Ectopic pregnancy: a hospital-based case-control study of risk factors and duration of hospital stay

Abstract

To evaluate the association between ectopic pregnancy and two potential risk factors (blood group 0 and underweight) and to examine the duration of hospital stay by patients with ectopic pregnancy, a hospital-based case-control study was undertaken. Also included in the study were other risk factors, namely pelvic inflammatory disease, cigarette smoking, the use of an intrauterine device, tubal ligation, prior ectopic pregnancy, abortion, marital status, oral contraceptive use, prior abdominal surgery and cigarette smoking. The investigation included 72 cases diagnosed from January 1, 1976 through December 31, 1987 at the University

Hospital, Kansas City, Missouri and 72 pair-matched controls selected from live-birth deliveries at the same hospital. The risk factors, demographic profiles and ii

clinical care data were obtained via medical record abstract. Certain clinical characteristics of the 72 cases and 10 additional cases were also evaluated in a descriptive manner. Univariate McNemar analyses revealed no

statistically significant association between blood group 0 and EP although white cases were almost twice as likely to have blood group 0 as their matched controls. Likewise, being underweight revealed no statistically significant association with EP although it appeared to be protective among black cases (McNemar OR = 0.6). Four variables revealed strong risk factors for EP: having eversmoked (OR = 2.1, P = 0.04), previous fetal loss (OR = 4, P = 0.0001), previous EP

(OR = 14.3, P = 0.0001), and previous abdominal/pelvic surgery (OR = 6.3, p = 0.0001). There was no association demonstrated between abortion, stillbirth, oral contraception and EP. Being unmarried w~s found to be protective (OR = 0.4, p = 0.01). Whereas this was statistically significant among white cases (p < 0.02),

it was not among black cases (p = 0.4). Following stepwise logistic regression analysis (using: parity, eversmoking, underweight and prO), only one variable, pro, remained strong (p = 0.009). This showed that pro

iii

was a confounder in eversmoking. Using paired t test, cases were found to stay longer in the hospital than controls (p = 0.001). Those cases who had not been

seen by an obstetrician stayed longer in the hospital than those who had been seen by an obstetrician. Although nearly 50% of cases had ruptured fallopian tubes by time of admission, this did not make them stay longer in the hospital. A statistically significant protective factor in the study appeared to be the status of being unmarried.