4 months of age. 3 Our analysis of reported pertussis deaths in the US suggests that pertussis deaths have increased among infants <4 months of age in parallel with the increase in reported cases. Pertussis mortality rates continued to increase through the 1990s and in 2000, 17 pertussis deaths among infants were reported. 11 These data suggest that the US has experienced increased circulation of B. pertussis. Other developed countries have recently reported increasing rates of pertussis 12, 13 and an increasing number of pertussis deaths 14 among young infants.
Among older infants and preschool children in the United States, reported pertussis deaths declined in the 1990s. The very low number of pertussis deaths among older infants and children in the absence of severe underlying illness is likely a result of improved immunization rates among infants in the 1990s 15 and improved clinical care of infants and children with severe pneumonia. Improved recognition and reporting of pertussis among young infants may have contributed to the rise in reported cases and fatalities among infants <4 months of age in the 1990s. However, (1) the stable case-fatality rate for infant pertussis cases and (2) the stable reported incidence of pertussis and decline in reported fatalities among infants ≥4 months of age suggest that the increase in pertussis among young infants may not have been primarily a result of improved reporting.
Our estimate of the completeness of reporting of pertussis deaths to CDC during 1990 through 1999 that we found (65%) was higher than previous estimates (29% for 1989 through 1993 16 and 33% for 1985 through 1988 17) because we used different matching criteria. In particular we modified the age variable in the CDC data to correspond with the age variable in the NHCS database. This resulted in more matches (cases reported to both systems), and thus a lower estimate of the number of unreported deaths and a higher estimate of completeness of reporting. When we used age at onset in the CDC data to be consistent with previous analyses, we obtained an estimate of 35%. Thus our higher estimate with use of the modified age variable represents an improved estimate, not a true increase in reporting.
The capture-recapture estimate of the total number of pertussis deaths during the 1990s had several limitations. Passive surveillance systems such as the CDC pertussis surveillance systems and NCHS death certificate files will undercount pertussis deaths and are not likely to be as sensitive as active surveillance systems or enhanced laboratory surveillance systems. 18 In addition it is likely that we underestimated the number of unreported pertussis deaths because reporting to the NNDSS-SPSS system may not be independent of reporting to the NCHS data system and capture-recapture analyses using nonindependent data sources result in underestimates of total cases. 4 Although these limitations may have resulted in an underestimate of the overall pertussis mortality rate, our estimates of the completeness of reporting were relatively stable throughout the 1990s, also suggesting that the increased pertussis mortality for the 1990s may not have been primarily from increased reporting.
Pertussis deaths may go unreported because the diagnosis of pertussis is not considered or because routinely available diagnostic tests for pertussis (culture and direct fluorescent antibody) are insensitive. In our series pertussis was unsuspected at the time of death in three of the four out-of-hospital deaths and in several in-hospital fatalities. The clinical presentation of severe pertussis is often atypical in infants, and especially in neonates. 19–22 Apnea-bradycardia episodes can be the presenting sign, and cough can be minimal or absent. Pertussis has been found to be the cause of death in some cases of sudden unexpected infant deaths. 23 The increasing availability of PCR may eventually improve the diagnostic sensitivity for pertussis cases; however, only four of the deaths among infants in our series were confirmed by PCR, suggesting that PCR played a minor role in the increase in reported infant deaths.
Severe respiratory insufficiency caused by primary pertussis pneumonia, secondary bacterial pneumonia or both is the most common cause of death in cases of severe pertussis. At autopsy widespread mucus plugging and mucosal damage are found in the lungs and airways. 24 Among fatal pertussis cases in the US, evidence of respiratory coinfection or secondary infection was common at or soon after admission. More severe disease from coinfection with respiratory viruses and bacteria may be responsible for the higher number of reported pertussis deaths and the higher pertussis case-fatality ratio among infants in the winter months, although reported cases of pertussis among infants are highest in summer months. 3 Although coinfection was common, no respiratory pathogens other than B. pertussis were identified pre- or postmortem in most infants who died in the first week of hospitalization. Although diagnostic techniques for secondary pneumonia caused by either bacterial or viral pathogens may not be highly sensitive, these data suggest a significant contribution of primary pertussis pneumonia to contemporary pertussis mortality in the US.
On initial presentation infants with severe pertussis are often diagnosed as having presumed sepsis. As illustrated in the case histories, empiric antibiotic regimens used as the initial treatment of presumed sepsis among infants often do not include erythromycin, the preferred antibiotic for treating B. pertussis, or other macrolide antibiotics. However, severe pertussis can be fatal even if a macrolide antibiotic is administered at the time of admission. 11, 25
This case series also suggests the importance of severe pulmonary hypertension leading to hypotension and worsening hypoxia. Severe pulmonary hypertension in infant pertussis has been increasingly recognized. In 1993 fatal pertussis was described in three infants who developed severe pulmonary hypertension, low cardiac output and shock. 26 Other reports of fatal infant pertussis cases with severe pulmonary hypertension have been reported from several countries. 24, 27–31 Data from an international registry on patients receiving ECMO showed the survival rate among 27 children who received ECMO for pertussis (19%) was lower than the survival rate among 1225 children receiving ECMO for respiratory failure from other causes (61%). 29 Among infants with pertussis treated with ECMO, the youngest infants had the worst prognosis. The risk factors and optimal treatment for severe pulmonary hypertension associated with severe pertussis among infants are not clearly defined.
An analysis of pertussis deaths in the United States in 1992 and 1993 suggested that reported prematurity and Hispanic ethnicity may be risk factors for infant deaths. 25 Our study found that a disproportionate number of infant pertussis deaths were associated with these factors throughout the 1990s. A history of reported gestational age <37 weeks was particularly common among fatal cases among infants ≥2 months of age. Our data on gestational age was primarily obtained from discharge summaries and some of the infants in our series may have been misclassified as premature instead of small-for-gestational age. However, the proportion of infants in our series classified with greater degrees of prematurity (29% <35 weeks) is high and is less likely to be the result of misclassification. Data on race and ethnicity obtained from medical records are also subject to potential misclassification by clinicians; however, the large proportion of infants with fatal pertussis who were classified as Hispanic is unlikely to be the result of misclassification alone. The reasons for the disproportionate number of pertussis deaths among Hispanics remain unexplained but may reflect a disproportionate number of cases of pertussis among Hispanic infants. Native Americans had the highest pertussis mortality rate, although this rate was based on a small number of deaths. Further investigations are needed of possible factors responsible for ethnic and racial disparities in pertussis deaths.
Our data on contacts with cough illness are consistent with other reports that suggest that sources of pertussis for very young infants are primarily family members, especially parents. 10, 32 The increasing incidence of pertussis among adolescents and adults in the child-bearing years in the vaccine era may be producing increased morbidity and mortality among infants too young to be protected with immunization, as predicted a decade ago. 33 Improved information on the transmission setting for pertussis in very young infants is needed to guide control strategies; additional studies would be useful.
Eliminating pertussis deaths will require new strategies to control pertussis because most deaths now are occurring among infants <4 months of age and are not prevented by the current vaccination strategy. Additional strategies will need to focus on protecting very young infants and could include improved education of parents and clinicians about adult and infant pertussis, improved pertussis diagnosis and prophylactic treatment and the use of acellular pertussis vaccines among adult caregivers when preparations are licensed for this age group.
This analysis was supported by the US Centers for Disease Control and Prevention. We gratefully acknowledge the contributions of Barbara Bardenheier, Iain Hardy, Rina Shaikh, Pamela Srivastava, Rafi Tachdjian and Naomi Wortis for their hard work in maintaining surveillance for pertussis deaths over the decade; of Barry Sirotkin in obtaining and providing guidance in analyzing NCHS data; and of the treating physicians and disease surveillance personnel in all states for making this study possible.
1. Guris D, Strebel PM, Bardenheier B, et al. Changing epidemiology of pertussis in the United States
: increasing reported incidence among adolescents and adults, 1990–1996. Clin Infect Dis 1999; 28: 1230–7.
2. Farizo KM, Cochi SL, Zell ER, Brink EW, Wassilak SG, Patriarca PA. Epidemiological features of pertussis in the United States
, 1980–1989. Clin Infect Dis 1991; 14: 708–19.
3. Tanaka M, Vitek C, Pascual FB, Bisgard KM, Murphy T. Increasing incidence of pertussis among young infants in the United States
, 1980–98 [Abstract 494]. In: 38th Annual Meeting of the Infectious Diseases Society of America, New Orleans, September 7 to 10, 2000. Alexandria, VA: Infectious Diseases Society of America, 2000.
4. Hook EB, Regal RR. Capture-recapture methods in epidemiology: methods and limitations. Epidemiol Rev 1995;17:243–64 (Errata: Am J Epidemiol 1998;148:1219).
5. Chapman DG. Some properties of the hypergeometric distribution with applications to zoological sample censuses. vol 1. Berkeley, CA: University of California Publications in Statistics, 1951: 131–60.
6. Regal RR, Hook EB. Goodness-of-fit based confidence intervals for estimates of the size of a closed population. Stat Med 1984; 3: 287–91.
7. Monthly Postcensal Resident Population, by single year of age, sex, race, and Hispanic origin: National Population Estimates for the 1990s [database online]. Washington, DC: US Census Bureau. Available at :http://eire.census.gov/popest/archives/national/nat_90s_detail/nat_90s_1.php
. Accessed June 23, 2002.
8. Kleinbaum DG, Kupper LL, Muller KE. Applied regression analysis and other multivariable methods. 2nd ed. Boston: PWS-KENT, 1988: 497–512.
9. SAS Institute. The GENMOD procedure. In: SAS/STAT software: changes and enhancements through Release 6.12. Cary, NC: SAS Institute, 1997: 247–348.
10. Cherry JD, Heininger U. Pertussis and other Bordetella
infections. In: Feigin RD, Cherry JD, eds. Textbook of pediatric infectious diseases. 4th ed. Philadelphia: Saunders, 1998: 1423–40.
11. Plott K, Vitek C, Pascual FB, et al. Pertussis deaths among infants, United States
, 2000. MMWR 2002; 51: 616–8.
12. Gay NJ, Miller E. Pertussis transmission in England and Wales. Lancet 2000; 355: 1553–4.
13. Ntezayabo B, De Serres G, Duval B. Pertussis resurgence in Canada largely caused by a cohort effect. Pediatr Infect Dis J 2003; 22: 22–7.
14. Williams GD, Matthews NT, Choong RKC, Ferson MJ. Infant pertussis deaths in New South Wales 1996–1997. Med J Aust 1998; 168: 281–3.
15. CDC. National, state, and urban area vaccination coverage levels among children aged 19–35 months, United States
, 1999. MMWR 2000; 49: 585–9.
16. Shaikh R, Guris D, Strebel P, Wharton M. Underreporting of pertussis deaths in the United States
: need for improved surveillance. Pediatrics 1998; 101: 323.
17. Sutter RW, Cochi SL. Pertussis hospitalizations and mortality in the United States
, 1985–88: evaluation of the completeness of national reporting. JAMA 1992; 267: 386–91.
18. Crowcroft NS, Andrews N, Rooney C, Brisson M, Miller E. Deaths from pertussis are underestimated in England. Arch Dis Child 2002; 86: 336–8.
19. Sotomayer J, Weiner LB, McMillan JA. Inaccurate diagnosis in infants with pertussis: an eight year experience. Am J Dis Child 1985; 139: 724–7.
20. Heininger U, Stehr K, Cherry JD. Serious pertussis overlooked in infants. Eur J Pediatr 1992; 151: 342–3.
21. Hoppe JE. Neonatal pertussis. Pediatr Infect Dis J 2000; 19: 244–7.
22. Christie CDC, Baltimore RS. Pertussis in neonates. Am J Dis Child 1989; 143: 1199–202.
23. Heininger U, Stehr K, Schmidt-Schlapfer G, et al. Bordetella pertussis
infections and sudden unexpected deaths in children. Eur J Pediatr 1996; 155: 551–3.
24. Smith C, Vyas H. Early infantile pertussis: increasingly prevalent and potentially fatal. Eur J Pediatr 2000; 159: 898–900.
25. Wortis N, Strebel P, Wharton M, Bardenheier B, Hardy I. Pertussis deaths: report of 23 cases in the United States
, 1992 and 1993. Pediatrics 1996; 97: 607–12.
26. Goulin GD, Kaya KM, Bradley JS. Severe pulmonary hypertension associated with shock and death in infants infected with Bordetella pertussis
. Crit Care Med 1993; 21: 1791–4.
27. Sreenan CD, Osiovich H. Neonatal pertussis requiring extracorporeal membrane oxygenation. Pediatr Surg Int 2001; 17: 201–3.
28. Pierce C, Klein N, Peters M. Is leukocytosis a predictor of mortality in severe pertussis infection. Intensive Care Med 2000; 26: 1512–14.
29. Williams GD, Numa A, Sokol J, Tobias V, Duffy BJ. ECLS in pertussis: does it have a role. Intensive Care Med 1998; 24: 1089–92.
30. Von Rosenstiel IA, Plotz FB. ECLS in pertussis: does it have a role? Intensive Care Med 1999; 25: 1341–2.
31. Casano P, Odena MP, Cambra FJ, Martin JM, Palomeque A. Bordetella pertussis
causing pulmonary hypertension. Arch Dis Child 2002; 86: 453.
32. Bisgard KM, Cianfrini CL, Pascual FB, et al. Infant pertussis: who is the source? Prospective investigation of cases from GA, IL, MN, MA, January 1999–October 2000 [Abstract 615]. In: Abstracts of the Pediatric Academic Societies 2001 Annual Meeting, April 28 to May 1, 2001. Baltimore: Pediatric Academic Societies, 2001.
33. Mortimer EA. Pertussis and its prevention: a family affair. J Infect Dis 1990; 161: 473–9.