The effect of cigarette smoking, alcohol consumption and fruit and vegetable consumption on IVF outcomes: a review and presentation of original data

Since cigarette smoking appears to have a detrimental effect on fertility, it is reasonable to assume that cigarette smoking may affect the risk or requirement for IVF. Furthermore, it is entirely possible that female smoking during IVF treatment may adversely affect the subsequent IVF outcomes. Cigarette smoking is the most sufficiently researched lifestyle factor that could affect IVF outcomes, compared to nutrient intake and alcohol consumption. However, the literature is still very limited. Original unpublished data from our laboratory examined the effects of cigarette smoking on oocyte production, fertilization rates, pregnancy and pregnancy loss rates. Lifestyle data collected from these patients is presented in Table 1. The mean number of oocytes retrieved from women undergoing IVF treatment did not significantly differ between regular smokers (11.1 oocytes retrieved, SD 6.5), ex-smokers (11.8, SD 10.1) or non-smokers (11.2, SD 7.6) which indicated that smoking did not influence oocyte production.

As a mechanism to determine the influence of smoking status on ovarian reserve and as a consequence of the age of the presented data, we utilised basal FSH level as an accepted marker of ovarian reserve rather than AMH in smoking and non-smoking women [18]. It was found that oocyte production lowered along with increased basal FSH levels (p < 0.001), and basal FSH appeared to increase in tandem with years of cigarette smoking (β = 0.007, p = 0.035), even after adjusting for the effects of age and infertility status. These results suggested a negative relationship was apparent between years of cigarette smoking and oocyte production, although this was not represented in the mean number of oocytes collected between the smoking status groups. However, we speculate that this observation could be related to the rFSH dosage administered during stimulation before oocyte retrieval, which may compensate for any negative effects of cigarette smoke on oocyte production and ovarian function. Interestingly, the results from our laboratory also showed that fertilization rates were not influenced by current female smoking status (Table 2). However, fertilization rates did decrease as female years of smoking increased (p < 0.001), and there is an observable and significant interaction between fertilization rates and nicotine (Table 2).

In relation to male patients and applying multiple logistic regression analysis, data from our laboratory showed that male nicotine intake was significantly and positively associated with a first-trimester (up to 12-week) non-ectopic pregnancy loss (p = 0.008) (Table 3). The risk of pregnancy loss increased by 2.9 % with every 1 mg of nicotine intake in males and after adjusting for female age, the odds ratio (OR) of a first-trimester pregnancy loss for male smokers was 2.2 (95 % CI 1.1–4.3) when compared to non-smokers (Table 3). We speculate that this association between male smoking and pregnancy loss is likely due to the result of spermatozoal DNA damage that occurred pre-conception. However, this was contradictory to female smoking status, where there was no significant effect on pregnancy loss (Table 3).

Our original research has demonstrated that cigarette smoking influenced several aspects of fertility and female smoking was associated with increased basal FSH and decreased oocyte fertilization, while male smoking was positively correlated with first-trimester pregnancy loss. A range of other studies have been performed that corroborate some of our findings, while others illustrate the complex interplay between lifestyle factors and positive ART outcomes.

It has been demonstrated that actively smoking women had significantly lower serum AMH, when compared to non-smoking women undergoing to IVF, which indicated a negative impact on ovarian reserve [14]. In addition, it was reported that smoking couples had a reduction in the number of oocytes retrieved (by 40 %), while those with smoking male partners were reduced by a further 5 % [19]. However, other researchers have found that the mean number of oocytes retrieved for smokers (9.6 oocytes) and non-smokers (9.0 oocytes) was not significantly different [20]. Taken together, these outcomes echo those presented in our study above where female smoking was associated with increased basal FSH, but this did not manifest in significant differences in the number of oocytes collected or pregnancy loss rates. Consequently, this observation may be related to rFSH stimulation regimens, where some clinics apply a standard dosage based on female age, while others, including PIVET, applied a step-up dosage for those women not responding optimally [21].

Interestingly, a study conducted in Dutch IVF clinics examined the effect of smoking on the IVF success rate in 8457 women [20], and they showed that smoking decreased the live birth rate by 7.3 %. Furthermore, the spontaneous abortion rate per pregnancy was higher for smokers at 21.4 % compared with 16.4 % for non-smokers [20]. We observed that female smoking years and nicotine intake was not associated with first-trimester pregnancy loss, but the difference between the studies may be first explained in a practical sense by the larger sample size in the Dutch study [20]. Second, there is also the clear possibility that smoking women in our smaller cohort, who have become pregnant following IVF, would alter and probably cease their smoking habits upon being informed of successful pregnancy. It is therefore not astonishing that female smoking status does not have a significant impact on most IVF outcomes in our study as it is likely they will modify their behaviour. However, in other work, the risk ratio (RR) for smoking couples failing to become pregnant was doubled when compared to non-smoking couples [19]. Moreover, the RR increased by 4 % for each additional year of smoking. In addition, the authors reported that live birth rates were significantly decreased for smoking couples in comparison to non-smoking couples [19]. They also demonstrated that if a woman had smoked during her lifetime, she was twice as likely to fail becoming pregnant compared to non-smoking women. This risk eventually increased by 9 % with each additional year of smoking.

A meta-analysis conducted in 2008 included 21 studies that focused on reproductive outcomes and lifestyle factors [22]. They found that cigarette smokers had a significantly lower pregnancy and live birth rates per cycle in comparison with non-smokers. In the same meta-analysis, smoking was also significantly associated with an increased risk of spontaneous miscarriage. However, there was no association between smoking and fertilization rates in 17 studies that included fertilization rate parameters [22]. Conversely, Zitzmann et al. [23] reported that fertilization rates in IVF patients were influenced by female smoking. Importantly, in addition to active smoking status, an association between passive smoking and IVF outcome has also been shown to impair fertility parameters [24]. In one study, patients who were exposed to side stream smoke had lower pregnancy rates that were equal to mainstream smokers [24]. Since other studies have shown that increased levels of reactive oxygen species (ROS) in follicular fluid were correlated with poor oocyte quality [25], there remains the possibility that the decrease in oocyte production may be due to the increased levels of ROS promoted by cigarette smoking over a long period of time. The involvement of ROS in male and female fertility is described in more detail later in the present manuscript. However, despite several laudable studies investigating the effect of female smoking on IVF outcomes, the research is still very limited and much more work is required.

Comparably, the effect of male smoking alone on IVF outcomes is another area that is lacking. It has been indicated that for couples undergoing IVF (n = 148) and ICSI (n = 153), a woman with a smoking male partner had lower ICSI success rates (22 %), compared to 8 % for women with non-smoking partners [23]. Furthermore, the authors found similar results for IVF, where IVF success rate was 18 % for women with smoking partners and 32 % with non-smoking partners in 213 treatment cycles. Recent reports have suggested that smoking negatively affects semen quality parameters including morphology [26, 27] and motility [27], yet the biological mechanism remains unclear. However, one of the main constituents of cigarette smoke is the highly mutagenic and carcinogenic compound benzo[a]pyrene [28]. Research has found that sperm exposed to benzo[a]pyrene display a moderately increased amount of DNA fragmentation, along with increased benzo[a]pyrene adducts on DNA in comparison with non-exposed sperm [29]. Sperm DNA damage is associated with an increased risk of pregnancy loss after IVF [30], therefore exposure to benzo[a]pyrene and the resulting DNA damage could be a potential mechanism leading to increased risk for pregnancy loss with male smoking partners. Taken together, all of these results, including the previously unpublished results from our laboratory, imply that for women undergoing IVF treatment with a partner who smokes cigarettes, there may be a negative impact on the subsequent IVF outcome. Consequently, couples undergoing IVF treatment should consider the cessation of smoking to enhance likelihood of a successful IVF outcome.

The effect of alcohol consumption on reproductive outcomes has been documented in several studies over recent years [37‐39]. Although most work suggest an adverse relationship between alcohol consumption and reproductive outcomes [38, 39], some studies were not able to identify a direct relationship [37]. Consequently, the true impact of alcohol on IVF outcomes is complex and not entirely clear. Considering that alcohol consumption is a common lifestyle factor among many individuals, the effect of alcohol consumption on IVF outcomes needs to be determined. The data obtained from our laboratory demonstrated no evidence of a significant relationship between female alcohol consumption and fertility parameters such as, oocyte production, fertilization rate, β-hCG pregnancy rate or pregnancy loss (Tables 2 & 3). The outcomes were the same even after adjusting for several variables like patient age, BMI and oocyte quality. This could be because females tended to decrease their alcohol consumption before undergoing IVF treatment in order to increase their chance of a pregnancy.

Conversely, we observed that male alcohol consumption had a positive effect on fertilization rate (β = 0.074, P < 0.001) (Table 2) and did not impact significantly on first-trimester pregnancy loss (Table 3). The cause of this positive effect on fertilization is difficult to determine. However, we speculate that consumption of fruits and vegetables, which is also included in our work, may improve rates of fertilization, and our results suggested that there was a significant interaction between these two variables (Table 2). Furthermore, this indicated that a common factor derived from both may promote the observed positive effects on fertilization.

From a nutritional perspective, one potential common compound that may mediate these responses is folic acid. Alcoholic beverages such as beer, has a high folate content, and studies have revealed that there is a positive correlation between beer intake and blood folate [40]. In addition, dark leafy green vegetables, such as avocados, broccoli and asparagus, also have high folate content. Interestingly, low folate concentration in seminal plasma has been correlated with increased sperm DNA damage [41] and folate is known to have antioxidant properties that may counteract negative effects of ROS (Fig. 1) [42], which may in turn be one of the possible reasons for the positive effects of folate-containing food and beverages. The impact of folate on IVF outcomes is currently under consideration by our group.

A recent large study in 2545 couples demonstrated that following adjustment for age, BMI and cigarette use, women who consumed four or more alcoholic drinks per week (n = 237) had a 16 % lower chance of live birth in comparison to women consuming less (n = 994) [38]. For couples where both members consumed more than 4 alcoholic drinks, there was a 21 % lower chance compared with couples who consumed less [38]. This was contradictory to our results, which showed no evidence of a significant relationship between female alcohol consumption and pregnancy or pregnancy loss, but may be explained by the difference in sample size and population (n = 2545 versus n = 68; Table 3). However, they also demonstrated that the overall effect of male alcohol consumption and pregnancy loss in their partners was not significant and this was similar to our findings [38]. On the other hand, associations were found between alcohol consumption and fertilization. In men and women consuming four or more alcoholic drinks per week, there was a 48 % decreased chance of fertilization [38], but this was not observed in our study (Table 2).

Given the controversial beneficial effects of male alcohol consumption on IVF outcomes shown here, it is not surprising that other studies have failed to reveal any beneficial effect, and this further complicates the outlook in relation to alcohol consumption and IVF. Klonoff-Cohen et al. [39], demonstrated that alcohol consumption was associated with a 13 % decrease in number of oocytes retrieved, even after adjusting for other confounding factors such as age, race, reason for infertility and smoking status in 221 couples with female infertility. Female consumption of alcohol one month and one week before the procedure date increased the risk of not becoming pregnant in comparison to controls (women who did not consume alcohol). Moreover, there was a trend towards an increased risk of miscarriage in women who consumed alcohol one week before the procedure. However, similar to our results, there was no significant association between female alcohol consumption and live birth rates [39]. However, in this study, women who drank 12 g more of alcohol per day during the first week of the attempt, this resulted in a 59 % increase in fertilization. Due to the fact that women undergoing IVF treatment modify their behaviour to increase their chance of IVF success, this could have potentially caused the non-significant effect of alcohol consumption seen in our results. However, in that study only 6 % of the women consumed alcohol during the treatment, while approximately half of the patients in our study consumed alcohol during treatment, although the average consumption rate was less than one standard drink per day. Therefore, women who are highly motivated to have a successful IVF outcome are more likely to drink less alcohol or stop drinking altogether. Also it is important to take into account that most of our data was obtained through patient questionnaires. Patients may over-report behaviours seen as socially desirable and may under-report behaviours they deem not to be.

Interestingly, these researchers reported that male consumption of alcohol did not affect sperm count, motility, morphology, fertilization rate or chance of pregnancy [39]. However, more recent studies suggested that sperm concentration, total count and progressive motility were reduced in alcohol consumers in both IVF and control male patients [26], while DNA damage was increased in sperm derived from infertile man who consumed alcohol [43]. Others have also indicated that alcohol also impacts on semen volume, along with levels of male sex steroids [44], which could impact on IVF outcomes. The impact on DNA damage is discussed in more detail and against the backdrop of reactive oxygen species (ROS) below.

Taken together, these data suggest a possible relationship between alcohol consumption and several IVF outcomes and semen parameters (Fig. 1). Interestingly, the studies described above along with our data constitute some of the only research material in this area, and consequently much more research is required to fully understand the complex relationship between alcohol consumption and IVF outcomes.

The main nutritional factor measured in the study from our laboratory was fruit and vegetable consumption. Intake of fruits and vegetables may be an overall indicator of diet quality, and a diet of high quality most likely provides essential nutrients. In our study, the only IVF outcome found to be significantly influenced by fruit and vegetable consumption was fertilization rate and this was seen in males only, and not in females (Table 2). Male fruit and vegetable consumption had a positive effect on fertilization rate (β = 0.034, p < 0.001) (Table 2). Unfortunately, six males never completed the F&V section of the survey even after several attempts to contact them. This accounted for 2 % of the male population. The general response rate of males for F&V was good (98 %), but we speculate that these six individuals may have been in the lower quartile of F&V consumers, and may have not reported because they considered the low intake of F&V to be not socially desirable. In this case, inclusion of their data may have strengthened the logistic regression analysis showing that F&V consumption had a positive effect on fertilisation rates. However, if the opposite was true and they were high F&V consumers with low fertilization rates, it is unlikely that their outcomes would influence the highly significant correlation observed in our analysis (β = 0.034, p < 0.001) (Table 2). Fruit and vegetables contain a high level of antioxidants [66], consequently, these antioxidants may have potentially reversed or protected sperm from DNA damage induced by ROS, which subsequently increased fertilization rates. On the other hand, our female population may not have had enough ROS-related disorders to show any significant beneficial effects from the antioxidants contained in fruit and vegetables.

Fruit and vegetable consumption had no significant influence on first-trimester pregnancy loss (Table 3). This was evident for both male and female fruit and vegetable consumption. Consequently, these data indicate that the possible benefits of dietary intake relate to fertility parameters upstream of implantation and may associate with better semen or oocyte quality.

Generally, research on the effect of nutrition in IVF outcomes, especially fruit and vegetable intake, is limited. This is an important area of research that needs to be expanded upon given the current trends of increased infertility, particularly against the backdrop of over-nutrition in the existing population. The few studies that have focused on the relationship between nutrition and IVF demonstrate varying results. The associations between serum polyunsaturated fatty acids (PUFAs) and embryo implantation rates in women undergoing IVF were examined recently [67]. PUFAs derived from the diet that were measured included linoleic acid (LA) and alpha-linolenic acid (ALA). Their results found that no specific PUFA was associated with chance of pregnancy, but interestingly the ratio of LA to ALA was significant with regard to pregnancy. When adjusted for confounding factors such as age, endometriosis, peak oestradiol level and male factor fertility, it was found that there was a dose dependent increase in chance of pregnancy with increased LA:ALA ratios, and this was also weakly correlated with embryo implantation rates [67]. Although this study does suggest that supplementation of PUFA’s could possibly improve fertilization rates, further studies are needed to determine if such supplementation could increase pregnancy in couples undergoing IVF treatment.

In addition, an investigation conducted by Vujkovic et al. in 2010 demonstrated the impact of diet on IVF outcomes in 160 couples undergoing IVF/ICSI treatment at a fertility clinic in the Netherlands. They identified two types of dietary patterns in women who participated in the study. A ‘health conscious-low processed’ dietary pattern, which consisted of a high intake of vegetables, fruit, grains whole-grains, and a low intake of snacks, meat and mayonnaise. High adherence to this diet was correlated with an increase in red blood cell folate [68]. The other dietary pattern was identified as the “Mediterranean” diet, which consisted of a high intake of vegetable oils, vegetables, legumes, fish, and a low intake of food snacks or fast foods. High adherence to this diet was correlated with increased folate, vitamin B6 in the blood and follicular fluid. In terms of IVF success, neither diet were associated with embryo quality or fertilization rate, but a high adherence to the Mediterranean diet was associated with a 40 % increase in the probability of pregnancy after IVF/ICSI [68]. Interestingly, these observations were adjusted for confounding factors and were not significantly affected by lifestyle factors of the males including age, alcohol consumption, smoking and BMI. The authors suggested that the reason for the increase in pregnancy chance when adhering to the ‘Mediterranean’ diet was due to the high intake of vegetable oils, which are high in linoleic acid. At the molecular level, linoleic acid is a precursor for prostaglandins, which are important in growth and development of ovarian follicles, the process of ovulation in addition to the development of appropriate endometrial receptivity. Therefore, these results suggested that an increased intake of linoleic acid could be beneficial for implantation of a fertilized embryo. One significant limitation of this latter study, was that the diet of the male partner was not included and as demonstrated in our own results, increased fruit and vegetable consumption in males caused an increase in fertilization rate, ultimately suggesting that the male diet is also an important factor to consider (Table 2).

Interestingly, a similar study from the same group was conducted in males from 161 sub-fertile couples in the Netherlands [69]. Here, they identified two patterns of diets—a “health conscious” diet containing high intake of fruit, vegetables, fish, seafood, legumes and whole-grains, and low intakes of fatty sauces, meat, sugar, confectionary and refined grains, and the “Traditional Dutch” diet, which contained high intakes of potatoes, meat, whole grains, fatty sauces such as mayonnaise, and a low intake of alcohol, cereals, fruit, sugar and soup. They found that the “health conscious” diet was inversely proportional to sperm DNA damage, which was explained by high intakes of fruit and vegetables when examined individually [69]. Furthermore, we speculate that this inverse relationship between high intakes of fruits and vegetables and sperm DNA damage might be due to high levels of antioxidants contained within fruits and vegetables.

Increased antioxidant intake has been associated with high sperm numbers and motility [64] and reduction of the incidence of sperm DNA fragmentation [66]. Participants consuming the “Traditional Dutch” diet, showed increased sperm concentration explained by the low intake of alcohol and high intake of potatoes [69]. These results could explain the relationship between unhealthy diets and decreased semen quality in Westernized countries. However, more studies are required to determine the specific cause of these associations.

Increased fruit and vegetable consumption was associated with increased fertilization rates in our study, suggesting a possible relationship between male diet and some IVF outcome parameters (Table 2). However, as we are one of the few groups to show this relationship, there is a requirement for further research to elucidate precisely how the diet of males affects IVF outcomes in their partners. These studies would help to determine the relationship between IVF success and male diet, which could possibly help increase the success rate of IVF if the male patient knows how to adjust their own diet to accommodate for this. Further research also needs to be undertaken to determine the molecular compounds causing the nutritional effects in IVF, and it is possible that a diet can be designed for both men and women, which will provide the greatest benefit for women undergoing IVF treatment.