Do assisted reproductive technologies cause adverse fetal outcomes?

Abstract

Treatment for infertility using assisted reproductive technologies (ART) is highly successful and has been used to help a steadily growing number of couples worldwide. In 1999, in the United States, more than 86,000 treatment cycles were performed resulting in the birth of more than 30,000 babies. Despite this widespread application, few follow-up studies of children conceived through ART have been performed, and more rigorous investigation of this important issue has clearly been needed. In recent months, three studies linking ART with several complications have been published in high profile and widely read general medical journals: Schieve et al. (1Schieve L.A. Meikle S.F. Ferre C. Peterson H.B. Jeng G. Wilcox L.S. Low and very low birth weight in infants conceived with use of assisted reproductive technology.N Engl J Med. 2002; 346: 731-737Crossref PubMed Scopus (838) Google Scholar) reported that singletons conceived using ART were at an increased risk for low birth weight, whereas Hansen et al. (2Hansen M. Kurinczuk J.J. Bower C. Webb S. The risk of major birth defects after intracytoplasmic sperm injection and in vitro fertilization.N Engl J Med. 2002; 346: 725-730Crossref PubMed Scopus (916) Google Scholar) suggested an increased risk of major birth defects. Finally, Stromberg et al. (3Stromberg B. Dahlquist G. Ericson A. Finnstrom O. Koster M. Stjernqvist K. Neurological sequelae in children born after in-vitro fertilization a population-based study.Lancet. 2002; 359: 461-465Abstract Full Text Full Text PDF PubMed Scopus (448) Google Scholar) concluded that children conceived through IVF have an increased risk of neurological problems, especially cerebral palsy.The importance of these studies is obvious as they provide clues of possible risks associated with ART. However, they are all retrospective analyses of data collected through registries and therefore are vulnerable to biases inherent to such study design. We must be careful not to overinterpret the data by concluding that the use of ART, whether from gamete or embryo manipulation or use of medications, is the direct cause of the complications—the observed associations may simply be explained by one or more confounders, such as an underlying infertility-related condition in the treated women. This short communication is an attempt to place these three articles in proper perspective for the clinician and to provide the impetus for conducting further studies to determine the true nature of the described associations.At the outset, it should be stated that a major weakness of all three studies is the lack of proper controls. If the aim of a study is to determine whether a cause and effect relationship exists between the process of IVF and a specific outcome (i.e., low birth weight, major birth defects, neurological problems), the appropriate control population is that of babies born to infertile women achieving pregnancies by methods other than IVF. None of the three studies specifically included such a control population.Art and low birth weightSchieve et al. (1Schieve L.A. Meikle S.F. Ferre C. Peterson H.B. Jeng G. Wilcox L.S. Low and very low birth weight in infants conceived with use of assisted reproductive technology.N Engl J Med. 2002; 346: 731-737Crossref PubMed Scopus (838) Google Scholar) showed that singletons conceived using ART were at an increased risk for low birth weight, overall adjusted risk ratio of 1.8 (95% confidence interval [CI] 1.7, 1.9). They compared infants born with the use of IVF and intracytoplasmic sperm injection (ICSI) in the United States in 1996 and 1997, with the entire group of infants born in the United States in 1997. They concluded that “this study suggests that the increased risk… may be directly related to such treatments for infertility.” Although they also acknowledged that the mechanisms underlying the association remain uncertain, they argue that it is more likely that the relationship is due to the treatment rather than to the underlying condition in infertile women. Such conclusions are clearly not supported by the presented data.Their strongly stated conclusion appears to be based primarily on a subgroup analysis in which the sample was restricted to infants born to couples with a diagnosis of male factor infertility. The rationale for choosing this subgroup was to include in the analysis women presumably without infertility-related conditions. The relationship between ART and low birth weight in this subgroup remained significant (relative risk [RR] 1.7, 95% CI 1.5,1.9), suggesting that the association is unrelated to an underlying female factor but rather to ART itself. Nevertheless, we recommend caution in reaching this conclusion. First, the diagnosis of male infertility did not exclude the presence of other etiologies. It is common for infertile couples to have more than one infertility-related condition. However, the 1996 and 1997 SART/CDC (Society for Assisted Reproductive Technologies/Centers for Disease Control) data, unlike the 1999 data, used only single diagnoses to categorize patients. In 1999, SART/CDC changed the categorization of causes of infertility, adding a category for “multiple factors, female + male.” This combination was present in 17.5% of the cases, nearly the same as the “male factor” category, which was reported in the Schieve article as 19.5%. Second, the diagnosis of male factor can be nonspecific, as demonstrated by Guzick et al. (4Guzick D.S. Overstreet J.W. Factor-Litvak P. Brazil C.K. Nakajima S.T. Coutifaris C. et al.Sperm morphology, motility, and concentration in fertile and infertile men.N Engl J Med. 2001; 345: 1388-1393Crossref PubMed Scopus (913) Google Scholar), who showed that traditional semen evaluation criteria are not truly diagnostic of male infertility. Thus, using this subgroup from the 1996 and 1997 SART/CDC registries to conclude that ART causes an increased risk of low birth weight in normal women may be inaccurate.Further doubt on this conclusion is shed by the observation that infants delivered by gestational carriers were not at an increased risk for low birth weight (RR 1.2, 95% CI 0.6,1.8). Because gestational carriers are unlikely to have any infertility-related conditions, the finding that they were not at an increased risk strongly argues that ART may not be the mechanism responsible for the observed relationship in the larger sample. It is true that this subanalysis had a small sample size (n = 180) and may have been underpowered to observe the relationship. Nevertheless, this observation significantly weakens the investigators’ overall conclusion.Finally, all of the analyses were stratified by the number of infants delivered (i.e., singletons, twins, triplets, or higher order). This evaluation revealed that only singletons were at an increased risk, whereas twins had a risk ratio of 1.0 (95% CI 1.0,1.1). This finding is also contrary to the theory that ART is responsible. If embryos conceived through ART are truly at an increased risk for stunted growth, we would have expected this risk to be accentuated in the relatively compromised environment of a uterus with multiple gestations. This statement can be corroborated by the intrauterine growth retardation generally seen in multiple gestations. In addition, it has previously been suggested that multiple developing gestations in early pregnancy limit placental development and ultimate fetal growth (5Dickey R.P. Taylor S.N. Lu P.Y. Sartor B.M. Storment J.M. Rye P.H. et al.Spontaneous reduction of multiple pregnancy incidence and effect on outcome.Am J Obstet Gynecol. 2002; 186: 77-83Abstract Full Text Full Text PDF PubMed Scopus (172) Google Scholar). Indeed, Schieve et al. (1Schieve L.A. Meikle S.F. Ferre C. Peterson H.B. Jeng G. Wilcox L.S. Low and very low birth weight in infants conceived with use of assisted reproductive technology.N Engl J Med. 2002; 346: 731-737Crossref PubMed Scopus (838) Google Scholar) confirmed previous reports showing that presence of multiple gestations during the first trimester is highly associated with growth retardation, whether the number of gestations is reduced or not. They attempted to control for this effect by comparing pregnancies in which the number of fetal hearts noted on ultrasonography did not exceed the number of infants born. However, this approach did not control for the possible adverse effect of anembryonic gestations, which would have been present even if a fetal heart did not develop and thus was not detected. Because this is more likely in ART where multiple embryos are transferred, there is potential for bias skewing the findings in favor of the alternative hypothesis.Art and major birth defectsHansen et al. (2Hansen M. Kurinczuk J.J. Bower C. Webb S. The risk of major birth defects after intracytoplasmic sperm injection and in vitro fertilization.N Engl J Med. 2002; 346: 725-730Crossref PubMed Scopus (916) Google Scholar) investigated the incidence of birth defect after ICSI and IVF in Western Australia between 1993 and 1997, and found an increase in major birth defects (MBD) diagnosed by 1 year of age when compared to a population control group, overall adjusted odds ratio of 2.0 (95% CI 1.3,3.2). Also, the study demonstrated that children conceived with ART were more likely to have multiple MBD. They suggest that the increase in risk may be attributed to the treatment; however, they acknowledge that confounding due to underlying causes of infertility could not be eliminated. An appropriate control population of babies born to infertile women achieving pregnancy without the use of ART would have eliminated this major weakness.Beyond this, the study has several potential problems. Most notably, the control population was significantly different from the ART patients: the controls were much younger (28.2 vs. 32.6 years for ICSI and 34.1 for IVF), were less likely to live in a metropolitan area, consisted of a smaller percentage of whites, and had a higher mean parity. Although they attempted to control for age and parity in their analysis, they could not adjust for socioeconomic status and environmental differences. Many birth defects have been linked to nutritional deficiencies and environmental teratogens, which may differ between the geographic areas. Thus, the controls may not represent the same population, and this may have biased the results. Second, they attempted to control for increased diagnostic vigilance among ART infants, which may have resulted in an increased rate of MBD detection. However, the potential for a detection bias cannot be entirely eliminated.Finally, the clinical significance of these findings should be considered. Of note, many of the MBD are amenable to medical or surgical correction and do not have significant long-term sequelae. It would be useful to compare the rates of both clinically severe and correctable MBD between the groups. Such information would be useful in counseling couples considering ART.Art and neurological problemsStromberg et al. (3Stromberg B. Dahlquist G. Ericson A. Finnstrom O. Koster M. Stjernqvist K. Neurological sequelae in children born after in-vitro fertilization a population-based study.Lancet. 2002; 359: 461-465Abstract Full Text Full Text PDF PubMed Scopus (448) Google Scholar) used data from Swedish registries for the period between 1982 and 1995. They compared rates of neurological problems in children born through ART and population-based controls and found an association between cerebral palsy and the use of ART. Cerebral palsy is a difficult clinical entity to define and has been associated with a long list of etiologies including prematurity and low birth weight. Although adjusting for low birth weight still resulted in an increased risk (odds ratio [OR] 2.1, 95% CI 1.0,4.5), these studies are unable to control for the myriad of conditions associated with cerebral palsy including subclinical infection in utero. In addition, the clinical impact of these findings is questionable. Doubling of a very small risk results in a minor increase in the overall incidence of cerebral palsy.They themselves conclude that a much more important risk factor is multiple gestations. Similar to the article by Schieve et al. (1Schieve L.A. Meikle S.F. Ferre C. Peterson H.B. Jeng G. Wilcox L.S. Low and very low birth weight in infants conceived with use of assisted reproductive technology.N Engl J Med. 2002; 346: 731-737Crossref PubMed Scopus (838) Google Scholar) concerning low birth weight, multiple gestations were at a higher risk for neurological complications; however, twins conceived by ART did not demonstrate an increase in risk above the controls. Thus, we again suggest that the observed associations may be due to a potential confounder, such as an underlying condition in infertile women. This, once again, highlights the importance of the use of a proper control population, which should have been babies born to infertile couples as a result of treatment other than IVF.ConclusionsThe three articles we reviewed have suggested an association between use of ART and increased risk of problems in the resulting children. However, the validity of these associations and certainly the implication of a cause and effect relationship are weakened by potential limitations in the quality of the data inherent to registries. Furthermore, as highlighted, a major flaw of all three studies is the lack of appropriate controls, which ideally should have been babies born to infertile couples achieving pregnancy without the use of ART.These studies should be applauded, as they have provided the groundwork and direction for future research. Carefully constructed, prospective, multicenter studies using the appropriate control patient population are clearly needed. It is imperative to determine whether use of ART results in increased risk to the children born after these procedures, as the answer may have profound clinical implications. For now we should make our patients aware of the possible associations (not risks) as part of obtaining informed consent for the treatment. However, we do not believe that altering patient selection for ART and reducing the treatment’s utilization is justified at this time. Finally, it should not be forgotten that even if the described associations are accurate, infertile couples may still choose to accept the risk, as the chances of having a normal infant remain greatly in their favor, and lack of progeny may represent a greater personal tragedy than a potentially small risk of giving birth to an affected child.Do the clinical and laboratory protocols used in ART cause adverse outcomes in infants conceived through these procedures? The present data suggest associations, but do not prove cause and effect relationships. We recommend that, until better evidence is available, our consent procedures should be modified, but our clinical practice should not. Treatment for infertility using assisted reproductive technologies (ART) is highly successful and has been used to help a steadily growing number of couples worldwide. In 1999, in the United States, more than 86,000 treatment cycles were performed resulting in the birth of more than 30,000 babies. Despite this widespread application, few follow-up studies of children conceived through ART have been performed, and more rigorous investigation of this important issue has clearly been needed. In recent months, three studies linking ART with several complications have been published in high profile and widely read general medical journals: Schieve et al. (1Schieve L.A. Meikle S.F. Ferre C. Peterson H.B. Jeng G. Wilcox L.S. Low and very low birth weight in infants conceived with use of assisted reproductive technology.N Engl J Med. 2002; 346: 731-737Crossref PubMed Scopus (838) Google Scholar) reported that singletons conceived using ART were at an increased risk for low birth weight, whereas Hansen et al. (2Hansen M. Kurinczuk J.J. Bower C. Webb S. The risk of major birth defects after intracytoplasmic sperm injection and in vitro fertilization.N Engl J Med. 2002; 346: 725-730Crossref PubMed Scopus (916) Google Scholar) suggested an increased risk of major birth defects. Finally, Stromberg et al. (3Stromberg B. Dahlquist G. Ericson A. Finnstrom O. Koster M. Stjernqvist K. Neurological sequelae in children born after in-vitro fertilization a population-based study.Lancet. 2002; 359: 461-465Abstract Full Text Full Text PDF PubMed Scopus (448) Google Scholar) concluded that children conceived through IVF have an increased risk of neurological problems, especially cerebral palsy. The importance of these studies is obvious as they provide clues of possible risks associated with ART. However, they are all retrospective analyses of data collected through registries and therefore are vulnerable to biases inherent to such study design. We must be careful not to overinterpret the data by concluding that the use of ART, whether from gamete or embryo manipulation or use of medications, is the direct cause of the complications—the observed associations may simply be explained by one or more confounders, such as an underlying infertility-related condition in the treated women. This short communication is an attempt to place these three articles in proper perspective for the clinician and to provide the impetus for conducting further studies to determine the true nature of the described associations. At the outset, it should be stated that a major weakness of all three studies is the lack of proper controls. If the aim of a study is to determine whether a cause and effect relationship exists between the process of IVF and a specific outcome (i.e., low birth weight, major birth defects, neurological problems), the appropriate control population is that of babies born to infertile women achieving pregnancies by methods other than IVF. None of the three studies specifically included such a control population. Art and low birth weightSchieve et al. (1Schieve L.A. Meikle S.F. Ferre C. Peterson H.B. Jeng G. Wilcox L.S. Low and very low birth weight in infants conceived with use of assisted reproductive technology.N Engl J Med. 2002; 346: 731-737Crossref PubMed Scopus (838) Google Scholar) showed that singletons conceived using ART were at an increased risk for low birth weight, overall adjusted risk ratio of 1.8 (95% confidence interval [CI] 1.7, 1.9). They compared infants born with the use of IVF and intracytoplasmic sperm injection (ICSI) in the United States in 1996 and 1997, with the entire group of infants born in the United States in 1997. They concluded that “this study suggests that the increased risk… may be directly related to such treatments for infertility.” Although they also acknowledged that the mechanisms underlying the association remain uncertain, they argue that it is more likely that the relationship is due to the treatment rather than to the underlying condition in infertile women. Such conclusions are clearly not supported by the presented data.Their strongly stated conclusion appears to be based primarily on a subgroup analysis in which the sample was restricted to infants born to couples with a diagnosis of male factor infertility. The rationale for choosing this subgroup was to include in the analysis women presumably without infertility-related conditions. The relationship between ART and low birth weight in this subgroup remained significant (relative risk [RR] 1.7, 95% CI 1.5,1.9), suggesting that the association is unrelated to an underlying female factor but rather to ART itself. Nevertheless, we recommend caution in reaching this conclusion. First, the diagnosis of male infertility did not exclude the presence of other etiologies. It is common for infertile couples to have more than one infertility-related condition. However, the 1996 and 1997 SART/CDC (Society for Assisted Reproductive Technologies/Centers for Disease Control) data, unlike the 1999 data, used only single diagnoses to categorize patients. In 1999, SART/CDC changed the categorization of causes of infertility, adding a category for “multiple factors, female + male.” This combination was present in 17.5% of the cases, nearly the same as the “male factor” category, which was reported in the Schieve article as 19.5%. Second, the diagnosis of male factor can be nonspecific, as demonstrated by Guzick et al. (4Guzick D.S. Overstreet J.W. Factor-Litvak P. Brazil C.K. Nakajima S.T. Coutifaris C. et al.Sperm morphology, motility, and concentration in fertile and infertile men.N Engl J Med. 2001; 345: 1388-1393Crossref PubMed Scopus (913) Google Scholar), who showed that traditional semen evaluation criteria are not truly diagnostic of male infertility. Thus, using this subgroup from the 1996 and 1997 SART/CDC registries to conclude that ART causes an increased risk of low birth weight in normal women may be inaccurate.Further doubt on this conclusion is shed by the observation that infants delivered by gestational carriers were not at an increased risk for low birth weight (RR 1.2, 95% CI 0.6,1.8). Because gestational carriers are unlikely to have any infertility-related conditions, the finding that they were not at an increased risk strongly argues that ART may not be the mechanism responsible for the observed relationship in the larger sample. It is true that this subanalysis had a small sample size (n = 180) and may have been underpowered to observe the relationship. Nevertheless, this observation significantly weakens the investigators’ overall conclusion.Finally, all of the analyses were stratified by the number of infants delivered (i.e., singletons, twins, triplets, or higher order). This evaluation revealed that only singletons were at an increased risk, whereas twins had a risk ratio of 1.0 (95% CI 1.0,1.1). This finding is also contrary to the theory that ART is responsible. If embryos conceived through ART are truly at an increased risk for stunted growth, we would have expected this risk to be accentuated in the relatively compromised environment of a uterus with multiple gestations. This statement can be corroborated by the intrauterine growth retardation generally seen in multiple gestations. In addition, it has previously been suggested that multiple developing gestations in early pregnancy limit placental development and ultimate fetal growth (5Dickey R.P. Taylor S.N. Lu P.Y. Sartor B.M. Storment J.M. Rye P.H. et al.Spontaneous reduction of multiple pregnancy incidence and effect on outcome.Am J Obstet Gynecol. 2002; 186: 77-83Abstract Full Text Full Text PDF PubMed Scopus (172) Google Scholar). Indeed, Schieve et al. (1Schieve L.A. Meikle S.F. Ferre C. Peterson H.B. Jeng G. Wilcox L.S. Low and very low birth weight in infants conceived with use of assisted reproductive technology.N Engl J Med. 2002; 346: 731-737Crossref PubMed Scopus (838) Google Scholar) confirmed previous reports showing that presence of multiple gestations during the first trimester is highly associated with growth retardation, whether the number of gestations is reduced or not. They attempted to control for this effect by comparing pregnancies in which the number of fetal hearts noted on ultrasonography did not exceed the number of infants born. However, this approach did not control for the possible adverse effect of anembryonic gestations, which would have been present even if a fetal heart did not develop and thus was not detected. Because this is more likely in ART where multiple embryos are transferred, there is potential for bias skewing the findings in favor of the alternative hypothesis. Schieve et al. (1Schieve L.A. Meikle S.F. Ferre C. Peterson H.B. Jeng G. Wilcox L.S. Low and very low birth weight in infants conceived with use of assisted reproductive technology.N Engl J Med. 2002; 346: 731-737Crossref PubMed Scopus (838) Google Scholar) showed that singletons conceived using ART were at an increased risk for low birth weight, overall adjusted risk ratio of 1.8 (95% confidence interval [CI] 1.7, 1.9). They compared infants born with the use of IVF and intracytoplasmic sperm injection (ICSI) in the United States in 1996 and 1997, with the entire group of infants born in the United States in 1997. They concluded that “this study suggests that the increased risk… may be directly related to such treatments for infertility.” Although they also acknowledged that the mechanisms underlying the association remain uncertain, they argue that it is more likely that the relationship is due to the treatment rather than to the underlying condition in infertile women. Such conclusions are clearly not supported by the presented data. Their strongly stated conclusion appears to be based primarily on a subgroup analysis in which the sample was restricted to infants born to couples with a diagnosis of male factor infertility. The rationale for choosing this subgroup was to include in the analysis women presumably without infertility-related conditions. The relationship between ART and low birth weight in this subgroup remained significant (relative risk [RR] 1.7, 95% CI 1.5,1.9), suggesting that the association is unrelated to an underlying female factor but rather to ART itself. Nevertheless, we recommend caution in reaching this conclusion. First, the diagnosis of male infertility did not exclude the presence of other etiologies. It is common for infertile couples to have more than one infertility-related condition. However, the 1996 and 1997 SART/CDC (Society for Assisted Reproductive Technologies/Centers for Disease Control) data, unlike the 1999 data, used only single diagnoses to categorize patients. In 1999, SART/CDC changed the categorization of causes of infertility, adding a category for “multiple factors, female + male.” This combination was present in 17.5% of the cases, nearly the same as the “male factor” category, which was reported in the Schieve article as 19.5%. Second, the diagnosis of male factor can be nonspecific, as demonstrated by Guzick et al. (4Guzick D.S. Overstreet J.W. Factor-Litvak P. Brazil C.K. Nakajima S.T. Coutifaris C. et al.Sperm morphology, motility, and concentration in fertile and infertile men.N Engl J Med. 2001; 345: 1388-1393Crossref PubMed Scopus (913) Google Scholar), who showed that traditional semen evaluation criteria are not truly diagnostic of male infertility. Thus, using this subgroup from the 1996 and 1997 SART/CDC registries to conclude that ART causes an increased risk of low birth weight in normal women may be inaccurate. Further doubt on this conclusion is shed by the observation that infants delivered by gestational carriers were not at an increased risk for low birth weight (RR 1.2, 95% CI 0.6,1.8). Because gestational carriers are unlikely to have any infertility-related conditions, the finding that they were not at an increased risk strongly argues that ART may not be the mechanism responsible for the observed relationship in the larger sample. It is true that this subanalysis had a small sample size (n = 180) and may have been underpowered to observe the relationship. Nevertheless, this observation significantly weakens the investigators’ overall conclusion. Finally, all of the analyses were stratified by the number of infants delivered (i.e., singletons, twins, triplets, or higher order). This evaluation revealed that only singletons were at an increased risk, whereas twins had a risk ratio of 1.0 (95% CI 1.0,1.1). This finding is also contrary to the theory that ART is responsible. If embryos conceived through ART are truly at an increased risk for stunted growth, we would have expected this risk to be accentuated in the relatively compromised environment of a uterus with multiple gestations. This statement can be corroborated by the intrauterine growth retardation generally seen in multiple gestations. In addition, it has previously been suggested that multiple developing gestations in early pregnancy limit placental development and ultimate fetal growth (5Dickey R.P. Taylor S.N. Lu P.Y. Sartor B.M. Storment J.M. Rye P.H. et al.Spontaneous reduction of multiple pregnancy incidence and effect on outcome.Am J Obstet Gynecol. 2002; 186: 77-83Abstract Full Text Full Text PDF PubMed Scopus (172) Google Scholar). Indeed, Schieve et al. (1Schieve L.A. Meikle S.F. Ferre C. Peterson H.B. Jeng G. Wilcox L.S. Low and very low birth weight in infants conceived with use of assisted reproductive technology.N Engl J Med. 2002; 346: 731-737Crossref PubMed Scopus (838) Google Scholar) confirmed previous reports showing that presence of multiple gestations during the first trimester is highly associated with growth retardation, whether the number of gestations is reduced or not. They attempted to control for this effect by comparing pregnancies in which the number of fetal hearts noted on ultrasonography did not exceed the number of infants born. However, this approach did not control for the possible adverse effect of anembryonic gestations, which would have been present even if a fetal heart did not develop and thus was not detected. Because this is more likely in ART where multiple embryos are transferred, there is potential for bias skewing the findings in favor of the alternative hypothesis. Art and major birth defectsHansen et al. (2Hansen M. Kurinczuk J.J. Bower C. Webb S. The risk of major birth defects after intracytoplasmic sperm injection and in vitro fertilization.N Engl J Med. 2002; 346: 725-730Crossref PubMed Scopus (916) Google Scholar) investigated the incidence of birth defect after ICSI and IVF in Western Australia between 1993 and 1997, and found an increase in major birth defects (MBD) diagnosed by 1 year of age when compared to a population control group, overall adjusted odds ratio of 2.0 (95% CI 1.3,3.2). Also, the study demonstrated that children conceived with ART were more likely to have multiple MBD. They suggest that the increase in risk may be attributed to the treatment; however, they acknowledge that confounding due to underlying causes of infertility could not be eliminated. An appropriate control population of babies born to infertile women achieving pregnancy without the use of ART would have eliminated this major weakness.Beyond this, the study has several potential problems. Most notably, the control population was significantly different from the ART patients: the controls were much younger (28.2 vs. 32.6 years for ICSI and 34.1 for IVF), were less likely to live in a metropolitan area, consisted of a smaller percentage of whites, and had a higher mean parity. Although they attempted to control for age and parity in their analysis, they could not adjust for socioeconomic status and environmental differences. Many birth defects have been linked to nutritional deficiencies and environmental teratogens, which may differ between the geographic areas. Thus, the controls may not represent the same population, and this may have biased the results. Second, they attempted to control for increased diagnostic vigilance among ART infants, which may have resulted in an increased rate of MBD detection. However, the potential for a detection bias cannot be entirely eliminated.Finally, the clinical significance of these findings should be considered. Of note, many of the MBD are amenable to medical or surgical correction and do not have significant long-term sequelae. It would be useful to compare the rates of both clinically severe and correctable MBD between the groups. Such information would be useful in counseling couples considering ART. Hansen et al. (2Hansen M. Kurinczuk J.J. Bower C. Webb S. The risk of major birth defects after intracytoplasmic sperm injection and in vitro fertilization.N Engl J Med. 2002; 346: 725-730Crossref PubMed Scopus (916) Google Scholar) investigated the incidence of birth defect after ICSI and IVF in Western Australia between 1993 and 1997, and found an increase in major birth defects (MBD) diagnosed by 1 year of age when compared to a population control group, overall adjusted odds ratio of 2.0 (95% CI 1.3,3.2). Also, the study demonstrated that children conceived with ART were more likely to have multiple MBD. They suggest that the increase in risk may be attributed to the treatment; however, they acknowledge that confounding due to underlying causes of infertility could not be eliminated. An appropriate control population of babies born to infertile women achieving pregnancy without the use of ART would have eliminated this major weakness. Beyond this, the study has several potential problems. Most notably, the control population was significantly different from the ART patients: the controls were much younger (28.2 vs. 32.6 years for ICSI and 34.1 for IVF), were less likely to live in a metropolitan area, consisted of a smaller percentage of whites, and had a higher mean parity. Although they attempted to control for age and parity in their analysis, they could not adjust for socioeconomic status and environmental differences. Many birth defects have been linked to nutritional deficiencies and environmental teratogens, which may differ between the geographic areas. Thus, the controls may not represent the same population, and this may have biased the results. Second, they attempted to control for increased diagnostic vigilance among ART infants, which may have resulted in an increased rate of MBD detection. However, the potential for a detection bias cannot be entirely eliminated. Finally, the clinical significance of these findings should be considered. Of note, many of the MBD are amenable to medical or surgical correction and do not have significant long-term sequelae. It would be useful to compare the rates of both clinically severe and correctable MBD between the groups. Such information would be useful in counseling couples considering ART. Art and neurological problemsStromberg et al. (3Stromberg B. Dahlquist G. Ericson A. Finnstrom O. Koster M. Stjernqvist K. Neurological sequelae in children born after in-vitro fertilization a population-based study.Lancet. 2002; 359: 461-465Abstract Full Text Full Text PDF PubMed Scopus (448) Google Scholar) used data from Swedish registries for the period between 1982 and 1995. They compared rates of neurological problems in children born through ART and population-based controls and found an association between cerebral palsy and the use of ART. Cerebral palsy is a difficult clinical entity to define and has been associated with a long list of etiologies including prematurity and low birth weight. Although adjusting for low birth weight still resulted in an increased risk (odds ratio [OR] 2.1, 95% CI 1.0,4.5), these studies are unable to control for the myriad of conditions associated with cerebral palsy including subclinical infection in utero. In addition, the clinical impact of these findings is questionable. Doubling of a very small risk results in a minor increase in the overall incidence of cerebral palsy.They themselves conclude that a much more important risk factor is multiple gestations. Similar to the article by Schieve et al. (1Schieve L.A. Meikle S.F. Ferre C. Peterson H.B. Jeng G. Wilcox L.S. Low and very low birth weight in infants conceived with use of assisted reproductive technology.N Engl J Med. 2002; 346: 731-737Crossref PubMed Scopus (838) Google Scholar) concerning low birth weight, multiple gestations were at a higher risk for neurological complications; however, twins conceived by ART did not demonstrate an increase in risk above the controls. Thus, we again suggest that the observed associations may be due to a potential confounder, such as an underlying condition in infertile women. This, once again, highlights the importance of the use of a proper control population, which should have been babies born to infertile couples as a result of treatment other than IVF. Stromberg et al. (3Stromberg B. Dahlquist G. Ericson A. Finnstrom O. Koster M. Stjernqvist K. Neurological sequelae in children born after in-vitro fertilization a population-based study.Lancet. 2002; 359: 461-465Abstract Full Text Full Text PDF PubMed Scopus (448) Google Scholar) used data from Swedish registries for the period between 1982 and 1995. They compared rates of neurological problems in children born through ART and population-based controls and found an association between cerebral palsy and the use of ART. Cerebral palsy is a difficult clinical entity to define and has been associated with a long list of etiologies including prematurity and low birth weight. Although adjusting for low birth weight still resulted in an increased risk (odds ratio [OR] 2.1, 95% CI 1.0,4.5), these studies are unable to control for the myriad of conditions associated with cerebral palsy including subclinical infection in utero. In addition, the clinical impact of these findings is questionable. Doubling of a very small risk results in a minor increase in the overall incidence of cerebral palsy. They themselves conclude that a much more important risk factor is multiple gestations. Similar to the article by Schieve et al. (1Schieve L.A. Meikle S.F. Ferre C. Peterson H.B. Jeng G. Wilcox L.S. Low and very low birth weight in infants conceived with use of assisted reproductive technology.N Engl J Med. 2002; 346: 731-737Crossref PubMed Scopus (838) Google Scholar) concerning low birth weight, multiple gestations were at a higher risk for neurological complications; however, twins conceived by ART did not demonstrate an increase in risk above the controls. Thus, we again suggest that the observed associations may be due to a potential confounder, such as an underlying condition in infertile women. This, once again, highlights the importance of the use of a proper control population, which should have been babies born to infertile couples as a result of treatment other than IVF. ConclusionsThe three articles we reviewed have suggested an association between use of ART and increased risk of problems in the resulting children. However, the validity of these associations and certainly the implication of a cause and effect relationship are weakened by potential limitations in the quality of the data inherent to registries. Furthermore, as highlighted, a major flaw of all three studies is the lack of appropriate controls, which ideally should have been babies born to infertile couples achieving pregnancy without the use of ART.These studies should be applauded, as they have provided the groundwork and direction for future research. Carefully constructed, prospective, multicenter studies using the appropriate control patient population are clearly needed. It is imperative to determine whether use of ART results in increased risk to the children born after these procedures, as the answer may have profound clinical implications. For now we should make our patients aware of the possible associations (not risks) as part of obtaining informed consent for the treatment. However, we do not believe that altering patient selection for ART and reducing the treatment’s utilization is justified at this time. Finally, it should not be forgotten that even if the described associations are accurate, infertile couples may still choose to accept the risk, as the chances of having a normal infant remain greatly in their favor, and lack of progeny may represent a greater personal tragedy than a potentially small risk of giving birth to an affected child.Do the clinical and laboratory protocols used in ART cause adverse outcomes in infants conceived through these procedures? The present data suggest associations, but do not prove cause and effect relationships. We recommend that, until better evidence is available, our consent procedures should be modified, but our clinical practice should not. The three articles we reviewed have suggested an association between use of ART and increased risk of problems in the resulting children. However, the validity of these associations and certainly the implication of a cause and effect relationship are weakened by potential limitations in the quality of the data inherent to registries. Furthermore, as highlighted, a major flaw of all three studies is the lack of appropriate controls, which ideally should have been babies born to infertile couples achieving pregnancy without the use of ART.