Fig. 2.

Subgroup meta-analysis. For some of the published studies, results from disjunctive subgroups (e.g. men and women) are reported. If so, subgroup estimates were used instead of overall results. For this reason, the number of subgroup results included need not to sum up to total number of selected studies. For published studies, asthma latency is defined as the minimal allowed latency by the design of a case–control study. Within the ILCCO pooled analysis, asthma latency was calculated for each study participant. Age, age at LC diagnosis (cases) or interview (controls); latency: asterisk indicates latency of 0–2 years; smoking: non-smokers, never- + former smokers + non-smokers (if so specified); ever smokers, former + current smokers + ever smokers (if so specified); heavy smokers, as defined in the original publication; histology: sqCLC, squamous cell carcinoma; non-SCLC, all types of LC apart from SCLC; LCLC, large cell lung carcinoma.