- Email: [email protected]
Cigarette smoking as a risk factor for ectopic pregnancy
Gynecology
Cigarette smoking as a risk factor for ectopic pregnancy Mona Saraiya, MD, MPH,a, b Cynthia J. Berg, MD, MPH,b Juliette S. Kendrick, MD,b Lilo T. Strauss, MA,b Hani K. Atrash, MD, MPH,b and Young W. Ahn, MDc Atlanta, Georgia OBJECTIVE: Our purpose was to assess the risk of ectopic pregnancy among women who smoke cigarettes. STUDY DESIGN: We used data from a case-control study of ectopic pregnancy conducted from October 1988 to August 1990 at an inner-city hospital in Georgia. Cases were 196 non-Hispanic black women with a surgically confirmed ectopic pregnancy. Controls were non-Hispanic black women who had delivered either a live or a stillborn infant weighing at least 500 gm (n = 882) or who were pregnant and seeking an induced abortion (n = 237). RESULTS: After we adjusted for parity, douching history, history of infertility, and age, the odds ratio for ectopic pregnancy was 1.9 (95% confidence interval 1.4 to 2.7) for women who smoked during the periconception period compared with women who did not smoke at that time. After stratification by the amount of daily smoking during the periconception period, the odds ratio rose from 1.6 (95% confidence interval 0.9 to 2.9) for women who smoked 1 to 5 cigarettes to 1.7 (95% confidence interval 1.1 to 2.8) for women who smoked 6 to 10 cigarettes to 2.3 (95% confidence interval 1.3 to 4.0) for women who smoked 11 to 20 cigarettes, and to 3.5 (95% confidence interval 1.4 to 8.6) for women who smoked >20 cigarettes per day. CONCLUSION: In this inner-city population, cigarette smoking was an independent, dose-related risk factor for ectopic pregnancy among black women. The public health and medical care communities should inform the public of this additional risk associated with cigarette smoking and intensify intervention strategies to reduce cigarette smoking among women of reproductive age. (Am J Obstet Gynecol 1998;178:493-8.)
Key words: Ectopic pregnancy, cigarette smoking
This study assesses whether cigarette smoking is a potential risk factor for ectopic pregnancy and whether there is a dose-response effect. In 1992 the incidence of ectopic pregnancies was estimated to be 19.7 per 1000 reported pregnancies,1 a threefold increase from the incidence in 1970. Established risk factors for ectopic pregnancy include maternal age, low parity, low gravidity, pelvic inflammatory disease, sexually transmitted diseases (STDs), prior tubal surgery, history of infertility, intrauterine contraceptive device (IUD) use, and prior ectopic pregnancy.2 Previous studies that have examined the relationship between cigarette smoking and ectopic Office,a
From the Epidemic Intelligence Service, Epidemiology Program and the Division of Reproductive Health,b National Center for Chronic Diseases Prevention and Health Promotion, Centers for Disease Control and Prevention, and the Department of Gynecology and Obstetrics, Emory University School of Medicine.c This study was conducted by the Emory University School of Medicine, with funding from the Centers for Disease Control and Prevention. Received for publication May 12, 1997; revised July 10, 1997; accepted October 6, 1997. Reprint requests: Mona Saraiya, MD, MPH, Centers for Disease Control and Prevention, Division of Tuberculosis Elimination/International Activities, National Center for HIV, STD, and TB Prevention, Mailstop E10, 1600 Clifton Road NE, Atlanta, GA 30333. 6/1/86700
pregnancy have shown inconsistent results, with either no statistically significantly increased relative risk of ectopic pregnancy3-5 or increased relative risks ranging from 1.4 to 4.0.6-12 Only two studies5, 6 out of seven2, 4-6, 8-10 that have examined the number of cigarettes smoked at various times before or during pregnancy have shown a dose-response effect. Few of these studies simultaneously addressed issues of proper selection of control subjects, adequate sample size, possible dose-response effect, and adequate control for potential confounders.2, 10 Material and methods We examined data from a case-control study conducted at an inner-city hospital in Atlanta, Georgia. Eligible case women included all women aged 18 to 44 years who had a surgically confirmed ectopic pregnancy and were admitted to the hospital between October 1988 and August 1990. These women were identified through ongoing surveillance of the admission log of the gynecology service. Control women were a systematic random sample of two groups of women served by the hospital: (1) women who were delivered of a live or stillborn infant weighing ≥500 gm at the same hospital during the same time period (obstetric controls) and (2) women 493
494 Saraiya et al.
March 1998 Am J Obstet Gynecol
Table I. Number of women eligible, reasons not interviewed, number interviewed, and exclusions*
Eligible Reason for not interviewing Refused Illness Unavailable Other reason Total not interviewed Interviewed Unreliable interview Available for analysis Reason for exclusion* Previous ectopic pregnancy Previous tubal surgery Other race or ethnicity (includes Hispanic, white, and others) Current IUD use Abdominal pregnancy Missing data on smoking during periconception period Total excluded Included in analysis *Numbers
Cases
Controls
364
1957
35 2 6 27 70 294 0 294
137 0 403 112 652 1305 4 1301
35 35 24
24 7 151
2 1 1
0 0 0
98 196
182 1119
are based on hierarchical exclusions.
seeking an induced abortion who were identified at the preabortion visit to the ambulatory surgical clinic of their hospital (abortion controls). These groups of controls were selected in the 4:1 proportion of pregnant women in the population served by the hospital who do and do not seek to carry their pregnancies to term, as recommended by Weiss et al.13 To best represent pregnant women, the obstetric controls and the abortion controls in this study were combined into one group for all subsequent analyses. Details of the study’s control selection14 and characteristics for each control population15 have been described previously. A total of 81% of the eligible cases and 67% of eligible controls were interviewed Mondays through Fridays. A 1hour questionnaire that covered demographics, obstetric and medical histories, and personal habits was administered to the participants, and these self-reported data were used for analysis. For women who had smoked ≥100 cigarettes during their lifetime, the following information was obtained: age at initiation of smoking on a regular basis, total duration of cigarette use, average number of cigarettes usually smoked per day during the years smoked, and smoking status and number of cigarettes smoked during the periconception period. We defined the periconception period as the 6 months before the last normal menstrual period (LMP) plus the month after the LMP, the time when the purported effects of cigarette smoking might have the greatest impact on the implantation site of the pregnancy. No information was collected from the medical records.
Women who had a history of ectopic pregnancy, tubal surgery (including tubal ligation), or current IUD use were excluded because these are strong risk factors for ectopic pregnancy and we wanted to isolate the exposure of interest (Table I). Women who were of other than black race or who were Hispanic were excluded in this analysis, because smoking behaviors vary across racial and ethnic groups and the numbers in these groups were too small for separate analyses. One case of an abdominal pregnancy and one case with missing information on smoking during the periconception period were also excluded. After exclusions, 196 case women and 1119 control women (882 obstetric and 237 abortion controls) remained in our analysis. Possible confounding variables identified a priori from review of the literature included maternal age, education, marital status, household income, previous parity, previous gravidity, former IUD use, prior pelvic surgery, history of induced abortions, history of spontaneous abortions, history of infertility (having visited a physician for infertility problems), number of sexual partners, history of vaginal douching, history of pelvic inflammatory disease, and history of any STDs. Crude odds ratios by case-control status were calculated for these variables, as well as for smoking. Those variables that were significant in the univariate analysis were then included in the multivariate model. Unconditional logistic regression with the “change-inestimate” confounder selection strategy16 and the Hosmer-Lemeshow goodness-of-fit test17 were used to adjust for the potential confounders and determine the final model. To check for possible effect modification, we compared stratum-specific odds ratios for cigarette smoking and ectopic pregnancy,18 using biologically plausible interaction terms in the logistic regression models for age (18 to 24, 25 to 44 years), previous parity (continuous), history of infertility, and history of pelvic inflammatory disease. We examined the overall risk of ectopic pregnancy and smoking during the periconception period and then examined the possible dose-response effect of number of cigarettes smoked. We first divided the study population into two groups: women who had never smoked (never smokers) and women who had smoked 100 cigarettes in their lifetime (ever smokers). As the crude odds ratio for ever smokers was more than twice that of never smokers, we divided ever smokers into two groups: women who smoked in the past but who did not smoke during the periconception period (former smokers) and women who smoked during the periconception period (current smokers). The crude odds ratio for former smokers was not significantly different from that for never smokers. Because the small number of former smokers was inadequate for separate analysis, we combined former smokers with never smokers for further analyses.
Saraiya et al. 495
Volume 178, Number 3 Am J Obstet Gynecol
Table II. Characteristics of ectopic pregnancy cases and controls Cases (n = 196) Characteristic Maternal age ≥25 yr* Education <12 yr† Marital status—Ever married or living together* Household income† <$10,000 $10,000-$29,999 ≥$30,000 Previous parity‡ 0 1 ≥2 Previous gravidity§ 0 1 ≥2 Former IUD use* Prior pelvic surgery† History of induced abortion History of spontaneous abortion History of infertility* No. of sexual partners >4* History of douching* History of pelvic inflammatory disease* History of STDs† Ever smoking* Smoking during periconception periodll
Controls (n = 1119)
No.
%
No.
%
106 52 130
54.1 26.5 66.3
403 283 593
36.0 25.3 53.0
94 69 14
53.1 39.0 7.9
591 335 69
59.4 33.7 6.9
60 70 66
30.6 35.7 33.7
275 372 472
24.6 33.2 42.2
36 54 106 27 80 54 40 27 114 190 41 118 85 80
18.4 27.6 54.1 13.8 43.0 27.6 20.4 13.8 58.8 96.9 20.9 60.8 43.4 40.8
172 285 662 91 427 343 182 39 502 998 142 652 293 251
15.4 25.5 59.2 8.1 40.3 30.7 16.3 3.5 44.9 89.2 12.7 58.7 26.2 22.4
*Difference is significant at p < 0.05. †Missing data: education, 1 control;
household income, 19 cases and 124 controls; prior pelvic surgery, 10 cases and 59 controls; history of STDs, 2 cases and 9 controls. ‡Difference is significant at p = 0.06. §Percentages do not add to 100.0 because of rounding. llPericonception period defined as the 6 months before LMP plus the month after the LMP.
To examine dose response, we classified smoking during the periconception period into five categories: (1) nonsmokers (former smokers and never smokers), (2) women who smoked 1 to 5 cigarettes per day, (3) women who smoked 6 to 10 cigarettes per day, (4) women who smoked 11 to 20 cigarettes per day, and (5) women who smoked ≥21 cigarettes per day. The four groups of current smokers were chosen to evenly distribute the number of smokers by pack categories such as quarter pack days, half pack days, or pack days. The study proposal, consent form, and questionnaire were reviewed and approved by the institutional review boards of Emory University and the Centers for Disease Control and Prevention. Results Approximately one fifth of the eligible cases and one third of the eligible controls were not interviewed because of refusals, illness, unavailability, or other reasons (Table I). One third of the interviewed cases and about one seventh of the interviewed controls were excluded, as previously discussed, which left 196 cases and 1119 controls for use in the analysis.
Bivariate analysis showed that women with ectopic pregnancy were significantly older than the controls and were more likely to have been married at some time (Table II). Compared with controls, cases were of lower previous parity and were more likely to have used an IUD, seen a physician for infertility in the past, and reported pelvic inflammatory disease. Cases were also more likely than controls to have had more than four sexual partners in their lifetime, ever douched vaginally, ever smoked (43% cases vs 26% controls), and smoked during the periconception period (41% vs 22% controls). We found no statistically significant difference between cases and controls in their education, income, previous gravidity, history of pelvic surgery, induced abortions, spontaneous abortions, or history of STDs. The distribution of number of cigarettes smoked among women who smoked during the periconception period was largely skewed to those smoking fewer cigarettes; 27% reported smoking between 1 and 5 cigarettes, 42% reported smoking 6 to 10 cigarettes, 25% reported smoking 11 to 20 cigarettes, and 6% reported smoking >20 cigarettes (data not shown). The crude odds ratio of 2.4 (95% confidence interval
496 Saraiya et al.
March 1998 Am J Obstet Gynecol
Table III. Smoking during periconception period* and risk of ectopic pregnancy Cases Smoking during periconception period Never or former smokers 1-5 Cigarettes/day 6-10 Cigarettes/day 11-20 Cigarettes/day >20 Cigarettes/day
Controls
No.
%
No.
%†
Odds ratio‡ and 95% confidence interval
116 18 30 23 9
59.2 9.2 15.3 11.7 4.6
868 70 108 59 14
77.6 6.3 9.7 5.3 1.2
1.0 (Referent) 1.6 (0.9-2.9) 1.7 (1.1-2.8) 2.3 (1.3-4.0) 3.5 (1.4-8.6)
*Periconception period defined as the 6 months before the LMP plus the month after the LMP. †Percentages do not add to 100.0 because of rounding. ‡Adjusted for age, parity, infertility, and douching; Hosmer-Lemeshow goodness-of-fit test p = 0.5799.
1.7 to 3.3) for ectopic pregnancy for women who smoked during the periconception period compared with those who did not smoke during the periconception period changed to an odds ratio of 1.9 (95% confidence interval 1.4 to 2.7) when adjusted for age, previous parity, history of infertility, and vaginal douching. No effect modification was seen by age, previous parity, history of infertility, or history of pelvic inflammatory disease. To assess whether there was a dose-response effect, we stratified by the amount of smoking during the periconception period. The adjusted odds ratio of 1.6 for women who smoked 1 to 5 cigarettes per day rose to 1.7 for those smoking 6 to 10 cigarettes per day, to 2.3 for those women smoking 11 to 20 cigarettes per day, and to 3.5 for those smoking >20 cigarettes per day (Table III). We tested for the significance of the dose-response trend by replacing smoking categories in the final model with the number of cigarettes smoked as a continuous variable. Smoking entered as a continuous variable was statistically significant (p = 0.0002), indicating an increase in risk of ectopic pregnancy with increasing number of cigarettes smoked. When history of pelvic inflammatory disease was added to the model, the odds ratio for smoking, stratified by number of cigarettes smoked, did not change and the Hosmer-Lemeshow goodness-of-fit test worsened; thus pelvic inflammatory disease was not included in the final model. Comment In this inner-city population of black women, cigarette smoking was an independent and dose-related risk factor for ectopic pregnancy after we controlled for other important risk factors such as age, previous parity, history of infertility, and vaginal douching. The elevated risk ranged from 1.6 to 3.5 times that of nonsmokers, depending on the number of cigarettes smoked. In addition, because smoking as a continuous variable showed a significantly increased risk with increasing number of cigarettes, we are confident in calling this a dose-response effect. Although data from several other studies support
the hypothesis that cigarette smoking is a risk factor for ectopic pregnancy, this study’s strengths lie in its study population, its control selection, its exposure definition, and its control for confounders. This study is one of the first to examine cigarette smoking as a risk factor for ectopic pregnancy among a population at higher risk for ectopic pregnancy but at a lower risk of cigarette smoking (i.e., black women), yet we showed an increased twofold risk of ectopic pregnancy among cigarette smokers, similar to that found in other studies among predominantly white women.6, 8, 9 The ectopic pregnancy rate has been reported to be one third higher among black and other minority women than among white women whereas the risk of death from an ectopic pregnancy among minority women is three times as high as that among white women.19 Recent national data report a smoking prevalence of 24% among black women of reproductive age in 1993 (comparable to smoking prevalence of 22% among our controls) compared with 30% among white women.20 Although cigarette smoking might not explain the increased incidence of ectopic pregnancy seen among black women, our study documents the role that a modifiable risk factor such as cigarette smoking might play in ectopic pregnancy. The control group we used consisted of women who were delivered of live or stillborn infants and women who were seeking induced abortion, in a ratio similar to that found in the hospital population. Combining these two groups, as others have suggested, allowed us to make a better estimate of smoking in the entire population of pregnant women who use this hospital than the use of either control group alone would have done.12 The exposure, cigarette smoking, has been defined in different ways for different studies. Several studies have used smoking at the time of the interview2, 7, 10, 11 as a proxy for smoking during the periconception period, resulting in a different period of reference for the cases than the controls. More important, such an exposure measure can result in an inaccurate indication of the
Saraiya et al. 497
Volume 178, Number 3 Am J Obstet Gynecol
smoking status at time of implantation, because many women quit smoking during pregnancy.21 Because we specifically asked whether subjects smoked during the 6 months before the LMP and the month after the LMP, our exposure measure was consistent for cases and controls and included the crucial time frame for implantation of a pregnancy. We were also able to control for many of the established risk factors for ectopic pregnancy found in other studies, as well as for potential but unestablished risk factors such as vaginal douching.5, 22, 23 Pelvic inflammatory disease is a well-known risk factor for ectopic pregnancy, and whether to adjust for a history of pelvic inflammatory disease in the context of cigarette smoking is an important methodologic issue. Some researchers, convinced that pelvic inflammatory disease is in the causal pathway of cigarette smoking and ectopic pregnancy, favor the exclusion of pelvic inflammatory disease as a covariate; others, convinced that pelvic inflammatory disease is an independent risk factor, favor its inclusion. In this analysis the inclusion or exclusion of the pelvic inflammatory disease variable in the final model did not affect the odds ratio of cigarette smoking or any other covariates, and pelvic inflammatory disease was not considered a significant confounder. Moreover, the Hosmer-Lemeshow goodness-of-fit value was improved when pelvic inflammatory disease was not included. Similar reasoning could be applied to the history of infertility serving as an intermediary between cigarette smoking and ectopic pregnancy. However in this case infertility did affect the odds ratio of other covariates with a worsened HosmerLemeshow goodness-of-fit value when the infertility covariate was left out and therefore was considered a significant confounder and kept in the model. The decision to include or exclude pelvic inflammatory disease or infertility in the model might have been different if we had been able to use clinical data that might have captured “silent” pelvic inflammatory disease or true infertility instead of the self-reported clinical pelvic inflammatory disease and self-reported infertility available in this analysis. A limitation of this epidemiologic study, as well as many others, is the reliance on self-report of various exposures that could potentially be affected by recall bias. Previous work has demonstrated that self-reported data underestimate smoking24 and pelvic inflammatory disease prevalences.25 Biochemical validation of self-reported smoking with markers such as serum or urine cotinine and validation of pelvic inflammatory disease history with laparoscopy would have been useful. Because these were past exposures, measurement of these exposures was not feasible with our case-control study. However, our outcome, ectopic pregnancy, was well documented. Another limitation is our inability to analyze never and former smokers separately because of the small number
of former smokers. One study of the risk of ectopic pregnancy among former smokers concluded that the effect of cigarette smoking was the result of short-term mechanisms, whereas another implicated long-term mechanisms.8, 9 Our results suggest smoking as an etiologic factor in ectopic pregnancy, but we were not able to truly discern a more chronic versus a more acute effect of tobacco smoke although the lack of increased risk among the small number of former smokers would tend to argue for the latter. On the basis of laboratory studies in humans and animals, researchers have postulated several mechanisms by which cigarette smoking might play a role in ectopic pregnancies. These mechanisms include one or more of the following: delayed ovulation, altered tubal and uterine motility, or altered immunity.24-30 To date, no study has supported a specific mechanism by which cigarette smoking affects the occurrence of ectopic pregnancy. We know already that the effects of maternal tobacco use during and after pregnancy include increased risk of spontaneous abortion, abruptio placentae, preterm premature rupture of membranes, low birth weight, sudden infant death syndrome, perinatal death, and childhood respiratory disorders.22, 31, 32 With an estimated 26% prevalence of cigarette smoking among women of reproductive age,18 increased awareness and effective counseling techniques can potentially influence a woman’s decision to smoke before she becomes pregnant, especially when the safety of a future pregnancy is at risk. Recently released publications33, 34 that provide specific recommendations for physicians and other health care providers to help patients, especially women, quit smoking may be useful. More important, efforts should be directed at preventing young women from ever beginning to smoke.32 REFERENCES
1. Centers for Disease Control. Ectopic pregnancy—United States, 1990-1992. MMWR Morb Mortal Wkly Rep 1995;4:46-8. 2. Chow WH, Daling JR, Cates W, Greenberg RS. Epidemiology of ectopic pregnancy. Epidemiol Rev 1987;9:70-93. 3. Zahniser SC, Gupta SC, Kendrick JS, Lee NC, Spirtas R. Tubal pregnancy and cigarette smoking: is there an association? J Wom Health 1994;3(5):329-36. 4. Levin AA, Schoenbaum SC, Stubblefield PG, Zimicki S, Monson RR, Ryan KJ. Ectopic pregnancy and prior induced abortion. Am J Public Health 1982;72:253-6. 5. Parazzini F, Tozzi L, Ferraroni M, Bocciolone L, La Vecchia C, Fedele L. Risk factors for ectopic pregnancy: an Italian case-control study. Obstet Gynecol 1992;80:821-6. 6. Coste J, Job-Spira N, Fernandez H, Papiernik E, Spira A. Risk factors for ectopic pregnancy: a case-control study in France, with special focus on infectious factors. Am J Epidemiol 1991;133:839-49. 7. Handler A, Davis F, Ferre C, Yeko T. The relationship of smoking and ectopic pregnancy. Am J Public Health 1989;79:1239-42. 8. Campbell OM, Gray RH. Smoking and ectopic pregnancy: a multinational case-control study. In: Rosenberg MD, editor. Smoking and reproductive health. Littleton (MA): PSG; 1987. p. 70-4. 9. Chow WH, Daling JR, Weiss NS, Voigt LF. Maternal cigarette
498 Saraiya et al.
10.
11.
12.
13.
14.
15.
16. 17.
18.
19.
20.
21.
22.
smoking and tubal pregnancy. Obstet Gynecol 1988;71:16770. Stergachis A, Scholes D, Daling JR, Weiss NS, Chu J. Maternal cigarette smoking and the risk of tubal pregnancy. Am J Epidemiol 1991;133:332-7. Phillips RS, Tuomala RE, Feldblum PJ, Schachter J, Rosenberg MJ, Aronson MD. The effect of cigarette smoking, Chlamydia trachomatis infection, and vaginal douching on ectopic pregnancy. Obstet Gynecol 1992;79:85-90. Kalandidi A, Doulgerakis M, Tzonos A, Hsieh CC, Aravandinos D, Trichopoulos D. Induced abortions, contraceptive practices, and tobacco smoking as risk factors for ectopic pregnancy in Athens, Greece. Br J Obstet Gynaecol 1991;98:207-13. Weiss NS, Daling JR, Chow WH. Control definition in case-control studies of ectopic pregnancy. Am J Public Health 1985;75:67-8. Kendrick JS, Tierney EF, Lawson HW, Strauss LT, Klein L, Atrash HK. Previous cesarean delivery and the risk of ectopic pregnancy. Obstet Gynecol. 1996;87:297-301. Atrash HK, Strauss LT, Kendrick JS, Skjeldestad FE, Ahn YW. The relation between induced abortion and ectopic pregnancy. Obstet Gynecol. 1997;89:512-8. Maldonado G, Greenland S. Simulation study of confounder-selection strategies. Am J Epidemiol 1993;138:923-6. Lemeshow S, Hosmer DW. A review of goodness of fit statistics for use in the development of logistic regression models. Am J Epidemiol 1982;115:92-106. Kleinbaum DG, Kupper LL, Morgenstern H. Epidemiological research: principles and quantitative methods. New York: Van Nostrand Reinhold; 1982. Goldner TE, Lawson HW, Xia S, Atrash HK. Surveillance for ectopic pregnancy—United States, 1970-1989. MMWR Morb Mortal Wkly Rep 1993;42:73SS-85SS. CDC Surveillance Summaries, 1993 Dec. Publication No.: SS-6. Husten CG, Chrismon JH, Reddy MN. Trends and effects of cigarette smoking among girls and women in the United States. 1965-1993. J Am Med Wom Assoc 1996;51:11-8. Floyd RL, Zahniser SC, Gunter EP, Kendrick JS. Smoking during pregnancy: prevalence, effects, and intervention strategies. Birth 1991;18:47-52. Daling JR, Weiss NS, Schwartz SM, Stergachis A, Wang SP, Foy H,
March 1998 Am J Obstet Gynecol
23.
24.
25.
26.
27. 28. 29. 30.
31.
32.
33.
34.
et al. Vaginal douching and the risk of tubal pregnancy. Epidemiology 1991;2:40-8. Kendrick JS, Atrash HK, Strauss LT, Garguillo PM, Ahn YW. Vaginal douching: an important risk factor for ecotpic pregnancy among black women. Am J Obstet Gynecol 1997;176:991-7. DiFranza JR, Lew RA. Effect of maternal cigarette smoking on pregnancy complications and sudden infant death syndrome. J Fam Pract 1995;40:385-94. Westrom L. Incidence, prevalence, and trends of acute pelvic inflammatory disease and its consequences in industrialized countries. Am J Obstet Gynecol 1980;138:880-92. Campbell O. Ectopic pregnancy and smoking: confounding or causality? In: Poswillo D, Alberman E, editors. Effects of smoking on the fetus, neonate, and child. New York: Oxford University Press; 1992. p. 23-44. Mattison DR. The effects of smoking on fertility from gametogenesis to implantation. Environ Res 1982;28:410-33. Macmahon B, Trichopoulos D, Cole P, Brown J. Cigarette smoking and urinary estrogens. N Engl J Med 1982;307:1063-5. Yoshinaga K, Rice C, Krenn J, Pilot RL. Effects of nicotine on early pregnancy in the rat. Biol Reprod 1979;20:294-303. Neri A, Eckerling B. Influence of smoking and adrenaline (epinephrine) on the uterotubal insufflation test (Rubin test). Fertil Steril 1969;20:818-28. Evans J, Golding J. Parental smoking and respiratory problems in childhood. In: Poswillo D, Alberman E, editors. Effects of smoking on the fetus, neonate, and child. New York: Oxford University Press; 1992. p. 121-37. Nicholl J, O’Cathain A. Antenatal smoking, postnatal passive smoking, and the sudden infant death syndrome. In: Poswillo D, Alberman E, editors. Effects of smoking on the fetus, neonate, and child. New York: Oxford University Press; 1992. p. 138-49. Fiore MC, Bailey WC, Cohen SJ, Dorfman SF, Goldstein MC, Gritz ER, et al. Smoking cessation: a guide for primary care clinicians. Clinical practice guideline. Quick reference guide for primary care clinicians. No.: 18. Rockville (MD): Dept. of Health and Human Services (US), Public Health Service, Agency for Health Care Policy and Research and Centers for Disease Control and Prevention; 1996 Apr. AHCPR Publication No.: 96-0693. Kendrick JS, Merrit RA. Women and smoking: an update for the 1990s. Am J Obstet Gynecol 1996;175:528-35.
Cigarette smoking as a risk factor for ectopic pregnancy Mona Saraiya, MD, MPH,a, b Cynthia J. Berg, MD, MPH,b Juliette S. Kendrick, MD,b Lilo T. Strauss, MA,b Hani K. Atrash, MD, MPH,b and Young W. Ahn, MDc Atlanta, Georgia OBJECTIVE: Our purpose was to assess the risk of ectopic pregnancy among women who smoke cigarettes. STUDY DESIGN: We used data from a case-control study of ectopic pregnancy conducted from October 1988 to August 1990 at an inner-city hospital in Georgia. Cases were 196 non-Hispanic black women with a surgically confirmed ectopic pregnancy. Controls were non-Hispanic black women who had delivered either a live or a stillborn infant weighing at least 500 gm (n = 882) or who were pregnant and seeking an induced abortion (n = 237). RESULTS: After we adjusted for parity, douching history, history of infertility, and age, the odds ratio for ectopic pregnancy was 1.9 (95% confidence interval 1.4 to 2.7) for women who smoked during the periconception period compared with women who did not smoke at that time. After stratification by the amount of daily smoking during the periconception period, the odds ratio rose from 1.6 (95% confidence interval 0.9 to 2.9) for women who smoked 1 to 5 cigarettes to 1.7 (95% confidence interval 1.1 to 2.8) for women who smoked 6 to 10 cigarettes to 2.3 (95% confidence interval 1.3 to 4.0) for women who smoked 11 to 20 cigarettes, and to 3.5 (95% confidence interval 1.4 to 8.6) for women who smoked >20 cigarettes per day. CONCLUSION: In this inner-city population, cigarette smoking was an independent, dose-related risk factor for ectopic pregnancy among black women. The public health and medical care communities should inform the public of this additional risk associated with cigarette smoking and intensify intervention strategies to reduce cigarette smoking among women of reproductive age. (Am J Obstet Gynecol 1998;178:493-8.)
Key words: Ectopic pregnancy, cigarette smoking
This study assesses whether cigarette smoking is a potential risk factor for ectopic pregnancy and whether there is a dose-response effect. In 1992 the incidence of ectopic pregnancies was estimated to be 19.7 per 1000 reported pregnancies,1 a threefold increase from the incidence in 1970. Established risk factors for ectopic pregnancy include maternal age, low parity, low gravidity, pelvic inflammatory disease, sexually transmitted diseases (STDs), prior tubal surgery, history of infertility, intrauterine contraceptive device (IUD) use, and prior ectopic pregnancy.2 Previous studies that have examined the relationship between cigarette smoking and ectopic Office,a
From the Epidemic Intelligence Service, Epidemiology Program and the Division of Reproductive Health,b National Center for Chronic Diseases Prevention and Health Promotion, Centers for Disease Control and Prevention, and the Department of Gynecology and Obstetrics, Emory University School of Medicine.c This study was conducted by the Emory University School of Medicine, with funding from the Centers for Disease Control and Prevention. Received for publication May 12, 1997; revised July 10, 1997; accepted October 6, 1997. Reprint requests: Mona Saraiya, MD, MPH, Centers for Disease Control and Prevention, Division of Tuberculosis Elimination/International Activities, National Center for HIV, STD, and TB Prevention, Mailstop E10, 1600 Clifton Road NE, Atlanta, GA 30333. 6/1/86700
pregnancy have shown inconsistent results, with either no statistically significantly increased relative risk of ectopic pregnancy3-5 or increased relative risks ranging from 1.4 to 4.0.6-12 Only two studies5, 6 out of seven2, 4-6, 8-10 that have examined the number of cigarettes smoked at various times before or during pregnancy have shown a dose-response effect. Few of these studies simultaneously addressed issues of proper selection of control subjects, adequate sample size, possible dose-response effect, and adequate control for potential confounders.2, 10 Material and methods We examined data from a case-control study conducted at an inner-city hospital in Atlanta, Georgia. Eligible case women included all women aged 18 to 44 years who had a surgically confirmed ectopic pregnancy and were admitted to the hospital between October 1988 and August 1990. These women were identified through ongoing surveillance of the admission log of the gynecology service. Control women were a systematic random sample of two groups of women served by the hospital: (1) women who were delivered of a live or stillborn infant weighing ≥500 gm at the same hospital during the same time period (obstetric controls) and (2) women 493
494 Saraiya et al.
March 1998 Am J Obstet Gynecol
Table I. Number of women eligible, reasons not interviewed, number interviewed, and exclusions*
Eligible Reason for not interviewing Refused Illness Unavailable Other reason Total not interviewed Interviewed Unreliable interview Available for analysis Reason for exclusion* Previous ectopic pregnancy Previous tubal surgery Other race or ethnicity (includes Hispanic, white, and others) Current IUD use Abdominal pregnancy Missing data on smoking during periconception period Total excluded Included in analysis *Numbers
Cases
Controls
364
1957
35 2 6 27 70 294 0 294
137 0 403 112 652 1305 4 1301
35 35 24
24 7 151
2 1 1
0 0 0
98 196
182 1119
are based on hierarchical exclusions.
seeking an induced abortion who were identified at the preabortion visit to the ambulatory surgical clinic of their hospital (abortion controls). These groups of controls were selected in the 4:1 proportion of pregnant women in the population served by the hospital who do and do not seek to carry their pregnancies to term, as recommended by Weiss et al.13 To best represent pregnant women, the obstetric controls and the abortion controls in this study were combined into one group for all subsequent analyses. Details of the study’s control selection14 and characteristics for each control population15 have been described previously. A total of 81% of the eligible cases and 67% of eligible controls were interviewed Mondays through Fridays. A 1hour questionnaire that covered demographics, obstetric and medical histories, and personal habits was administered to the participants, and these self-reported data were used for analysis. For women who had smoked ≥100 cigarettes during their lifetime, the following information was obtained: age at initiation of smoking on a regular basis, total duration of cigarette use, average number of cigarettes usually smoked per day during the years smoked, and smoking status and number of cigarettes smoked during the periconception period. We defined the periconception period as the 6 months before the last normal menstrual period (LMP) plus the month after the LMP, the time when the purported effects of cigarette smoking might have the greatest impact on the implantation site of the pregnancy. No information was collected from the medical records.
Women who had a history of ectopic pregnancy, tubal surgery (including tubal ligation), or current IUD use were excluded because these are strong risk factors for ectopic pregnancy and we wanted to isolate the exposure of interest (Table I). Women who were of other than black race or who were Hispanic were excluded in this analysis, because smoking behaviors vary across racial and ethnic groups and the numbers in these groups were too small for separate analyses. One case of an abdominal pregnancy and one case with missing information on smoking during the periconception period were also excluded. After exclusions, 196 case women and 1119 control women (882 obstetric and 237 abortion controls) remained in our analysis. Possible confounding variables identified a priori from review of the literature included maternal age, education, marital status, household income, previous parity, previous gravidity, former IUD use, prior pelvic surgery, history of induced abortions, history of spontaneous abortions, history of infertility (having visited a physician for infertility problems), number of sexual partners, history of vaginal douching, history of pelvic inflammatory disease, and history of any STDs. Crude odds ratios by case-control status were calculated for these variables, as well as for smoking. Those variables that were significant in the univariate analysis were then included in the multivariate model. Unconditional logistic regression with the “change-inestimate” confounder selection strategy16 and the Hosmer-Lemeshow goodness-of-fit test17 were used to adjust for the potential confounders and determine the final model. To check for possible effect modification, we compared stratum-specific odds ratios for cigarette smoking and ectopic pregnancy,18 using biologically plausible interaction terms in the logistic regression models for age (18 to 24, 25 to 44 years), previous parity (continuous), history of infertility, and history of pelvic inflammatory disease. We examined the overall risk of ectopic pregnancy and smoking during the periconception period and then examined the possible dose-response effect of number of cigarettes smoked. We first divided the study population into two groups: women who had never smoked (never smokers) and women who had smoked 100 cigarettes in their lifetime (ever smokers). As the crude odds ratio for ever smokers was more than twice that of never smokers, we divided ever smokers into two groups: women who smoked in the past but who did not smoke during the periconception period (former smokers) and women who smoked during the periconception period (current smokers). The crude odds ratio for former smokers was not significantly different from that for never smokers. Because the small number of former smokers was inadequate for separate analysis, we combined former smokers with never smokers for further analyses.
Saraiya et al. 495
Volume 178, Number 3 Am J Obstet Gynecol
Table II. Characteristics of ectopic pregnancy cases and controls Cases (n = 196) Characteristic Maternal age ≥25 yr* Education <12 yr† Marital status—Ever married or living together* Household income† <$10,000 $10,000-$29,999 ≥$30,000 Previous parity‡ 0 1 ≥2 Previous gravidity§ 0 1 ≥2 Former IUD use* Prior pelvic surgery† History of induced abortion History of spontaneous abortion History of infertility* No. of sexual partners >4* History of douching* History of pelvic inflammatory disease* History of STDs† Ever smoking* Smoking during periconception periodll
Controls (n = 1119)
No.
%
No.
%
106 52 130
54.1 26.5 66.3
403 283 593
36.0 25.3 53.0
94 69 14
53.1 39.0 7.9
591 335 69
59.4 33.7 6.9
60 70 66
30.6 35.7 33.7
275 372 472
24.6 33.2 42.2
36 54 106 27 80 54 40 27 114 190 41 118 85 80
18.4 27.6 54.1 13.8 43.0 27.6 20.4 13.8 58.8 96.9 20.9 60.8 43.4 40.8
172 285 662 91 427 343 182 39 502 998 142 652 293 251
15.4 25.5 59.2 8.1 40.3 30.7 16.3 3.5 44.9 89.2 12.7 58.7 26.2 22.4
*Difference is significant at p < 0.05. †Missing data: education, 1 control;
household income, 19 cases and 124 controls; prior pelvic surgery, 10 cases and 59 controls; history of STDs, 2 cases and 9 controls. ‡Difference is significant at p = 0.06. §Percentages do not add to 100.0 because of rounding. llPericonception period defined as the 6 months before LMP plus the month after the LMP.
To examine dose response, we classified smoking during the periconception period into five categories: (1) nonsmokers (former smokers and never smokers), (2) women who smoked 1 to 5 cigarettes per day, (3) women who smoked 6 to 10 cigarettes per day, (4) women who smoked 11 to 20 cigarettes per day, and (5) women who smoked ≥21 cigarettes per day. The four groups of current smokers were chosen to evenly distribute the number of smokers by pack categories such as quarter pack days, half pack days, or pack days. The study proposal, consent form, and questionnaire were reviewed and approved by the institutional review boards of Emory University and the Centers for Disease Control and Prevention. Results Approximately one fifth of the eligible cases and one third of the eligible controls were not interviewed because of refusals, illness, unavailability, or other reasons (Table I). One third of the interviewed cases and about one seventh of the interviewed controls were excluded, as previously discussed, which left 196 cases and 1119 controls for use in the analysis.
Bivariate analysis showed that women with ectopic pregnancy were significantly older than the controls and were more likely to have been married at some time (Table II). Compared with controls, cases were of lower previous parity and were more likely to have used an IUD, seen a physician for infertility in the past, and reported pelvic inflammatory disease. Cases were also more likely than controls to have had more than four sexual partners in their lifetime, ever douched vaginally, ever smoked (43% cases vs 26% controls), and smoked during the periconception period (41% vs 22% controls). We found no statistically significant difference between cases and controls in their education, income, previous gravidity, history of pelvic surgery, induced abortions, spontaneous abortions, or history of STDs. The distribution of number of cigarettes smoked among women who smoked during the periconception period was largely skewed to those smoking fewer cigarettes; 27% reported smoking between 1 and 5 cigarettes, 42% reported smoking 6 to 10 cigarettes, 25% reported smoking 11 to 20 cigarettes, and 6% reported smoking >20 cigarettes (data not shown). The crude odds ratio of 2.4 (95% confidence interval
496 Saraiya et al.
March 1998 Am J Obstet Gynecol
Table III. Smoking during periconception period* and risk of ectopic pregnancy Cases Smoking during periconception period Never or former smokers 1-5 Cigarettes/day 6-10 Cigarettes/day 11-20 Cigarettes/day >20 Cigarettes/day
Controls
No.
%
No.
%†
Odds ratio‡ and 95% confidence interval
116 18 30 23 9
59.2 9.2 15.3 11.7 4.6
868 70 108 59 14
77.6 6.3 9.7 5.3 1.2
1.0 (Referent) 1.6 (0.9-2.9) 1.7 (1.1-2.8) 2.3 (1.3-4.0) 3.5 (1.4-8.6)
*Periconception period defined as the 6 months before the LMP plus the month after the LMP. †Percentages do not add to 100.0 because of rounding. ‡Adjusted for age, parity, infertility, and douching; Hosmer-Lemeshow goodness-of-fit test p = 0.5799.
1.7 to 3.3) for ectopic pregnancy for women who smoked during the periconception period compared with those who did not smoke during the periconception period changed to an odds ratio of 1.9 (95% confidence interval 1.4 to 2.7) when adjusted for age, previous parity, history of infertility, and vaginal douching. No effect modification was seen by age, previous parity, history of infertility, or history of pelvic inflammatory disease. To assess whether there was a dose-response effect, we stratified by the amount of smoking during the periconception period. The adjusted odds ratio of 1.6 for women who smoked 1 to 5 cigarettes per day rose to 1.7 for those smoking 6 to 10 cigarettes per day, to 2.3 for those women smoking 11 to 20 cigarettes per day, and to 3.5 for those smoking >20 cigarettes per day (Table III). We tested for the significance of the dose-response trend by replacing smoking categories in the final model with the number of cigarettes smoked as a continuous variable. Smoking entered as a continuous variable was statistically significant (p = 0.0002), indicating an increase in risk of ectopic pregnancy with increasing number of cigarettes smoked. When history of pelvic inflammatory disease was added to the model, the odds ratio for smoking, stratified by number of cigarettes smoked, did not change and the Hosmer-Lemeshow goodness-of-fit test worsened; thus pelvic inflammatory disease was not included in the final model. Comment In this inner-city population of black women, cigarette smoking was an independent and dose-related risk factor for ectopic pregnancy after we controlled for other important risk factors such as age, previous parity, history of infertility, and vaginal douching. The elevated risk ranged from 1.6 to 3.5 times that of nonsmokers, depending on the number of cigarettes smoked. In addition, because smoking as a continuous variable showed a significantly increased risk with increasing number of cigarettes, we are confident in calling this a dose-response effect. Although data from several other studies support
the hypothesis that cigarette smoking is a risk factor for ectopic pregnancy, this study’s strengths lie in its study population, its control selection, its exposure definition, and its control for confounders. This study is one of the first to examine cigarette smoking as a risk factor for ectopic pregnancy among a population at higher risk for ectopic pregnancy but at a lower risk of cigarette smoking (i.e., black women), yet we showed an increased twofold risk of ectopic pregnancy among cigarette smokers, similar to that found in other studies among predominantly white women.6, 8, 9 The ectopic pregnancy rate has been reported to be one third higher among black and other minority women than among white women whereas the risk of death from an ectopic pregnancy among minority women is three times as high as that among white women.19 Recent national data report a smoking prevalence of 24% among black women of reproductive age in 1993 (comparable to smoking prevalence of 22% among our controls) compared with 30% among white women.20 Although cigarette smoking might not explain the increased incidence of ectopic pregnancy seen among black women, our study documents the role that a modifiable risk factor such as cigarette smoking might play in ectopic pregnancy. The control group we used consisted of women who were delivered of live or stillborn infants and women who were seeking induced abortion, in a ratio similar to that found in the hospital population. Combining these two groups, as others have suggested, allowed us to make a better estimate of smoking in the entire population of pregnant women who use this hospital than the use of either control group alone would have done.12 The exposure, cigarette smoking, has been defined in different ways for different studies. Several studies have used smoking at the time of the interview2, 7, 10, 11 as a proxy for smoking during the periconception period, resulting in a different period of reference for the cases than the controls. More important, such an exposure measure can result in an inaccurate indication of the
Saraiya et al. 497
Volume 178, Number 3 Am J Obstet Gynecol
smoking status at time of implantation, because many women quit smoking during pregnancy.21 Because we specifically asked whether subjects smoked during the 6 months before the LMP and the month after the LMP, our exposure measure was consistent for cases and controls and included the crucial time frame for implantation of a pregnancy. We were also able to control for many of the established risk factors for ectopic pregnancy found in other studies, as well as for potential but unestablished risk factors such as vaginal douching.5, 22, 23 Pelvic inflammatory disease is a well-known risk factor for ectopic pregnancy, and whether to adjust for a history of pelvic inflammatory disease in the context of cigarette smoking is an important methodologic issue. Some researchers, convinced that pelvic inflammatory disease is in the causal pathway of cigarette smoking and ectopic pregnancy, favor the exclusion of pelvic inflammatory disease as a covariate; others, convinced that pelvic inflammatory disease is an independent risk factor, favor its inclusion. In this analysis the inclusion or exclusion of the pelvic inflammatory disease variable in the final model did not affect the odds ratio of cigarette smoking or any other covariates, and pelvic inflammatory disease was not considered a significant confounder. Moreover, the Hosmer-Lemeshow goodness-of-fit value was improved when pelvic inflammatory disease was not included. Similar reasoning could be applied to the history of infertility serving as an intermediary between cigarette smoking and ectopic pregnancy. However in this case infertility did affect the odds ratio of other covariates with a worsened HosmerLemeshow goodness-of-fit value when the infertility covariate was left out and therefore was considered a significant confounder and kept in the model. The decision to include or exclude pelvic inflammatory disease or infertility in the model might have been different if we had been able to use clinical data that might have captured “silent” pelvic inflammatory disease or true infertility instead of the self-reported clinical pelvic inflammatory disease and self-reported infertility available in this analysis. A limitation of this epidemiologic study, as well as many others, is the reliance on self-report of various exposures that could potentially be affected by recall bias. Previous work has demonstrated that self-reported data underestimate smoking24 and pelvic inflammatory disease prevalences.25 Biochemical validation of self-reported smoking with markers such as serum or urine cotinine and validation of pelvic inflammatory disease history with laparoscopy would have been useful. Because these were past exposures, measurement of these exposures was not feasible with our case-control study. However, our outcome, ectopic pregnancy, was well documented. Another limitation is our inability to analyze never and former smokers separately because of the small number
of former smokers. One study of the risk of ectopic pregnancy among former smokers concluded that the effect of cigarette smoking was the result of short-term mechanisms, whereas another implicated long-term mechanisms.8, 9 Our results suggest smoking as an etiologic factor in ectopic pregnancy, but we were not able to truly discern a more chronic versus a more acute effect of tobacco smoke although the lack of increased risk among the small number of former smokers would tend to argue for the latter. On the basis of laboratory studies in humans and animals, researchers have postulated several mechanisms by which cigarette smoking might play a role in ectopic pregnancies. These mechanisms include one or more of the following: delayed ovulation, altered tubal and uterine motility, or altered immunity.24-30 To date, no study has supported a specific mechanism by which cigarette smoking affects the occurrence of ectopic pregnancy. We know already that the effects of maternal tobacco use during and after pregnancy include increased risk of spontaneous abortion, abruptio placentae, preterm premature rupture of membranes, low birth weight, sudden infant death syndrome, perinatal death, and childhood respiratory disorders.22, 31, 32 With an estimated 26% prevalence of cigarette smoking among women of reproductive age,18 increased awareness and effective counseling techniques can potentially influence a woman’s decision to smoke before she becomes pregnant, especially when the safety of a future pregnancy is at risk. Recently released publications33, 34 that provide specific recommendations for physicians and other health care providers to help patients, especially women, quit smoking may be useful. More important, efforts should be directed at preventing young women from ever beginning to smoke.32 REFERENCES
1. Centers for Disease Control. Ectopic pregnancy—United States, 1990-1992. MMWR Morb Mortal Wkly Rep 1995;4:46-8. 2. Chow WH, Daling JR, Cates W, Greenberg RS. Epidemiology of ectopic pregnancy. Epidemiol Rev 1987;9:70-93. 3. Zahniser SC, Gupta SC, Kendrick JS, Lee NC, Spirtas R. Tubal pregnancy and cigarette smoking: is there an association? J Wom Health 1994;3(5):329-36. 4. Levin AA, Schoenbaum SC, Stubblefield PG, Zimicki S, Monson RR, Ryan KJ. Ectopic pregnancy and prior induced abortion. Am J Public Health 1982;72:253-6. 5. Parazzini F, Tozzi L, Ferraroni M, Bocciolone L, La Vecchia C, Fedele L. Risk factors for ectopic pregnancy: an Italian case-control study. Obstet Gynecol 1992;80:821-6. 6. Coste J, Job-Spira N, Fernandez H, Papiernik E, Spira A. Risk factors for ectopic pregnancy: a case-control study in France, with special focus on infectious factors. Am J Epidemiol 1991;133:839-49. 7. Handler A, Davis F, Ferre C, Yeko T. The relationship of smoking and ectopic pregnancy. Am J Public Health 1989;79:1239-42. 8. Campbell OM, Gray RH. Smoking and ectopic pregnancy: a multinational case-control study. In: Rosenberg MD, editor. Smoking and reproductive health. Littleton (MA): PSG; 1987. p. 70-4. 9. Chow WH, Daling JR, Weiss NS, Voigt LF. Maternal cigarette
498 Saraiya et al.
10.
11.
12.
13.
14.
15.
16. 17.
18.
19.
20.
21.
22.
smoking and tubal pregnancy. Obstet Gynecol 1988;71:16770. Stergachis A, Scholes D, Daling JR, Weiss NS, Chu J. Maternal cigarette smoking and the risk of tubal pregnancy. Am J Epidemiol 1991;133:332-7. Phillips RS, Tuomala RE, Feldblum PJ, Schachter J, Rosenberg MJ, Aronson MD. The effect of cigarette smoking, Chlamydia trachomatis infection, and vaginal douching on ectopic pregnancy. Obstet Gynecol 1992;79:85-90. Kalandidi A, Doulgerakis M, Tzonos A, Hsieh CC, Aravandinos D, Trichopoulos D. Induced abortions, contraceptive practices, and tobacco smoking as risk factors for ectopic pregnancy in Athens, Greece. Br J Obstet Gynaecol 1991;98:207-13. Weiss NS, Daling JR, Chow WH. Control definition in case-control studies of ectopic pregnancy. Am J Public Health 1985;75:67-8. Kendrick JS, Tierney EF, Lawson HW, Strauss LT, Klein L, Atrash HK. Previous cesarean delivery and the risk of ectopic pregnancy. Obstet Gynecol. 1996;87:297-301. Atrash HK, Strauss LT, Kendrick JS, Skjeldestad FE, Ahn YW. The relation between induced abortion and ectopic pregnancy. Obstet Gynecol. 1997;89:512-8. Maldonado G, Greenland S. Simulation study of confounder-selection strategies. Am J Epidemiol 1993;138:923-6. Lemeshow S, Hosmer DW. A review of goodness of fit statistics for use in the development of logistic regression models. Am J Epidemiol 1982;115:92-106. Kleinbaum DG, Kupper LL, Morgenstern H. Epidemiological research: principles and quantitative methods. New York: Van Nostrand Reinhold; 1982. Goldner TE, Lawson HW, Xia S, Atrash HK. Surveillance for ectopic pregnancy—United States, 1970-1989. MMWR Morb Mortal Wkly Rep 1993;42:73SS-85SS. CDC Surveillance Summaries, 1993 Dec. Publication No.: SS-6. Husten CG, Chrismon JH, Reddy MN. Trends and effects of cigarette smoking among girls and women in the United States. 1965-1993. J Am Med Wom Assoc 1996;51:11-8. Floyd RL, Zahniser SC, Gunter EP, Kendrick JS. Smoking during pregnancy: prevalence, effects, and intervention strategies. Birth 1991;18:47-52. Daling JR, Weiss NS, Schwartz SM, Stergachis A, Wang SP, Foy H,
March 1998 Am J Obstet Gynecol
23.
24.
25.
26.
27. 28. 29. 30.
31.
32.
33.
34.
et al. Vaginal douching and the risk of tubal pregnancy. Epidemiology 1991;2:40-8. Kendrick JS, Atrash HK, Strauss LT, Garguillo PM, Ahn YW. Vaginal douching: an important risk factor for ecotpic pregnancy among black women. Am J Obstet Gynecol 1997;176:991-7. DiFranza JR, Lew RA. Effect of maternal cigarette smoking on pregnancy complications and sudden infant death syndrome. J Fam Pract 1995;40:385-94. Westrom L. Incidence, prevalence, and trends of acute pelvic inflammatory disease and its consequences in industrialized countries. Am J Obstet Gynecol 1980;138:880-92. Campbell O. Ectopic pregnancy and smoking: confounding or causality? In: Poswillo D, Alberman E, editors. Effects of smoking on the fetus, neonate, and child. New York: Oxford University Press; 1992. p. 23-44. Mattison DR. The effects of smoking on fertility from gametogenesis to implantation. Environ Res 1982;28:410-33. Macmahon B, Trichopoulos D, Cole P, Brown J. Cigarette smoking and urinary estrogens. N Engl J Med 1982;307:1063-5. Yoshinaga K, Rice C, Krenn J, Pilot RL. Effects of nicotine on early pregnancy in the rat. Biol Reprod 1979;20:294-303. Neri A, Eckerling B. Influence of smoking and adrenaline (epinephrine) on the uterotubal insufflation test (Rubin test). Fertil Steril 1969;20:818-28. Evans J, Golding J. Parental smoking and respiratory problems in childhood. In: Poswillo D, Alberman E, editors. Effects of smoking on the fetus, neonate, and child. New York: Oxford University Press; 1992. p. 121-37. Nicholl J, O’Cathain A. Antenatal smoking, postnatal passive smoking, and the sudden infant death syndrome. In: Poswillo D, Alberman E, editors. Effects of smoking on the fetus, neonate, and child. New York: Oxford University Press; 1992. p. 138-49. Fiore MC, Bailey WC, Cohen SJ, Dorfman SF, Goldstein MC, Gritz ER, et al. Smoking cessation: a guide for primary care clinicians. Clinical practice guideline. Quick reference guide for primary care clinicians. No.: 18. Rockville (MD): Dept. of Health and Human Services (US), Public Health Service, Agency for Health Care Policy and Research and Centers for Disease Control and Prevention; 1996 Apr. AHCPR Publication No.: 96-0693. Kendrick JS, Merrit RA. Women and smoking: an update for the 1990s. Am J Obstet Gynecol 1996;175:528-35.