Although non-Hodgkin lymphoma (NHL) has not been considered to be a smoking
-related malignancy, recent investigations suggest otherwise. We evaluated
this association in a cohort of 37,336 women, aged 55-69 years, who reporte
d in a mailed questionnaire in 1986 information regarding smoking history a
s well as demographic, medical history and dietary factors. Cancer and mort
ality experience through 1996 was determined by linkage to the Iowa Cancer
Registry and other databases; there were 200 incident cases of NHL during t
he 380,231 total person-years of follow-up. Compared to never smokers, form
er (age-adjusted RR = 1.0; 95% CI 0.8-1.5) and current smokers (age-adjuste
d RR = 1.0; 95% CI 0.7-1.5) were not at elevated risk of NHL, and there was
no trend with pack-years smoked (p(trend) = 0.3). Multivariate adjustment
for other NHL risk factors did not alter these findings. Age-adjusted analy
sis by NHL subtype revealed a suggestive positive association of smoking wi
th follicular NHL [(RRformer= 1.3; 95% CI 0.6-28), (RRcurrent= 1.8; 95% CI
0.8-3.8)], which strengthened after multivariate adjustment [(RRformer= 1.6
; 95% CI 0.7-3.4), (RRcurrent = 2.3; 95% CI 1.0-5.0)]; there was no associa
tion for diffuse or small cleaved-cell NHL. Our study findings, which are c
onsistent with other recent investigations, suggest that smoking may be ass
ociated with an increased risk of follicular NHL.