Evidence of deteriorating semen quality in the United Kingdom: birth cohort study in 577 men in Scotland over 11 years

BMJ 1996; 312 doi: (Published 24 February 1996) Cite this as: BMJ 1996;312:467
  1. Stewart Irvine, consultanta,
  2. Elizabeth Cawood, scientific officera,
  3. David Richardson, research officera,
  4. Eileen MacDonald, medical laboratory scientific officerb,
  5. John Aitken, professora
  1. a Medical Research Council Reproductive Biology Unit, Centre for Reproductive Biology, Edinburgh EH3 9EW
  2. b Reproductive Medicine Laboratories, University of Edinburgh, Centre for Reproductive Biology, Edinburgh EH3 9EW
  1. Correspondence and requests for reprints to: Dr Irvine.
  • Accepted 5 February 1996


Objective: To determine whether the quality of semen has changed in a group of over 500 Scottish men born between 1951 and 1973.

Design: Retrospective review of data on semen quality collected in a single laboratory over 11 years and according to World Health Organisation guidelines.

Setting: Programme of gamete biology research funded by Medical Research Council.

Subjects: 577 volunteer semen donors. Of these, 171 were born before 1959, 120 were born in 1960-4, 171 in 1965-9, and 115 in 1970-4.

Main outcome measures: Conventional criteria of semen quality including semen volume (ml), sperm concentration (106/ml), overall motility (% motile), total number of sperm in the ejaculate (106), and total number of motile sperm in the ejaculate (106).

Results: When the four birth cohort groups were compared a later year of birth was associated with a lower sperm concentration, a lower total number of sperm in the ejaculate, and a lower number of motile sperm in the ejaculate. The median sperm concentration fell from 98x106/ml among donors born before 1959 to 78x106/ml among donors born after 1970 (P=0.002). The total number of sperm in the ejaculate fell from 301x106 to 214x106 (P=0.0005), and the total number of motile sperm in the ejaculate fell from 169.7x106 to 129.0x106 (P=0.0065).

Conclusion: This study provides direct evidence that semen quality is deteriorating, with a later year of birth being significantly associated with a reduced number of sperm in adult life.

Key messages

  • Key messages

  • When men born in the 1970s were compared with men born in the 1950s, the total number of motile sperm in the ejaculate was reduced by almost 25%

  • These data confirm previously published data from other countries that semen quality is changing, declining by about 2.1% per year

  • Research is urgently required to examine the function as well as the number of sperm and to assess whether these changes are affecting human health and male fertility


Although the meta-analysis by Carlsen et al attracted much attention,1 the suggestion that the quality of human semen may be deteriorating is not new.2 3 4 5 6 7 Examining data on 14947 men that had been published in 61 papers between 1938 and 1991, Carlsen et al observed a clearly significant decline in average sperm concentration (0.94 x 106/ml/year) corresponding to a decline from 113x106/ml in 1940 to 66x106/ml in 1990. Furthermore, they drew attention to the fact that this change was happening in association with an increase in the reported incidence of congenital malformations of the male genital tract, such as cryptorchidism8 and hypospadias,9 10 and a striking increase in the rates of registration of testicular cancer.11 12 These developments led Sharpe and Skakkebaek to postulate that the observed changes may have a common origin in perinatal life, perhaps mediated through exposure to environmental xenoestrogens.13 More recently, Auger et al reported a decline in the semen quality of a large group of French men, noting that both older age (at ejaculation) and a later year of birth were associated with a decline in the conventional criteria of semen quality.14

The meta-analysis of Carlsen et al has been criticised because the studies included were undertaken in different countries and at different times and therefore bias in subject recruitment or changes in methods of semen analysis may have affected the results.15 In a similar way, the study by Auger et al was criticised for selection bias (only men of proved fertility were studied) and because only one semen sample from each man was included.16 Although there is, as yet, no clear evidence that male fertility, as opposed to semen quality, is declining, the issue of whether later birth cohorts have poorer quality semen is important.17 18 We examined semen quality in a large group of unselected men contributing semen samples to a programme of gamete biology research in the United Kingdom.

Subjects and methods

Subjects were volunteer donors who had offered to provide semen samples for this unit's programme of gamete biology research. Recruitment was by several approaches, including discussions at antenatal parent craft groups, advertisements to undergraduate populations, and personal contact by existing donors. Each potential donor was seen and counselled by the research support and clinical staff of the programme. All participants gave written informed consent to the use of their semen for research according to local guidelines, which incorporated those promulgated by the British Andrology Society19 and the Human Fertilisation and Embryology Authority.

Before the submission of their first sample donors were not selected on the basis of proved fertility or the absence of factors associated with impaired quality of semen. This study is based on a cohort of 577 donors who joined the programme between 1984 and 1995. The donor's date of birth, date of submission of first sample, and results of semen analysis were recorded. The median age of this group of donors at the time of providing their first sample was 27 (10th-90th centile 20-36). Over a quarter (144) were professional men (social classes I and II), 14.3% (82) were in social class IIIN (non-manual), 14.1% (81) were in manual occupations, and only 5.2% (30) were unemployed. Information on occupation was unavailable for 16 (2.8%) subjects. Less than 40% (224) were students and almost half (43.5%, 251) were of proved fertility.20 Information on fertility was unavailable for 11 (1.99%) subjects.


All semen samples were analysed in one laboratory according to a standardised method. The normal reference range was that defined for our local population (sperm concentration >/=20x106/ml and overall motility >/=40%)21 and remained unchanged throughout the period of the study.15 The first sample submitted by each volunteer was analysed, and donors were instructed to abstain from ejaculation for three or four days before giving it. Samples were collected by masturbation into sterile plastic containers.

The sample was allowed to liquefy at 37°C for 30 minutes before analysis, which was within 90 minutes of ejaculation. In general, samples were analysed according to the guidelines of the World Health Organisation.22 23 24 The volume of the ejaculate (ml) was determined by aspirating the liquefied sample into a graduated disposable pipette. To determine the concentration of sperm (106/ml), 10 µl of gently mixed semen was thoroughly mixed with 190 µl of sperm diluting fluid (50 g sodium bicarbonate in 10 ml of 35% formaldehyde per litre of water and loaded in duplicate into the chambers of a haemocytometer (Improved Neubauer, BDH, Lutterworth, Leicestershire). The haemocytometer was examined under a microscope (Ortholux, Leitz, Wetzlar, Germany) at a final magnification of 400 times, and the mean of the counts obtained in the two chambers was calculated. Motility was examined by placing a 20 µl drop of mixed semen on to a prewarmed microscope slide, which was covered with a coverslip (19x19 mm). This preparation was examined at a magnification of 400 times under phase contrast illumination; with the aid of an eyepiece graticule the slide was scanned and at least 100 sperm were examined in four to six randomly chosen fields.

Overall motility was determined as the proportion of sperm showing evidence of movement (WHO grades a, b, and c) to the total number of spermatozoa counted (WHO grades a, b, c, and d).24 The percentage of morphologically normal sperm was not routinely determined in the first samples from this group of donors.


Data were collected in standard microcomputer spreadsheets and analysed using both SPSS version 6.0 for Windows (SPSS, Chicago, Illinois, USA) or STATISTICA-MAC version 4.1 (Statsoft, Tulsa, Oklahoma, USA). The distribution of variables was examined and when appropriate, variables were normalised before analysis by log transformation. Ejaculate volume, sperm concentration, and the derived variables of total number of sperm in the ejaculate (ejaculate volumexsperm concentration;x106 sperm) and total number of motile sperm in the ejaculate (ejaculate volumexsperm concentrationxmotility/100;x106 sperm) were not normally distributed.

In general, data are presented as medians (10th-90th centile), except when indicated otherwise. Relations between variables were examined using linear and stepwise multiple linear regression. In addition, donors were divided into four roughly equal cohorts of five years according to year of birth: 1955-9, 1960-4, 1965-9, and 1970-4. For the purposes of this analysis, the 65 (11.3%) donors born before 1955 were included in the first birth cohort. Differences between groups were examined by analysis of variance or Kruskal-Wallis analysis of variance with comparison of medians as appropriate.


When the age of donors was examined with respect to the year of their first donation, there was no evidence that the age of donors had changed during the period of data collection. Table 1 shows the age, fertility, and semen quality of the study population. As expected, few men in the younger birth cohorts were of proved fertility.

Table 1

Quality of semen of 577 volunteer semen donors. Values are means (SD) unless stated otherwise

View this table:

When semen quality was examined in relation to donors' year of birth, several significant negative relations were observed, suggesting that donors born later tended to have poorer quality semen. Ejaculate volume did not correlate with either year of birth or age at donation. In contrast, sperm concentration decreased by 2.1% per year, the total number of sperm in the ejaculate by 2.01%, and the total number of motile sperm in the ejaculate by 2.04% per year (table 2). Overall motility was weakly positively related to a later year of birth, increasing by 0.18% per year. Because the donors born earlier are older, similar relations but with opposite sign were observed between age at donation and semen quality (table 2) There were no relations between the year of donation and any measures of semen quality, with the exception of overall motility, which increased by 1.2% per year (95% confidence interval 0.87 to 1.62, P<0.0001).

Table 2

Relations between measures of semen quality and year of donor's birth and age at donation (linear regression analysis)

View this table:

Because the process of aging is associated with a deterioration in semen quality,25 a series of stepwise multiple linear regression analyses was undertaken in which both age at donation and year of birth were entered against the various measures of semen quality. On each occasion only one variable, usually year of birth, was selected and was negatively related to semen quality. In the case of the total number of motile sperm in the ejaculate, age at donation was selected because of its stronger positive relation with this variable. In other words, it was not possible in this dataset to observe a negative relation between age and semen quality that was independent of year of birth.

Table 3 and figures 1 and 2 show the semen quality of the four birth cohorts. The median sperm concentration (x106/ml) among donors born in the 1950s was 98.0 (10th-90th centile 38.6-218.4), falling to 78.0 (21.0-166.4) among those born in the 1970s (P=0.002) (fig 1). The overall percentage of motile sperm did not show any change from the 1950s cohort to the 1970s cohort (61.0 (41.6-76) v 64.0 (38.1-83.9); fig 1), but the total number of motile sperm in the ejaculate (x106) fell from 169.7 (52.3-503.9) to 129.0 (29.0-325.8), a fall of almost 24% (fig 2). Similar results were observed when the log transformed variables were examined by analysis of variance (figs 1 and 2).

Table 3

Median values (with 10th-90th centiles) for measures of sperm quality in 577 volunteer semen donors

View this table:
Fig 1
Fig 1

Box plots for raw and log transformed sperm concentration with overall motility in 577 volunteer semen donors according to birth cohort. Changes in sperm concentration were significant (P=0.002) by Kruskal-Wallis test, changes in overall motility were not

Fig 2
Fig 2

Box plots for raw and log transformed total sperm counts in ejaculate and for total number of motile sperm in ejaculate in 577 volunteer semen donors according to birth cohort. Changes in total number of sperm and total number of motile sperm were significant (P=0.0005 and P=0.0065 respectively) by Kruskal-Wallis test


So far as we know, these data are the first evidence that the quality of semen is deteriorating in the United Kingdom. Our findings support previous reports that the quality of human semen seems to be falling.1 14 In particular, we have observed a decline in sperm concentration and the total number of sperm and of motile sperm in the ejaculate in association with a later year of birth, such that men born in the 1970s are producing some 24% fewer motile sperm in their ejaculate than are men born in the 1950s. Auger et al observed that the decline in sperm concentration was some 2.6% per annum with later year of birth,14 while we found that this figure was 2.1%. Unlike Auger et al, we could not show the independent effect of older age on semen quality, reflecting the larger number of subjects, greater age range, and longer data collection period in their study.

As the within subject coefficient of variation for the conventional criteria of semen quality is high,16 it has been suggested that it may be more valid to use multiple semen samples from each donor. In this study we included only the first sample provided by each volunteer to minimise the effects of selection bias resulting from the later exclusion of donors with subnormal semen quality.15 Although we did not record the duration of abstinence separately, all donors were asked to abstain for three to four days before giving their first sample. The duration of abstinence affects several measures of semen quality, most notably ejaculate volume,26 and younger donors would be expected to have shorter periods of abstinence, confounding the assessment of semen quality. However, no association was seen between either age (at ejaculation) or year of birth and ejaculate volume.


Previous studies addressing the issue of secular changes in semen quality have tended to concentrate on selected groups of men, which has raised concern about the possible influence of selection bias. For example, Nelson and Bunge reported data on patients presenting for vasectomy.2 These men had an average sperm concentration of 48x106/ml, with 20% being below 20x106/ml and 7% above 100x106/ml.2 They contrasted these observations with those of MacLeod and Gold in a group of 1000 men of proved fertility; these men had an average sperm concentration of 107x106/ml, only 5% being below 20x106/ml and 38% above 100x106/ml.27

Data have also been collected on men donating semen for therapeutic donor insemination. For example, Leto and Frensilli reported data on 275 such donors and noted that the average sperm concentration (a mean of 12 replicates for each man) fell from 120x106/ml in 1973 to almost 90x106/ml in 1980.4 Most commonly studied, and most subject to changing selection bias, are the male partners of couples presenting with infertility. MacLeod and Wang, for example, reported data on the male partners of 1300 infertile partnerships and saw no appreciable secular trend over three decades.28 However, Bostofte et al noted a fall in the median sperm concentration from 73.4x106/ml to 54.5x106/ml when they compared the semen quality of 1077 Danish men examined in 1952 with that of 1000 similar men examined in 1972.5 In a similar study in Sweden Osser et al compared 185 men in 1980-1 with a similar number of age matched controls from 1960-1.7 They observed that the mean sperm concentration fell from 125.4x106/ml in 1960 to 78x106/ml in 1980.


We emphasise that there is, as yet, no evidence that male fertility, as opposed to semen quality, is declining. Not only are the conventional criteria of semen quality poor indicators of fertility29 30 but it is difficult to judge from the available epidemiological evidence whether the prevalence of infertility is changing.31 Our data indicate that a later year of birth is associated with poorer semen quality in adult life. As such, it is consistent with the hypothesis advanced by Sharpe and Skakkebaek that environmental or other factors acting during fetal and perinatal life can have profound effects on subsequent adult reproductive function,13 but it would also be consistent with environmental factors acting much later in life. A recent report has drawn attention to the possibility that xenoestrogens may be implicated,18 and data from experiments in animals support the hypothesis that perinatal exposure to known environmental xenoestrogens results in alterations in testicular size and sperm production in adulthood.32 Urgent research is required to measure these apparent changes in human semen quality more widely, to examine apparent regional variations, to examine sperm function as opposed to simply sperm number, and to assess whether such changes are having any impact on human health and male fertility.


  • Funding This work is part of the Medical Research Council Reproductive Biology Unit's core funded programme of gamete biology research.

  • Conflict of interest None.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
View Abstract