Infertility, blood mercury concentrations and dietary seafood consumption: a case–control study
* Dr C. M. Y. Choy, Department of Obstetrics and Gynaecology, Prince of Wales Hospital, Shatin, N.T., Hong Kong, SAR, China.
Objective To compare blood mercury concentrations of infertile couples with those of fertile couples in Hong Kong, and to examine the relationship between blood mercury concentrations and seafood consumption.
Design Case–control study.
Setting In vitro fertilisation (IVF) Unit and Antenatal Unit of a university teaching hospital.
Sample One hundred fifty-seven infertile couples attending IVF treatment and 26 fertile couples attending antenatal care without known occupational exposure to mercury.
Methods Mercury concentrations in whole blood were measured by cold vapour atomic absorption spectrophotometry. A dietitian recorded the quantity of seafood consumption among infertile couples via a food-frequency questionnaire. Blood mercury concentrations and quantity of seafood consumption were compared between infertile and fertile couples.
Main outcome measures Whole blood mercury concentrations, quantity of seafood consumption.
Results Infertile couples had higher blood mercury concentrations than fertile couples. ‘Infertile males with abnormal semen’ and ‘infertile females with unexplained infertility’ also had higher blood mercury concentrations than their fertile counterparts. Blood mercury concentrations were positively correlated with quantity of seafood consumption. Infertile subjects with elevated blood mercury concentrations consumed a larger amount of seafood.
Conclusion Higher blood mercury concentration is associated with male and female infertility. Higher seafood consumption is associated with elevated blood mercury concentrations in our infertile population.
Infertility is a common clinical problem and affects at least 10–15% of couples worldwide1. Exposure to environmental toxins has been implicated as a potential cause of infertility2,3, and mercury is one of these potential reproductive toxins. Animal studies have demonstrated an adverse effect of mercury on spermatogenesis4,5. However, data on the relationship between mercury and male infertility in humans have been contradictory. In vitro studies have shown that mercury is capable of inducing sperm abnormality6. Other in vivo studies failed to correlate mercury concentrations in blood or seminal plasma with abnormal semen parameters, but suffered from the limitations of a small sample size7, a narrow range of semen parameters7 and a narrow range of mercury concentrations8. In an earlier pilot study involving 59 male partners of infertile couples in Hong Kong, we found that 35% had abnormally high blood mercury concentrations, whereas blood concentrations of lead, aluminium, selenium and cadmium were normal, suggesting that this population may be at risk of mercury toxicity9.
In females, studies on the effect of mercury on fertility have been scarce and inconclusive. Studies in rodents showed that elevated blood mercury concentrations were associated with reduced female fertility10. However, human occupational studies of female dental assistants revealed a confusing U-shaped dose response of fecundability to mercury vapour exposure11. We could not identify any study on the effect of mercury exposure via non-occupational routes on female fertility.
Among different environmental sources of mercury, the aquatic food chain is an important site of bioaccumulation. However, the extent to which seafood intake contributes to bodily accumulation of mercury varies between countries, from 20% in Belgium to 85% in the USA12. In Hong Kong, the frequency of fishmeals has been shown to be associated with hair mercury concentrations13, suggesting that our fish-eating population may be particularly at risk of mercury toxicity.
The aim of this study was to compare blood mercury concentrations of infertile couples with those of fertile couples. The study also aimed to quantify seafood consumption and examine its relationship with blood mercury concentrations among infertile couples. Blood was used as the biomarker for mercury exposure, as elevated blood mercury concentrations have been shown to correlate with the severity of clinical mercury toxicity14. In addition, mercury concentrations in blood, rather than in hair or urine, reflect exposure to both methylmercury and inorganic mercury15. Inorganic and organic mercury were not measured separately due to their overlapping spectra of reproductive toxicity as shown in human in vitro studies6.
A case–control study was designed. The infertile group was comprised of male and female partners of married couples undergoing in vitro fertilisation (IVF) treatment at the IVF Unit of the Prince of Wales Hospital, the Chinese University of Hong Kong. The characteristics of infertility were recorded. Within this infertile group, an ‘abnormal semen group’ was defined for those males who had two consecutive abnormal semen samples according to the guidelines of World Health Organization on sperm concentration, motility and morphology16, more than three months apart. A female subgroup of ‘unexplained infertility’ was also defined for those females from the infertile group, whose husbands had normal semen analysis and who had not been found to suffer from anovulation, endometriosis, tubal malfunction or peritoneal adhesions.
For the control group, the subjects were fertile married couples, where the wife was currently pregnant without a history of infertility. The couples were recruited during their early second trimester antenatal care at the Prince of Wales Hospital. The husbands of the control group confirmed fatherhood verbally, although they did not verify their fertility status by giving a semen sample. The difficulty in obtaining semen samples from volunteers in Asia has been well reported in the literature, and is due to cultural beliefs within its population7.
All subjects were recruited at the Prince of Wales Hospital over a six-month period between year 2000 and 2001 after informed consent. Over 90% of the study subjects lived in the same catchment area of the hospital, which served the Northeastern New Territories of Hong Kong. Demographic characteristics of the subjects including age, occupation and smoking habits were compared between the groups. Individuals with known occupational exposure to mercury were excluded as they might be exposed to much higher levels of inorganic mercury than the average population. Infertile males with known aetiology for semen abnormality such as chromosomal abnormality, testicular malformation or surgery were excluded from the study. The Clinical Ethics Committee of the Chinese University of Hong Kong approved this study.
Mercury concentrations were measured in whole blood from each study subject by cold vapour atomic absorption spectrophotometry (Flow Injection Mercury System, Perkin Elmer, Norwalk, Connecticut, USA) at the Department of Chemical Pathology, Prince of Wales Hospital, Hong Kong. All samples were analysed in duplicate. The detection limit of the measurement was 0.25 nmol/L, and the coefficient of variation was 5.7% at 70 nmol/L. A blood mercury concentration exceeding 50 nmol/L was regarded as abnormal17.
A dietitian recorded the quantity of seafood consumption among infertile subjects in a modified food-frequency questionnaire, via a personal interview. The questionnaire recorded the typical consumption quantity and frequency of 50 food items with the aid of visual food models, per day, per week and per month, with an emphasis on fish and shellfish intake. This type of food-frequency questionnaire has been used previously in assessing population exposure to methylmercury18. Other sources of exposure to mercury including the intake of fish pill supplements or Chinese herbal medicine, and the use of dental amalgam or skin-lightening cosmetic creams, were also recorded.
Data were analysed using the Statistical Package for Social Sciences (SPSS, Chicago, Illinois, USA). Blood mercury concentrations were log-transformed and presented as geometric means and geometric standard deviations. Student's t tests were used to compare log-transformed blood mercury concentrations of the infertile group with the control group. Logistic regression analysis was used to investigate the effect of age and blood mercury concentrations on fertility (Dependent variable = fertile or not). Spearman's rank correlation test was used to determine the relationship between blood mercury concentrations and the quantity of seafood consumption. χ2 tests were used to analyse categorical variables where appropriate. The level of significance was set at P < 0.05.
A total of 183 couples were recruited into our study, of which 157 couples were in the infertile group and 26 couples were in the control group. Two of the females and seven of the males in the infertile group failed to have their blood taken. The final analysis was performed on 150 males and 155 females in the infertile group, compared with 26 males and 26 females in the control group. Among the infertile group, 95 couples suffered from primary infertility; 30 couples had no known cause for their infertility and 40 males had abnormal semen. The average duration of infertility was 5.7 years.
The participation rates were 90% in the infertile group and 20% in the control group. The low participation rate of the control group was due to their reluctance in participating in research studies, when they were healthy and did not require any routine blood taking. We anticipated this difficulty prior to the study and therefore did not involve the control group when investigating the source of mercury exposure; otherwise the participation rate may have been even lower. The difference in participation rate may have been a potential source of bias. However, the authors believe that because the two groups were living in the same residential area and their occupations as well as smoking habits were similar (Table 1), our findings were not biased by this difference. The difference in age between the two groups (Table 1) is intrinsic in the problem of infertility, as IVF treatment is often the final resort for these infertile patients. In order to control this possible confounding factor, logistic regression analysis was performed to control for the effect of age on the difference between blood mercury concentrations of the infertile group and the control group.
Table 1. Demographic characteristics of the infertile group and the control group. Values are given as mean [SD] or n (%).
|Age (years)||38.0 [4.2]||34.3 [5.8]||<0.01||34.8 [3.3]||31.0 [5.1]||<0.01|
|Unemployed||4 (2.7)||0 (0)||NS||51 (32.9)||12 (46.2)||NS|
| Industrial worker||5 (3.3)||1 (3.8)||NS||1 (0.6)||1 (3.8)||NS|
| Labourer||45 (30)||14 (53.8)||NS||4 (2.6)||0 (0)||NS|
| Office worker||96 (64)||11 (42.3)||NS||99 (63.9)||13 (50.0)||NS|
|Smokers||49 (32.7)||9 (34.6)||NS||9 (5.8)||0 (0)||NS|
Overall, the infertile group had significantly higher blood mercury concentrations than the control group (Table 2). After controlling the difference in age between the two groups by logistic regression analysis, blood mercury concentration remains predictive of the status of fertility (P < 0.01), implying that the relationship between blood mercury concentrations and infertility was independent of age. In the subgroup analysis of ‘infertile males with abnormal semen’ and ‘infertile females with unexplained infertility’, infertile subjects also had significantly higher blood mercury concentrations than their fertile counterparts (Table 2). A higher proportion of subjects in the infertile group had abnormally elevated blood mercury concentrations, although this difference reached statistical significance in the female group only (Table 2).
Table 2. Blood mercury concentrations and number of subjects with abnormally high blood mercury concentrations among the infertile group and the control group. Values are given as geometric mean [SD] or n (%).
|Males (n= 176)|
|Blood Hg (mmol/L)||40.6 [1.7]1||44.2 [1.7]2||31.2 [1.1]|
|Subjects with elevated blood Hg||53 (35.3)3||15 (37.5)3||6 (15.0)|
| ||Infertile group (n= 155)||Unexplained infertility group (n= 30)||Control group (n= 26)|
|Females (n= 181)|
|Blood Hg (mmol/L)||33.2 [1.7]4||37.0 [1.1]4||17.5 [2.1]|
|Subjects with elevated blood Hg||36 (33.2)2||10 (33.3)5||1 (3.8)|
Blood mercury concentrations were positively correlated with the quantity of seafood consumption (r= 0.21, P < 0.001). When comparing infertile subjects having abnormally high blood mercury concentrations with those having normal blood mercury concentrations, seafood consumption was the only different source of exposure to mercury among various other sources (Table 3).
Table 3. Environmental sources of exposure to mercury (Hg) among infertile subjects having elevated blood mercury concentrations and among those having normal blood mercury concentrations. Values are given as mean [SD] or n (%).
|Seafood consumption (g/day)||108.1 [59.5]||90.3 [57.4]||0.02*|
|Taking fish pill supplement||4 (4.5)||15 (6.9)||NS|
|Taking Chinese herbal medicine||34 (38.2)||108 (50.0)||NS|
|Having dental amalgams||60 (67.4)||160 (74.1)||NS|
|Using skin-lightening cream||20 (22.4)||66 (30.6)||NS|
|Unemployed||15 (16.9)||52 (24.1)||NS|
|Industrial worker||2 (2.2)||6 (2.8)||NS|
|Labourer||18 (20.2)||45 (20.8)||NS|
|Office worker||61 (68.5)||158 (73.1)||NS|
Our finding that a substantial proportion (35%) of males from the infertile group had abnormally high blood mercury concentrations is consistent with that from our previous pilot study9. When male infertility was defined specifically by abnormal semen analysis, the association between blood mercury concentrations and infertility remained significant. We have attempted to overcome the limitations of previous studies by having a larger sample size and a more sensitive method of assaying mercury. The association between in vivo exposure to mercury and male infertility found in our study is biologically plausible to be a causal relationship. It has been shown that membranes of acrosomal cap, the midpiece and the tail of the human sperm are potential binding sites for mercury19. Subsequently, disruptions of sperm membrane permeability, mitochondrial function, DNA synthesis by the microtubules and motion generation by the microtubule sliding assembly are possible mechanisms of mercury toxicity19–21. Other than the sperm themselves, supporting cells in the testis22, in the epididymis23 and in the seminal vesicles24 are also possible targets of mercury toxicity. The final result may then be semen abnormality and clinical infertility.
Our study showed that 23% of the females in the infertile group had abnormally high blood mercury concentrations. This was significantly more common than that in the control group, and the relationship remained significant when comparing the subgroup of ‘females with unexplained infertility’ with the control group. The question of whether chronic mercury exposure is an aetiology of female infertility in this subgroup of patients remains to be investigated. Menstrual irregularity has been reported to be more common when women had occupational exposure to mercury25. However, less than 2% of our study population had anovulation, implying that anovulation could not explain the association between blood mercury concentrations and female infertility in our study. It has been shown that mercury destructs ova chromosomes of rodents in vitro26, an effect similar to that in the male epididymis. Thus, it is possible that mercury may induce cytotoxic or genotoxic injury in the ovaries leading to poor oocyte quality, a mechanism mirroring that of mercury toxicity in the male reproductive system. Further investigations are required to test this hypothesis.
The positive correlation between quantity of seafood consumption and blood mercury concentrations suggests that higher seafood consumption may contribute to higher blood mercury concentrations. This is demonstrated among our infertile subjects with abnormally elevated blood mercury concentrations, who consumed a larger amount of seafood. Seafood contaminated with mercury is a possible source of excessive mercury exposure in our infertile population. Contamination of the marine waters around Hong Kong with heavy metals is common27. A recent testing in Thailand on 10 samples of shark's fin, some of which were exported from Hong Kong to Thailand, revealed that six out of the 10 samples had mercury concentrations exceeding the upper safety limit for human consumption, as recommended by the World Health Organization28. Three of these samples exceeded the limit by 10 times. Other popular dietary species such as tuna and swordfish are predatory fish and may also accumulate high concentrations of mercury. Reduction in dietary consumption of seafood is a measure that may be effective in controlling the accumulation of mercury. However, this should be balanced against the beneficial effects of other components of fish, such as those of 3-omega fatty acids and selenium.
Higher blood mercury concentration is associated with male and female infertility, and higher seafood consumption is associated with elevated blood mercury concentrations in our infertile population. Further studies should be performed to investigate the mechanism of mercury toxicity on the male and female reproductive system, as well as to define the safety of seafood consumption in Hong Kong.