Time trends in waiting time to pregnancy among Danish twins

Tina Kold Jensen1,5, Mike Joffe2, Thomas Scheike3, Axel Skytthe4, David Gaist4 and Kaare Christensen4

1 Department of Environmental Medicine and 4 Department of Epidemiology, University of Southern Denmark, Institute of Public Health, 5000 Odense C, 3 Department of Biostatistics, University of Copenhagen, 2200 Copenhagen, Denmark and 2 Department of Epidemiology and Public Health, Faculty of Medicine (St Mary's Campus), Imperial College, London, UK

5 To whom correspondence should be addressed at: University of Southern Denmark, Institute of Public Health, Department of Environmental Medicine, Winsloewsparken 17, 5000 Odense C, Denmark. Email: tkjensen{at}health.sdu.dk


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
BACKGROUND: Little is known about time trends in fecundity because few population-based data are available. In a survey among female twins born from 1953 to 1976, their time to pregnancy did not differ from singletons and can be considered to represent the fecundity of the general population. METHODS: Information was collected by interview about waiting time to first pregnancy (TTP) and any periods of subfecundity among both male and female twins born between 1931 and 1952. Trends were analysed by considering the year of birth of the index person (birth cohort effect) or year at which the first attempt started (period effect). RESULTS: Eighty-five percent of male and 81.3% of female eligible twins participated. A total of 1598 male twins and 1653 female twins reported a TTP value and 1671 men and 1715 women had a value for their first attempt to conceive including unsuccessful attempts. No overall trends in either male or female TTP were observed with increasing year of birth or of starting time, after adjustment for confounders, but for attempts fecundity increased among female twins by year of birth or of starting the attempt. A decreasing risk of severe infertility with increasing year of birth or year for start of the attempt was observed among male twins, but this observation was based on only 81 men and was not seen among female twins. CONCLUSION: No decreasing trend in fecundity was observed among Danish twins born between 1931 and 1952 who had completed their reproduction. Female twins had a slight increase in fecundity, and men a decrease in severe infertility. In addition, TTP was apparently well reported and recalled for up to 50 years among both male and female twins.

Key words: fertility/time to pregnancy/time trends/twins


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Semen quality may have declined considerably during recent decades, especially in parts of north-west Europe including Denmark (Auger et al., 1995Go; Swan et al., 2000Go), and significant geographical differences exist (Jouannet et al., 2001Go). It is uncertain whether the decline is accompanied by a fall in fecundity, which is defined as the probability of a couple conceiving in a menstrual cycle. Fecundity can be measured by using the time taken to conceive (time to pregnancy, TTP) (Baird et al., 1986Go). TTP has been proven to be a valuable tool to measure fecundity, it is easy to obtain information about and well recalled (Zielhuis et al., 1992Go; Joffe et al., 1995Go), even many years after the pregnancy (Joffe et al., 1995Go).

Little is known about the true extent of and time trend in fecundity because population-based data are generally not available. Some population studies have reported no secular decline in fecundity (Mosher and Pratt, 1990Go; Templeton et al., 1991Go). In Britain, a cross-sectional study showed significantly increasing fecundity over time (Joffe, 2000Go); two transient dips in fecundity were observed. A Swedish study among 400 000 women found a decrease in subfecundity (defined as TTP >1 year) from 1983 to 1993 which followed a birth cohort pattern (Akre et al., 1999Go). However, the analysis did not take into account truncation bias at both ends of the study period, which occurs when a calendar time cut-off date is imposed (e.g. when data were collected), after which pregnancies are no longer included (Jensen et al., 2000Go). It is therefore preferable to study couples who have finished their reproductive career in order to avoid this type of bias.

Apart from semen quality, fecundity is determined by numerous factors, many of which can change over time: maternal caffeine (Bolumar et al., 1997Go), alcohol intake (Jensen et al., 1998bGo), smoking (Augood et al., 1998Go), timing of sexual intercourse and regularity of the menstrual cycle (Kolstad et al., 1999Go), body mass index (Jensen et al., 1999Go; Pasquali et al., 2003Go), recent use of oral contraceptives (Harlap and Baras, 1984Go) and occupational exposures (Rowland et al., 1992Go; Larsen et al., 1998Go). In particular, various sexually transmitted diseases (STDs) with adverse effects on reproduction have been detected and treated with increasing success during recent decades. Furthermore, women postpone childbirth, and increasing age at first birth reduces fecundity years (Friedmann, 1977Go; van Noord-Zaadstra et al., 1991Go).

The Danish Twin Register is population based, and a recent survey among female twins born between 1953 and 1976 showed that for females, TTP did not differ from singletons and can therefore be considered to represent the fecundity of the general population (Christensen et al., 1998Go). No singleton data were available for males, but male TTP did not vary according to zygosity (Christensen et al., 2003Go). We collected detailed information about waiting time to first pregnancy and any periods of subfecundity among both male and female twins born between 1931 and 1952, since they had already completed their reproduction and we therefore avoided truncation bias. The aim of the study was to examine period and birth cohort effects in fecundity in this large population-based cohort. Given the uncertain state of the literature, with a possible decline in semen quality but an observation of increasing fecundity in the UK (Joffe, 2000Go), we did not have an a priori hypothesis about the direction of the trend.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
We identified pairs of twins born between 1931 and 1952 through the Danish Twin Registry (Skytthe et al., 2003Go). The twins had responded to a brief mailed questionnaire (response rate 77%) in 1997 and had declared their willingness to participate in future studies (90% of responders). This questionnaire included items on similarity of the twins, based on which we assigned zygosity, a method found to result in misclassification in <5% (Christiansen et al., 2003Go). Within each of the 22 birth cohorts, we randomly identified 40 pairs of each zygosity [monozygotic (MZ), and dizygotic same and opposite sex (DZss and DZos)]. Half of the retrieved MZ and DZss twin pairs were female. Owing to insufficient numbers of MZ pairs in the birth cohort of 1933, 1934 and 1936, we retrieved an additional 11 MZ pairs from the birth cohorts of 1931 and 1935. A total of 2640 pairs were identified. Shortly before the survey, we obtained demographic information on these twins from a continuously updated nationwide population registry, the Danish Civil Registration System. Ninety-one twins (1.7%) had died or emigrated in the 2 year period preceding the study, leaving 5189 twins eligible for participation in the study. A pilot study testing all procedures resulted in minor changes only. We compared participants and non-participants with regard to age, gender, urban residency and marital status using data from the Civil Registration System. For details, see Gaist et al. (2000)Go.

The twins participated in detailed interviews in 1998–1999. The questionnaires included a section about reproduction, which differed for male and female twins. Both were asked whether they had ever tried to conceive and if they had ever attempted for >12 months. If they responded positively to that question, they were asked to state when they experienced that period for the first time, whether they eventually succeeded and for how long they tried to conceive in total. They provided information about their total number of live births and stillbirths. They were asked about their first pregnancy leading to a live birth: whether the pregnancy was planned and, if so, for how long they attempted to become pregnant before they succeeded. No information about TTP leading to non-birth pregnancy outcomes was obtained, since recall relating to miscarriages is generally believed to be less valid after a long period of recall.

Both male and female twins stated their own and their partner's age at the start of the attempt, and provided information about their own and their partner's education (none; trainee; or college) and length of school education (Table I). Furthermore, the questionnaire included a section about mothers' smoking habits when they were pregnant with the index person and smoking habits in the home during childhood (none smoked; one parent smoked; or both parents smoked). Female twins were asked whether they used oral contraception before the attempt, their height in centimetres and weight in kilograms before the conception [to calculate the body mass index (weight/height2 in kg/m2)], and if they smoked during pregnancy (used as an indicator of smoking during the attempt). Male twins were asked to state their height and weight at age 25 years and their current smoking habit, which was used as an indicator of smoking habits during the pregnancy attempt (Table I).


View this table:
[in this window]
[in a new window]
 
Table I. Distribution of factors affecting fecundity (in percent) among male and female twins stratified for year of birth

 
Both questionnaires included information on previous and/or current genital diseases. Female twins were asked whether they had ever experienced diseases or operations making it difficult to conceive, especially STDs, ruptured appendix, endometriosis, salpingitis, tubal cysts or infections. Male twins were questioned about history of cryptorchidism, parotitis as an adult, operations for inguinal hernia or varicocele, epididymitis and STDs. We determined whether the diseases had occurred before the first attempt to conceive.

We defined three outcomes: TTP to first live birth, length of first pregnancy attempt leading or not leading to a pregnancy, and severe infertility. Waiting time to pregnancy to first live birth was defined among twins who reported that their pregnancy was planned. First attempts were defined as unsuccessful (attempt lasting >12 months) if this occurred before the waiting time to first live birth. If not, TTP to first live birth was used for the first attempt value. Severely infertile twins were defined as twins who reported that they tried to conceive but did not have any children.

Couples with unplanned pregnancies are excluded in TTP studies, which may introduce bias. This is particularly important in the present study that spans a period of time during which large changes in reproductive behaviour occurred. We therefore examined the proportion of unplanned pregnancies by year of birth and year of start of attempt. In addition, a sensitivity analysis was carried out in which unplanned pregnancies were included and assigned a TTP value of 1 month, to test whether that affected our findings. These two measures would uncover any bias resulting from changes in planning or in its reporting.

Statistics
Trends in TTP were analysed by considering the year of birth of the index person (birth cohort effect) or the year in which the first attempt or TTP started (period effect) as regression variables in the logistic regression model. We did this by using the period or cohort variable as continuous regression variables (trend analysis) or categorizing them into intervals. These intervals were chosen so they included an approximately equal number of events each.

Fecundability is the probability of obtaining a clinically recognized pregnancy in a menstrual cycle among couples not pregnant in the previous cycles. We analyswed the TTP by survival analysis techniques for discretely observed survival times. We implemented this analysis as a logistic regression on the total number of observed months with the outcome ‘pregnant/ not pregnant’ (Scheike and Jensen, 1997Go). The fecundability odds ratio (FOR) measures the odds of a conception within a cycle among couples with one birth or start year category divided by the odds for the baseline stratum. This is equivalent to the fecundability ratio (FR) derived from the Cox's proportional hazard regression previously used in many publications (Jensen et al., 1998aGo; Joffe, 2000Go). We used reporting in months instead of cycles as the former is more readily recalled.

First the unadjusted FOR in different birth and start years were compared to describe trends in ‘raw’ fecundity, and then a multiple logistic regression was undertaken, taking into account the different distribution of covariates over time to adjust for changes in known risk factors such as smoking habits and age at first attempt. The birth or start years were entered as categorical or continuous explanatory variables to test whether this had a significant effect on the FOR. This was done by comparing the differences in the 2-log likelihood (by {chi}2 test) in the models with and without the categorical variable.

Potential covariates were examined in bivariate analyses by comparing birth year and starting time distributions among couples with and without the exposure. The covariates were included in the multiple logistic regression analysis and excluded stepwise if the point estimates of the association between birth and period effects and fecundability changed by <10% after exclusion. FORs are presented with 95% confidence intervals (CIs).

TTP was censored after 12 months, which is the time where infertility treatment is normally started, and also because sparse values beyond 12 months interfere with the statistical analysis and recall of longer TTP is less reliable (Joffe et al., 1995Go). TTP values of 0 and 1 were grouped, as respondents may have difficulty making a clear distinction between them when conceptions occur very quickly.

Severe infertility was coded as a dichotomous variable (yes; no) and analysed in a multiple logistic regression to calculate the OR. The only covariate inserted was age at start of the attempt and diseases in reproductive organs. An OR <1.0 implies higher fertility (less infertility) in these analyses, in contrast to the analyses of TTP and all attempts, in which an FOR <1 implies reduced fertility.


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Eighty-one percent of female and 85% of male eligible twins, respectively, participated. Participants were slightly younger, more likely to be married and less frequently lived in urban areas compared with non-participants (for details, see Gaist et al., 2000Go).

A total of 2116 and 2198 female and male twins answered the reproductive questions and 2010 and 2002, respectively, had tried to conceive (Figure 1a and b). Among these, 100 women and 125 men had unsuccessful attempts, of whom 65 (3.2%) and 81 (3.8%) never conceived (were part of a severely infertile couple). Furthermore, five women and eight men had an unsuccessful attempt before their first waiting time to a live birth. A total of 1917 women and 1911 men had children, and 1754 (91%) women and 1766 (92%) men had planned their pregnancies. Of these, 1653 (94.2%) women and 1598 (90.5%) men reported a TTP. Finally, 1715 women and 1671 men had a value for their first attempt including unsuccessful attempts (Figure 1a and b).



View larger version (28K):
[in this window]
[in a new window]
 
Figure 1. Selection procedure for female (a) and male (b) twins with values for TTP and attempts. Some twins with no children had not attempted for >12 months (28 female and 10 male twins), and they were excluded.

 
Before adjustment, fecundity decreased with increasing year of starting the attempt (continuous variable) among the female twins, but no similar findings were observed among male twins (Figure 2a and b, Table II). No trend in length of first attempt was found with birth cohort effect among male or female twins before adjustment (Figure 2ad, Table II) although women born between 1937 and 1942 seemed to have a higher fecundity than the other birth cohorts. The risk of severe infertility increased with increasing birth and time period before adjustment among the female twins. The distribution of covariates among men and women born in different years can be seen in Table I. The age at the start of the first pregnancy attempt showed no clear trend in relation to year of birth among either female or male twins or their partners. The level of education and the percentages exposed to smoking in utero increased with increasing year of birth. Body mass index decreased with increasing birth year, especially among women. Among the female twins, the use of oral contraceptives 3 months before the attempt increased, but no change in diseases in reproductive organs were seen.



View larger version (25K):
[in this window]
[in a new window]
 
Figure 2. (a–d) TTP distribution (unadjusted) by year of birth and year of starting the attempt among male (b,d) and female (a,c) twins.

 

View this table:
[in this window]
[in a new window]
 
Table II. Adjusted fecundability odds ratio (FOR) for TTP and first attempts in relation to year of birth and starting time among female and male twins born from 1931 to 1952

 
FORs for all attempts and for TTP only were calculated, separately for female and male twins. No overall trends in either female or male TTP were observed with increasing year of birth or starting time, after adjustment for female age at the start of the attempt, diseases in reproductive organs before the attempt among the index persons, index person exposed to smoking in utero, for male twins current smoking and for female twins smoking in pregnancy and use of oral contraceptives 3 months before TTP (Table II).

For all attempts, a slight increase in fecundity over time was seen among female twins, after adjustment for the index person's age at the start of the attempt, diseases in reproductive organs before the attempt and index person exposed to smoking in utero (Table II). This was compatible with either a period or a birth cohort effect. In male twins, also adjusted for current smoking, there was no clear trend (although a weak non-significant upward trend is discernible for a period effect). The unadjusted increased risk of severe infertility among female twins reversed after adjustment for age at the start of the attempt and diseases in reproductive organs.

As sensitivity analyses, the analyses were repeated without combining the values 0 and 1 months, and with elimination of both from the analyses. In addition, different cut-offs were used for censoring (6 and 14 months) which did not change the findings. Monozygotic and dizygotic twins were also analysed separately, with similar findings.

It is possible that the percentage of unplanned pregnancies could have changed over time. There was no change in the percentage of unplanned pregnancies with year of birth among either gender, but a decrease with increasing year of start of the attempt (Figure 3a and b). The twins with unplanned pregnancies were younger, slimmer, more often non-smokers and not using oral contraeption before the attempt. Including the unplanned pregnancies with a TTP of 1 month in the analyses did not change the main findings among either men or women.



View larger version (25K):
[in this window]
[in a new window]
 
Figure 3. The percentage of unplanned pregnancies in relation to birth year and starting year among male and female twins.

 
Non-response to an item (the percentage of missing answers to the TTP question) and digit preference (a preference for numbers of months corresponding to whole or half years) are likely to be related to the length of recall. Item non-response fell with increasing year of birth, from 8 to 6% among women and from 12 to 8% among men. For digit preference, we examined trends in the percentages of participants answering that their TTP was 6, 12, 18, 24, 30, 36, 42, 48, 54, 60, 66, 72, 78, 84 and 90 months. There were no trends in the percentages of answers in these categories over period or birth year (data not shown). In addition, the digit preference around 6 months seemed to be similar in all time groups for both males and females (Figure 2).

The OR for being severely infertile in relation to year of birth and year of starting the attempt among male and female twins can be seen in Table 2. Male twins were significantly less infertile as their year of birth and starting the attempt increased, after controlling for diseases in reproductive organs before the attempt started and age at the start of the attempt. No such trend was observed for female twins. The analyses were repeated without controlling for diseases in reproductive organs, as they may be the cause of the infertility, which did not change the findings. These analyses were, however, based on only 65 and 81 reported cases, respectively, of severe infertility among female and male twins.


    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
We found no trend in TTP with birth cohort or time period among female and male twins born between 1931 and 1952. For all attempts, a slight increase in fertility was seen among female twins, as either a period or a birth cohort effect, but no similar trend was observed among the male twins. A decreasing risk of severe infertility with increasing year of birth or year for start of the attempt was observed among male twins, this observation being statistically significant despite being based on only 81 men; it was not seen among the female twins.

We believe that our findings are trustworthy and can be generalized to represent the fecundity of the general population for the following reasons. (i) A previous study among female twins born between 1953 and 1976 found that the fecundity of twins did not differ from that of singletons (Christensen et al., 1998Go, 2003Go). (ii) We studied a sample of >1700 men and women, which provides enough power to study TTP taking into account the effect of covariates. However, the number of severely infertile couples was low, thus not allowing a clear conclusion about time trends in severe infertility to be drawn. (iii) The twins had practically all finished their reproduction and therefore truncation bias does not have to be considered (Jensen et al., 2000Go). (iv) The validity of the TTP question was apparently high in spite of the long period of recall, a question which we now examine in detail.

The quality of the reproductive questions needs to be considered given that the length of recall was up to 50 years. Studies have shown that women give valid information about TTP (Zielhuis et al., 1992Go; Joffe et al., 1995Go) and that the quality of recall does not decline for at least 20 years at a group level (Joffe et al., 1995Go). Occupational studies have shown that men are also capable of providing TTP information following the same distributions as women, although with slightly more missing values and digit preference (Joffe, 1989Go). In the 1958 British birth cohort, the same analytical results were obtained from the TTP values given by male and female respondents, but with a slightly higher rate of item non-response among men (Joffe and Li, 1994bGo). In the British study of trends in TTP, closely similar findings were observed from replies given by male and female respondents (Joffe, 2000Go).

In our data, the TTP distribution reported by female and male twins was very similar (Figure 2), with 80 and 88% of the female twins and 80 and 89% of male twins conceiving in 6 and 12 months, respectively. The proportion of twins conceiving within 3 months was 67 and 68%, respectively, which is higher than other TTP studies. The proportion of couples who had never conceived (3.2 and 3.8%), and the proportion of pregnancies that were reported as planned (91 and 92%) were also very similar for women and men. The similarity between male and female respondents strongly suggests that these self-reports correspond to the true values, especially as they are plausible for this population in the light of other evidence (Larsen et al., 1998Go; Jensen et al., 1999Go; Karmaus et al., 1999). There was no time trend in digit preference, and digit preference for 6 months was of similar magnitude in all time categories. The percentage item non-response increased in the older subjects, which is reasonable given the length of the recall period, but the response rate was still high. We therefore believe that the quality of the responses was sound.

The effect of covariates in our study was similar to previously published studies. Fecundity decreased with increasing maternal age (Friedmann, 1977Go; van Noord-Zaadstra et al., 1991Go), maternal smoking (Augood et al., 1998Go), exposure to mother's smoking in pregnancy (Jensen et al., 1998), diseases in reproductive organs among both male and female twins and use of oral contraceptives before the attempt, increased the conception probability during 0–3 months after cessation but may be due to the fact that women taking OC comprise a more fertile subgroup (Harlap and Baras, 1984Go). This demonstrates that the accuracy of reporting in this study is comparable with those with considerably shorter duration of recall.

Some limitations of our study need to be considered. The study period spanned 22 years for the birth cohort effect, and most of the second half of the 20th century for the period effect, a period during which many societal changes occurred. We obtained information on most relevant changes, such as contraceptive behaviour/planning, female smoking, parental ages and educational level. However, because of the long period of recall, we did not find it feasible to obtain information about some possible covariates, such as frequency of sexual intercourse, alcohol and caffeine intake and length of the menstrual cycle. This may have affected our findings, if the distribution of unmeasured covariates changed over time. The men did not provide information about their own and their partners' smoking habits during their time of attempt, and the women were asked about smoking habits in pregnancy, not during the attempt. In addition, information about the first unsuccessful attempt was only recorded if it lasted >1 year and less information about covariates was asked than for the first TTP, making it more difficult to adjust for covariates in the multiple regression analyses.

A possible problem with the extreme categories for the period effect analyses also needs to be considered, because the earliest category would have included only relatively young twins, and the last category would have comprised relatively old twins. Because female fecundity decreases with age, this could introduce an apparent downward trend in fecundity over time, but adjusting for female partner's age would remove this bias. Furthermore, TTP is moderately correlated for MZ twins (Christensen et al., 2003Go) but it has been shown that the primary effect of clustering of twins pairs is on the standard errors and not parameter estimates, and we therefore slightly underestimated the CIs, but even so we found no significant trend in fecundity.

In addition, some more general limitations in studies of TTP need to be considered (Baird et al., 1986Go; Weinberg et al., 1993Go, 1994Go; Joffe, 2003Go). Pregnancies reported as unplanned were excluded. More fertile couples are more likely to have birth control failures or accidental pregnancies, and are consequently more likely to be excluded from analyses of pregnancy attempts. Some couples may also have changed their attitude to the pregnancy and reported it as planned even though it occurred as a result of a birth control failure. This is likely to have occurred in this study, since the period of recall was long and the percentages of unplanned pregnancies were lower than in previous studies. If time trends in unplanned pregnancies, or in unplanned pregnancies reported as planned, occurred, this could have biased our findings. No time trends in the percentages of unplanned pregnancies with year of birth were observed, but more unplanned pregnancies occurred in the earlier years of starting the attempts. This could have masked a period effect, a trend of declining fecundity with time. In addition, non-planners were younger and therefore more fecund. However, the analyses were repeated after exclusion of women who became pregnant during the first month, and conversely by inclusion of unplanned pregnancies (setting TTP=1), which did not change our findings, suggesting that these were not important sources of bias.

Both male and female twins were more likely to report having had chlamydia or gonorrhoea in the later study period. This may underestimate the underlying fecundity of the later cohorts, making it less likely to find an increase in fecundity. However, early treatment of STDs may have the effect of protecting fecundity, thereby overestimating the underlying fecundity of the late cohorts. We did, however, control for STDs in the multiple regression analysis. Mean age at start of the first attempt has increased in the general population during the study period but we were unable to study trends in age in these data because the twins were sampled by cohorts and the more likely to be younger in the first period (1948–60) and older in the last period (1971–). Postponing childbirth may reduce fecundity, especially in women (Friedmann, 1977Go; van Noord-Zaadstra et al., 1991Go). However, we did take into account age at the start of the attempt in our analyses but were unable to control for the fact that women attempting for the first time at age 30 years in the early 1980s may not be comparable with women aged 30 attempting in the early 1960s.

Infertility treatment has also increased in later cohorts, e.g. clomiphene (IVF treatment was not performed before the early 1980s in Denmark). Successful treatment may lead to a shift in the TTP distribution towards lower fecundity since couples who were otherwise infertile and absent from the sample are now included, and they have long TTP values. So any impact of effective treatment would obscure a trend towards increasing fecundity in the later periods in the TTP data. It would, however, not affect our data on attempts. If the success rate increases over time, more couples would be expected eventually to have a child and therefore fewer will be severely infertile. We only included pregnancies leading to a birth, since we believed that it was not feasible to obtain valid information about miscarriages due to the long period of recall. TTP is, however, not strongly related to miscarriage risk, and since we make internal comparisons we do not believe that this can have affected our findings (Joffe and Li, 1994aGo).

Few studies have examined trends in biological fertility, and these were outlined in the introduction (Akre et al., 1999Go; Joffe, 2000Go). Semen quality may have deteriorated in several populations during recent decades (Auger et al., 1995Go; Swan et al., 1997Go; Bonde et al., 1998bGo), but this applies probably mostly to cohorts born since 1960, whereas our cohorts were born from 1931 to 1952. Unchanged fecundity does not exclude the possibility of declining sperm count, as TTP is associated with sperm concentration only below a threshold of ~40 x 106/ml (Bonde et al., 1998aGo); and it is estimated that 12 000 TTP values need to be included in a study to detect a decrease in sperm count of 21% over a 15 year period (Slama et al., 2004Go).

In conclusion, we found no decline in fecundity among >4000 Danish twins born between 1931 and 1952 who had completed their reproduction. There was a weak increase in fecundity over time among female twins, and a fall in severe infertility among men. In addition, TTP was apparently well reported and recalled for up to 50 years among both male and female twins.


    Acknowledgements
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
This study was supported by grants from the Danish national Research Foundation (Grant no. SV1736), Sygekassernes Helsefond Grant 11/2 12-97.


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Akre O, Cnattingius S, Bergström R, Kvist U, Trichopoulos D and Ekbom A (1999) Human fertility does not decline: evidence from Sweden. Fertil Steril 71, 1066–1069.[CrossRef][ISI][Medline]

Auger J, Kunstmann JM, Czyglik F and Jouannet P (1995) Decline in semen quality among fertile men in Paris during the past 20 years. N Engl J Med 332, 281–285.[Abstract/Free Full Text]

Augood C, Duckitt K and Templeton AA (1998) Smoking and female infertility: a systematic review and meta-analysis. Hum Reprod 13, 1532–1539.[Abstract]

Baird DD, Wilcox AJ and Weinberg C (1986) Use of time to pregnancy to study environmental exposures. Am J Epidemiol 124, 470–480.[Abstract]

Bolumar F, Olsen J, Rebagliato M, Bisant L and the European Study Group of Infertility and Subfecundity (1997) Caffeine intake and delayed conception: a European multicenter study on infertility and subfecundity. Am J Epidemiol 145, 324–334.[Abstract]

Bonde JP, Ernst E, Jensen TK, Hjollund NHI, Kolstad H, Henriksen TB, Scheike T, Giwercman A, Olsen J and Skakkebæk NE (1998a) The relation between semen quality and fertility. A population based study of 430 first-pregnancy planners. Lancet 352, 1172–1177.[CrossRef][ISI][Medline]

Bonde JP, Jensen TK, Larsen SB, Abell A, Scheike T, Hjollund NHI, Kolstad H, Ernst E, Giwercman A, Skakkebæk NE, Keiding N and Olsen J (1998b) Year of birth and sperm count in 10 Danish occupational studies. Scand J Work Environ and Health 24, 407–413.

Christensen K, Basso O, Kyvik KO, Juul S, Boldsen J, Vaupel JW and Olsen J (1998) Fecundability of female twins. Epidemiology 9, 189–192.[ISI][Medline]

Christensen K, Kohler HP, Basso O, Olsen J, Vaupel JW and Rodgers JL (2003) The correlation of fecundability among twins: evidence of a genetic effect on fertility? Epidemiology 14, 60–64.[CrossRef][ISI][Medline]

Christiansen L, Frederiksen H, Schousboe K, Skytthe A, Wurmb-Schwark N, Christensen K and Kyvik K (2003) Age- and sex-differences in the validity of questionnaire-based zygosity in twins. Twin Res 6, 275–278.[CrossRef][ISI][Medline]

Friedmann S (1977) Artificial donor insemination with frozen human semen. Fertil Steril 28, 1230–1233.[ISI][Medline]

Gaist D, Bathum L, Skytthe A, Jensen TK, McGue M, Vaupel JW and Christensen K (2000) Strength and anthropometric measures in identical and fraternal twins: no evidence of masculinization of females with male co-twins. Epidemiology 11, 340–343.[CrossRef][ISI][Medline]

Harlap S and Baras M (1984) Conception-waits in fertile women after stopping oral contraceptives. Int J Fertil 29, 73–80.[ISI][Medline]

Jensen TK, Henriksen TB, Hjollund NHI, Scheike T, Kolstad H, Giwercman A, Ernst E, Bonde JP, Skakkebæk NE and Olsen J (1998a) Adult and prenatal exposure to tobacco smoke as risk indicators of fertility among 430 Danish couples. Am J Epidemiol 148, 992–997.[Abstract]

Jensen TK, Hjollund NH, Henriksen TB, Scheike T, Kolstad H, Giwercman A, Ernst E, Bonde JP, Skakkebaek NE and Olsen J (1998b) Does moderate alcohol consumption affect fertility? Follow up study among couples planning first pregnancy. Br Med J 317, 505–510.[Abstract/Free Full Text]

Jensen TK, Scheike T, Keiding N, Schaumburg I and Grandjean P (1999) Fecundability in relation to body mass and menstrual cycle patterns. Epidemiology 10, 422–428.[CrossRef][ISI][Medline]

Jensen TK, Keiding N, Scheike T, Slama R and Spira A (2000) Declining human fertility? Fertil Steril 73, 421–423.

Joffe M (1989) Feasibility of studying subfertility using retrospective self reports. J Epidemiol Community Health 43, 268–274.[Abstract]

Joffe M (2000) Time trends in biological fertility in Britain. Lancet 355, 1961–1965.[CrossRef][ISI][Medline]

Joffe M (2003) Invited commentary: the potential for monitoring of fecundity and the remaining challenges. Am J Epidemiol 157, 89–93.[Abstract/Free Full Text]

Joffe M and Li Z (1994a) Association of time to pregnancy and the outcome of pregnancy. Fertil Steril 62, 71–75.[ISI][Medline]

Joffe M and Li Z (1994b) Male and female factors in fertility. Am J Epidemiol 140, 921–929.[Abstract]

Joffe M, Villard L, Li Z, Plowman R and Vessey M (1995) A time to pregnancy questionnaire designed for long term recall: validity in Oxford, England. J Epidemiol Community Health 49, 314–319.[Abstract]

Jouannet P, Wang C, Eustache F, Jensen TK and Auger J (2001) Semen quality and male reproductive health: the controversy about human sperm concentration decline. APMIS 109, 333–344.[CrossRef][ISI][Medline]

Karmaus W, Juul S and European Infertility and Subfecundity Group (1999) Infertility and Subfecundity in population-based samples from Denmark, Germany, Italy, Poland and Spain. Eur J Public Health 9, 229–235.[Abstract]

Kolstad HA, Bonde JP, Hjollund NHI, Jensen TK, Henriksen TB, Ernst E, Giwercman A, Skakkebæk NE and Olsen J (1999) Menstrual cycle patterns and fertility. A prospective follow-up study of pregnancy and early embryonal loss in 295 couples who were planning their first pregnancy. Fertil Steril 71, 490–496.[CrossRef][ISI][Medline]

Larsen SB, Joffe M and Bonde JP (1998) Time to pregnancy and exposure to pesticides in Danish farmers. ASCLEPIOS Study Group. Occup Environ Med 55, 278–283.[Abstract]

Mosher WD and Pratt WF (1990) Fecundity and Infertility in the United States, 1965–1988. Advanced Data from Vital and Health Statistics. National Center for Health Statistics, Hyattsville.

Pasquali R, Pelusi C, Genghini S, Cacciari M and Gambineri A (2003) Obesity and reproductive disorders in women. Hum Reprod Update 9, 359–372.[Abstract/Free Full Text]

Rowland AS, Baird DD, Weinberg CR, Shore DL, Shy CM and Wilcox AJ (1992) Reduced fertility among women employed as dental assistants exposed to high levels of nitrous oxide. N Engl J Med 327, 993–997.[Abstract]

Scheike T and Jensen TK (1997) A discrete survival model with random effects: an application to time to pregnancy. Biometrics 53, 318–329.[ISI][Medline]

Skytthe A, Pedersen NL, Kaprio J, Stazi MA, Hjelmborg JV, Iachine I, Vaupel JW and Christensen K (2003) Longevity studies in GenomEUtwin. Twin Res 6, 448–454.[CrossRef][ISI][Medline]

Slama R, Kold-Jensen T, Scheike T, Ducot B, Spira A and Keiding N (2004) How would a decline in sperm concentration over time influence the probability of pregnancy? Epidemiology 15, 458–465.[CrossRef][ISI][Medline]

Swan SH, Elkin EP and Fenster L (1997) Have sperm densities declined? A reanalyses of global trend data. Environ Health Perspect 105, 1228–1232.[ISI][Medline]

Swan SH, Elkin EP and Fenster L (2000) The question of declining sperm density revisited: an analysis of 101 studies published in 1931–1996. Environ Health Perspect 108, 961–966.[ISI][Medline]

Templeton A, Fraser C and Thompson B (1991) Infertility—epidemiology and referral practice. Hum Reprod 6, 1391–1394.[Abstract]

van Noord-Zaadstra BM, Looman CWN, Alsbach H, Habbema JDF, te Velde ER and Karbaat J (1991) Delaying childbearing: effect of age on fecundity and outcome of pregnancy. Br Med J 302, 1361–1365.[ISI][Medline]

Weinberg CR, Baird DD and Rowland AS (1993) Pitfalls inherent in retrospective time-to-event studies: the example of time to pregnancy. Stat Med 12, 867–879.[ISI][Medline]

Weinberg CR, Baird DD and Wilcox AJ (1994) Sources of bias in studies of time to pregnancy. Stat Med 13, 671–681.[ISI][Medline]

Zielhuis GA, Hulscher MEJL and Florack EIM (1992) Validity and reliability of a questionnaire on fecundability. Int J Epidemiol 21, 1151–1156.[Abstract]

Submitted on October 28, 2004; resubmitted on December 8, 2004; accepted on December 8, 2004.