- Design
- Systematic review and network meta-analysis of observational studies, with life-table modelling
- Population
- 34 publications, 1,459,685 participants, of whom 966,144 had normal spirometry
- Primary outcome
- Mortality hazard ratio by GOLD grade versus normal spirometry
- Effect
- GOLD 1 HR 1.09 (95% CI 0.97–1.22); GOLD 2 1.57 (1.41–1.75); GOLD 3 2.39 (2.03–2.83); GOLD 4 4.54 (3.44–6.01)
Thirty-four publications covering 1,459,685 participants — 966,144 of them with normal spirometry — were pooled by network meta-analysis to give mortality hazard ratios by GOLD grade at baseline, all referenced to normal spirometry.
The gradient is monotonic and steep at the top: GOLD 1, hazard ratio 1.09 (95% CI 0.97–1.22); GOLD 2, 1.57 (1.41–1.75); GOLD 3, 2.39 (2.03–2.83); GOLD 4, 4.54 (3.44–6.01). Combined with German life tables, the derived one-year mortality probabilities span from 0.15% for a woman with GOLD 1 aged 45–49 to 64.8% for a man with GOLD 4 aged 85 or over.
The GOLD 1 result deserves attention because its confidence interval crosses 1.0. Mild airflow obstruction on spirometry, in isolation, is not associated with clearly raised mortality in this synthesis. That is not an argument for ignoring it — early COPD is where smoking cessation changes trajectory — but it is a useful corrective to a diagnosis label that patients hear as a death sentence. The purpose of the work was to feed a German health-economic model, and the mortality probabilities are German; the hazard ratios are the transferable part.
- Quote the grade, not the diagnosis, when discussing prognosis — GOLD 1 and GOLD 4 are entirely different conversations.
- Use GOLD 1 as an opportunity for smoking cessation rather than a prognostic warning.
- Do not transfer the absolute German mortality probabilities to Indian patients; the hazard ratios are what generalise.
- Remember that GOLD grade is spirometric severity alone and does not capture symptom burden or exacerbation history, which drive treatment choice.
- Combine the grade with exacerbation frequency and comorbidity when estimating an individual's risk.
Why it matters
"You have COPD" carries a prognosis in a patient's mind that GOLD 1 does not support.
Don't overread it
The absolute mortality probabilities are model-based and specific to Germany; only the hazard ratios transfer.
The statistics, in plain English
The GOLD 1 hazard ratio of 1.09 with an interval from 0.97 to 1.22 includes 1.0, meaning these data are compatible with no excess mortality at that grade. Each successive grade's interval sits clearly above 1 and clear of the one below, so the gradient itself is solid. The one-year mortality probabilities are model-derived — pooled hazard ratios applied to German life tables — rather than directly observed, so they inherit the assumptions of both.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for pulmonology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free