Our study underwent 1-year, 3-year and long-term follow-up with a median follow-up time of 41 months and a maximum follow-up time of 74 months. 697 patients met the inclusion criteria and 134 patients were lost to follow-up. A total of 194 (27.8%) patients died during the entire follow-up. Of these, 75 (10.7%) patients died within one year. The number of early smoking patients was 384 patients, and late smoking patients was 313. After adjusting according to the smoking index, there were 149 patients in each group. Among the patients who quit smoking, there were 206 early smoking patients and 253 late smoking patients. After adjusting for smoking index, there were 104 patients in each group.

Demographic characteristics

The overall median age for the population was 68 years with 97.6% being male. The median age of smoking initiation in early smoking patients was 18 years and 98.7% are male. The median age of late smoking COPD patients was 33 years and 96.2% were male. Comparing the baseline data at admission between the two groups (Supplementary Table 1), early smoking COPD patients were currently younger than late smoking COPD patients (66 years old vs. 72 year old, P < 0.001). However, we could not find any other differences in the other demographic data between the two groups.

Early smoking leading to severer cigarette addiction

Compared to late smoking patients, the early smoking COPD population had longer smoking histories and higher smoking indices (50 vs. 30 packets/year, P < 0.0001), despite early smoking COPD population were at younger age (Supplementary Table 1). In addition, early smoking COPD patients had lower rates of smoking cessation than late smoking COPD patients. (53.5% vs. 83.9%, P < 0.0001, Fig. 2), suggesting that early smoking might lead to heavier smoking burden and severer cigarette addiction.

Fig. 2

Correlation analysis between age of smoking initiation and smoking index (A). Comparison of smoking cessation rates in early smoking or late smoking COPD patients (B)

Early smoking associated with worse condition at the beginning of AE

In the whole hospitalized population for AECOPD population, early smoking patients were found to have poorer lung function (FEV1/FVC, 35.95% vs. 38.61%, P = 0.011), higher frequency of adverse events in the past 12 months before admission (2 vs. 1, P = 0.031), and higher PaCO2 (51 vs. 49 mm/Hg, P = 0.023, Supplementary Table 1) at the beginning of admission. We performed propensity score matching to account for the effect of smoking index factors and found that adjusted early smoking COPD cohort had poorer lung function (FEV1/FVC, 34% vs. 38.89%, P = 0.023), a higher frequency of adverse events in the past 12 months before hospitalization (2 vs. 1, P = 0.011) and a greater reliance on triple therapy during stabilization (38.9% vs. 27.5%, P = 0.049, Table 1).

Table 1 Baseline characteristics (after adjusting smoking index)Early smoking and comorbidity

In terms of comorbidities, there was no significant difference in the prevalence of comorbidities such as coronary artery disease, hypertension, diabetes, pneumonia, bronchiectasis, respiratory failure, prior pulmonary TB, and cor pulmonale between the two groups of patients (Table 2). However, the ultrasound results showed that left ventricular diameters were thicker in early smoking COPD patients than late smoking COPD patients (44 vs. 42 mm, P = 0.016, Supplementary Table 2), and after adjusted the differences seemed to be more significant (44 vs. 42 mm, P = 0.003, Table 2) .

Table 2 The effect of early or late smoking on comorbidities and left heart function in COPD patients (after adjusting smoking index)Early smoking patients and all-cause mortality

Although the direct comparison could not find the differences in all-cause mortality between early smoking and late smoking COPD patients (Supplementary Table 3), after adjusting the possible bias factor, smoking index, exhibited elevated three-year (24.2% vs. 11.4%, P = 0.004, Table 3) and long-term mortality rates (32.9% vs. 15.4%, P < 0.001, Table 3) along with reduced survival durations in early smoking COPD patients (Fig. 3).

Table 3 The effect of early or late smoking on prognosis in COPD patients (after adjusting smoking index)Fig. 3

Kaplan-Meyer survival curves for the effect of early or late smoking on one-year (A), three-year (B), and long-term (C) mortality in paired COPD patients

Early smoking leads to poorer smoking cessation outcomes

In the entire cohort of COPD patients who ceased smoking, early smoking patients were found to have poorer lung function (FEV1% predicted: 27.9% vs. 32%, P = 0.007; FEV1/FVC: 35% vs. 38%, P = 0.017, Supplementary Table 4), and thicker left ventricular diameters than late smoking patients (44 vs. 42 mm, P = 0.025, Supplementary Table 5). We performed propensity score matching to account for the effect of smoking index factors and found that poorer lung function (FEV1% predicted: 26% vs. 31.1%, P = 0.013; FEV1/FVC: 34.1% vs. 38%, P = 0.010, Table 4), and thicker left ventricular diameters (44 vs. 42 mm, P = 0.003, Table 5), again survival analysis revealed higher long-term mortality (32.9% vs. 15.4%, P = 0.010, Table 6) and shorter survival time (P = 0.0128, Fig. 4) in early smoking patients, rather than late smoking patients.

Table 4 Early smoking impaired smoking cessation (after adjusting smoking index)Table 5 The impact of smoking cessation on comorbidities and left heart function in early or late smoking COPD patients (after adjusting smoking index)Table 6 The impact of smoking cessation on short- and long-term prognosis in early or late smoking COPD patients (after adjusting smoking index)Fig. 4

Kaplan-Meyer survival curves for the effect of smoking cessation on one-year (A), three-year (B), and long-term (C) mortality in paired patients who were early or late smoking COPD patients