Risk of cervical cancer may be slightly elevated for users of injectables

Women who have used the injectable contraceptive depot medroxyprogesterone acetate (DMPA) may have a slightly increased risk of cervical cancer, according to an analysis of data from the World Health Organization's Collaborative Study of Neoplasia and Steroid Contraceptives, a large multinational, hospital-based study.(1) Ever-users were 43% more likely than nonusers to have developed carcinoma in situ. Their risk increased with duration of DMPA use, but not indefinitely: Cancer risk appeared to decline among long-term users 10 years after first use of DMPA and five years after last use.

The Collaborative Study was conducted to assess the relationship between DMPA and cancers of the breast, cervix uteri, ovary, corpus uteri and liver. For an examination of the risk of early-stage cervical cancer, the investigators drew on data from 1,217 women with carcinoma in situ and 8,956 controls who had been admitted to three hospitals in Thailand and one in Mexico. Women who had had a hysterectomy or who were admitted for complications of pregnancy or childbirth were excluded from the analysis, as were women under age 20, those who had never menstruated or had sexual intercourse and those who did not know whether they had used DMPA. Participants had to have been born after 1925 or 1930 (i.e., to have been of reproductive age when oral contraceptives first became available locally) and to have lived for at least one year in the geographic area served by a participating hospital.

The cancer patients were selected from among new admissions to the wards in each hospital where cervical cancer was treated. The diagnosis of carcinoma in situ was made by a pathologist working in each of the hospitals. All diagnoses were then confirmed by a pair of independent pathologists. Controls were chosen from among patients admitted to selected wards of each hospital for a wide variety of conditions. Women who were admitted for an obstetric or gynecologic condition, benign breast disease, cardiovascular disease or a previously diagnosed cancer were ineligible to serve as controls.

The accrual of cases began in October 1970; it continued in Mexico until 1985 and in Thailand until 1988. Information on known and suspected risk factors was gathered through a standardized questionnaire filled out during a personal interview. In 1984, a supplemental question was introduced in both countries to obtain information on participants' history of smoking and of genital warts. In addition, serologic samples to test for antibodies against herpes simplex virus and cytomegalovirus were collected in Thailand beginning in 1983, and information on the sexual behavior of each woman's current husband was collected there beginning in 1986.

On average, women with carcinoma in situ were younger than the controls (37.7 years vs. 40.9 years of age). The ratio of cancer patients to controls varied among the four hospitals, as did the prevalence of DMPA use. Thus, the investigators included age and hospital in all logistic models used to estimate the relative risk of cancer among DMPA users. They also evaluated the effects of other potential confounding variables and included in their final model the two variables--number of pregnancies and use of oral contraceptives--that altered relative risk estimates by more than 5%. The frequency of Pap smears was also included in the final logistic regression model.

Women with cervical carcinoma in situ were more likely than women in the control group to have ever used DMPA (23% vs. 15%). When five confounding variables--age, hospital, number of pregnancies, ever use of oral contraceptives and frequency of Pap smears--were taken into account, ever-users of DMPA had a 43% higher risk of cancer than did nonusers (a relative risk of 1.43).

Among women who joined the study during the time in which a smoking history was obtained, ever-users of DMPA were 18% more likely than nonusers to have cervical cancer (a relative risk of 1.18). However, in this subgroup, taking ever-users' smoking status into account did not appreciably alter their risk of cancer (relative risk of 1.19). Similarly, in Thailand, ever-users of DMPA who joined the study after serologic samples began to be collected were 45% more likely than never-users to have developed cancer (relative risk of 1.45), but their risk was not appreciably affected when prior herpes or cytomegalovirus infection were taken into account.

Moreover, the sexual behavior of a woman's current husband--such as whether he used condoms when he visited a prostitute, the total number of visits he made to prostitutes and his total number of episodes of sexually transmitted disease--did not appreciably alter the estimate of the risk of cancer among ever-users of DMPA. The investigators conclude, therefore, that their analyses "are not likely [to be] confounded by smoking, a prior...infection, or the sexual behavior of the women's husbands."

The researchers also estimated the cervical cancer risk separately for women with symptoms and for those without symptoms. They found that ever-users of DMPA who were asymptomatic at diagnosis were 92% more likely than nonusers to have carcinoma in situ (relative risk of 1.92), while those with symptomatic disease were 25% more likely (relative risk of 1.25); both of these increases in risk were statistically significant at p

Further analyses showed that as the length of time women had used DMPA increased, so did the likelihood of cervical cancer, and the increase in risk was higher among those who were asymptomatic when the cancer was detected. For example, asymptomatic women who had used DMPA for longer than five years were three times as likely as never-users to have carcinoma in situ (relative risk of 3.07), while among symptomatic women the comparable relative risk was 1.77.

There also was a significant upward trend in cancer risk with time since first use: Ever-users who had initiated DMPA use 5-10 years prior to the study were 69% more likely than nonusers to have symptomatic disease at diagnosis (a relative risk of 1.69), and were 89% more likely to have asymptomatic disease.

To reduce the chance that screening bias influenced risk estimates among asymptomatic women, the investigators limited their examination of the impact of long-term DMPA use on the risk of cancer to symptomatic women. They found that women who had first used DMPA 5-10 years before joining the study were more than twice as likely as nonusers to have symptomatic carcinoma in situ (a relative risk of 2.19). However, those who had first used DMPA more than 10 years before were not significantly more likely than nonusers to have cervical cancer. Additionally, symptomatic women who had used DMPA for more than five years and had stopped using the method 1-5 years before were twice as likely as nonusers to have carcinoma in situ (relative risk of 2.06); in contrast, those who had used DMPA for more than five years but who had last used the method more than five years before were not significantly more likely than nonusers to have developed cervical carcinoma.

The researchers also estimated the risk of symptomatic disease, based on whether a woman had had a Pap smear six months or more prior to being interviewed for the study. Ever-users who had not had a Pap smear were slightly more likely to have symptomatic cervical cancer than were those who had had a prior Pap smear (relative risks of 1.30 and 1.13, respectively). That relative risks were higher among unscreened women than among tested women suggests, according to the investigators, that "the relative risk of symptomatic disease in relation to DMPA use may have been somewhat underestimated."

The researchers comment that the apparent increase in risk associated with DMPA use was unlikely to be explained by overdiagnosis of cervical carcinoma in tissue samples from users, since all diagnoses of in situ disease were confirmed by two pathologists who had no knowledge of the women's contraceptive experience. They also observe that the results were not likely to have been affected by a failure to diagnose cancer among controls who were not screened for the disease, since it is relatively rare.

The investigators conclude that among long-term DMPA users, the seeming increase in the risk of carcinoma in situ appears to decline with time since last use. A previous study from the same group found that DMPA use was not associated with an increased risk of invasive cervical cancer.(2) These findings suggest "either that the small increase in risk of in situ disease is an artifact due to uncontrolled confounding or unidentified sources of bias, or represents the induction of a condition that is reversible or does not proceed to invasive disease," the researchers conclude. They also recommend that DMPA users receive periodic Pap smears.

References

1. D. B. Thomas et al., "Cervical Carcinoma in situ and Use of Depot-Medroxyprogesterone Acetate (DMPA)," Contraception, 51:25-31, 1995.

2. WHO Collaborative Study of Neoplasia and Steroid Contraceptives, "Depot-Medroxyprogesterone Acetate (DMPA) and Risk of Invasive Squamous Cell Cervical Cancer," Contraception, 45:299-312, 1992.

Copyright The Alan Guttmacher Institute Jul 1995
Provided by ProQuest Information and Learning Company. All rights Reserved