Smoking is largest contributor to reduced life expectancy in Danish HIV cohort

Nathan Geffen, CSSR

A recent analysis from the Danish HIV Cohort Study, published as an access article in Clinical Infectious Diseases has reported that smoking is the biggest cause of life-years lost in Danish people with HIV, and that this is a much bigger cause of death than HIV-related illnesses. [1]

This long-established cohort provides excellent observational data to estimate life-expectancy and mortality differences in HIV positive versus HIV negative controls. Denmark provides free ART, high-quality care, at specialist centres, with low loss to follow-up, so health systems failures are less likely to confound results. For example, participants are routinely seen every few months and (since 2004) tobacco and alcohol use is recorded annually.

This analysis included all patients receiving care from 1995 to 2010 who were older than 16 at HIV diagnosis and who had smoking data available. Time was calculated from the date of first available smoking status, age 35 years or one year after the date of HIV diagnosis, whichever came last, until death, emigration, or 1 September 2010.

Current smokers were defined at enrolment as people who smoked any type of tobacco at least once a week. Previous smokers were defined as people who were smokers but had given up before enrolment. All others were defined as never-smokers. Smoking status was not changed during the study.

Of 5,300 people with HIV in the cohort, over 2,400 were excluded because of injection drug use (567), missing data on smoking (1,497) or age less than 35 years at the end of the study period or at censoring (363). This left just under 3,000 people in the study.

Participants were then matched by sex and year of birth with over 10,600 HIV negative controls. A separate analysis that included injection drug users (IDU) was also performed. IDU were excluded from the main analysis as the extremely high smoking rates (only 7/567 didn’t smoke) and lower life-expectancy than the general HIV positive population, would both confound the results.

At baseline, smoking was both more common for HIV positive people and individual cigarette use was higher. Among HIV vs controls respectively, rates were 47% vs 21% for current, 18% vs 33% for previous and 35% vs 47% for never smokers. The median number of cigarettes smoked a day by current smokers was 20 (IQR: 10-20) vs 15 (IQR: 10-20) in the positive vs negative groups.

In the HIV positive group, viral load, AIDS diagnosis at baseline, years of ART and years since diagnosis were similar across smoking categories, but hepatitis C status was 9.8%, 5.4% and 4.8% respectively. The HIV positive population was ethnically diverse while the case controls were entirely Danish. However a sensitivity analysis found that neither this, nor gender, had a marked impact on the results.

The two groups were followed for over 14,000 and 45,000 patient years, respectively. The median follow-up time was 4.2 years (IQR: 3.1-5.5) for HIV patients and 4.1 years (IQR, 2.9-5.8) for population controls.

Factors in the multivariable analysis included age, year of HIV diagnosis, excess consumption of alcohol, body mass index, CD4 cell count, and viral load at baseline. Smoking was the factor associated with the highest risk of death and did not interact with these other variables.

The excess mortality associated with smoking was much higher among HIV patients compared to the case controls but the relative risk of death associated with smoking did not differ. This is because the smoking rate was much higher in the HIV-positive cohort.

The main results of the study were:

  • Life expectancy for a 35-year-old HIV positive person was calculated as 62.6 years (95% CI: 59.9-64.6) for smokers, 69.1 years (95% CI, 67.5-71.2) for previous smokers and 78.4 years (95% CI: 70.8-84.0) for never-smokers. The loss of life-years associated with smoking was twice as high as HIV-related causes.
  • HIV positive smokers had higher all-cause mortality (mortality rate ratio [MRR] 4.4; 95% CI: 3.0-6.7).
  • HIV positive smokers had higher non-AIDS-related mortality (MRR: 5.3; 95% CI: 3.2-8.8).
  • HIV positive smokers who smoked 30 or more cigarettes a day had higher mortality than those who smoked less than 30 (MRR 4.2; 95% CI: 2.6-6.9).
  • The excess mortality rate per 1,000 person-years among HIV positive current vs HIV positive never smokers was 17.6 (95%CI: 13.3-21.9). For smokers versus never smokers without HIV this was 4.8 (95% CI: 3.2-6.4).
  • The population-attributable risk of death associated with smoking was 61.5% among HIV patients and 34.2% among controls.
  • AIDS-related deaths were also more likely among smokers and previous smokers versus never-smokers: 5.2 (95%CI: 3.7-7.3) vs 6.0 (95%CI: 4.0-10) vs 1.4 (95%CI: 0.7-3.0), respectively.
  • More than 60% of deaths in the HIV positive cohort were due to factors associated with smoking.

The MRR for cancer deaths was 5.1 (95%CI: 3.6-7.2) vs 5.1 (95%CI: 3.0-8.6) vs 1.7 (95%CI 0.8-3.3), for current vs previous vs never smokers, respectively. The corresponding rates for cardiovascular disease were 2.7 (95%CI: 1.7-4.3), 0.7 (95%CI: 0.2-2.9) and 0.8 (0.3-2.2). The authors explained that while the risk of cardiovascular disease for previous smokers diminishes quickly to be similar to non-smokers, the higher cancer risk remains. Overall mortality was halved in previous smokers compared to current smokers, emphasising the importance of successful cessation interventions.

There was a trend to higher rates of violent deaths among current and previous smokers versus never smokers, indicating that smokers and alcohol use was higher in smokers, but although this was not statistically significant, this indicates that there may be some confounding in the study’s main findings (also with social-economic status). The authors also speculated that the much higher contribution to mortality by smoking in HIV positive people versus the general population might be due to nicotine causing inflammation.


The association of smoking to mortality was so large that it is almost certainly a significant cause of lost life-years in people with HIV, even allowing for confounding that is possible in any observational study.

It is unclear whether the greater association with death due to smoking in HIV positive people compared HIV negative controls is, as the authors speculate, due to HIV-related factors, or because HIV positive smokers are more likely to have other risks.

Nevertheless, this study is clear evidence that for well-resourced HIV care centres with cohorts that have good access to ART, finding ways to reduce smoking should be a priority. This is easier said than done. Nicotine is more addictive than alcohol, heroin, meta-amphetamines, cocaine and marijuana. [2]

Additional research is needed on effective smoking cessation interventions as a Cochrane review shows that while some smoking cessation interventions help, their effects are modest and they have side effects. [3]


  1. Helleberg M et al. 2012. Mortality attributable to smoking among HIV-1-infected individuals: a nationwide, population-based cohort study. Clinical Infectious Diseases Advance Access, published December 18, 2012.
  2. Widmaier E et al. 2008. Vander’s Human Physiology. McGraw-Hill. p. 245.
  3. Cahill K et al. 2012. Can nicotine receptor partial agonists, including cytisine, dianicline and varenicline, help people to stop smoking? Cochrane Database of Systematic Reviews 2012, Issue 4. Art. No.: CD006103. DOI: 10.1002/14651858.CD006103.pub6

Links to other websites are current at date of posting but not maintained.