| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH |
From the Reproductive Epidemiology Section, California Department of Health Services, Emeryville, California.
Address reprint requests to: Gayle C. Windham, PhD California Department of Health Services Environmental Health Investigations Branch Reproductive Epidemiology Section 1515 Clay Street, 17th Floor Oakland, CA 94612 E-mail: gwindham{at}hw1.cahwnet.gov
 |
Abstract |
|---|
Top
 Abstract Materials and MethodsResultsDiscussionReferences |
|---|
Methods: In a prospective study, 408 women collected urine daily for one to seven menstrual segments (cycles), maintained daily diaries, and completed detailed interviews. Smoking data from the diaries were averaged over each segment and verified by cotinine assay. Urine samples were analyzed for metabolites of steroid hormones to define the day of ovulation and various menstrual characteristics, including: 1) segment, follicular, luteal phase, and menses length, 2) variabililty, and 3) anovulation.
Results: Heavy smoking (at least 20 cigarettes per day) was associated with nearly four times the risk of short segment (less than 25 days) as was nonsmoking (adjusted odds ratio 3.8, 95% confidence limits 1.1, 12.7). Mean segment length was on average 2.6 days shorter with heavy versus no smoking (95% confidence limits 0.14, 5.0), due almost entirely to shortening of the follicular phase. Women who smoked an average of ten or more cigarettes per day had significantly more variable segment and menses lengths than nonsmokers. Based on small numbers, the data suggested that with greater smoking, there was a possible increased risk of anovulation and short luteal phase. Segments of exsmokers with ten or more pack-years of exposure were more likely to be short and have shorter luteal phases than those of never smokers.
Conclusion: The effects found in this study of smoking on the menstrual cycle might explain in part associations of smoking with other reproductive endpoints, such as subfecundity and early menopause.
Cigarette smoke is a known developmental toxicant.1–3 Smoking has been associated with adverse reproductive outcomes such as infertility, subfecundity, and younger age at menopause.4–8 A few studies have suggested that cigarette smoking increases menstrual disorders, with an intermediate effect among exsmokers.9–12 These studies found that smokers were more likely than nonsmokers to have oligomenorrhea and other aspects of abnormal menstruation, including irregular periods and intermenstrual bleeding. The purpose of this study was to examine the effects of smoking on menstrual cycle function by using biologic measures rather than self-reporting.
|
Materials and Methods |
|---|
|
Top
AbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Almost 6500 members of the Kaiser Permanente Medical Care Program in northern California were screened by short telephone interviews to identify women at risk of pregnancy (to study fetal loss), and willing to collect and freeze first-morning urine samples daily for up to 6 months. The eligibility criteria included the following: English-speaking, aged 18–39 years, married, neither partner sterilized, not using oral contraceptives or intrauterine devices, local residence, current Kaiser membership, not currently pregnant, less than 3 months of unprotected intercourse, and a menstrual period within the lst 6 weeks. Participants were enlisted between May 1990 and June 1991. Of 1092 eligible women, 553 agreed to participate, but 89 dropped out during urine collection and 61 became ineligible. Eight women with tubal ligations, who served as controls for pregnancy-related analyses, are included (excluding them did not affect the results). Before urine collection, participants completed a detailed baseline interview by telephone, which asked about past and current cigarette smoking and numerous potential confounders. Women filled out daily diaries during the urine collection phase, recording vaginal bleeding (number of pads or tampons/day), intercourse, and number of cigarettes smoked each day.
Definition of Menstrual Characteristics
Each urine sample was analyzed for metabolites of estrogen (estrone conjugates) and progesterone (pregnanediol-3-glucuronide) by enzyme-linked immunoassay14 and adjusted for creatinine. Menstrual segments were identified by bleeding patterns (not all segments represent ovulatory cycles). Determination of ovulatory status was based on a sufficient relative rise in progesterone over baseline levels.13 The day of ovulation was estimated using an algorithm, previously validated against luteinizing hormone, which generally selects the day after the peak of the estrogen-to-progesterone ratio.13,15 Steroid assays were repeated in segments with abnormalities (eg nonovulatory, short luteal phase) to confirm results. In cases of discordant results, the more normal assays were analyzed to be conservative. Steroid levels were also examined graphically, and in a small portion of segments (5.6%), the day of ovulation was recoded to correspond better to the steroid patterns.
Segment length was calculated from the first day of menses to the day before the next menses. Follicular phase length included the day of ovulation. We examined mean segment and phase lengths, and categorized them on the basis of the fifth and 95th percentiles of their distributions. Short and long segments were defined as less than 25 days and greater than 35 days, respectively, and compared with normal-length segments (25–35 days). Long follicular and short luteal phases were expected to be most clinically relevant and were defined as more than 23 days and less than 11 days, respectively. Menses length was derived from the number of days of reported pad or tampon use for the main (first) bleeding episode, with greater than 7 days defined as long. All segments were not eligible for calculating each endpoint, because urine collection dates were not timed to menses dates.13 For example, last segments for which urine collection was not complete (including pregnancies) were excluded from analyses of segment and luteal phase length, but might have been included in analyses of menses or follicular phase length. Therefore, 1658 segments were available for segment length analyses, 1541 for follicular phase, 1450 for luteal phase, and 1749 for menses analyses.
For measures of menstrual variability, we calculated the within-woman standard deviation and the range (maximum minus minimum) of segment, follicular phase, luteal phase, and menses lengths for each woman with at least two eligible segments. Women with ranges greater than 7 days for segment and follicular phase length and greater than 3 days for luteal phase and menses length were considered irregular (or highly variable). We also analyzed anovulation by woman, rather than segment, because a single anovulatory episode could be spread across several segments, or could occur within one long segment, depending on whether the woman experienced any breakthrough bleeding. Anovulation was defined as more than 36 consecutive days elapsed without a rise in progesterone above the anovulation threshold.13 Seven women with nonovulatory segments defined by the original ovulation algorithm, but missing information to evaluate completely the full 36 days, were considered possibly anovulatory. Thirty-one additional women could not have anovulatory status determined because of insufficient data.
Independent Variables and Statistical Analysis
The average number of cigarettes smoked per day during each segment was calculated from the daily diary and classified as none, less than 20, or 20 or more. The average smoked per day for each woman during her entire urine collection phase also was calculated. Because of smaller numbers at the woman level, we classified smoking as none, less than 10, or 10 or more cigarettes per day in these analyses. Lifetime pack-years were calculated for current and exsmokers from usual amount and duration of smoking, recorded on the baseline interview, and represent the equivalent of years that one pack per day was smoked (without accounting for quit attempts).
To validate self-reported smoking, pooled urine samples (first 5 days of segment) from one to three segments per woman were assayed for nicotine and its metabolite, cotinine.16 All of the nonsmokers, but none of the women smoking at least one cigarette per day, had average urinary cotinine levels less than 25 ng/mL. Two potential misreporters (eg, who reported only occasional smoking but had cotinine levels greater than 200 ng/mL) were excluded from further analyses, as was one nonsmoking woman who reported using nicotine gum, leaving 408 women.
The appropriate unit of analysis for the length parameters is the menstrual segment.17,18 Because of the expected correlation within women, mixed models that account for repeated measures were used.19,20 The compound symmetry covariance structure, which assumes that all repeated units (eg, segments) within a woman are correlated equally, fits the data best and was used in all models. For woman-level analyses of anovulation and highly variable ranges of length, we used logistic regression to model the effects of smoking, and linear models for analysis of the within-woman standard deviation.
The data were analyzed first with smoking alone in a model to calculate univariate estimates of effect, controlling only for repeated measures. Numerous covariates from the baseline interview were examined initially as potential confounders, including demographics (ie, age, race-ethnicity, education, and employment status); reproductive history (ie, number of pregnancies, pregnancy losses, elective abortions, and contraceptive practices); lifestyle factors (ie, caffeine and alcohol intake, exercise, body mass index [BMI], and stress); and employment. Variables with evidence of confounding or that were associated strongly with endpoints or smoking (eg, age, race, history of pregnancy loss and elective abortion, BMI [kg/m2], exercise, caffeine and alcohol intake, and tubal ligation status) were included in all adjusted models for ease of presentation, which may decrease precision. The analyses were rerun only in ovulatory segments to preclude any differences due to the potential association of smoking and anovulation. The findings were similar or stronger in these restricted analyses.
|
Results |
|---|
|
Top
AbstractMaterials and MethodsResultsDiscussionReferences |
|---|
, participants tended to be white, parous, and well-educated. The mean age at entry was 31 years (standard deviation [SD] 4.2). On average, each woman contributed 5.6 segments to the study, of which 3.6 were complete segments. On the baseline questionnaire, 9.2% of the women reported being current regular smokers, and 20.7% reported being exsmokers. The women’s daily diaries indicated that 10% smoked an average of one or more cigarettes per day, and another 5% smoked less frequently. Compared with nonsmokers, smokers were significantly less educated, drank more alcoholic and caffeinated beverages, and were more likely to have had prior pregnancies, prior pregnancy losses, or therapeutic abortions (Table 1). Five of the eight women with tubal ligations were smokers.
|
). Heavily exposed segments were not more likely to be long or have long follicular phases than nonexposed segments. Short luteal phase was more than twice as likely at the heavy smoking level, but the confidence limits were wide (Table 2). Similarly, the average segment length of 29.3 days with no smoking decreased by 2.6 days (95% confidence limits 0.14, 5.0) with heavy smoking, after adjustment. Limiting the analysis to the 1439 segments that met the eligibility criteria for all three length parameters, the adjusted difference in segment length by heavy smoking was -2.1 days (CLs -4.8, 0.64), due entirely to a shortening of the mean follicular phase length. The mean menses length of 5.3 days did not vary by smoking level.
|
). Effects were similar for follicular and luteal phase length, but of a much smaller magnitude. The risk of experiencing anovulation was nearly doubled in moderate smokers, but the numbers were too small to assess this endpoint adequately (Table 3). Adding the seven women with possible anovulatory episodes suggested greater associations (adjusted OR 2.8), but the confidence limits remained wide (0.48, 16.5).
|
). Furthermore, current smokers with at least 10 pack-years had segments an average of 3 days shorter than those of never smokers, whereas heavily exposed exsmokers had segments about 1 day shorter. Stratifying exsmokers by time since quitting (less than 5 years versus more) showed that the effects of greater exposure were stronger among more recent quitters for both mean segment length and short segment. In segments of exsmokers with 10 or more pack-years of exposure, the likelihood of having short luteal phases was significantly more than that of never smokers (adjusted OR 5.4; confidence limits 1.2, 24.4), irrespective of time since quitting.
|
). Exsmokers did not appear to have the increased variability in menses length seen among heavily exposed current smokers. |
Discussion |
|---|
|
Top
AbstractMaterials and MethodsResultsDiscussionReferences |
|---|
A few other studies examined menstrual function in relation to smoking, related primarily to self-reported symptoms. A community survey in Los Angeles found that the prevalence of menstrual disorders for which a physician was seen (varied diagnoses including dysmenorrhea and oligomenorrhea) was higher among heavier smokers (at least 15 cigarettes per day).10 A postal survey in England found six of seven aspects of abnormal menstruation to be reported more frequently by smokers, including frequent and irregular periods,9 consistent with our findings. That study also found prolonged and heavy menses among smokers, which is not consistent with our data. In another study, the 1-year prevalence of secondary amenorrhea was reported to be somewhat higher in smokers and exsmokers than in nonsmokers.12 A recent study assessed endpoints more comparable to ours, but was based only on a menstrual diary.11 The strongest findings were increased dysmenorrhea and amount of bleeding (self-reported four-point scale) as well as shorter duration of bleeding among smokers. We did not find a decrease in mean menses length among smokers, neither have other studies.9,10 In this recent study,11 the heaviest smokers (more than ten per day) had a 22% greater mean SD in cycle length than nonsmokers (not significant), in the direction of our results. That study had limited power to examine higher smoking levels, and the sample was rather selective, because it consisted of women 37–39 years old whose mothers had participated in a randomized clinical trial of diethylstilbestrol during pregnancy. A small study21 in which hormone levels were measured during a single cycle found that heavy smokers had cycles averaging 1.6 days shorter than nonsmokers, with a mean follicular phase 1.4 days shorter (not adjusted), similar to our results.
Our study has several strengths compared with these previous studies of menstrual function and smoking. The primary exposure data were prospective recordings of daily smoking habits, and there was little evidence of misreporting, based on cotinine levels. The associations noted were consistent with two sources of smoking data (usual amount at baseline and daily diary). The menstrual parameters were based on biologic measures, rather than self-reporting only, over multiple cycles per participant. This allowed us to examine endpoints not studied previously in relation to smoking, including follicular and luteal phase length and variability, and anovulation. Abnormalities were confirmed, and our results for mean cycle and phase lengths and rates of anovulation are consistent with those found in the literature.22,23 Many potential confounders were considered in these analyses.
There also are some limitations of the study. Because of the labor-intensive nature of data collection, less than 40% of women initially eligible completed the study. Thus, the sample may not be representative, as evidenced by the high education level. The study also did not have adequate power to assess some of the endpoints, particularly for heavier smoking and woman-level characteristics. The most extreme abnormalities were likely excluded because the eligibility criteria were designed to increase the likelihood of pregnancy, and because the more normal or typical hormone patterns were included when duplicate assays were discordant.
The exact mechanisms by which smoking might exert effects on menstrual function remain to be clarified. Tobacco smoke contains thousands of compounds, some of which might be ovarian toxicants.4,24 Other evidence suggests that smoking affects steroid hormone levels and metabolism.25–27 The ramifications of menstrual disturbances on other health endpoints, such as fertility, were not examined in this report, but might be numerous. Short cycles might indicate abnormal folliculogenesis, and a short luteal phase might indicate an inadequate progesterone response. Women with variable cycle lengths might have difficulty trying to conceive, because timing of ovulation will be less predictable. Anovulation has obvious relevance to time to conception. These possible effects are consistent with evidence that smoking is associated with decreased fertility.5,28 On the other hand, shorter cycles might lead to more rapid depletion of oocytes, shortening the reproductive lifespan.29,30 Smokers have been found to reach menopause an average of two years earlier than nonsmokers,7,8 and early menopause appears to be associated with other health problems.31 Women with short cycles also might be at higher risk for breast cancer.32
Our report suggests that smoking affects menstrual function, using the most objective data available to date. Because of limited previous work, these findings should be confirmed in other prospective studies. If confirmed, they would show effects of smoking on endpoints of high prevalence, which have other ramifications for reproductive health, and provide more reasons for women to reduce smoking.
|
Footnotes |
|---|
Received March 25, 1998. Received in revised form June 29, 1998. Accepted July 9, 1998.
|
References |
|---|
|
Top
AbstractMaterials and MethodsResultsDiscussionReferences |
|---|
2. Hogue CJR, Sappenfield W. Smoking and low birth weight: Current concepts. In: Rosenberg MJ, ed. Smoking and reproductive health. Littleton, MA: PSG Publishing Co., 1987:97–103.
3. Windham GC. Effects of tobacco smoke on risk of spontaneous abortion. Current issues in smoking and reproductive health. Clinical proceedings. Washington, DC: Association of Reproductive Healthy Professionals, 1996.
4. Stillman RJ, Rosenberg MJ, Sachs BP. Smoking and reproduction. Fertil Steril 1986;46:545–66.[Medline]
5. Baird DD. Evidence for reduced fecundity in female smokers. In: Poswillo D, Alberman E, eds. Effects of smoking on the fetus, neonate and child. Oxford, United Kingdom: Oxford University Press, 1992.
6. Thomford PJ, Mattison DR. The effect of cigarette smoking on female reproduction. J Ark Med Soc 1986;82:597–604.[Medline]
7. Kaufman DW, Slone D, Rosenberg L, Miettinen OS, Shapiro S. Cigarette smoking and age at natural menopause. Am J Public Health 1980;70:420–2.
8. Midgette AS, Baron JA. Cigarette smoking and the risk of natural menopause. Epidemiology 1990;1:474–9.[Medline]
9. Brown S, Vessey AM, Stratton I (Imperial Cancer Research Fund Cancer Epidemiology and Clinical Trials Unit). The influence of method of contraception and cigarette smoking on menstrual patterns. Br J Obstet Gynaecol 1988;95:905–10.[Medline]
10. Sloss EM, Frerichs RR. Smoking and menstrual disorders. Int J Epidemiol 1983;12:107–9.
11. Hornsby PP, Wilcox AJ, Weinberg CR. Cigarette smoking and disturbance of menstrual function. Epidemiology 1998;9:193–8.[Medline]
12. Pettersson F, Fries H, Nillius SJ. Epidemiology of secondary amenorrhea. Am J Obstet Gynecol 1973;117:80–6.[Medline]
13. Waller K, Swan SH, Windham GC, Fenster L, Elkin EP, Lasley BL. Use of urine biomarkers to evaluate menstrual function in healthy, premenopausal women. Am J Epidemiol 1998;11:1071–80.
14. Munro CJ, Stabenfeldt GH, Cragun JR, Addiego LA, Overstreet JW, Lasley BL. Relationship of serum estradiol and progesterone concentrations to the excretion profiles of their major urinary metabolites as measured by enzyme immunoassay and radioimmunoassay. Clin Chem 1991;37:838–44.
15. Baird DD, Weinberg CR, Wilcox AJ, McConnaughey DR, Musey PI. Using the ratio of urinary estrogen and progesterone metabolites to estimate day of ovulation. Stat Med 1990;10:255–66.
16. Jacob P III, Yu L, Wilson M, Benowitz NL. Selected ion monitoring method for determination of nicotine, cotinine, and deuteriumlabelled analogs. Absence of an isotope effect in the clearance of (S)-nicotine-3',3'-d2 in humans. Biol Mass Spectrom 1991;20:247–52.[Medline]
17. Martinez-Schenell B, Wilcox LS, Peterson HB, Jamison PM, Hughes JM. Evaluating the effects of tubal sterilization on menstrual function: Selected issues in data analysis. Stat Med 1993;12: 355–63.[Medline]
18. Harlow SD, Zeger SL. An application of longitudinal methods to the analysis of menstrual diary data. J Clin Epidemiol 1991;44: 1015–25.[Medline]
19. Laird N, Ware J. Random effects models for longitudinal data. Biometrics 1982;38:963–74.[Medline]
20. Zeger S, Liang KY. Longitudinal data analysis for discrete and continuous outcomes. Biometrics 1986;42:121–30.[Medline]
21. Zumoff B, Miller L, Levit CD, Miller EH, Heinz U, Kalin M, et al. The effect of smoking on serum progesterone, estradiol, and luteinizing hormone levels over a menstrual cycle in normal women. Steroids 1990;55:507–11.[Medline]
22. Harlow SD, Ephross SA. Epidemiology of menstruation and its relevance to women’s health. Epidemiol Rev 1995;17:265–86.
23. Baird DD, McConnaughey DR, Weinberg CR, Musey PI, Collins DC, Kesner JS, et al. Application of a method for estimating day of ovulation using urinary estrogen and progesterone metabolites. Epidemiology 1995;6:547–50.[Medline]
24. Mattison DR, Plowchalk DR, Meadows MJ, Miller MM, Malek A, London S. The effects of smoking on oogenesis, fertilization and implantation. Semin Reprod Endocrinol 1989;7:291–304.
25. Baron JA, La Vecchia C, Levi F. The antiestrogenic effect of cigarette smoking in women. Am J Obstet Gynecol 1990;162:502–14.[Medline]
26. Michnovicz JJ, Herschcopf RJ, Naganuma H, Bradlow HL, Fishman J. Increased 2-hydroxylation of estradiol as a possible mechanism for the anti-estrogenic effect of cigarette smoking. N Engl J Med 1986;315:1305–9.[Abstract]
27. Canick JA, Barbieri RL. The effect of smoking on hormone levels in vivo and steroid hormone biosynthesis in vitro. In: Wald N, Baron J, eds. Smoking and hormone-related disorders. Oxford, United Kingdom: Oxford University Press, 1990:209–16.
28. Bolumar F, Olsen J, Boldsen J, and the European Study Group on Infertility and Subfecundity. Smoking reduces fecundity: A European multicenter study on infertility and subfecundity. Am J Epidemiol 1996;143:578–87.
29. Bromberger JT, Matthews KA, Kuller LH, Wing RR, Meilahn EN, Plantinga P. Prospective study of the determinants of age at menopause. Am J Epidemiol 1997;145:124–33.
30. Whelan EA, Sandler DP, McConnaughey DR, Weinberg CR. Menstrual and reproductive characteristics and age at natural menopause. Am J Epidemiol 1990;131:625–32.
31. Sowers MFR, La Pietra MT. Menopause: Its epidemiology and potential association with chronic diseases. Epidemiol Rev 1995; 17:287–302.
32. Kelsey JL, Gammon MD, John EM. Reproductive factors and breast cancer. Epidemiol Rev 1993;15:36–47.
This article has been cited by other articles:
|
|
 |
|
  J. B. Stanford and D. B. Dunson Effects of Sexual Intercourse Patterns in Time to Pregnancy Studies Am. J. Epidemiol., May 1, 2007; 165(9): 1088 - 1095. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. L. Holt, K. L. Cushing-Haugen, and J. R. Daling Risk of Functional Ovarian Cyst: Effects of Smoking and Marijuana Use according to Body Mass Index Am. J. Epidemiol., March 15, 2005; 161(6): 520 - 525. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Liu, E. B. Gold, B. L. Lasley, and W. O. Johnson Factors Affecting Menstrual Cycle Characteristics Am. J. Epidemiol., July 15, 2004; 160(2): 131 - 140. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. C. Windham, C. Bottomley, C. Birner, and L. Fenster Age at Menarche in Relation to Maternal Use of Tobacco, Alcohol, Coffee, and Tea during Pregnancy Am. J. Epidemiol., May 1, 2004; 159(9): 862 - 871. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Santoro, S. L. Crawford, J. E. Allsworth, E. B. Gold, G. A. Greendale, S. Korenman, B. L. Lasley, D. McConnell, P. McGaffigan, R. Midgely, et al. Assessing menstrual cycles with urinary hormone assays Am J Physiol Endocrinol Metab, March 1, 2003; 284(3): E521 - E530. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Smoking and Menstrual Abnormalities Journal Watch Women's Health, February 1, 1999; 1999(201): 4 - 4. [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
