As shown in Fig. 1, data were retrieved from the on-going population-based study Good Aging in Skåne (GÅS)14. Briefly, 60- to 93-year-old subjects living in Skåne, Sweden, are randomly invited using the population register. Participants are offered a thorough physical, medical, and psychological examination and are invited to follow-up examinations at regular intervals until death. To encourage participation of frail adults, the study team performs home visits following the same study protocol. Three waves have been fully recruited with an initial participation rate of approximately 60%. Participants recruited during the first wave in 2001–2004 were 60, 66, 72, 78, 81, 84, 87, 90, or 93 years old. Participants recruited during the second wave in 2006–2013 were 60, 66, or 81 years old. Participants recruited during third wave in 2012–2016 were 60 or 81 years old. A fourth wave is being recruited since 2019 including participants who are 60 or 81 years old. The GÅS study is conducted according to the Declaration of Helsinki and Good Clinical Practice guidelines and is approved by the Lund University Ethics Review Board (LU 744–00). All participants provide written informed consent. In this study, we aim to study early onset progression, and thus to minimize bias participants with a COPD diagnosis at baseline were excluded from the analysis.

Participants diagnosed with COPD, with missing spirometry or unable to perform three successful manoeuvres, at baseline were excluded from this study. During follow-up 143 participants received a COPD diagnosis in clinical practice.

Spirometry assessments were performed using a Vitalograph 2120 spirometer (Vitalograph Ltd, Buckingham, UK) according to the American Thoracic Society guidelines15. Bronchodilators were not administrated during the first wave baseline visit. Subjects received 1.0 mg of β2-receptor agonist terbutaline 10 min prior to the spirometry at all other visits.

Comorbidities were identified during the medical examination, and by retrieving medical records and diagnosis codes from the Skåne Healthcare Registry16 (see Supplementary Table 1). Information about prescribed medicines was obtained from medical records and self-reported. The COPD diagnosis was retrieved from medical records. Clinically, diagnosis of COPD in the region of Skåne is based on three criteria: spirometry verified obstructivity (FEV1/FVC < 0.7 after bronchodilation), current airway symptoms, and a history of a risk factor for COPD. COPD diagnosis also included emphysema and chronic bronchitis.

COPD was graded according to the Global Strategy for the Diagnosis, Management, and Prevention of Chronic Obstructive Lung Disease (GOLD) classification17 using the spirometries collected during the study visits. In some cases, participants diagnosed with COPD had a normal spirometry during the study visit. Therefore, an additional category level (category 0) was included to accommodate diagnosed subjects whose FEV1s/FVC ratio was ≥ 0.75. The spirometry performed at the study visit was used for the COPD GOLD classification. The reference equations for lower limit of normal are only available for subjects younger than 95 years old and for this calculation, age was truncated at 95 years.

The aim of the statistical analysis was to estimate the average annual decline rate in FEV1s in COPD patients at early COPD onset and to compare it to the decline rate in participants without COPD. The annual change in FEV1s was defined as the difference in FEV1s between two consecutive study visits, divided by the time passed between the visits. This model assumes that the annual decline rate is constant over time. A mixed model for repeated measures with random intercept (participants) was implemented. The variables FEV1s, age, sex, smoking status, education, body mass index, heart disease, cerebrovascular disease, asthma, and diabetes (at baseline) were included to mitigate confounding.

Estimating the difference in annual decline rate between participants with and without a COPD diagnosis is challenging since the two groups should only differ in their COPD status. Patients in clinical practice do not perform spirometry examinations in a regular way. Therefore, it could be more likely for a participant with a low FEV1s to get a spirometry examination and thus a COPD diagnosis compared to a participant with higher FEV1s values regardless of their true COPD status. The previous value of FEV1s may affect the likelihood of getting a COPD diagnosis and thus time-varying confounding may be present. In order to minimize bias, sensitivity analyses using a marginal structural mixed model were performed (see Supplementary Table 5). The risk for misclassification in the COPD diagnosis is acknowledged. We, therefore, presented a Supplementary Note about sensitivity analyses where the COPD status was assigned using the spirometry results obtained at the study visits. The statistical software Stata IC 14.2 (StataCorp LLC, Texas, USA) was used.

Further information on research design is available in the Nature Research Reporting Summary linked to this article.