Human Reproduction Update Advance Access published online on October 15, 2008
Human Reproduction Update, doi:10.1093/humupd/dmn046
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Effects of cigarette smoking upon clinical outcomes of assisted reproduction: a meta-analysis
1 University of Sheffield School of Medicine and Biomedical Sciences, Beech Hill Road, Sheffield S10 2RX, UK 2 Department of Reproductive Medicine, Sheffield Teaching Hospitals, The Jessop Wing, Tree Root Walk, Sheffield S10 2TJ, UK 3 Health Services Research Section, ScHARR, Regent Court, 30 Regent Street, Sheffield S1 4DA, UK 4 School of Health and Related Research, Regent Court, 30 Regent Street, Sheffield S1 4DA, UK 5 Academic Unit of Reproductive and Developmental Medicine, Jessop Wing Hospital, Tree Root Walk, Sheffield S10 2TJ, UK
To whom correspondence should be addressed at: 6 Correspondence address. E-mail: mda03aw{at}sheffield.ac.uk
 |
Abstract |
|---|
BACKGROUND: The aim of this meta-analysis was to investigate whether any difference exists in success rate of clinical outcomes of assisted reproductive technologies (ART) between women who actively smoke cigarettes at the time of treatment and those who do not.
METHODS: An intensive computerized search was conducted on published literature from eight databases, using search terms related to smoking, assisted reproduction and outcome measures. Eligible studies compared outcomes of ART between cigarette smoking patients and a control group of non-smoking patients and reported on live birth rate per cycle, clinical pregnancy rate per cycle, ectopic pregnancy rate per pregnancy or spontaneous miscarriage rate per pregnancy, and 21 studies were included in the meta-analyses. Pooled odds ratios (OR) and 95% confidence intervals (CI) were calculated for the data, and statistical heterogeneity was tested for using 
2 and I2 values. A systematic review examined the effect of smoking upon fertilization rates across 17 studies.
RESULTS: Smoking patients demonstrated significantly lower odds of live birth per cycle (OR 0.54, 95% CI 0.30–0.99), significantly lower odds of clinical pregnancy per cycle (OR 0.56, 95% CI 0.43–0.73), significantly higher odds of spontaneous miscarriage (OR 2.65, 95% CI 1.33–5.30) and significantly higher odds of ectopic pregnancy (OR 15.69, 95% CI 2.87–85.76). A systematic literature review revealed that fertilization rates were not significantly different between smoking and non-smoking groups in most studies.
CONCLUSIONS: This meta-analysis provides compelling evidence for a significant negative effect of cigarette smoking upon clinical outcomes of ART and should be presented to infertility patients who smoke cigarettes in order to optimize success rates.
Key words: assisted reproduction / cigarette / IVF / meta-analysis / smoking
|
Introduction |
|---|
Cigarette smoking is well documented to be a significant health risk and major cause of premature mortality, yet it remains a prevalent lifestyle habit in today's society.
Although the negative effects of cigarette smoking on general health are well known, the effects of smoking on assisted reproduction technologies (ART) are less well documented, and conflicting reports continue to be published regarding the effect. While many studies suggest that cigarette smoking has a negative influence upon outcomes of ART, the exact mechanism of this effect is poorly understood. A recent systematic review of the evidence revealed that smoking may affect rates of oocyte retrieval, IVF, embryo transfer, pregnancy and live birth delivery (Klonoff-Cohen, 2005
).
Contrary to these results, not all studies have demonstrated statistically significant negative effects of smoking on ART outcome (Trapp et al., 1986; Hughes et al., 1994; Sharara et al., 1994; Sterzik et al., 1996; Weigert et al., 1999; Wright et al., 2006).
Research has shown that providing ART patients with information detailing the negative impact of cigarette smoking upon their fertility can be highly effective in achieving smoking cessation (Hughes et al., 2000). Therefore, if a significant effect of smoking upon clinical outcomes of ART does in fact exist, clinicians can make a measurable impact upon ART success by sharing this data with their patients.
Two previous meta-analyses have investigated the effect of cigarette smoking on clinical pregnancy rates for IVF (Feichtinger et al., 1997; Augood et al., 1998). The relatively small number of studies pooled and the large body of evidence that has been published since this time means that there is a need for an up-to-date meta-analysis to be completed to provide clear evidence on any effects of cigarette smoking on ART.
The aim of this meta-analysis is to combine the results from the growing body of evidence regarding ART success rates in smoking and non-smoking patients.
|
Materials and Methods |
|---|
Criteria for including studies for this review
Types of studies
All human studies reporting outcomes of ART compared female cigarette smoking patients with a control group of non-smoking patients who report on a minimum of one the following outcomes:
- Live birth rate per cycle
- Clinical pregnancy rate per cycle
- Ectopic pregnancy rate per pregnancy
- Spontaneous miscarriage rate per pregnancy
- Fertilization rate per cycle
Due to the nature of the research question, we did not expect to retrieve any randomized control trials from the literature search, and accordingly only cohort and case–control studies were included in this review. Studies that only investigated the effect of passive cigarette smoking were excluded. For the present review, only published material was included to ensure quality of studies and results.
Types of intervention
ART included in the review were IVF, intracytoplasmic sperm injection (ICSI), gamete intrafallopian transfer (GIFT), and zygote intrafallopian transfer (ZIFT). Studies reporting results for spontaneous pregnancy or intra-uterine insemination were excluded.
Types of participant
Study inclusion was limited to those reporting outcomes of assisted reproduction in active cigarette smokers versus non-smoking patients at the time of treatment.
Where studies differentiated between non-smokers and ex-smokers, only the results for non-smokers were used. Although evidence has been published suggesting that the fertility of an ex-smoker resembles that of a non-smoker (Phipps et al., 1987), there is no standardized definition of an ‘ex-smoker’ and publications rarely stated their definition in terms of the time since stopping smoking. In an attempt to eliminate this likely cause of heterogeneity between studies, we did not include ex-smokers in the meta-analysis.
Where studies reported outcomes after oocyte donation, the smoking status of the recipient was used for analysis rather than that of the donor, as the outcomes measured were following successful embryo transfer.
Types of outcome measure
The outcome measures for this meta-analysis were:
- Live birth rate: defined as the birth of one or more infants that show signs of life.
- Clinical pregnancy rate: defined as the presence of a sonographically visible gestational sac in the uterus.
- Spontaneous miscarriage rate: defined as loss of pregnancy up to 20 weeks gestation.
- Ectopic pregnancy rate: defined as the presence of an extra-uterine pregnancy diagnosed by ultrasound or laparoscopy.
- Fertilization rate: defined as the proportion of oocytes fertilized.
Search strategy for identification of studies
An intensive computerized search was conducted on published literature from eight databases in May 2007. The databases searched (inclusive dates) were Medline (1950–2007), Embase (1980–2007), Cochrane library (1991–2007), PsycINFO (1806–2007), Science Citation Index (1900–2007), Social Science Citation Index (1900–2007), CINAHL (1982–2007) and Medline In-Process and Other Non-Indexed Citations. No date or language restrictions were used.
The search strategy used combinations of search terms related to smoking, assisted reproduction and outcome measures. Search terms used related to smoking were smoking, cigarette and tobacco. Terms relating to assisted reproduction were in vitro fertilization, IVF, assisted reproduction, assisted conception, artificial reproduction, fertility, infertility, subfertility, fecundity and subfecundity. Terms related to outcome measures were pregnancy rate, birth rate, ovarian reserve, ovulation induction, fertilization rate, conception rate, embryo quality, cycle cancellation rate, miscarriage rate and ectopic pregnancy rate.
The literature search (see Fig 1) retrieved 356 references for possible inclusion in the meta-analysis. After reading the title and abstract where necessary of these references, 312 articles were excluded for not addressing the primary research question. A further 11 publications were excluded on this basis after retrieving the full article for further study. Of the remaining 33 publications identified as addressing the primary research question, a further 12 studies were excluded: four studies were not published in English and therefore did not allow their methodological quality to be fully evaluated (Hruba et al., 1999; pa-Martynow et al., 2005, 2006; Triopon et al., 2006), three articles (Hughes et al., 1992; Van Voorhis et al., 1992; Crha et al., 2000) presented data from the same cohort of patients published in other included studies (Hughes et al., 1994; Van Voorhis et al., 1996; Crha et al., 2001) and two studies reported clinical pregnancy rates per embryo transfer and not per cycle ART (Neal et al., 2005; Motejlek et al., 2006). One study could not be included due to reporting results in the form of risk ratios without raw data that could be pooled by meta-analysis (Klonoff-Cohen et al., 2001), one analysed results with regard to follicular fluid cotinine values alone and not smoking status (Zenzes et al., 1997), and the final study analysed results with regard to male partner smoking status only (Zitzmann et al., 2003).
|
The remaining 21 studies met all inclusion criteria for this meta-analysis. Bibliographies of relevant studies were hand searched but did not identify any other potentially relevant studies. The characteristics of studies included in the meta-analysis are summarized in Table I.
|
Statistical analysis
Dichotomous data were extracted from the individual studies. Revman software [Review Manager (RevMan), Version 4.2 for Windows, Copenhagen: the Nordic Cochrane Centre, The Cochrane Collaboration, 2003] was used to express these results as combined odds ratio (OR) with 95% confidence intervals (CI), and to combine these results for meta-analysis where primary outcomes were reported by two or more studies. Statistical heterogeneity between studies was identified by a chi-squared value greater than its degree-of-freedom, or an I2 value >25%. Where this was the case, a random effects model was used for meta-analysis. A fixed effect method for meta-analysis was used to evaluate risk of ectopic pregnancy: in this case there was no significant evidence of heterogeneity.
We initially combined all studies together to calculate pooled OR for the outcomes regardless of the method of ART, and then performed a subgroup analysis based on method of ART intervention. This allowed us to perform a subgroup analysis of outcomes from IVF treatment only, but we were unable to perform subgroup analyses for other ART methods due to insufficient numbers of studies reporting outcomes for these interventions.
A sensitivity analysis was performed adjusting for age as a potential confounding factor, and then repeated after excluding studies using oocyte donation cycles in order to assess heterogeneity between these and other patients.
|
Results |
|---|
Twenty-one studies were included in the meta-analysis (Trapp et al., 1986
; Harrison et al., 1990; Elenbogen et al., 1991; Pattinson et al., 1991; Agnani et al., 1994; Hughes et al., 1994; Sharara et al., 1994; Maximovich and Beyler, 1995; Gustafson et al., 1996; Sterzik et al., 1996; Van Voorhis et al., 1996; Feichtinger et al., 1997; El-Nemr et al., 1998; Weigert et al., 1999; Crha et al., 2001; Winter et al., 2002; Crha et al., 2003; Tiboni et al., 2004; Lintsen et al., 2005; Wright et al., 2006; Soares et al., 2007). Four of these studies reported data on live birth rates per cycle (Pattinson et al., 1991; Sharara et al., 1994; Gustafson et al., 1996; Lintsen et al., 2005). Eighteen studies reported data on clinical pregnancy rates per cycle (Trapp et al., 1986; Harrison et al., 1990; Elenbogen et al., 1991; Pattinson et al., 1991; Agnani et al., 1994; Hughes et al., 1994; Sharara et al., 1994; Gustafson et al., 1996; Sterzik et al., 1996; Van Voorhis et al., 1996; Feichtinger et al., 1997; El-Nemr et al., 1998; Weigert et al., 1999; Crha et al., 2001, 2003; Tiboni et al., 2004; Wright et al., 2006; Soares et al., 2007). Seven studies reported data on miscarriage rates per pregnancy (Harrison et al., 1990; Pattinson et al., 1991; Hughes et al., 1994; Maximovich and Beyler, 1995; Gustafson et al., 1996; Winter et al., 2002; Soares et al., 2007), and three studies examined the risk of ectopic pregnancy (Elenbogen et al., 1991; Agnani et al., 1994; Gustafson et al., 1996).
Fifteen studies reported data on the effect of female cigarette smoking on IVF alone, allowing for subgroup analysis (Trapp et al., 1986; Elenbogen et al., 1991; Pattinson et al., 1991; Agnani et al., 1994; Hughes et al., 1994; Sharara et al., 1994; Maximovich and Beyler, 1995; Gustafson et al., 1996; Sterzik et al., 1996; Feichtinger et al., 1997; El-Nemr et al., 1998; Weigert et al., 1999; Crha et al., 2001, 2003; Lintsen et al., 2005). Of the remaining studies, one reported on oocyte donation cycles (Soares et al., 2007), two combined data for IVF and ICSI (Neal et al., 2005; Wright et al., 2006), one combined data for IVF and GIFT (Harrison et al., 1990), one reported results for IVF, GIFT and ZIFT (Van Voorhis et al., 1996), one reported results after IVF, ICSI and GIFT (Winter et al., 2002) and the final study used ICSI and FIVET (Tiboni et al., 2004).
Eighteen of the 21 included studies had a primary aim of investigating the effect of cigarette smoking (with or without other lifestyle factors) on outcomes of ART (Trapp et al., 1986; Harrison et al., 1990; Elenbogen et al., 1991; Pattinson et al., 1991; Agnani et al., 1994; Hughes et al., 1994; Sharara et al., 1994; Maximovich and Beyler, 1995; Sterzik et al., 1996; Van Voorhis et al., 1996; Feichtinger et al., 1997; El-Nemr et al., 1998; Weigert et al., 1999; Crha et al., 2001; Winter et al., 2002; Lintsen et al., 2005; Wright et al., 2006; Soares et al., 2007). The remaining three studies reported the effect of cigarette smoking although this was not the primary aim: one study investigated ascorbic acid supplementation in ART (Crha et al., 2003), one investigated hyperandrogenism (Gustafson et al., 1996) and one investigated concentrations of fat soluble vitamins and micronutrients between smokers and non-smokers (Tiboni et al., 2004).
Of the 28 studies included in the systematic review and meta-analysis, 9 included information regarding male partner smoking status (Pattinson et al., 1991; Hughes et al., 1994; Joesbury et al., 1998; Klonoff-Cohen et al., 2001; Zitzmann et al., 2003; Neal et al., 2005; Soares et al., 2007; Wright et al., 2006; Zenzes et al., 1997).
Fertilization rates were reported in many studies as continuous data with mean percentage oocytes fertilized and standard deviation, omitting the total number of oocytes retrieved and the number of cycles undertaken for each group. We therefore consider it to be statistically naïve to attempt to combine these results between studies, and accordingly a meta-analysis for fertilization rates has not been attempted. Data for this outcome has been included in the systematic review.
A study with the highest quality for the purpose of this review would be of prospective design with sample sizes based upon a power calculation. There would be no significant difference in age between smoking and non-smoking groups, and patients above the age of 40 would not be included. Results would be reported for only the first cycle for each patient in order to eliminate the confounding effect of several attempts: continuation of IVF depends on predictors of success observed in the first cycle (Stolwijk et al., 1996). Studies would also report the smoking status of the male partner. Table II displays the quality of the included studies based upon these criteria.
|
Live birth rate per cycle
Data from four studies were combined in the meta-analysis to determine the pooled OR for live birth rate regardless of method of assisted reproduction. The four studies included in this meta-analysis reported data on 3252 cycles for smokers and 4213 cycles for non-smoking controls. The sample size varied across the trials from 100 to 6903 cycles. Smoking patients demonstrated a significantly decreased live birth rate per cycle (OR 0.54, 95% CI 0.30–0.99).
Clinical pregnancy rate per cycle
The eighteen studies included in this meta-analysis reported data on cycles for 1284 smokers and 3959 matched controls. The sample size varied across the trials from 41 to 834 cycles. The clinical pregnancy rate per cycle was significantly lower for smokers (OR 0.56, 95% CI 0.43–0.73) (Fig. 2).
|
Spontaneous miscarriage rate per clinical pregnancy
The seven studies included reported data on pregnancies of 211 smokers and 1688 matched controls. The sample size varied across the trials from 23 to 1196. A significant increase in the odds of miscarriage per pregnancy was observed in the group of smokers (OR 2.65, 95% CI 1.33–5.30) (Fig. 3).
|
Ectopic pregnancy rate per clinical pregnancy
The data from three studies provided no evidence for statistical heterogeneity and therefore a fixed effect model was used. The studies reported data on pregnancies for 10 smokers and 42 non-smoking controls. The ectopic pregnancy rate per clinical pregnancy was significantly higher in smokers (OR 15.69, 95% CI 2.87–85.76), although the large CI is a likely reflection of the small sample size.
|
Subgroup analysis |
|---|
Subgroup analysis was carried out for patients undergoing IVF treatment only.
The treatment intervention for all studies reporting live birth rates used IVF and therefore no subgroup analysis was required.
Clinical pregnancy rate per cycle
The subgroup analysis included data for 1042 smokers and 2037 non-smokers. The clinical pregnancy rate per cycle remained significantly lower for smokers (OR 0.57, 95% CI 0.42–0.77).
Spontaneous miscarriage rate per cycle
Data on cycles for 52 smokers and 158 non-smoking controls was pooled. There was no longer evidence of a significant difference in odds of miscarriage (OR 2.99, 95% CI 0.94–9.47).
Ectopic pregnancy rate per cycle
All studies reporting ectopic pregnancy rates per clinical pregnancy used conventional IVF so no subgroup analysis was required.
|
Sensitivity analysis: age |
|---|
A sensitivity analysis was performed based on the age of patients, and the meta-analysis repeated after excluding all studies where age was likely to be a confounding factor. Three studies were excluded because of a statistically significant difference in age between smokers and non-smokers (Van Voorhis et al., 1996
; Weigert et al., 1999; Wright et al., 2006). A further six studies were excluded for failing to state whether any significant difference in age between study groups existed (Trapp et al., 1986; Harrison et al., 1990; Agnani et al., 1994; Feichtinger et al., 1997; Tiboni et al., 2004; Lintsen et al., 2005). Of the remaining studies, a further three studies were excluded for including patients over the age of 40 (Maximovich and Beyler, 1995; Winter et al., 2002; Soares et al., 2007). The results for one study (El-Nemr et al., 1998) were stratified by age and could therefore be adjusted to include only those patients below 40 years of age. The studies excluded by the sensitivity analysis for age included all those that did not use conventional IVF treatment only. Therefore, no further subgroup analysis was required.
Three studies were combined for live birth rate per cycle for 203 smokers and 359 non-smokers. There was no longer any evidence for a difference in live birth rate between smokers and non-smokers (OR 0.41, 95% CI 0.14–1.14).
Clinical pregnancy rate per cycle
Data from nine studies was pooled, and reported on the clinical pregnancy rate in 621 cycles of smokers and 859 cycles of non-smokers. The clinical pregnancy rate per cycle remained significantly lower for smokers (OR 0.51, 95% CI 0.32–0.79).
Spontaneous miscarriage rate per clinical pregnancy
Three studies reported on miscarriage rates for pregnancies in a total of 37 smokers and 93 non-smokers. There was no longer any evidence for a difference in miscarriage rate per pregnancy (OR 1.88, 95% CI 0.55–6.37).
Ectopic pregnancy rate per pregnancy
Two studies were available to evaluate the odds of ectopic pregnancy per pregnancy after sensitivity analysis. These reported on pregnancies of six smokers and 22 non-smokers. The ectopic pregnancy rate remained significantly higher for smokers (OR 14.70, 95% CI 1.53–141.15).
Sensitivity analysis had differing effects on the evidence of statistical heterogeneity for each outcome of ART, as evidenced by chi-squared statistic, P-value and the I2 statistic. There was little change in evidence of statistical heterogeneity for live birth rate and clinical pregnancy rate. There was a substantial decrease in statistical heterogeneity for miscarriage rate, and a slight increase for ectopic pregnancy rate.
|
Sensitivity analysis: oocyte donation cycles |
|---|
 TOPAbstractIntroductionMaterials and MethodsResultsSubgroup analysisSensitivity analysis: age Sensitivity analysis: oocyte... Fertilization rateMale partner smoking statusDiscussionReferences |
|---|
A further sensitivity analysis was conducted to exclude the only study that reported on clinical pregnancy rate and spontaneous miscarriage rate for oocyte donation cycles (Soares et al., 2007
), in order to assess heterogeneity between these and other patients. Clinical pregnancy rate per cycle
Sensitivity analysis had a negligible effect upon the odds of achieving clinical pregnancy between smokers and non-smokers (OR 0.57, 95% CI 0.43–0.75). There was very little change in evidence of statistical heterogeneity for this measure.
Spontaneous miscarriage rate per clinical pregnancy
Five studies reported on miscarriage rates per pregnancy of 196 smokers and 1337 non-smokers. The odds of miscarriage increased by 0.4 after sensitivity analysis (OR 3.05, 95% CI 1.74–5.73), and there was a small decrease observed in the measures of statistical heterogeneity (2 = 7.17, df = 5, I2 = 30.3%).
|
Fertilization rate |
|---|
Data for fertilization rate has been included in the systematic review and can be seen in Table III. Of the 17 studies that reported fertilization rates per cycle ART, the majority failed to find any significant difference in fertilization rates between smoking and non-smoking groups (Trapp et al., 1986
; Pattinson et al., 1991; Hughes et al., 1994; Sharara et al., 1994; Sterzik et al., 1996; Van Voorhis et al., 1996; El-Nemr et al., 1998; Joesbury et al., 1998; Weigert et al., 1999; Neal et al., 2005; Wright et al., 2006). A significant decrease in fertilization rates for smokers was observed in four studies (Elenbogen et al., 1991; Rosevear et al., 1992; Crha et al., 2001; Tiboni et al., 2004), while one study found a significantly higher fertilization rate for smokers (Crha et al., 2003). One study found a significantly higher fertilization rate only for smokers over the age of 35 (Zenzes et al., 1997).
|
|
Male partner smoking status |
|---|
Male partner smoking status is another factor responsible for clinical heterogeneity between studies; however, insufficient data was reported to allow for sensitivity or subgroup analysis. Only nine of the 28 studies included in the systematic review and meta-analysis contained any information regarding male partner smoking status (Pattinson et al., 1991
; Hughes et al., 1994; Zenzes et al., 1997; Joesbury et al., 1998; Klonoff-Cohen et al., 2001; Zitzmann et al., 2003; Neal et al., 2005; Wright et al., 2006; Soares et al., 2007). Of these, one reported a non-significant effect upon live birth rate (Pattinson et al., 1991). Five studies reported the effect of male smoking upon clinical pregnancy rate (Pattinson et al., 1991; Joesbury et al., 1998; Klonoff-Cohen et al., 2001; Zitzmann et al., 2003; Neal et al., 2005), and of these three found evidence of a significant difference in pregnancy rate between male smokers and non-smokers (Joesbury et al., 1998; Zitzmann et al., 2003; Neal et al., 2005). Of the three studies that reported fertilization rates between male partner smokers and non-smokers (Hughes et al., 1994; Joesbury et al., 1998; Neal et al., 2005), none found any evidence of a significant effect. Of the remaining three studies, one required all participating male partners to be non-smokers to allow examination of female smoking effect only (Soares et al., 2007), and the remaining two studies stated the smoking status of the males but undertook no analysis of the effect of male smoking, reporting insufficient data to allow this to be done (Zenzes et al., 1997; Wright et al., 2006).
A funnel plot was generated to detect any publication bias. This was generated for the outcome of clinical pregnancy rate only as there were insufficient numbers of studies reporting on other outcome measures. The funnel plot generated suggests some presence of publication bias: although there are equal numbers of studies either side of the vertical line, the plot is not symmetrical (Fig. 4). While publication bias is one explanation for this asymmetry it is important to also consider other sources of selection bias, heterogeneity of studies, methodological limitations of studies and chance (Higgins and Green, 2005).
|
|
Discussion |
|---|
The results of this meta-analysis provide evidence of a significant negative effect of cigarette smoking by women at the time of infertility treatment, upon clinical outcomes of ART.
Compared with non-smokers, cigarette smokers demonstrated a significantly lower live birth rate per ART cycle (OR 0.54, 95% CI 0.30–0.99). Surprisingly, only four studies reported results for this outcome. Future studies should endeavour to report live birth rates because the goal of ART from the perspective of the patient is to overcome infertility by achieving live birth. While it is important for physicians to be aware of proximate treatment-orientated success rates that are often reported, such as oocyte retrieval or fertilization rates, it is also important that more distal patient-orientated outcomes such as live birth rates are reported. Without this information, prospective patients cannot make informed decisions about lifestyle factors that could affect their ART success.
Results from one study included in the meta-analysis for live birth rate (Lintsen et al., 2005) are particularly significant, and the high quality design of this study adds significant weight to the evidence. The authors examined the success rate of IVF in 8457 women from all IVF clinics in the Netherlands, included patients undergoing first IVF cycle only, and classified patients as smokers only if they had a stable smoking habit for at least 1 year prior to treatment. In this study, the odds of live birth for smokers was 0.73 that of non-smokers. This study shows the narrowest CI (0.64–0.84) and was given the highest weight (43.14) in the meta-analysis.
Although not included in the meta-analysis for not reporting raw data, one other study investigated the effect of cigarette smoking on live birth rate and found the risk of smokers not achieving a live birth to be significantly higher than non-smokers [Relative risk (RR) 2.51, 95% CI 1.11–5.67] (Klonoff-Cohen et al., 2001).
In order to examine the effect of cigarette smoking on ART it is also important to report clinical pregnancy rate. Although a treatment-orientated rather than patient-orientated outcome, clinical pregnancy rate is not confounded by the complications of smoking observed in the second and third trimesters of both spontaneous and ART pregnancies, that can affect live birth rate. Therefore, measuring the clinical pregnancy rate allows an objective evaluation of the efficacy of the ART cycle itself. The meta-analysis detected a significantly lower clinical pregnancy rate per cycle ART cycle among smokers (OR 0.56, 95% CI 0.43–0.73). It is interesting to note that this result is very similar to that reported in previous meta-analyses (Feichtinger et al., 1997; Augood et al., 1998). Augood et al. also reported a significantly lower clinical pregnancy rate amongst smokers (OR 0.66, 95% CI 0.49–0.88), and Feichtinger et al. found that smokers were required to undergo 1.79 (95% CI 1.24–2.59) IVF–embryo transfer cycles compared with non-smokers in order to achieve pregnancy.
Four additional studies investigated clinical pregnancy rates, but could not be included in the meta-analysis: one study did not report raw data, but found the risk of a smoker never achieving a pregnancy compared with a non-smoker to be significantly increased (RR 2.71, 95% CI 1.37–5.35) (Klonoff-Cohen et al., 2001). Three other studies reported clinical pregnancy rates per embryo transfer rather than per cycle ART. One found a significant difference in pregnancy rate [20% in 39 smokers compared with 48.3% in 146 non-smokers (Neal et al., 2005)], and two found no significant difference between smoking and non-smoking groups (Maximovich and Beyler, 1995; Motejlek et al., 2006).
Of patients that achieved a pregnancy following ART, smokers had significantly higher odds of spontaneous miscarriage (OR 2.65, 95% CI 1.33–5.30), and the meta-analysis also found an increased risk of ectopic pregnancy among smokers (OR 15.69, 95% CI 2.87–85.76). While statistically significant, the evidence for the higher OR observed for ectopic pregnancy rate is less reliable, as evidenced by the very large CI. Since there is no evidence for statistical heterogeneity for this outcome, this is likely to be due to the small sample size as a result of the paucity of studies reporting ectopic pregnancy rates.
The cyclic nature of ART gives rise to many difficulties in defining the exact effect of smoking upon outcomes of ART: the success rate of any outcome is not independent of other outcomes and is a reflection of previous stages in that cycle. Furthermore, the exact mechanisms by which cigarette smoking may exert a detrimental effect upon outcomes of ART are currently relatively unknown. A strong body of evidence indicates that the negative effect of cigarette smoking on fertility is evident in every system involved in the reproductive process (Soares and Melo, 2008). Research to date suggests that there are likely to be several mechanisms taking place at different stages of the assisted reproduction cycle that underlie these effects, with effects on the ovaries, the oocyte and the uterus. Current research suggests that in addition to a reduced ovarian reserve observed in smokers (Sharara et al., 1994), cigarette smoking may also impair folliculogenesis by causing an imbalance between pre- and anti-oxidants resulting in oxidative stress in the Graafian follicle (Paszkowski et al., 2002). Damage to the oocyte itself is likely to occur as a result of mutagens such as cadmium, cotinine and polycyclic aromatic hydrocarbons that are present in cigarette smoke and have been found in the follicular fluid of female cigarette smokers. Adducts of these compounds have the potential to damage DNA, alter the meiotic spindle of oocytes and affect the function and viability of oocyte granulosa cell complexes (Zenzes et al., 1998; Zenzes, 2000). Cigarette smoking has also been shown to decrease the number of mature oocytes retrieved following ovulation induction in ART (Zenzes et al., 1997): this may be caused by an increased concentration of vascular endothelial growth factor antagonists impairing angiogenesis of oocytes (Motejlek et al., 2006). The decreased fertilization and pregnancy rates often observed in these studies may be due to an increase in the zona pellucida thickness of oocytes observed in cigarette smokers (Shiloh et al., 2004), in addition to a reduced uterine receptiveness (Soares et al., 2007). Previous studies have also identified antioestrogenic effects of cigarette smoking that may account for increased miscarriage rates: inhibition of granulosa cell aromatase activity in smokers can result in corpus luteal deficiency (Shiverick and Salafia, 1999). Finally, studies in animal models have found tobacco to alter the function of Fallopian tubes therefore increasing the risk of ectopic pregnancy (Saraiya et al., 1998).
It is important to acknowledge the limitations of meta-analyses that pool data from observational studies rather than randomized controlled trials. There is more potential for results to be confounded, and it is difficult to ensure that retrospective studies are a true representation of the population. Unfortunately, due to the nature of the research question there is no alternative, and results from these meta-analyses still make a valid contribution to current knowledge and understanding of these issues.
All studies included in this meta-analysis share the common design of comparing a group of patients who smoke cigarettes with a control group that do not. Using this information it is possible to report a general effect of smoking on the outcomes of ART. Two forms of heterogeneity (statistical and clinical) exist between studies and must be accounted for. In this meta-analysis, statistical heterogeneity between studies was compensated for using a random effects model where appropriate.
One of the main factors responsible for clinical heterogeneity was the inclusion of significantly different age groups of smoking and non-smoking women in some studies. Age is one of the most important confounding factors in ART success and has been shown to be a strong predictor of this in numerous studies, with both pregnancy and live birth rates declining with increasing age of the patient (Winston and Handyside, 2007). In addition to the negative effect of age on the ovary, a negative effect of age has also been observed upon uterine receptiveness: ART outcomes were found to be considerably worse in older recipients of oocyte donation cycles (Soares et al., 2005). In an attempt to account for this heterogeneity, a sensitivity analysis was performed to exclude all studies that either reported a significant difference in age between groups, failed to state whether a significant age difference existed, or included patients over the age of 40. Surprisingly, when this sensitivity analysis was performed for clinical pregnancy rate, there was little effect on the statistical evidence of heterogeneity for all outcomes. This may be due to the sensitivity analysis significantly decreasing the sample sizes for outcomes, therefore reducing the reliability and power of the chi-squared test to detect heterogeneity. In order to eliminate this confounding factor in future studies, researchers should endeavour to compare outcomes between groups without statistically significant differences in age, and to either exclude older patients from their analyses or to stratify results by age thus allowing separate analyses to be conducted.
Male partner smoking status was identified to be a likely cause of clinical heterogeneity between studies. The vast majority of studies did not report male smoking status, and it was therefore not possible to undertake a sensitivity analysis for this confounding variable. The systematic review revealed conflicting evidence on the effect of male smoking upon clinical pregnancy rates, however, no evidence was found of a significant effect upon either live birth rates or fertilization rates. It is difficult to define the exact effect of male partner smoking status upon outcomes of ART, as the negative sequelae are likely to be as a result of both the direct effect upon spermatozoa and the indirect effect of passive smoking upon the female partner. Research suggests that both factors are significant influences, and that male partner smoking damages DNA of spermatozoa thereby hindering future embryonic development (Rubes et al., 1998; Zenzes et al., 1999), as well as the detrimental effect to female fertility of passive smoking from an actively smoking male partner (Neal et al., 2005).
The definition of smoking varied widely between studies, and this variation is another likely cause of heterogeneity. Some studies defined smokers as those smoking at least fifteen cigarettes per day (Elenbogen et al., 1991), or required a stable habit for a year before treatment (Lintsen et al., 2005). Other authors defined smokers to be those smoking 0-1 cigarettes per day with no minimum duration of smoking habit (Crha et al., 2001). Definitions of smoking in other studies varied between these extremes. It is therefore not possible to comment on outcomes of ART based on degree of smoking habit (number smoked per day), but rather to report results for only current cigarette smoking versus non-cigarette smoking at the time of treatment.
One of the studies included in the meta-analysis reported outcomes in smoking and non-smoking patients for oocyte donation cycles (Soares et al., 2007). The inclusion of this study is questionable due to the uncertainty of the site of the negative effect of smoking, whether it be ovarian, uterine or both. To assess whether this was a source of heterogeneity, a sensitivity analysis was conducted in which this study was excluded. The results showed a negligible effect upon clinical pregnancy rate and a small increase in the odds of miscarriage, indicating that the inclusion of this study did not have a significant effect upon the results.
Other potential confounding factors were rarely reported in studies and did not therefore allow for sensitivity analysis. One such example is BMI: only three of the 21 studies included in this meta-analysis reported BMI for both smoking and non-smoking patients (Gustafson et al., 1996; Wright et al., 2006; Soares et al., 2007), and it was therefore not possible to carry out a sensitivity analysis for this factor. Future studies should report this information along with details of any other potential confounding factors in order to identify and account for significant differences between groups.
Despite the heterogeneity between studies and causes of bias mentioned above, the large sample size gained in this meta-analysis by pooling results between studies compensates to a large extent for this variation.
We performed a subgroup analysis for method of ART in an attempt to further account for clinical heterogeneity between studies. This allowed us to compare outcomes between studies using conventional IVF treatment only, but there were insufficient studies to pool results for any other subgroup of method of ART. This analysis resulted in decreased statistical evidence for heterogeneity, thus confirming that treatment intervention type may be a factor in the heterogeneity of the initially analysed studies.
Subgroup analysis did not result in exclusion of any of the studies originally pooled for live birth rate and ectopic pregnancy rate. After subgroup analysis for IVF treatment only, statistically significant results were observed for clinical pregnancy rate (OR 0.57, 95% CI 0.42–0.77), although the results for spontaneous miscarriage rate were no longer significant (OR 2.99, 95% CI 0.94–9.47). This is likely to be due to excluding the two studies with the largest sample sizes, therefore significantly decreasing the power of meta-analysis.
In conclusion, the current meta-analysis provides evidence for a significantly negative effect of cigarette smoking on clinical outcomes of ART. There is particularly overwhelming evidence for a decreased clinical pregnancy rate amongst smokers, in addition to the strong implication of a negative effect on live birth rate, miscarriage rate, ectopic pregnancy rate and fertilization rate. In order to improve success rates for ART this evidence should be presented to actively smoking women seeking treatment for infertility, along with strong advice to cease smoking.
|
References |
|---|
Agnani G, Roux C, Roth P, Khalife H, Joanne C. Effects of smoking and pelvic inflammatory lesions on the outcome of in vitro fertilization/embryo transfer. Arch STD/HIV Res (1994) 8:271–275.
Augood C, Duckitt K, Templeton AA. Smoking and female infertility: a systematic review and meta-analysis. Hum Reprod (1998) 13:1532–1539.
Crha I, Hruba D, Fiala J, Ventruba P, Zakova J. Reduced numbers of retrieved oocytes in smoking women. Scripta Medica (2000) 73:299–304.
Crha I, Hruba D, Fiala J, Ventruba P, Zakova J, Petrenko M. The outcome of infertility treatment by in-vitro fertilisation in smoking and non-smoking women. Cent Eur J Public Health (2001) 9:64–68.[Medline]
Crha I, Hruba D, Ventruba P, Fiala J, Totusek J, Visnova H. Ascorbic acid and infertility treatment. Cent Eur J Public Health (2003) 11:63–67.[Medline]
El-Nemr A, Al-Shawaf T, Sabatini L, Wilson C, Lower AM, Grudzinskas JG. Effect of smoking on ovarian reserve and ovarian stimulation in in-vitro fertilization and embryo transfer. Hum Reprod (1998) 13:2192–2198.
Elenbogen A, Lipitz S, Mashiach S, Dor J, Levran D, Ben-Rafale Z. The effect of smoking on the outcome of in-vitro fertilization-embryo transfer. Hum Reprod (1991) 6:242–244.
Feichtinger W, Papalambrou K, Poehl M, Krischker U, Neumann K. Smoking and in vitro fertilization: a meta-analysis. J Assist Reprod Genet (1997) 14:596–599.[CrossRef][Web of Science][Medline]
Gustafson O, Nylund L, Carlstrom K. Does hyperandrogenism explain lower in vitro fertilization (IVF) success rates in smokers? Acta Obstet Gynecol Scand (1996) 75:149–156.[Web of Science][Medline]
Harrison KL, Breen TM, Hennessey JF. The effect of patient smoking habit on the outcome of IVF and GIFT treatment. Aust N Z J Obstet Gynaecol (1990) 30:340–342.[Web of Science][Medline]
Higgins J, Green S, Publication bias and funnel plots. Cochrane Handbook for Systematic Reviews of Interventions 4.2.5 [updated May 2005]; section 8. Chichester, UK: John Wiley & Sons, Ltd. 2005.
Hruba D, Fiala J, Crha I, Rezl P. Assisted reproduction in female smokers and non-smokers. Hygiena (1999) 44:8–14.
Hughes EG, Younglai EV, Ward SM. Cigarette smoking and outcomes of in-vitro fertilization and embryo transfer: a prospective cohort study. Hum Reprod (1992) 7:358–361.
Hughes EG, Yeo J, Claman P, Younglai EV, Sagle MA, Daya S, Collins JA. Cigarette smoking and the outcomes of in vitro fertilization: measurement of effect size and levels of action. Fertil Steril (1994) 62:807–814.[Web of Science][Medline]
Hughes EG, Lamont DA, Beecroft ML, Wilson DM, Brennan BG, Rice SC. Randomized trial of a ‘stage-of-change’ oriented smoking cessation intervention in infertile and pregnant women. Fertil Steril (2000) 74:498–503.[CrossRef][Web of Science][Medline]
Joesbury KA, Edirisinghe WR, Phillips MR, Yovich JL. Evidence that male smoking affects the likelihood of a pregnancy following IVF treatment: application of the modified cumulative embryo score. Hum Reprod (1998) 13:1506–1513.
Klonoff-Cohen H. Female and male lifestyle habits and IVF: What is known and unknown. Hum Reprod Update (2005) 11:179–203.[Medline]
Klonoff-Cohen H, Natarajan L, Marrs R, Yee B. Effects of female and male smoking on success rates of IVF and gamete intra-Fallopian transfer. Hum Reprod (2001) 16:1382–1390.
Lintsen AME, Pasker-de Jong PCM, de Boer EJ, Burger CW, Jansen CAM, Braat DDM, van Leeuwen FE, Kortman M, te Velde ER, Macklon N, et al. Effects of subfertility cause, smoking and body weight on the success rate of IVF. Hum Reprod (2005) 20:1867–1875.
Maximovich A, Beyler SA. Cigarette smoking at time of in vitro fertilization cycle initiation has negative effect on in vitro fertilization-embryo transfer success rate. J Assist Reprod Genet (1995) 12:75–77.[CrossRef][Web of Science][Medline]
Motejlek K, Palluch F, Neulen J, Grummer R. Smoking impairs angiogenesis during maturation of human oocytes. Fertil Steril (2006) 86:186–191.[CrossRef][Web of Science][Medline]
Neal MS, Hughes EG, Holloway AC, Foster WG. Sidestream smoking is equally as damaging as mainstream smoking on IVF outcomes. Hum Reprod (2005) 20:2531–2535.
pa-Martynow M, Pawelczyk L, Taszarek-Hauke G, Josiak M, Derwich K, Jedrzejczak P. The effect of smoking on infertility treatment in women undergoing assisted reproduction cycles [Polish]. Przeglad Lekarski (2005) 62:973–975.[Medline]
pa-Martynow M, Jedrzejczak P, Taszarek-Hauke G, Josiak M, Pawelczyk L. The impact of cigarette smoking on oocytes and embryos quality during in vitro fertilization program [Polish]. Przeglad Lekarski (2006) 63:838–840.[Medline]
Paszkowski T, Clarke RN, Hornstein MD. Smoking induces oxidative stress inside the Graafian follicle. Hum Reprod (2002) 17:921–925.
Pattinson HA, Taylor PJ, Pattinson MH. The effect of cigarette smoking on ovarian function and early pregnancy outcome of in vitro fertilization treatment. Fertil Steril (1991) 55:780–783.[Web of Science][Medline]
Phipps WR, Cramer DW, Schiff E. The association between smoking and female infertility as influenced by cause of the infertility. Fertil Steril (1987) 48:377–382.[Web of Science][Medline]
Rosevear SK, Holt DW, Lee TD, Ford WCL, Wardle PG, Hull MGR. Smoking and decreased fertilisation rates in vitro. Lancet (1992) 340:1195–1196.[CrossRef][Web of Science][Medline]
Rubes J, Lowe X, Moore D, Perreault S, Slott V, Evenson D, Selevan SG, Wyrobek AJ. Smoking cigarettes is associated with increased sperm disomy in teenage men. Fertil Steril (1998) 70:715–723.[CrossRef][Web of Science][Medline]
Saraiya M, Berg CJ, Kendrick JS, Strauss LT, Atrash HK, Ahn YW. Cigarette smoking as a risk factor for ectopic pregnancy. Am J Obstet Gynecol (1998) 178:493–498.[CrossRef][Web of Science][Medline]
Sharara FI, Beatse SN, Leonardi MR, Navot D, Scott RT. Cigarette-smoking accelerates the development of diminished ovarian reserve as evidenced by the clomiphene citrate challenge test. Fertil Steril (1994) 62:257–262.[Web of Science][Medline]
Shiloh H, Baratz SL, Koifman M, Ishai D, Bidder D, Weiner-Meganzi Z, Dirnfeld M. The impact of cigarette smoking on zona pellucida thickness of oocytes and embryos prior to transfer into the uterine cavity. Hum Reprod (2004) 19:157–159.
Shiverick KT, Salafia C. Cigarette smoking and pregnancy. I: Ovarian, uterine and placental effects. Placenta (1999) 20:265–272.[CrossRef][Web of Science][Medline]
Soares SR, Melo MA. Cigarette smoking and reproductive function. Curre Opin Obstet Gynecol (2008) 20:281–291.
Soares SR, Troncoso C, Bosch E, Serra V, Simon C, Remohi J, Pellicer A. Age and uterine receptiveness: predicting the outcome of oocyte donation cycles. J Clin Endocrinol Metab (2005) 90:4399–4404.
Soares SR, Simon C, Remohi J, Pellicer A. Cigarette smoking affects uterine receptiveness. Hum Reprod (2007) 22:543–547.
Sterzik K, Strehler E, De SM, Trumpp N, Abt M, Rosenbusch B, Schneider A. Influence of smoking on fertility in women attending an in vitro fertilization program. Fertil Steril (1996) 65:810–814.[Web of Science][Medline]
Stolwijk AM, Zielhuis GA, Hamilton CJ, Straatman H, Hollanders JM, Goverde HJ, van Dop PA, Verbeek AL. Prognostic models for the probability of achieving an ongoing pregnancy after in-vitro fertilization and the importance of testing their predictive value. Hum Reprod (1996) 11:2298–2303.
Tiboni GM, Bucciarelli T, Giampietro F, Sulpizio M, Di IC. Influence of cigarette smoking on vitamin E, vitamin A, beta-carotene and lycopene concentrations in human pre-ovulatory follicular fluid. Int J Immunopathol Pharmacol (2004) 17:389–393.[Web of Science][Medline]
Trapp M, Kemeter P, Feichtinger W. Smoking and in-vitro fertilization. Hum Reprod (1986) 1:357–358.
Triopon G, Tailland ML, Faillie JL, Bolzinger E, Mercier G, Ripart S, Boyer JC, Fabbro-Peray P, Poirey S, Archambault A, et al. In vitro fertilization and smoking: use of urinary cotinine and expired air carbon monoxide measurements [French]. Gynecol Obstet Fertil (2006) 34:1043–1050.[CrossRef][Medline]
Van Voorhis BJ, Syrop CH, Hammitt DG, Dunn MS, Snyder GD. Effects of smoking on ovulation induction for assisted reproductive techniques. Fertil Steril (1992) 58:981–985.[Web of Science][Medline]
Van Voorhis BJ, Dawson JD, Stovall DW, Sparks AET, Syrop CH. The effects of smoking on ovarian function and fertility during assisted reproduction cycles. Obstet Gynecol (1996) 88:785–791.[CrossRef][Web of Science][Medline]
Weigert M, Hofstetter G, Kaipl D, Gottlich H, Krischker U, Bichler K, Poehl M, Feichtinger W. The effect of smoking on oocyte quality and hormonal parameters of patients undergoing in vitro fertilization-embryo transfer. J Assist Reprod Genet (1999) 16:287–293.[CrossRef][Web of Science][Medline]
Winston RMI, Handyside AH. New challenges in human in vitro fertilization. Science (2007) 260:932–936.
Winter E, Wang J, Davies MJ, Norman R. Early pregnancy loss following assisted reproductive technology treatment. Hum Reprod (2002) 17:3220–3223.
Wright KP, Trimarchi JR, Allsworth J, Keefe D. The effect of female tobacco smoking on IVF outcomes. Hum Reprod (2006) 21:2930–2934.
Zenzes MT. Smoking and reproduction: gene damage to human gametes and embryos. Hum Reprod Update (2000) 6:122–131.
Zenzes MT, Reed TE, Casper RF. Effects of cigarette smoking and age on the maturation of human oocytes. Hum Reprod (1997) 12:1736–1741.
Zenzes MT, Puy LA, Bielecki R. Immunodetection of benzo[a]pyrene adducts in ovarian cells of women exposed to cigarette smoke. Mol Hum Reprod (1998) 4:159–165.
Zenzes MT, Puy LA, Bielecki R, Reed TE. Detection of benzo[a]pyrene diol epoxide-DNA adducts in embryos from smoking couples: Evidence for transmission by spermatozoa. Mol Hum Reprod (1999) 5:125–131.
Zitzmann M, Rolf C, Nordhoff V, Schrader G, Rickert-Fohring M, Gassner P, Behre HM, Greb RR, Kiesel L, Nieschlag E. Male smokers have a decreased success rate for in vitro fertilization and intracytoplasmic sperm injection. Fertil Steril (2003) 79(Suppl_3):1550–1554.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||