Jaime Mendiola, Division of Preventive Medicine and Public Health, University of Murcia School of Medicine, Espinardo (Murcia) 30100 Spain. E-mail: email@example.com
Several studies have investigated temporal trends in semen quality in Northern Europe, but none has examined this question in Southern Europe. A prior study conducted in Almeria Province (Southern Spain) reported higher sperm count and concentration among Spanish young men recruited from 2001 to 2002 compared with young men from Northern Europe. The aim of this new study was to examine whether semen quality has changed among Spanish young men in the last decade. In this cross-sectional study, questionnaires and semen samples were collected from 215 healthy young university students from Murcia Region between 2010 and 2011. The 273 men from the Almeria study previously studied were included in a trend analysis of the two populations from Southern Spain. Descriptive statistics were calculated for the Murcia study population and these and semen variables for the Murcia and Almeria study populations were compared. Study methods and population characteristics were similar across the two studies. Therefore, we used multiple linear regression analyses on the combined population (controlling for study centre, age, ejaculation abstinence time, season, smoking, medication during the last 3 months, Body mass index (BMI), presence of varicocoele and prenatal exposure to tobacco) to look for a birth-cohort effect over the combined study period (2001–2011). Sperm concentration and total sperm count declined significantly with year of birth in the pooled analysis (β = −0.04 and β = −0.06, respectively, both p < 0.01). Sperm counts were significantly lower in Murcia study subjects than in the Almeria participants; sperm concentration median (5th–95th) = 44.0 (8.9–129) million/mL vs. 51.0 (5.0–206) million/mL; p < 0.01 and total sperm count = 121 (17.8–400) million vs. 149 (8.0–599) million; p < 0.01. Other semen variables did not differ significantly between the two studies. Our study suggests that total sperm count and sperm concentration may have declined in young Spanish men over the last decade.
In 1992, Carlsen and colleagues using historical data suggested a temporal decline in human sperm concentration between 1938 and 1991 (Carlsen et al., 1992). This was later corroborated in a review including 47 more studies published between 1934 and 1996 and showed a large annual decline in sperm concentration in European men (−2.3%) and a smaller decline in US (−0.8%) (Swan et al., 2000). Following, several newer cross-sectional studies have shown the existence of geographical differences in semen quality among men who were not selected because of fertility or infertility (Jørgensen et al., 2002, 2011; Punab et al., 2002; Richthoff et al., 2002; Paasch et al., 2008; Axelsson et al., 2011; Fernandez et al., 2012). Besides, it has also been suggested that an impaired development of foetal testes might lead to increased risks of hypospadias, cryptorchidism, testis cancer or decreased spermatogenesis (Skakkebæk et al., 2001). The concept, called testicular dysgenesis syndrome (TDS), links the pathogenesis of the four disorders together, but does not entail that all affected men develop all four symptoms (Jørgensen et al., 2010).
Nordic studies have examined recent trends in semen quality of young men from general populations. In Sweden no change was observed between 2000 and 2010 (Axelsson et al., 2011), whereas a minor increase in total sperm counts and sperm concentration but not in morphology or motility have been shown for Danish men (between 1996 and 2010) (Jørgensen et al., 2012). In contrast, sperm concentration, total sperm counts and frequency of morphologically normal spermatozoa have decreased for Finnish men between 1998 and 2006 (Jørgensen et al., 2011). Recently, it has been shown adverse trends and values of sperm concentration and morphology in a large sample of men close to the general population between 1989 and 2005 in France (Rolland et al., 2012). However, all these studies indicated relatively low semen quality levels in the study populations when the results were evaluated against fecundability data in human subjects (Bonde et al., 1998; Guzick et al., 2001; Slama et al., 2002).
A study from the Almeria Province in the Southern Spain (Almeria study) carried out between 2001 and 2002 indicated that Spanish young men had higher sperm counts than men from Northern Europe (Fernandez et al., 2012). The aim of this new study was to examine whether semen quality has changed among Spanish young men in the last decade.
Materials and methods
University students, 18–23 years old, in the Murcia Region (Southern Spain) were included in the ‘Murcia Young Men's Study’ (Murcia study) between 6 October 2010 and 29 November 2011. Written informed consent was obtained from all subjects, and the study was approved by The Research Ethics Committee of the University of Murcia.
Flyers stating, ‘Young healthy male university students wanted for research project’ were posted at university campuses to invite students to participate in this study. To be included, the men had to be university students, born in Spain after 31 December 1987, and able to contact their mother and ask her to complete a questionnaire. A total of 240 students contacted us, among which 17 did not fulfil the inclusion criteria (five not been born in Spain: Nine been born before 31 December 1987, and three not able to contact their mothers). Therefore, 223 were eligible, but eight did not show up for their scheduled appointment, and thus 215 participated in the study.
On the day of attendance the men underwent a physical examination, provided a semen sample and completed questionnaires on lifestyle. The young men completed the questionnaire during the clinical appointment. However, in case they did not know the answer to a particular question they contacted their mothers. The questionnaire included information on demographics, previous and current conditions or diseases, tobacco smoking, medication and drug consumption and sexual development. Participants were rewarded for their participation (€50 gift card).
Body weight and height were measured using a digital scale (Tanita SC 330-S, London, UK). BMI was calculated as weight in kilograms divided by squared height in metres. The same specialized examiner (J.M.) carried out all the physical exams in our study centre. Testes sizes were measured using a Prader orchidometer (Andrology Australia, Clayton, Victoria, Australia). Presence of varicocoele or other scrotal abnormalities and testicular location were also evaluated.
The men had been asked to abstain from ejaculation for at least 48 h before sample collection. Nonetheless, subjects were not excluded if they had not abstained for that period of time (n = 30, 14%). Abstinence time (hours) was recorded as the time between current and previous ejaculation as reported by the study subject. Men collected semen samples by masturbation at the study centre. Semen volumes were estimated by specimen weight, assuming a semen density of 1.0 g/mL. For motility assessment, spermatozoa were classified as either motile or immotile (World Health Organization, 1999) to report the percentage of motile spermatozoa. Briefly, a 10 μL of well-mixed semen was placed on a clean glass slide that had been kept at 37 °C and covered with a 22 × 22 mm coverslip. The preparation was placed on the heating stage of a microscope at 37 °C and immediately examined at ×400 magnification. Sperm concentration was assessed using a haemocytometer (Improved Neubauer; Hauser Scientific Inc., Horsham, PA, USA). For that samples were diluted in a solution of 0.6 m NaHCO3 and 0.4% (v⁄v) formaldehyde in distilled water. The haemocytometer chamber was loaded with the dilution and the spermatozoa were allowed to settle in a humid chamber. From the same dilution, two chambers of the haemocytometer were assessed and at least 200 spermatozoa per replicate were counted. Total sperm count (volume × sperm concentration) was also calculated. Smears for morphology were made, air dried, fixed, Papanicolaou stained and assessed using strict criteria (Menkveld et al., 1990). The same specialized biologist carried out all the semen analyses (L.S.C.).
To increase consistency and international comparability (inter-laboratory variation) five sets of duplicate semen samples (600 μL each) (for sperm concentration only) were sent by mail during the study period from the University of Copenhagen's Department of Growth and Reproduction to the Murcia Andrology Laboratory. The specimens were blinded undiluted fresh sperm samples from regular semen donors that were preserved by adding 10 μL of a 3 m sodium azide solution per 1 mL of the ejaculate after liquefaction. No systematic difference was shown and the mean inter-examiner coefficient of variation was 4.0% with a range for the sets between 1.7 and 7.1%. The same quality control procedure was carried out in the Almeria study (Fernandez et al., 2012), which allows, to a high degree, comparability between our results on sperm concentrations.
Historical comparison study
The methods employed in the Almeria study from November 2001 to December 2002 have been described elsewhere (Fernandez et al., 2012). Briefly, the study was advertised via media (TV, radio, etc.) and at the University of Almeria ‘Open Days’ at the beginning of the academic course and aimed at undergraduate populations in the study area. The inclusion criteria were that the mothers of the men were born in Spain, the men were born in the Almeria province (Southern Spain) and living there at time of participation in the study. The study population comprised 273 men. On the day of attendance, the men returned a completed questionnaire, underwent a physical examination and provided a semen sample. Participants were rewarded for their participation (€20).
Descriptive statistics are presented using untransformed data. Linear regression models were used to test the effect of potential covariates. For that, semen volume, sperm concentration, total sperm count and percentage of normal sperm morphology were ln-transformed before analysis because of a skewed distribution. Covariate assessment included: age, BMI (kg/m2), ejaculation abstinence time (hours), presence of varicocoele (yes/no), recent fever (yes/no), smoking (current smoker vs. not current smoker), time to start semen analysis (minutes) (only for motility) and season (winter vs. spring, summer or fall). There was a significant positive association between abstinence time and semen volume (β = 0.001, p = 0.03) and abstinence time and sperm concentration and total sperm count (both p values <0.05); and a significant negative association between time to start semen analysis and percentage of motile spermatozoa (β = −0.15; p = 0.02). None of the other covariates were statistically significantly related to semen variables in the Murcia study population.
Murcia and Almeria study
Data from Almeria (Fernandez et al., 2012) and Murcia studies were pooled for statistical analysis. Descriptive statistics are presented using untransformed data for both studies. Student's t-test, Pearson chi-squared test and analysis of covariance (ancova) were used to examine whether sperm variables and covariates for men from the Murcia and Almeria study population were significantly different. Study methods and population characteristics were similar across the two studies. Therefore, the two populations were combined and multiple linear regression analyses performed controlling for appropriate covariates (age, ejaculation abstinence time, season, study centre, smoking, medication during the last 3 months, BMI, presence of varicocoele and pre-natal exposure to tobacco) to look for a year-of-birth effect over the combined study period (2001–2011). When inclusion of a potential covariate resulted in a change in the β-coefficient of <10%, the variable was not retained in final models. Ejaculation abstinence time was the only variable retained in the final models. All tests were two tailed and the level of statistical significance was set at 0.05. Statistical analyses were performed with the statistical package IBM spss 20.0 (IBM Corporation, Armonk, New York, USA).
Table 1 shows a general description of the Murcia study population and physical examination compared with Almeria study (Fernandez et al., 2012). Age, BMI, ejaculation abstinence time and percentage of current smokers did not differ between the studies. The two populations were quite similar, but a few variables differed. Murcia study subjects presented relatively larger testicular volume, reported less cryptorchidism, more previously diagnosed varicocoele and lower medication intake (p values <0.05). Medication intake was mainly antibiotics or medication for allergies in both populations.
Table 1. Descriptive characteristics of the participants and the physical examination for Murcia and Almeria study populations (Fernandez et al., 2012)
SD: standard deviation; (5–95): 5th–95th percentile.
Age calculated as difference between day of attendance in study and self-reported day of birth.
Ejaculation abstinence period calculated as difference between time of current ejaculation and self-reported time of previous ejaculation.
Size assessed by palpation.
Question was ‘How would you describe your own health?’
Cryptorchidism [not born with both testicles in scrotum (irrespective of spontaneous descend, treatment or still cryptorchid)]
Questions were ‘Have you ever had any of the following?’: Long-lasting disease (including diabetes/thyroid disease), sexually transmitted diseases (diagnosed with epididymitis, chlamydia or gonorrhoea) or inguinal hernia. Self-reported answers.
Taken any medication during 3 months prior to participation in study. Twenty per cent of the men in Murcia and 29% in Almeria had taken either antibiotics or medication against allergy.
Between-group differences for continuous and categorical variables were tested using Student's t-test or the Pearson chi-squared test respectively.
Table 2 presents the semen variables of the two studies. Sperm concentration and total count were significantly lower in Murcia study participants compared with the men studied in Almeria. Those significant differences remained after controlling for covariates. Besides, those analyses were replicated in a subset of men in both populations who had not taken any medicine and were without any previous or current andrological diseases and we obtained very similar results. Other semen variables did not differ significantly between the two studies. Sperm concentration and total sperm count declined significantly with year of birth in the pooled analysis (β = −0.04 and β = −0.06 respectively; both p values <0.01) (Fig. 1 A,B). No temporal trend was seen by year of birth for sperm motility or morphology (data not shown). In addition, we reran our statistical models excluding 14 and 6 outliers (sperm concentration >2 SD of the mean value) for the Almeria and Murcia study populations, respectively, and observed similar results (data available on request). Among the subjects with ejaculation abstinence time above 48 h, 40 and 45% of the Almeria and Murcia study populations, respectively, had a sperm concentration below 40 million/mL.
Table 2. Descriptive characteristics of the participants’ semen variables for Murcia and Almeria study populations (Fernandez et al., 2012)
SD: standard deviation; (5–95): 5th–95th percentile.
Between-group differences were tested using Student's t-test and ancova.
Semen volume (mL)
Sperm concentration (million/mL)
Total sperm count (million)
% Motile spermatozoa (A + B + C)
% Normal morphology
We calculated the expected change in total sperm count (95% CI) for a typical study participant at the beginning and end of the study period. For a man born in 1974, the expected total sperm count, using our final regression model, was 175 (141, 214) million. For a man born in 1993, the expected sperm count was 109 (90, 131) million. Therefore, the 20-year interval from the oldest to the youngest man (from 1974 to 1993) was associated with a decrease in total sperm count of 66 million (38%), or in average 1.9% (1.8, 2.0) per year.
Our results indicate that total sperm count and sperm concentrations may have declined in young Spanish men over the last decade. Several studies have investigated temporal trends in semen quality in Northern Europe, but the current study is the first to examine this question in Southern Europe. Therefore, our study expands the current data on temporal trends of semen quality in young men with unknown or untested fertility.
The interpretation of a temporal trend seems not to be biassed by comparison of two different study populations. The Murcia study subjects were similar to the Almeria study ones (Fernandez et al., 2012) in terms of age and general health conditions. Study methods with regard to target population, invitation procedures and economic rewards for participating in the study, physical examination and semen analysis procedures, including quality controls, were also very similar. Almeria and Murcia are adjacent provinces located in the Southern part of Spain, sharing similar historic, ethnic, lifestyles and climate conditions (Borrell et al., 2005). Although we cannot exclude that our results are biassed by regional differences between the two study sites, we regard this as less likely. Similarly, it is not possible to rule out that the differences we see are because of the data coming from two distinct studies. Even if demographics and methods are (to the extent possible) the same in both, this remains a possible explanation.
The changes we report between young men in the Almeria and Murcia studies may suggest a decline in sperm concentration, which would be consistent with a decline in the magnitude reported by Carlsen et al. (1992). In fact, the yearly rate of sperm decline calculated from these two studies is 1.9%, close to that estimated by Swan et al. (2000) for Europe (−2.3%). Moreover, it is noteworthy mentioning that the very same rate of sperm concentration decline (1.9%) has recently been reported by Rolland et al. (2012) in a large sample of French men between 1989 and 2005. It should also be mentioned that even though semen variables are within the normal ranges according to the World Health Organization, 2010 criteria, nonetheless between 40 and 45% of the study populations had a sperm concentration below 40 million/mL, an accepted cut-off for subfecundity (Bonde et al., 1998; Guzick et al., 2001; Slama et al., 2002).
Recent studies have seen no significant secular trend in sperm counts in young Swedish men (Axelsson et al., 2011), a slight increase in Danish (Jørgensen et al., 2012) and decrease in semen quality among Finns (Jørgensen et al., 2011). The latter, observed median sperm concentrations (5th–95th) were 60 (3–185) million/mL, 54 (5–167) and 50 (1–141) for birth cohorts between 1979 and 1981, 1982 and 1983 and 1987 respectively (Jørgensen et al., 2011). On the other hand, neither statistical differences nor temporal trend was observed in our pooled study with regard to sperm motility and morphology, similar to other results recently reported on young men (Jørgensen et al., 2012).
It remains unclear why the semen quality may have declined for young Spanish men. We still cannot resolve if the reasons for the downwards trend might be exposure to environmental or occupational toxins, or differences in lifestyles of the individuals. As Fernandez et al. (2012) pointed out: ‘The Southern Spain has experienced an increasing industrialization and modernization in agriculture practices in the recent years and with this, an increased risk of adverse exposures’. The decreasing trend in sperm counts suggests that the testicular function of young men may have become affected. We can speculate that both environmental exposures and lifestyles were different when Almeria and Murcia young men mother's became pregnant, not currently, at the time of recruitment. Epidemiological studies have shown associations between increasing risk of testicular cancer and decreasing semen quality at a population basis (Jørgensen et al., 2002, 2011; Punab et al., 2002; Richthoff et al., 2002; Huyghe et al., 2007; Paasch et al., 2008; Chia et al., 2010). In general, Spanish men have a low risk of testicular cancer which, however, seems to be increasing (Llanes González et al., 2008). Thus, the apparent association between semen quality and testicular risk in a population also seems to be true for the Spanish population. These results would also be in agreement with the hypothesis based on the TDS concept, in which reduced semen quality and testicular cancer would be two of the clinical manifestations of TDS (Skakkebæk et al., 2001).
Several studies report that prenatal tobacco exposure is associated with lower total sperm counts (Jensen et al., 2004; Ravnborg et al., 2011; Virtanen et al., 2012). Our results did not show a relationship between mothers smoking during pregnancy and sperm counts in their sons (results not shown). However, only 18% of the Murcia and 6% of Almeria study subjects were exposed to tobacco in utero, limiting our power to detect an effect of prenatal exposure to tobacco. Nevertheless, this low rate of exposure suggests that prenatal tobacco exposure is unlikely to explain the decline we report.
Environmental exposure to lead and cadmium has been associated with altered sperm quality in men attending infertility clinics in Southern Spain (Mendiola et al., 2011). However, the levels of exposure were relatively low and unlikely to have increased appreciably during the last 10 years. Environmental exposures to endocrine disruptors, such as phthalate esters and bisphenol A have also been associated with impaired reproductive male function (Duty et al., 2003; Hauser et al., 2006; Meeker et al., 2010; Mendiola et al., 2010, 2012). Although our study populations were exposed to these non-persistent compounds on a continuous basis (Koch & Calafat, 2009), we have little evidence that these substantially increased over the study period.
Our study has several potential weaknesses. There may have been differential participation between the two populations, which could produce selection bias. However, it is unlikely that participation rates were related to fertility status or semen quality as both were probably unknown to these young men. Our subjects were student volunteers who were compensated for their participation, which was not advertised as a fertility study. We cannot exclude the possibility that our findings reflect geographic rather than temporal differences in sperm concentrations. However, as stated earlier, the two study areas are in adjacent provinces in Southern Spain sharing a similar global environment. There is a significantly higher prevalence of cryptorchidism in the Almeria study. However, we cannot rule out that the self-reported responses to that question may alter those results. Nevertheless, sperm concentration and TSC were not statistically different between men who did and did not report cryptorchidism in our complete study population. Lastly, we saw the very same tendencies when we analysed the subpopulations without any important events in their medical history, including cryptorchidism. Therefore, this unexplained difference in rates of cryptorchidism does not substantially change our findings.
Quality control procedures were carried out for sperm concentration, thus inter-laboratory variation is unlikely to explain our findings. However, we were unable to carry out a quality control study of the assessment of the sperm motility and morphology, consequently those results should be interpreted with this reservation. Testicular volume appears to be different between the two studies and was measured with a Prader orchidometer. Nonetheless, the medians were from a clinical point of view very close, and the biological interpretation would be the same. We could not determine whether this is a result of the relatively coarse measurements available with the orchidometer and therefore associated with inter-examiner variation. The participants completed their questionnaires during the clinical appointment so that some of them may have been answered what they thought was correct, and therefore did not ask their mothers, where the actual answer might have been different if the mothers had answered.
In conclusion, our study suggests that there may be an adverse temporal trend in sperm counts and concentrations among young Spanish men during the last decade. Southern Spain has recently gone through a growing innovation and industrialization in many areas, and with this, an increased risk of potential adverse exposures, which might affect reproductive variables in men.
The authors gratefully acknowledge Dr. M. Roca, C. Ruiz, E. Belmonte, F. Mas and all the USP Dexeus Murcia clinic staff for their assistance in data collection; and the young men of the study for their participation. We also thank E. Estrella for database management. This work was supported by ‘Fundación Séneca, Región de Murcia, Agencia Regional de Ciencia y Tecnología, grant no: 08808/PI/08’, ‘Ministerio de Ciencia e Innovación, Instituto de Salud Carlos III (FIS) grant no: PI10/00985’ and Rigshospitalet, Copenhagen (Denmark) grant no: 961506336 to NJ.
Authors' contribution section
A.M.T.C., N.O., M.F.F, N.J. and S.H.S. were involved in study conception and design. L.M.A, G.V.S., J.M., M.F.F., K.J.R.R and L.S.C were involved in study execution and acquisition of data. L.M.A, N.J., S.H.S., J.J.L.E. and J.M contributed to data analysis and interpretation. L.M.A, J.M., J.J.L.E., A.M.T.C. drafted the manuscript. All authors provided substantial intellectual contributions and approved the final version of the manuscript.
Conflict of interest
The authors have no competing interests to declare.