Research Article

Cigarette smoking and human papillomavirus in patients with reported cervical cytological abnormality.

BMJ 1993; 306 doi: (Published 20 March 1993) Cite this as: BMJ 1993;306:749
  1. M P Burger,
  2. H Hollema,
  3. A S Gouw,
  4. W J Pieters,
  5. W G Quint
  1. Department of Obstetrics and Gynaecology, University Hospital, Groningen, Netherlands.


    OBJECTIVE--To assess the relation between two risk factors for cervical neoplasia: smoking and infection with oncogenic human papillomavirus. It has been suggested that smoking causes a local immunological defect, which could facilitate the infection and persistence of human papillomavirus. DESIGN--Cross sectional epidemiological study. Completion of a structured questionnaire by the patients, analysis of cervical scrapes for human papillomavirus, and morphological examination of biopsy specimens. SETTING--Outpatient gynaecological clinic. SUBJECTS--181 women with a report of cervical cytological abnormality. MAIN OUTCOME MEASURES--Prevalence of infection with oncogenic human papillomavirus and smoking habits. RESULTS--Oncogenic human papillomavirus was found in the cervix of 26 (41%) of the 63 women who did not smoke, 22 (58%) of the 38 who smoked 1-10 cigarettes a day, 28 (61%) of the 46 who smoked 11-20 cigarettes a day, and 26 (76%) of the 34 who smoked > or = 21 cigarettes a day. The prevalence of the virus thus increased in accordance with the number of cigarettes smoked (p = 0.001). This relation remained after adjustment for age at first intercourse and lifetime number of sexual partners. Of the 63 non-smokers, 23 had previously smoked at least 10 cigarettes a day at some time. Of these 23 women, 14 (61%) had oncogenic human papillomavirus in their cervix. Of the 40 women who had never smoked at least 10 cigarettes a day, 12 (30%) had the virus. The prevalence of oncogenic human papillomavirus in non-smokers therefore depended on previous smoking habits (p = 0.03). CONCLUSION--The dose dependent effect of cigarette smoking on the occurrence of oncogenic human papillomavirus favours a causal relation between these risk factors for cervical neoplasia.