Table 1.
Age started smoking regularly versus cause-specific mortality rate ratio (RR) among urban and rural men in about 2010 (second study) and about 1995 (first study)
| Breath CO, ppm |
All causes |
Lung cancer |
Ischaemic stroke or ischaemic heart disease |
COPD |
||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Number of deaths | RR* (95% group-specific CI) | Number of deaths | RR* (95% group-specific CI) | Number of deaths | RR* (95% group-specific CI) | Number of deaths | RR* (95% group-specific CI) | |||
| Urban men, about 2010 (2006–14) | ||||||||||
| Age started smoking (mean), years | ||||||||||
| Age <20 (16·9) | 16·5 | 843 | 1·98 (1·85–2·13) | 127 | 3·78 (3·15–4·54) | 200 | 2·03 (1·75–2·35) | 43 | 9·09 (6·60–12·50) | |
| Age 20–24 (21·0) | 15·2 | 797 | 1·61 (1·50–1·73) | 127 | 3·17 (2·67–3·77) | 182 | 1·50 (1·30–1·74) | 24 | 3·89 (2·62–5·79) | |
| Age ≥25 (30·4) | 12·6 | 754 | 1·44 (1·34–1·55) | 96 | 2·23 (1·82–2·73) | 196 | 1·49 (1·29–1·72) | 24 | 2·89 (1·93–4·32) | |
| All smokers | 14·9 | 2394 | 1·65 (1·59–1·73) | 350 | 2·98 (2·66–3·33) | 578 | 1·63 (1·49–1·77) | 91 | 4·61 (3·71–5·71) | |
| Non-smokers | 4·5 | 924 | 1·00 (0·94–1·07) | 69 | 1·00 (0·79–1·27) | 242 | 1·00 (0·88–1·14) | 18 | 1·00 (0·62–1·60) | |
| Trend p value† | <0·0001 | 0·0002 | 0·0032 | <0·0001 | ||||||
| Rural men, about 2010 (2006–14) | ||||||||||
| Age started smoking (mean), years | ||||||||||
| <20 (16·8) | 15·1 | 2710 | 1·39 (1·34–1·45) | 271 | 2·91 (2·58–3·29) | 334 | 1·37 (1·23–1·53) | 315 | 1·88 (1·68–2·10) | |
| 20–24 (21·2) | 13·6 | 2788 | 1·26 (1·21–1·31) | 275 | 2·45 (2·17–2·75) | 397 | 1·35 (1·23–1·49) | 261 | 1·41 (1·24–1·59) | |
| ≥25 (31·3) | 11·8 | 2434 | 1·05 (1·01–1·10) | 177 | 1·63 (1·41–1·90) | 347 | 1·04 (0·93–1·16) | 233 | 1·07 (0·94–1·22) | |
| All smokers | 13·6 | 7932 | 1·22 (1·20–1·25) | 723 | 2·30 (2·13–2·48) | 1078 | 1·24 (1·17–1·32) | 809 | 1·41 (1·31–1·51) | |
| Non-smokers | 6·0 | 2031 | 1·00 (0·96–1·05) | 92 | 1·00 (0·81–1·23) | 316 | 1·00 (0·89–1·12) | 172 | 1·00 (0·86–1·16) | |
| Trend p value† | <0·0001 | <0·0001 | 0·0005 | <0·0001 | ||||||
| Urban men, about 1995 (1991–99) | ||||||||||
| Age started smoking (mean), years | ||||||||||
| <20 (16·7) | .. | 1158 | 1·60 (1·51–1·70) | 142 | 3·20 (2·69–3·80) | 178 | 1·58 (1·36–1·84) | 194 | 1·82 (1·57–2·21) | |
| 20–24 (21·2) | .. | 1358 | 1·37 (1·30–1·44) | 130 | 2·31 (1·95–2·75) | 208 | 1·37 (1·19–1·57) | 209 | 1·43 (1·25–1·64) | |
| ≥25 (30·6) | .. | 1136 | 1·11 (1·04–1·17) | 97 | 1·73 (1·42–2·12) | 178 | 1·17 (1·01–1·36) | 186 | 1·16 (1·00–1·34) | |
| All smokers | .. | 3652 | 1·32 (1·28–1·37) | 369 | 2·32 (2·08–2·59) | 564 | 1·35 (1·23–1·47) | 589 | 1·42 (1·30–1·55) | |
| Non-smokers | .. | 1381 | 1·00 (0·95–1·06) | 82 | 1·00 (0·80–1·25) | 240 | 1·00 (0·87–1·14) | 197 | 1·00 (0·87–1·15) | |
| Trend p value† | <0·0001 | <0·0001 | 0·0052 | <0·0001 | ||||||
| Rural men, about 1995 (1991–99) | ||||||||||
| Age started smoking (mean), years | ||||||||||
| <20 (16·8) | .. | 5107 | 1·20 (1·16–1·23) | 236 | 1·98 (1·73–2·26) | 435 | 1·12 (1·02–1·23) | 1373 | 1·40 (1·33–1·48) | |
| 20–24 (20·9) | .. | 6215 | 1·14 (1·11–1·17) | 280 | 1·82 (1·62–2·05) | 605 | 1·17 (1·08–1·27) | 1573 | 1·28 (1·22–1·34) | |
| ≥25 (29·6) | .. | 3901 | 1·04 (1·01–1·07) | 139 | 1·37 (1·16–1·62) | 385 | 1·08 (0·98–1·18) | 986 | 1·06 (0·99–1·13) | |
| All smokers | .. | 15 223 | 1·13 (1·11–1·15) | 655 | 1·76 (1·62–1·91) | 1425 | 1·13 (1·07–1·20) | 3932 | 1·25 (1·21–1·30) | |
| Non-smokers | .. | 4855 | 1·00 (0·97–1·03) | 150 | 1·00 (0·85–1·18) | 499 | 1·00 (0·91–1·09) | 1114 | 1·00 (0·94–1·06) | |
| Trend p value† | <0·0001 | <0·0001 | 0·0620 | <0·0001 | ||||||
Group-specific CI for the non-smoker RR of 1·00 reflects the variance of the log risk in non-smokers; for each of the other RRs the CI is correspondingly wider than the group-specific CI. COPD=chronic obstructive pulmonary disease.
*
RRs (smokers vs non-smokers) were adjusted for 5-year age group and region, alcohol, and education; additional adjustment for occupation made no material difference to the RRs.
†
Trend test in smokers, ignoring non-smokers; if trend tests in this table had included non-smokers, each would have yielded p<0·0001.