Delayed immunisation and risk of pertussis in infants: unmatched case-control study

BMJ 2003; 326 doi: (Published 19 April 2003) Cite this as: BMJ 2003;326:852
  1. Cameron C Grant, senior lecturer (cc.grant{at},
  2. Mavis Roberts, project managera,
  3. Robert Scragg, senior lecturerb,
  4. Joanna Stewart, biostatisticianb,
  5. Diana Lennon, professor of community paediatricsc,
  6. Denise Kivell, charge nursec,
  7. Rodney Ford, clinical associate professord,
  8. Rosalie Menzies, scientific officere
  1. a Department of Paediatrics, Faculty of Medicine and Health Sciences, University of Auckland, Private Bag 92019, Auckland, New Zealand
  2. b Department of Community Health, Faculty of Medicine and Health Sciences, University of Auckland
  3. c Child and Youth Health, South Auckland Health, Private Bag 93311, Otahuhu, Auckland, New Zealand
  4. d Community Paediatrics, Canterbury District Health Board, Private Bag 4345, Christchurch, New Zealand
  5. e Virology and Immunology, Auckland District Health Board, New Zealand
  1. Correspondence to: C C Grant
  • Accepted 11 November 2002

Pertussis remains a severe disease in infants. As about two thirds of infants with pertussis are admitted to hospital, factors that seem to be associated with an increased risk of pertussis may in fact be associated with an increased risk of hospital admission. 1 2 The admission rate for pertussis in New Zealand is five to 10 times higher than in England and Wales and the United States.3 We determined whether immunisation reduced the risk of admission to hospital for pertussis by comparing infants admitted with pertussis and infants admitted with other acute respiratory illnesses.

Odds ratios (95% confidence intervals) of catching pertussis associated with delays in giving pertussis vaccine

View this table:

Participants, methods, and results

We performed an unmatched case-control study during the 1995-7 pertussis epidemic in Auckland, New Zealand. Pertussis was defined as cough lasting at least two weeks, with coughing paroxysms, inspiratory “whoop,” or vomiting after coughing. The control group consisted of 98 infants admitted to hospital with a coughing illness who were culture negative for Bordetella pertussis and had no B pertussis DNA detected in their nasopharyngeal sample after amplification by polymerase chain reaction. We interviewed each infant's care giver and determined written confirmation of the infant's immunisation status from his or her health record book or the family doctor's records.

In New Zealand, immunisations are scheduled at age 6 weeks, 3 months, and 5 months. An immunisation was delayed if it had not been received within 30 days of its first being due.4 We used logistic regression to calculate odds ratios and 95% confidence intervals to determine the risk of pertussis associated with delayed immunisation. We defined socioeconomic status by the occupation of the household's main income earner. We measured social deprivation by using the 1996 New Zealand social deprivation index.5

We identified 179 infants with a diagnosis of pertussis at discharge and enrolled 97 (54%). These did not differ from the non-enrolled infants in age, sex, ethnicity, gestation, birth weight, or social deprivation score. To ensure that the controls were a representative sample of children with non-pertussis coughing illnesses we compared the 98 enrolled control infants with two other groups of infants admitted to hospital with such illnesses: 227 infants with cough who had not been approached for enrolment and 78 infants with cough who had been identified as eligible but for whom informed consent was not obtained. The controls and the other two groups did not differ in age, proportion of infants of non-European ethnicity, or social deprivation index.

We obtained nasopharyngeal samples for culture from 95 (98%) of the 97 infants with pertussis and for polymerase chain reaction from 83 (86%). We identified B pertussis by culture in 32 (34%), by polymerase chain reaction in 73 (75%), and by either method in 76 (80%) infants.

The infants with pertussis were younger than the controls and more likely to have mothers with only primary school education (odds ratio 11.78, 95% confidence interval 2.02 to 225.37) and to live in more crowded households (2.12, 1.17 to 3.89) and households in the most socially deprived fifth (2.21, 1.18 to 4.23). We found no differences between the two groups in other characteristics of the infant (gestation, birth weight, ethnicity, or breast feeding), mother (age, marital status, and smoking), or household (mobility, smokers, occupation, and socioeconomic status).

The table shows associations between delayed immunisations and risk of admission to hospital with pertussis. In the multivariate analysis we found an increased risk associated with delay in the first, second, or third immunisation or any combination of these (odds ratio 4.50). Analysis by individual dose of vaccine showed that an increased risk of pertussis was associated with delay in the third dose (odds ratio 6.09). Including all variables describing infant, maternal, and household characteristics in the model did not alter the importance of increased risk associated with any delayed immunisations (odds ratio 6.13, 1.13 to 47.07).


Delayed immunisation is a specific risk factor for admission to hospital with pertussis rather than being a marker of infants at increased risk of admission to hospital for any acute respiratory illness. Improving on-time delivery of immunisations can be expected to decrease the admission rate for pertussis in New Zealand.


Contributors: CCG enrolled participants, analysed the data, and drafted the paper. RS directed study design and data analysis. DL initiated and designed the project. RF enrolled cases and assisted with the ethical application. JS advised on study design, data management, and analysis. RM designed and performed all of the polymerase chain reaction assays. DK identified cases and controls. MR enrolled cases and controls and supervised other interviewers. All authors revised the paper, CCG wrote the paper and is the guarantor.


  • Funding This research was supported by grants from the National Child Health Research Foundation and the Health Research Council of New Zealand.

  • Competing interests None declared.

  • The regional ethics committee approved the study.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
View Abstract