OBJECTIVE: To conduct a meta-analysis of epidemiologic studies to derive a pooled estimate of the association between asthma and cancer incidence and between allergy and cancer incidence. DATA SOURCES: Five cohort studies and 1 case-control study of cancer incidence in asthma, 4 studies of cancer mortality in asthma, and 3 studies of cancer incidence in allergy. STUDY SELECTION: We searched the National Library of Medicine Gateway to identify observational studies of cancer incidence in asthma and included any case-control or cohort study of incident cancers or of cancer mortality that met the predefined inclusion criteria. RESULTS: There was no significant association between asthma and cancer incidence. A single large cohort study demonstrated a protective effect of asthma; the remaining cohort studies demonstrated a slightly elevated risk of cancer associated with asthma. Inconsistencies in study design features included control of smoking in analyses, use of a control group, and methods to identify asthma status. CONCLUSIONS: The method of asthma identification affects the validity of asthma classification and the severity of asthma examined. The inconsistency in study designs could have contributed to the variability in results and the underestimation of the impact of asthma on cancer incidence.
OBJECTIVE: To conduct a meta-analysis of epidemiologic studies to derive a pooled estimate of the association between asthma and cancer incidence and between allergy and cancer incidence. DATA SOURCES: Five cohort studies and 1 case-control study of cancer incidence in asthma, 4 studies of cancer mortality in asthma, and 3 studies of cancer incidence in allergy. STUDY SELECTION: We searched the National Library of Medicine Gateway to identify observational studies of cancer incidence in asthma and included any case-control or cohort study of incident cancers or of cancer mortality that met the predefined inclusion criteria. RESULTS: There was no significant association between asthma and cancer incidence. A single large cohort study demonstrated a protective effect of asthma; the remaining cohort studies demonstrated a slightly elevated risk of cancer associated with asthma. Inconsistencies in study design features included control of smoking in analyses, use of a control group, and methods to identify asthma status. CONCLUSIONS: The method of asthma identification affects the validity of asthma classification and the severity of asthma examined. The inconsistency in study designs could have contributed to the variability in results and the underestimation of the impact of asthma on cancer incidence.