A population-based, controlled study of the relation between HIV infection and cervical neoplasia

Br J Obstet Gynaecol. 1994 Nov;101(11):986-91. doi: 10.1111/j.1471-0528.1994.tb13045.x.

Abstract

Objective: To study a geographically defined population of women exposed to a drug-related risk for human immunodeficiency virus (HIV), together with carefully matched neighbourhood controls, in order to examine (1) the proportion of different groups having cervical cytopathology screening; (2) the association between HIV infection and cervical intraepithelial neoplasia; (3) the independent effect of CD4+ lymphocyte count and duration of HIV infection; and (4) the correspondence between cervical cytopathology and colposcopically directed biopsy.

Design: A population-based study.

Subjects: All women domiciled in Lothians with the following characteristics: between 1983 and 1987 they had a history of injection drug use or a seropositive partner with a history of injection drug use; they had a pregnancy after that exposure where their serostatus was known. In addition, neighbourhood controls were identified by computer matching for the following criteria--post-code sector, housing deprivation score, age, parity, pregnancy outcome, ethnic group, year of pregnancy, smoking.

Main outcome measures: Search was carried out for the cervical smear nearest in time to the index pregnancy end date, providing serostatus was known at that time. All identified smears were assessed by a cytopathologist without knowledge of clinical information, study group or serostatus.

Results: Of 376 women, appropriate cervical smears were identified for 336. The proportions screened in the different groups were similar. There were more abnormal smears in the seropositive group than in the drug-related seronegative (P < 0.01) or the neighbourhood control groups (P < 0.001). HIV-infected women with abnormal smears had lower CD4+ lymphocyte counts (P < 0.0005). There was a reasonable correspondence between cytopathological classification and histological grading, and this was of similar strength in cases and controls (weighted kappa 0.72, 0.74).

Conclusions: There is a definite relation between HIV infection and cervical intraepithelial neoplasia. This is related to immune depletion but whether this is the sole mechanism for the association is unclear.

MeSH terms

  • Adult
  • CD4 Lymphocyte Count
  • Female
  • HIV Infections / complications*
  • HIV Infections / epidemiology
  • HIV Infections / immunology
  • Humans
  • Population Surveillance
  • Risk Factors
  • Scotland / epidemiology
  • Substance Abuse, Intravenous / epidemiology
  • Uterine Cervical Dysplasia / epidemiology
  • Uterine Cervical Dysplasia / etiology*
  • Uterine Cervical Dysplasia / immunology
  • Uterine Cervical Neoplasms / epidemiology
  • Uterine Cervical Neoplasms / etiology*
  • Uterine Cervical Neoplasms / immunology