In a population of heroin smokers, we found a high burden of lung disease. In the previously published baseline data, 50% of heroin users had COPD or ACO, with a mean MRC score of 3.1 and CAT score of 22.9.¹⁷ At follow-up, participants’ respiratory symptoms had worsened significantly from baseline, with annual increases in both CAT score (1.60) and MRC score (0.46), and mean oxygen saturation dropping from 97% to 95% from baseline to follow-up. We found that lung function measured by FEV₁ declined by 90 mL annually, which was both statistically and clinically significant. The proportion of subjects classified as having severe or very severe disease increased from 25% to 36% over the 2-year follow-up period. Neither ongoing illicit drug use nor prescriptions of inhaled medication were associated with change in lung function.

The symptoms reported in this study are consistent with those of studies in this population, with increased dyspnea among heroin users being the common symptom.^12,13,26 The decline in health status measured by a CAT score increase of 1.60 annually is greater than the 1 unit change seen in patients with stable COPD.²⁷ The rate of decline in FEV₁ is considerably higher than both the 30 mL/y age-related decline seen in nonsmokers and in people with tobacco-related COPD (which is reported at 35-79 mL/y, of which all but one paper reported an annual decline of ≤ 69 mL).^26,28 To date, research on lung function in heroin smokers has focused on cross-sectional studies. The results from this longitudinal cohort study support and enhance previous cross-sectional studies that suggest heroin users are at a high risk of COPD and suggest that their decline is worse than that of tobacco smokers. Walker et al¹¹ found heroin smokers developed early onset emphysema, with a mean age of diagnosis being 41 years, suggesting likely early progression of disease compared with nonheroin smokers. In Amsterdam, The Netherlands, Buster et al¹⁴ reported a difference in FEV₁ from predicted values, finding that heroin smokers had an FEV₁ of 260 mL less than predicted FEV₁.

The rapid decline in FEV₁ and the increase in respiratory symptoms in this population suggest heroin smoking is a driver of decline in lung function. Similarly, once established, this decline appears to continue even in people who stop smoking drugs.

Although COPD hospital admissions vary greatly across the United Kingdom, patients with COPD tend to have high health-care usage, particularly in areas of high deprivation.^29,30 Previous research has also shown that heroin users with respiratory exacerbations are more likely to be readmitted with exacerbations than current/ex-tobacco smokers (OR, 1.00 vs 0.22/0.26, respectively).¹⁸ It is also clear that with the high levels of health-care access observed in this population, it is likely that ongoing trends toward inhaling heroin will further increase the use of, and burden on, the health system.^4,6

The strengths of our study include that we followed-up the participants over a 18- to 24-month period in a community clinic setting. We have shown that it is feasible to engage this client group in both baseline and follow-up spirometry allowing for a diagnosis to be made. The lost to follow-up rate is a major limitation of this study, reducing the power of statistical analysis and making stratification of our results by age or GOLD stage unfeasible. Given a larger group, this information would potentially be helpful for targeting care, and is an area for future investigation. This population tends to smoke a mix of heroin, crack, and tobacco, establishing a causal relationship with therefore difficult. The participants in the study were generally from a poor socioeconomic background, and there is potential that their living condition environment could contribute to the rate of decline. Without significant heterogeneity of such potentially confounding factors, we have been unable to address this question further. There is also potential for selection bias, with people who regularly attend methadone clinics and have concerns about their respiratory system more likely to participate in the study.

In summary, our findings show the significant respiratory impairment with which heroin smoking is implicated, and a concerning accelerated rate of decline over time. Future studies with larger cohorts, possibly in the context of a targeted public health intervention, are needed to understand if specific subgroups are especially vulnerable, and how the personal and health-care costs associated with chronic respiratory illness could be best averted. The study methodology is in support of it being feasible to colocate respiratory and drug services to one community location. Future studies may benefit from a parallel group of heroin users without spirometric abnormalities at baseline to determine their rate decline compared with patients with COPD. These results combined with previous studies support the call for enhanced screening for inhaled drug users.¹⁹ A pilot followed by clinical trial would be needed to assess if screening and treatment services would be clinically and cost-effective in this population.