Secondary Logo

Journal Logo

Original Research

Increased Infertility With Age in Men and Women

Dunson, David B. PhD*; Baird, Donna D. PhD; Colombo, Bernardo PhD

Author Information
doi: 10.1097/01.AOG.0000100153.24061.45
  • Free

Waiting until older ages to have children is much more common in the United States today than it was 20 years ago.1,2 This has focused attention on the relationship between age and infertility.3 Although descriptive studies show declines in fertility with age, this can be due, at least in part, to less frequent sexual intercourse in older couples.4–6 The classic study that demonstrated a decline in fertility with age for females that could not be attributed to reduced intercourse frequency used data from women undergoing artificial insemination.7 Inseminations are timed to coincide with ovulation; thus, frequency and timing of intercourse did not affect the estimates of age effects. Comparable data for men have not been available until recently, although descriptive studies showed a decline with age in male fertility starting in the 30s.8 The European Fecundability Study examined age effects for men as well as women, while controlling for the timing of intercourse in relation to ovulation, by collecting daily intercourse information from women who were trained to use natural family planning to identify their fertile days.9 In this study of 7,288 menstrual cycles from women enrolled at 7 European centers, fertility began to decline by the late 20s for women and by the late 30s for men.10

This study addresses the important question of whether declines with age in male and female fertility throughout the 20s and 30s are primarily attributable to an increasing proportion of sterile couples in older age groups or to most couples becoming gradually less fertile, where we define sterility as the inability to conceive a pregnancy naturally in the absence of clinical interventions. The relationship between age and the risk of sterility is unknown. Such information is important for the counseling of infertile couples and for clinical management decisions. Other important information that we present includes the difference between age groups in the length of time required to conceive, the expected rates of infertility for couples of different ages, and the proportion of infertile couples conceiving naturally during a second year of trying. This study uses data from the European Fecundability Study, including information from 1,428 menstrual cycles that were not incorporated in previous analyses due to missing data, to address these questions.


Data are drawn from a prospective cohort study of daily fecundability. The research protocol was reviewed and approved by the Institutional Review Boards of Fondazione Lanza (Padua, Italy) and Georgetown University. Details of the study design and demographics of the study participants have been published previously.9,10 Briefly, from 1992 through 1996, 782 women were recruited through 7 European centers (Milan, Verona, Lugano, Dusseldorf, Paris, London, and Brussels) providing services on fertility awareness and natural family planning. Recruitment procedures were subject to the discretion of the local principal investigator and typically involved finding fertility awareness teachers willing to cooperate and having them provide information to their clients about the study. No information is available on the percentages of teachers or clients agreeing to participate among those contacted. Most couples were trying to avoid pregnancy initially (although intentions often changed during follow-up), and women were between 18 and 40 years of age at admission, had at least one menses after cessation of breast feeding or delivery, and were not currently taking hormonal medication or fertility-altering drugs. Couples with known fertility problems were excluded. Study participants kept daily records of basal body temperature, cervical mucus characteristics, and occurrences of intercourse and menstrual bleeding. The occurrence of clinical pregnancy (yes or no) was recorded for each menstrual cycle. Infertility testing data were not collected.

By adjusting for the timing and frequency of intercourse in the statistical analysis, we are able to use data from menstrual cycles in which the couples were attempting to avoid pregnancy by using fertility awareness methods without biasing downward estimates of fecundability. Menstrual cycles with no reported intercourse in the fertile interval defined relative to the identified day of ovulation are noninformative and do not contribute to the results. No data were collected on the couple’s desire to conceive.

From the coital records, the timing and frequency of intercourse varied substantially across the different cycles in the study, providing ample information for estimating day-specific conception probabilities for different age groups.

Previous work10 focused on the 5,860 cycles from 770 women for which a basal body temperature–based estimate of ovulation day was available. This earlier study estimated day-specific probabilities of conception within the fertile interval for men and women in different age categories. In order to address the goals of the current study, we include data from an additional 1,428 menstrual cycles from the original 770 women to rule out sterility for couples who conceived in one of these cycles. Information on clinical pregnancy is available for these cycles, but daily records were not collected. The sample size is large enough to obtain precise estimates of the proportions sterile in different age groups and to have high power to detect clinically important increases with age.

For couples in which the female partner has regular ovulatory cycles and an intact reproductive tract and the male partner has not undergone vasectomy, there is no way to reliably distinguish sterility, defined here as the inability to conceive naturally, from subfertility. Because direct data on the proportions of sterile couples in different age groups can never be obtained, it is necessary to use a statistical model to study sterility rates.11,12 A commonly used approach for addressing this problem is to use a “mixture” model, which allows the cumulative pregnancy curves to represent a weighted average of those who would eventually conceive if attempting for long enough and those who would remain sterile.13,14 Such an approach seems more realistic than widely used fertility models that effectively assume no sterility.15 Closely related mixture models, referred to as cure rate models, have been used in the cancer literature for estimating the proportion of patients cured of cancer, and for distinguishing between factors related to longer times to remission and those related to cure (see, for example, Gordon16 and Chen et al17).

Although fecundability and sterility are not directly observable for any couple under study, we estimated the proportions of sterile couples in different subgroups along with the distribution of fecundability for nonsterile couples by fitting a mixture model. The statistical methods have been presented in detail previously.18,19 From the fecundability distribution, we calculate the distribution of time to pregnancy for each age group. Inferences on differences between groups are based on posterior probabilities.

The mixture model derives from traditional approaches to fecundability data. If all couples had the same chance of conceiving in each menstrual cycle, denoted by the probability p, then the number of menstrual cycles to conception, denoted by T, would follow a geometric distribution,20 implying that a couple conceives in cycle t with probability Pr(T = t) = p(1− p)t − 1. Variability in the per-menstrual-cycle conception probability, p, leads to increased variability in the time to pregnancy distribution relative to the geometric. In general, the probability of conceiving in cycle t can be expressed as Pr(T = t) = ∫ pi(1 − pi)t − 1 f(pi) dpi, where pi is the per-menstrual-cycle probability of conception for a randomly selected couple and f(pi) is the distribution of pi. This well known observation has long been used to account for and study variability in fecundability.21 Weinberg and Gladen22 suggested incorporating a “point mass” at pi = 0 to account for a sterile subgroup of couples and then using a beta distribution23 to characterize variability in fecundability for the remaining nonsterile couples.

We follow the approach used by Wilcox et al,24 Dunson et al,25,10 and Stanford et al26 (among others) and express the probability of conception in a menstrual cycle as the maximal probability of conception given optimal timing multiplied by 1 minus a product of day-specific probabilities for separate intercourse days around the time of ovulation.

The time to pregnancy for a particular couple depends on the per-menstrual-cycle probability of conception for that couple through a simple geometric distribution, and the mixture model characterizes variability in fecundability for each age group. Therefore, to obtain an estimate of the time to pregnancy distribution for each age group, we generated the distribution of time to pregnancy for each percentile of the distribution of fecundability for a given age category and integrated across the percentile distributions. We stress that these estimates represent extrapolations from our statistical model using fertility parameters estimated from the European data. Our model is biologically motivated and flexible, and there is no evidence in the data that the assumptions are violated. Infertility was defined as 12 cycles without conception. The impact of frequency of intercourse on the time to pregnancy distribution was investigated by assuming a frequency of twice per week and then repeating the analyses with frequencies of 1 or 3 days per week.


Women aged 19–26 years had significantly higher probabilities of pregnancy than women aged 27–29 years (P = .01). Women aged 30–34 years were similar to the 27 to 29-year-olds, but women aged 35–40 years had further reductions in their probabilities of pregnancy. The decreases in fecundability were not attributable to an increase in sterility. The overall proportion of sterile couples was estimated to be approximately 1%, with no evidence of an increase with age.

The cumulative probabilities of conception for women in different age categories having regular noncontracepting intercourse at a frequency of 2 days per week are plotted in Figure 1A. The pregnancy rates decrease steadily with increasing age of the woman, causing an increase with age in the average time to pregnancy. The proportion of women failing to conceive within 12 cycles (thus meeting the criterion for clinical infertility) ranges from 8% for 19- to 26-year-olds to 13–14% for 27- to 34-year-olds, to 18% for 35- to 39-year-olds. If frequency of intercourse is reduced to once per week, the rates of infertility increase substantially to 15%, 22–24%, and 29% for women aged 19–26, 27–34, and 35–39 years, respectively (Figure 1B). If frequency of intercourse is increased to 3 times per week, pregnancy rates are nearly the same as those derived assuming intercourse twice weekly.

Figure 1.
Figure 1.:
A) Cumulative probability of conceiving a clinical pregnancy by the number of menstrual cycles attempting to conceive for women in different age categories (assuming intercourse occurs at a frequency of 2 times week). B) Cumulative probability of conceiving a clinical pregnancy by the number of menstrual cycles attempting to conceive for women in different age categories (assuming intercourse occurs at a frequency of 1 time week).Dunson. Age and Infertility. Obstet Gynecol 2004.

As shown in Figure 2, age of the man also has a large effect on time to pregnancy and the proportion of couples classified as clinically infertile. For men younger than 35 years, there is no effect, but starting in the late 30s, the impact of male age becomes pronounced. In particular, among 35-year-old women, the proportion of couples failing to conceive within 12 cycles increases from 18% if the male partner is 35 years old to 28% if the male partner is 40. After 24 cycles, the proportions still not pregnant are 9% and 16%, respectively. The effect on fertility of a man aging from 35 to 40 is about the same as the effect seen when intercourse frequency drops from twice per week to once per week.

Figure 2.
Figure 2.:
Effect of father’s age and intercourse frequency. Cumulative probability of conceiving a clinical pregnancy by the number of menstrual cycles attempting to conceive for a woman aged 35 years with a partner the same age and for a woman aged 35 years with a partner 5 years older.Dunson. Age and Infertility. Obstet Gynecol 2004.

Many infertile couples will conceive naturally in the second year of trying (Figures 1 and 2). Table 1 provides estimates of the percentage of infertile couples of different ages who will conceive naturally during their second year of trying. This percentage ranges from 43% to 63% depending on age.

Table 1
Table 1:
Percentage of Infertile Couples of Different Ages Who Will Conceive Naturally in the Second Year of Trying


Fecundability declines with both male10 and female27,7,28, 10 age. In this study, we investigated whether this decline arises from an increase in the number of sterile couples or from a gradual decline for the majority of couples. Our results indicate that sterility did not contribute to the observed decline with age in fecundability. Rather, it appears that the majority of couples in their 30s experience a gradual decline. Our study enrolled women only up to age 40 years, so the possibility of increases in sterility for couples in their 40s requires further investigation. Couples with known fertility problems were excluded from our study, so our results do not apply to couples diagnosed with fertility conditions such as low sperm counts or endometriosis (although some of the couples in the study may have had these conditions without being diagnosed). Our focus was instead on the majority of couples that have no clinical information about their fertility.

We estimated that the probability of sterility for these outwardly healthy couples in their 20s and 30s was approximately 1%, which is consistent with the approximately 2% of women who are involuntarily childless.29 Although infertility increased markedly in the late 30s, sterility did not increase significantly, indicating that most of the age-related drop in fertility is due to a gradual decline. Although sterility was rare in our sample, the estimates were stable because of the large sample and significant subsample of women with long follow-up times.

This result has important implications for couples with unexplained infertility and for clinical management decisions. Couples who have difficulty achieving pregnancy and are given a diagnostic evaluation often appear normal on all of the clinical tests. Although such couples may be classified as clinically infertile based on not conceiving after a year or more of unprotected intercourse, it is relatively unlikely that these couples are truly sterile and will be unable to conceive a pregnancy naturally if attempting for a longer interval. This observation is in agreement with earlier work noting the low specificity of clinical guidelines for infertility,3 the high rate of eventually conceiving among women failing to have a live birth within 2 years of marriage,30 the sizable proportion of nonsterile couples with a long time to pregnancy,31 and the high rate of pregnancy among women who were diagnosed with fertility problems but were not treated.32

A primary goal of our study was to estimate age-specific infertility rates, ie, the proportion of couples requiring more than a year to conceive. The graphs presented provide estimates of expected infertility rates for women and men of different age groups. The data for these estimates were collected from couples in 6 different countries, and about half of the couples had never been pregnant before, so they had not tested their fertility. Therefore, we expect our results to be generalizable to other populations of couples in developed countries. The impact of male aging on infertility rates has not been appreciated, but our results suggest that by the late 30s the male effects are substantial.

Our investigation of changes in frequency of intercourse showed that increasing frequency from 2 to 3 times per week had relatively little effect on the number of menstrual cycles required to conceive. However, time to pregnancy increased substantially for couples having intercourse only once per week. This probably occurs because the fertile interval each menstrual cycle is 5–6 days,25 so couples having sexual intercourse only once a week can miss it completely.

The definition of infertility that is commonly used clinically, more than 12 months required to conceive, provides a convenient cutoff for when a thorough medical workup for fertility-related problems should be done. When nothing is found that can be treated directly, couples can attempt conception with appropriate assisted reproductive technologies such as ovulation induction, in vitro fertilization, and intracytoplasmic sperm injection. However, these technologies are associated with increased risks for the mother, the fetus, and the developing child,33–36 including accumulating evidence for increased risk of long-term cognitive deficiencies in the children.37 For some fertility problems, couples are nearly certain to remain sterile without such treatments, but for many the chance of natural conception is unpredictable, and couples can choose to continue to try without assisted reproductive technologies. The data from our study provide estimates of expected conception rates beyond the first year of trying, so that couples and their physicians can evaluate their prospects of pregnancy in a more informed way.

Additional studies are needed to verify the generalizability of these results to other populations and to assess the impact of aging during the 40s. Our study cohort consisted of users of natural family planning methods, excluding individuals with known infertility, and these individuals may differ in subtle ways from the general population. Changes with age during the early and late 40s, when sterility should become more of a factor, are of substantial interest.


1. Ventura SJ, Martin JA, Curtin SC, Mathews TJ, Park MM. Births: final data for 1998. Natl Vital Stat Rep 2000;48(3):1–100.
2. Ventura SJ, Mosher WD, Curtin SC, Abma JC, Henshaw S. Trends in pregnancy rates for the United States, 1976–97: an update. Natl Vital Stat Rep 2001;49(4):1–9.
3. Menken J, Trussell J, Larsen U. Age and infertility [published erratum appears in Science 1986;234(5775): 413]. Science 1986;233:1389–94.
4. Bachrach CA, Horn MC. Married and unmarried couples, United States, 1982. Monthly vital statistics report; Vol. 23, no. 15. Hyattsville (MD): National Center for Health Statistics; 1987.
5. Mulligan T, Moss CR. Sexuality and aging in male veterans: a cross-sectional study of interest, ability, and activity. Arch Sex Behav 1991;20:17–25.
6. Bottiglioni F, De Aloysio D. Female sexual activity as a function of climacteric conditions and age. Maturitas 1982;4:27–32.
7. Federation CECOS, Schwartz D, Mayaux MJ. Female fecundity as a function of age: results of artificial insemination in 2193 nulliparous women with azoospermic husbands. N Engl J Med 1982;306:404–6.
8. Olsen J. Subfecundity according to the age of the mother and father. Danish Med Bull 1990;37:281–2.
9. Colombo B, Masarotto G. Daily fecundability: first results from a new data base. Demogr Res 2000;3:[39] p.
10. Dunson DB, Colombo B, Baird DD. Changes with age in the level and duration of fertility in the menstrual cycle. Hum Reprod 2002;17:1399–403.
11. Larsen U, Menken J. Measuring sterility rates from incomplete birth histories. Demography 1989;26:185–201.
12. Wood JW, Holman DJ, Yashin AI, Petersen RJ, Weinstein M, Chang MC. A multistate model of fecundability and sterility. Demography 1994;31:403–26.
13. Guzick DS, Bross DS, Rock JA. A parametric method for comparing cumulative pregnancy curves following infertility therapy. Fertil Steril 1982;37:503–7.
14. Guzick DS, Bross DS. Convenient numerical procedures for estimating cumulative pregnancy curves. Fertil Steril 1992;57:85–91.
15. Cramer DW, Walker AM, Schiff I. Statistical methods evaluating the outcome of infertility therapy. Fertil Steril 1979;32:80–86.
16. Gordon NH. Application of the theory of finite mixtures for the estimation of ‘cure’ rates of treated cancer patients. Stat Med 1990;9:397–407.
17. Chen M-H, Harrington DP, Ibrahim JG. Bayesian cure rate models for malignant melanoma: a case-study of Eastern Cooperative Oncology Group trial E1690. J R Stat Soc Ser C Appl Stat 2002;51:135–50.
18. Dunson DB, Zhou H. Bayesian modeling of fecundability and sterility. J Am Stat Assoc 2000;95:1054–62.
19. Dunson DB. Bayesian modeling of the level and duration of fertility in the menstrual cycle. Biometrics 2001;57:1067–73.
20. Johnson NL, Kotz S, Kemp AW. Univariate discrete distributions. New York (NY): Wiley; 1993.
21. Sheps MC. On the time required for conception. Popul Stud (NY) 1964;18:85–97.
22. Weinberg CR, Gladen BC. The beta-geometric distribution applied to comparative fecundability studies. Biometrics 1986;42:547–60.
23. Johnson NL, Kotz S, Balakrishnan N. Continuous univariate distributions. 2nd ed. New York (NY): Wiley-Interscience; 1994.
24. Wilcox AJ, Weinberg CR, Baird DD. Timing of sexual intercourse in relation to ovulation. Effects on probability of conception, survival of the pregnancy, and sex of the baby. N Engl J Med 1995;333:1517–21.
25. Dunson DB, Baird DD, Wilcox AJ, Weinberg CR. Day-specific probabilities of clinical pregnancy based on two studies with imperfect measures of ovulation. Hum Reprod 1999;14:1835–9.
26. Stanford JB, Smith KR, Dunson DB. Vulvar mucus observations and the probability of pregnancy. Obstet Gynecol 2003;101:1285–93.
27. Stovall DW, Toma SK, Hammond MG, Talbert LM. The effect of age on female fecundity. Obstet Gynecol 1991;77:33–36.
28. van Noord-Zaadstra BM, Looman CW, Alsbach H, Habbema JD, te Velde ER, Karbaat J. Delaying childbearing: effect of age on fecundity and outcome of pregnancy. BMJ, 1991;302:1361–5.
29. Centers for Disease Control (1997) Fertility, family planning, and women’s health: new data from the 1995 national survey of family growth. Vital and Health Statistics.
30. Trussell J, Wilson C. Sterility in a population with natural fertility. Popul Stud J Demogr 1985;39:269–86.
31. Bongaarts JM. A method for the estimation of fecundability. Demography 1975;12:645.
32. Collins JA, Wrixon W,James LB, Wilson EH. Treatment-independent pregnancy among infertile couples. N Engl J Med 1983;309:1201–6.
33. Schieve LA, Meikle SF, Ferre C, Peterson HB, Jeng G, Wilcox LS. Low and very low birth weight in infants conceived with use of assisted reproductive technology. N Engl J Med 2002;346:731–7.
34. Keith L, Oleszczuk JJ. Iatrogenic multiple birth, multiple pregnancy and assisted reproductive technologies. Int J Gynaecol Obstet 1999;64:11–25.
35. Hansen M, Kurinczuk JJ, Bower C, Webb S. The risk of major birth defects after intracytoplasmic sperm injection and in vitro fertilization. N Engl J Med 2002;346:725–30.
36. Maman E, Lunenfeld E, Levy A, Vardi H, Potashnik G. Obstetric outcome of singleton pregnancies conceived by in vitro fertilization and ovulation induction compared with those conceived spontaneously. Fertil Steril 1998;70:240–5.
37. Stromberg B, Dahlquist G, Ericson A, Finnstrom O, Koster M, Stjernqvist K. Neurological sequelae in children born after in-vitro fertilisation: a population-based study. Lancet 2002;359:461–5.

Cited By

This article has been cited 1 time(s).

Obstetrics & Gynecology
Rate of Pregnancy After Using Drospirenone and Other Progestin-Containing Oral Contraceptives
Cronin, M; Schellschmidt, I; Dinger, J
Obstetrics & Gynecology, 114(3): 616-622.
PDF (770) | CrossRef
© 2004 The American College of Obstetricians and Gynecologists