Mortality among “never smokers” living with smokers: two cohort studies, 1981-4 and 1996-9

BMJ 2004; 328 doi: (Published 22 April 2004) Cite this as: BMJ 2004;328:988
  1. Sarah Hill, public health medicine registrar1,
  2. Tony Blakely (tblakely{at}, senior research fellow1,
  3. Ichiro Kawachi, professor of social epidemiology2,
  4. Alistair Woodward, professor of public health1
  1. 1Department of Public Health, Wellington School of Medicine and Health Sciences, PO Box 7343, Wellington South, New Zealand
  2. 2Department of Health and Social Behavior, Harvard School of Public Health, 677 Huntington Avenue, Boston, MA 02115, USA
  1. Correspondence to: T A Blakely
  • Accepted 5 March 2004


Few studies have examined the association between passive smoking and all cause mortality; most of these have had limited study power.14 We present results from two population cohorts of adults who had never smoked (“never smokers”), followed for three year mortality according to household exposure to secondhand smoke.

Participants, methods, and results

The two cohorts comprised all New Zealand adults aged 45-74 years who responded to the 1981 and 1996 censuses and who identified themselves as never smokers, lived in a private dwelling (that is, not a prison, hospital, or other institution), and had provided data on smoking status for all household members aged 15 and over (87.0% of never smokers in 1981 and 85.3% in 1996).

Never smokers living in households with one or more current smokers were regarded as being exposed to secondhand smoke in the home; those living in households with no current smokers were regarded as not exposed. Cohort members were followed for mortality in the three years after the census by means of anonymous probabilistic linkage with a national register of mortality records.5 Record linkage was complete for 71.0% of eligible mortality records during 1981-4 and for 78.2% during 1996-9. Data were weighted to adjust for potential linkage bias.5

We calculated mortality and standardised for age and ethnicity using the 1996 census population as the standard. We used Poisson regression to adjust for age, ethnicity, marital status, and socioeconomic position, using a more restricted cohort with full demographic data (82.3% of the 1981 cohort and 89.9% of the 1996 cohort).

In both cohorts and sexes, mortality among never smokers was greater in those living in households with a current smoker (table).

All cause mortality among adults who have never smoked, by household exposure to secondhand smoke, 1981-4 and 1996-9

View this table:


Among adults who had never smoked we found a modest but consistent association between exposure to secondhand smoke in the home and mortality. This association persisted after adjustment for age, ethnicity, marital status, and socioeconomic position. The finding of about 15% excess mortality in never smokers exposed to secondhand smoke at home is consistent with the previous largest study in this area.1

Mortality and mortality rate ratios were standardised by age and ethnicity, and further adjustment for marital status and socioeconomic position altered the results only slightly. This suggests that these factors were not important confounders (independent of age and ethnicity). We could not adjust directly for lifestyle characteristics as these data are not included in the census. However, lifestyle factors are unlikely to act as important confounders when there is no confounding by socioeconomic position.

We considered exposure to secondhand smoke in the home only. Our inability to measure exposure in other settings introduces a degree of exposure misclassification; mortality rate ratios will probably be underestimated as a consequence. We suspect that this misclassification will be greater for the 1981-4 cohort, as smoking and exposure to secondhand smoke outside the home were more prevalent in New Zealand in the early 1980s. This may explain the apparently stronger association between household exposure and mortality in the 1996-9 cohort compared with the 1981-4 cohort.

What is already known on this topic

Few studies have examined the link between exposure to secondhand smoke and mortality

What this study adds

Adults who had never smoked and who lived with smokers had about 15% higher mortality than never smokers living in a smoke-free household

This study strengthens the case for a causal association between secondhand smoke and mortality

The results from this study add to the weight of evidence of harm caused by passive smoking and support steps to reduce exposure to other people's smoke—in the home and in other settings.

Papers pp 977,989


We thank Jackie Fawcett and June Atkinson for technical help with data extraction and analysis.


  • Contributors SEH conceived the study, analysed the data, and drafted the manuscript. TAB conceived and led the New Zealand census-mortality study (NZCMS) from which data for this study were drawn; advised on study design, data analysis, and interpretation; and contributed to the manuscript. AW and IK advised on the design, analysis, and interpretation of the study and contributed to the manuscript. SEH and TAB will act as joint guarantors for this paper.

  • Funding The NZCMS is primarily funded by the Health Research Council of New Zealand, with further funding from the Ministry of Health. Funding for SEH's salary came from the New Zealand Population Health Charitable Trust.

  • Competing interests None declared.

  • Ethical approval Not needed—see security statement on

  • Embedded Image A security statement about the New Zealand census-mortality study (NZCMS) is on

    This article was posted on on 5 April 2004:


View Abstract