The prevalence of diabetes mellitus has increased worldwide and women of childbearing age are widely affected.1 Pregnancies complicated by pre-existing diabetes are, especially if inadequately regulated, at high risk of adverse outcomes, including congenital anomalies in general.2 In diabetic pregnancies, the placenta may develop pathologies, including anatomical placental abnormalities, increased measures of angiogenesis, and increased placental weight.3 Biochemical alterations including hyperandrogenism,4 hyperglycemia, and hyperinsulinemia5,6 can be found in cord blood. These alterations may affect critical periods of fetal development, and animal studies have revealed detrimental effects on testicular structure and function as well as reproductive hormones in offspring.7
Hypospadias (ventral displacement of the urethral meatus) and cryptorchidism (undescended testis) are the most common genital anomalies in boys. Hypospadias occurs due to incomplete fusion of the urethral folds during first trimester, whereas cryptorchidism results from disruption of the 2-phased process of testicular descent during gestational weeks 8–35. Previous studies on the association between maternal diabetes and hypospadias are conflicting; some indicate a higher risk of hypospadias,2,8–12 while others do not.13–25 Diabetes may also be a risk factor for cryptorchidism26 but most studies,10,14,27–36 including a recent meta-analysis,37 found no association. The majority of studies have assessed pregestational and gestational diabetes in combination. Yet gestational diabetes normally appear after organogenesis and, considering the different vulnerable time windows for the 2 genital anomalies, the associations could well differ between types of diabetes, as well as between the 2 genital anomalies.
We therefore hypothesized that pregestational diabetes was associated with higher occurrence of both genital anomalies, and that gestational diabetes was associated with higher occurrence of cryptorchidism; additionally, that poor glycemic control further increased the risks. We studied these associations in a large population-based cohort study while considering type and severity of diabetes, as well as timing of diagnosis in relation to birth.
METHODS
Study Population
We identified all live-born singleton boys in Denmark between 1 January 1978 and 31 December 2012 and Sweden between 1 January 1987 and 31 December 2012 from the Danish38 and Swedish Medical Birth Registers.39 To be eligible, mothers and sons had to have valid unique personal identification numbers that were used to link individual-level information from health registers.
Ascertainment of Exposure
Information on diabetes mellitus came from the Swedish Medical Birth Register,39 the Danish National Patient Register,40 and the Danish National Diabetes Register.41 In Denmark and Sweden, the health care system is government-funded and ensures equal and free of charge access to hospital care. The responsible doctor register all diagnoses associated with each hospital contact, and reporting is mandatory by Danish and Swedish legislation.
During the study period, 3 versions of the International Classification of Diseases (ICD) were used: the ICD-8 (Denmark; 1978–1993), ICD-9 (Sweden; 1987–1996), and ICD-10 (Denmark; 1994–2012 and Sweden; 1997–2012) (eTable 1; https://links.lww.com/EDE/B295). From 1978 to 1986 in Denmark and 1987 to 1996 in Sweden, type 1- and type 2-diabetes were recorded with the same code (Denmark: ICD-8: 250 and Sweden: ICD-9: 250 or 648A). From 1987 to 1993 in Denmark, the ICD-8 separated type 1-diabetes (ICD-8: 249) and type 2-diabetes (ICD-8: 250). From 1994 to 2012 in Denmark and 1997 to 2012 in Sweden, the ICD-10 differentiated between type 1-diabetes (E10 or O240) and type 2-diabetes (E11 or O241). Gestational diabetes has been coded independently throughout the study period (ICD-8 [Denmark; 1978–1993]: 63474, Y6449, ICD-9 [Sweden; 1987–1996]: 648W and ICD-10 [Denmark; 1994–2012 and Sweden; 1997– 2012]: O244, O249).
The Danish National Diabetes Register41 holds information on diabetes from 1995 and includes ICD-10 diagnosis (not gestational diabetes), ≥2 prescription redemptions of insulin or oral antidiabetic drugs from the Register of Medicinal Product Statistics and chiropody for diabetic patients, 5 glucose measurements within 1 year or 2 blood glucose measurements per year in 5 consecutive years from The National Health Insurance Service Registry. We did not use registration with glucose measurements alone, since this registration has low validity.42 Thus, we additionally classified women with ≥2 prescription redemptions of insulin as having type 1 diabetes, and ≥2 prescription redemptions of oral antidiabetic drugs and chiropody for diabetic patients as having type 2 diabetes.
The main exposure was pregestational diabetes, defined as diagnosis of either type 1 or type 2 diabetes before birth. We chose this cutoff, instead of only including diagnoses before conception, based on the rationale that pre-existing but still unregistered diabetes is reported to the register at antenatal care hospital contacts during pregnancy. We also assessed type of diabetes. If women were recorded with different diabetes types at various hospital contacts, we used the following hierarchy; type 1-, type 2-, and gestational diabetes. In Danish data, we had information on diagnoses recorded from 1978 to 2012. Using this, boys of mothers diagnosed with the unspecific diabetes mellitus code (ICD-8: 250, Denmark;1978–1986) before birth were reclassified as exposed to either type 1 or type 2 diabetes, depending on whether the mothers were diagnosed with a specific code for type 1 or type 2 diabetes later in time (eTable 1; https://links.lww.com/EDE/B295). Because we only had information on diagnoses until birth in the Swedish data, we could not reclassify individuals with an unspecific diabetes mellitus code based on later diagnoses.
As presence of diabetic complications likely, at least in part, reflects a severe or insufficiently regulated disease with poor glycemic control, we used Danish data to identify women with pregestational diabetic complications: diabetic coma, ketoacidosis, or diabetes with renal, ophthalmic, neurological, peripheral circulatory, unspecified, and multiple complications (ICD-8 [1978–1993]: 249.01–249.08, 250.01–250.08, and ICD-10 [1994–2012]: E10.0-E10.8, E11.0-E11.8 and H36.0) (eTable 1; https://links.lww.com/EDE/B295).
Ascertainment of Outcome
Information on genital anomalies was obtained from the Danish National Patient Register40 and the Swedish Patient Register.43
For the main analysis, boys with hypospadias were defined based on the following diagnostic codes; ICD-8: 75220, 75221, 75222, 75228, 75229, ICD-9: 752G and ICD-10: Q54, Q540, Q540A, Q541, Q542, Q543, Q548, and Q549. In a secondary analysis, we also considered boys who were diagnosed with hypospadias and had corrective surgery for hypospadias (eTable 2; https://links.lww.com/EDE/B295).
Boys with cryptorchidism were defined as those diagnosed with cryptorchidism (ICD-8: 75210, 75211, 75219, ICD-9: 752F and ICD-10: Q53, Q531, Q531A, Q532, Q532A or Q539) who also had corrective surgery for cryptorchidism (eTable 2; https://links.lww.com/EDE/B295).
Confounding and Mediating Factors
Based on directed acyclic graphs,44,45 we identified potential confounders and adjusted for maternal age at birth (<20, 20–29, 30–39, ≥40 years), parity (1, ≥1), maternal years of education (≤9, 10–14, ≥15 years), maternal nationality (Nordic, non-Nordic countries), calendar year at birth (Sweden, <1990, 1990–1994, 1995–1999, 2000–2005, >2005 and Denmark, >1985, 1985–1989, 1990–1994, 1995–1999, 2000–2005, >2005). Further, using Swedish data from 1992 to 2012, we additionally adjusted for maternal cigarette smoking in early pregnancy (none-smokers, <10, ≥10 cigarettes/day) and maternal body mass index (BMI) (<18.5, 18.5–24.9, 25.0–29.9, 30.0–34.9, ≥35.0 kg/m2).
Statistical Analyses
Missing Data and Multiple Imputation
We had complete information on variables in the main analyses for 97.6% of Danish boys and 97.7% of Swedish boys. To account for the few missing data, we applied multiple imputation using chained equations and imputed 50 datasets.46,47 For details on missing data and the multiple imputation method, please see eAppendix 1 (https://links.lww.com/EDE/B295).
Main Analyses
Boys were followed in the registers from birth until 31 December, 2012. Thus, the follow-up time ranged from 0 to 35 years (mean: 17.1 years) in Denmark and from 0 to 26 years (mean: 13.2 years) in Sweden. Although the genital anomalies are present at birth, a large fraction of boys are diagnosed throughout childhood, and we accounted for variations in follow-up time using time-to-event analyses. In the main analysis, we estimated adjusted hazard ratios (HRs) with 95% confidence intervals (CIs) by Cox regression in both Danish and Swedish data. We used robust standard errors to account for dependence between pregnancies of the same mother. The boys’ age was used as the time scale, and boys were followed from birth until the first diagnosis of the genital anomalies, death, emigration, or end of follow-up (31 December 2012), whichever came first. As we considered the 2 populations rather homogeneous, we calculated combined weighted estimates.
The proportional hazards assumption was visually inspected using log-minus-log-plots. We observed that the diagnostic pattern of the genital anomalies differed between exposed and unexposed. Among boys of women with pregestational diabetes, the diagnoses were registered closer to birth than among unexposed boys. Thus, the rates were not proportional in the first years of follow-up. In the main Cox regression analyses, the average HR during follow-up was estimated.48 Then, when estimating the HR in different follow-up intervals, we found that the HR of genital anomalies was higher in the first years than in the later years of follow-up, especially in the Swedish data subset. We hypothesized that this was a result of detection bias, hence, that it did not reflect the true risk but rather that pregnant diabetic woman and their offspring were followed more closely, leading to earlier diagnosis of the genital anomalies, as compared with the general population.
Therefore, to further investigate this, we used a different statistical method, the pseudo-observation method49. The pseudo-observation method estimates relative risks (RR) and does not rely on an assumption of proportional rates. It creates a transformation of the time-to-event data called pseudo-observations49 that are analyzed in a generalized linear regression model, taking death or emigration into account as competing risks.
Under the hypothesis of earlier diagnoses of the genital anomalies among boys of diabetic mothers, we would expect that the RRs were high at birth, decreased with follow-up and reached a plateau, if the follow-up was sufficiently long.
To quantify the impact on the RRs of an earlier diagnoses among boys of women with pregestational diabetes, we therefore estimated RRs at 0.5, 1, 2, …, 30 years for Denmark and at 0.5, 1, 2, …, 25 years for Sweden, using the pseudo-observation method.
Moreover, using the pseudo-observation method, we also assessed the robustness of the main results from the Cox regression model. With rare events and proportional rates, the HR is approximately equal to the RR50 and we therefore compared the HRs from the Cox regression model to the RRs obtained by the pseudo-observation method.49
Subanalyses
We conducted the following a priory planned subanalyses. We (1) performed logistic regression models to further check the robustness of the Cox regression model. Using Danish data, we (2) assessed timing of type 1 diabetes diagnosis in relation to birth. There may be some diagnostic delay of diabetes. Based on the hypothesis that women who were diagnosed shortly after pregnancy may have had a prediabetic condition during pregnancy, we thus considered timing of type 1 diabetes diagnosis categorized as no diabetes diagnosed, diagnosed pregestation, diagnosed ≤2 years after birth, diagnosed 2–5 years after birth, and diagnosed >5 years after birth. The 2 latter categories were used as a “negative control” in this analysis, with no expected increased risk. Using Danish data, we also (3) studied paternal diabetes as the exposure to explore potential confounding by genetic and family-related factors. Further, using Swedish data, we further assessed possible confounding and (4) additionally adjusted for maternal BMI and cigarette smoking among boys born from 1992 to 2012. Then, (5) as boys diagnosed with but not operated for cryptorchidism may represent true cryptorchidism with spontaneous descent, we included all diagnosed cases in the analysis and (6) we classified hypospadias as those who were both diagnosed and had surgical treatment, to assess the risk of more severe forms of hypospadias. Furthermore, (7) we repeated the analyses after excluding boys with other malformations or syndromes (ICD-8: 74000–75999, ICD-9: 740–759X, ICD-10: Q00–Q99) and studied the association between maternal diabetes and occurrence of isolated cryptorchidism and hypospadias. (8) We restricted the analyses to nulliparous women because a higher occurrence of the genital anomalies have been shown among nulliparous women and (9) performed analyses stratified by different birth year periods (Sweden, <1995, 1995–1999, 2000–2004, >2005 and Denmark, >1990, 1990–1994, 1995–1999, 2000–2005, >2005), and tested for interaction. Finally, we (10) fitted alternative imputation models and (11) performed complete case analyses. Data were analyzed using STATA 11.1 and 13.1. The study was approved by the Danish Data Protection Agency (No.2013-41-192096) and the Research Ethics Committee at Karolinska Institutet (No.2013/2192-32).
RESULTS
Main Findings
The study population consisted of 2,416,246 singleton live-born boys; 1,073,026 (44.4%) born in Denmark and 1,343,220 (55.6%) born in Sweden. Among these, 7,526 (0.7%) Danish boys and 6,086 (0.5%) Swedish boys were exposed to pregestational diabetes and 8,990 (0.8%) and 12,331 (0.9%) boys from Denmark and Sweden, respectively, were exposed to gestational diabetes. We identified 4,853 boys (4.5/1,000) with hypospadias in Denmark with a mean age at diagnosis of 2.8 years. In Sweden, 7,996 (6.0/1,000) boys were diagnosed with hypospadias with a mean age at diagnosis of 1.9 years. Further, 27,342 boys (25.5/1,000) in Denmark and 19,728 (14.7/1,000) boys in Sweden had cryptorchidism. The mean age at cryptorchidism diagnosis was 6.0 years in Denmark and 4.0 years in Sweden. Among boys with cryptorchidism in Denmark and Sweden, respectively, 17,186 (16.0/1,000) and 12,165 (9.1/1,000) underwent corrective surgery.
The proportion of mothers with pregestational diabetes increased over the study period. Diabetic women were on average older, had higher parity, shorter education, and higher BMI than nondiabetic women (Table 1).
;)
TABLE 1: Descriptive Characteristics According to Exposure Status Among 1,073,026 Danish Singleton Boys, Denmark, 1978–2012 and 1,343,220 Swedish Singleton Boys, Sweden, 1987–2012
Pregestational diabetes was associated with higher occurrence of hypospadias and cryptorchidism in both Denmark and Sweden (Tables 2, 3). Boys exposed to type 1 diabetes had the highest occurrence of hypospadias (adjusted combined HR, 1.83 [95% CI, 1.47, 2.28]) and cryptorchidism (adjusted combined HR, 1.44 [95% CI, 1.22, 1.70]). For type 2 diabetes, we observed a tendency toward higher occurrence based on few cases, and the magnitude of the association differed between countries. We observed slightly increased risks of both genital anomalies among boys exposed to gestational diabetes, highest for cryptorchidism.
TABLE 2: Hazard Ratios for Hypospadias Among 1,073,026 Danish Singleton Boys, Denmark, 1978–2012 and 1,343,220 Swedish Singleton Boys, Sweden, 1987–2012
Among women with pregestational diabetes, 21% had diabetic complications and among these, the risk of genital anomalies was even higher (Figure 1 and Table 4). For hypospadias, the adjusted HR was 2.33 (95% CI, 1.48, 3.66), and for cryptorchidism, it was 1.92 (95% CI, 1.39, 2.65).
TABLE 3: Hazard Ratios for Cryptorchidism Verified by Corrective Surgery Among 1,073,026 Danish Singleton Boys, Denmark, 1978–2012 and 1,343,220 Swedish Singleton Boys, Sweden, 1987–2012
TABLE 4: Hazard Ratios for Hypospadias and Cryptorchidism According to Maternal Pregestational Diabetic Complications Among 1,073,026 Danish Singleton Boys, Denmark, 1978–2012
FIGURE 1: Hazard ratios for cryptorchidism and hypospadias according to maternal pregestational diabetic complications among 1,073,026 Danish Singleton Boys, Denmark, 1978–2012.
We further used the pseudo-observation method to estimate RRs for the genital anomalies, comparing boys exposed to pregestational diabetes with unexposed, while also estimating how the RRs were affected by length of the follow-up time. First, we found that the main results from the Cox regression were robust to change in statistical model; the RRs at end of follow-up were comparable, yet slightly lower than the HRs both in Denmark (eFigure 1A; https://links.lww.com/EDE/B295) and Sweden (eFigure 1B; https://links.lww.com/EDE/B295). Thus, the HRs from the main analysis could be interpreted as RRs at end of follow-up.
Further, in Sweden, the RRs of genital anomalies among boys of mothers with pregestational diabetes were very high during the first years of follow-up (eFigure 1B; https://links.lww.com/EDE/B295). Then, after the first years, the RRs decreased and leveled off after 5–7 years (eFigure 1B; https://links.lww.com/EDE/B295). These patterns, although less pronounced, were also seen in the Danish data subset (eFigure 1A; https://links.lww.com/EDE/B295).
Subanalyses
In subanalyses, we (1) used logistic regression and found similar results (eTable 3; https://links.lww.com/EDE/B295), (2) assessed timing of type 1-diabetes diagnosis and also found higher occurrences of genital anomalies in boys of mothers diagnosed ≤2 years postpartum but no increased risk among boys of mothers diagnosed >2 years postpartum (Figure 2). Furthermore, we found no association between paternal diabetes and hypospadias (HR, 0.98 [95% CI, 0.66, 1.45]) and cryptorchidism (HR, 1.09 [95% CI, 0.86, 1.38]). Then, (4) adjustment for maternal cigarette smoking and BMI did not alter the associations for pregestational diabetes, but attenuated the association between gestational diabetes and hypospadias and also slightly for cryptorchidism (eTable 4; https://links.lww.com/EDE/B295). We then (5) included all boys diagnosed with cryptorchidism and found slightly stronger associations in the same direction (eTable 5; https://links.lww.com/EDE/B295) and (6) restricted analyses to hypospadias verified by surgery, and found slightly higher estimates. Furthermore, in Denmark 784 (16.2%) of the 4,853 boys with hypospadias and 4,284 (15.7%) of the 27,342 boys with cryptorchidism were also registered with genetic syndromes or other congenital abnormalities during follow-up. In Sweden, 1,677 (20.1%) of the 7,996 boys with hypospadias and 5,960 (30.2%) of the 19,728 boys with cryptorchidism had genetic syndromes or other congenital abnormalities. We then (7) restricted the analyses to isolated cases of hypospadias and cryptorchidism, which did not alter the results. Subsequently, (8) excluding multiparous women slightly attenuated the results. In (9) stratified analyses by birth year periods, the highest HRs were seen before the ICD-10 implementation, but with overlapping CI. Finally, (10) and (11) similar results were found in alternative imputation models and the complete case analyses.
FIGURE 2: Hazard ratios for cryptorchidism and hypospadias according to timing of maternal type 1 diabetes diagnosis in relation to birth among 1,073,026 Danish Singleton Boys, Denmark, 1978–2012.
DISCUSSION
In this large study of 2 Nordic populations, boys of mothers with type 1 diabetes had higher occurrence of hypospadias and cryptorchidism. Boys exposed to gestational diabetes only had some excess risk. The highest occurrence was found in boys of mothers with pregestational diabetic complications, indicating that poorly controlled diabetes is an additional risk factor. We also observed an increased risk among women diagnosed with type 1 diabetes up to 2 years postpartum, which could reflect diagnostic delay or that biochemical precursors of diabetes may have influenced the risk before an overt diabetes have been developed. Furthermore, we found no associations between the genital anomalies and maternal diabetes diagnosed more than 2 years postpartum or paternal diabetes. Taken together, these results may indicate that the observed associations could be caused by diabetes per se rather than being confounded by time stable genetic, family-related, or lifestyle factors.
The potential biologic mechanisms underlying the associations observed are unresolved but placental dysfunction or hormonal alterations following poor glycemic control could play a role. It is known that insufficient androgen action during the critical time window of genital development in early gestation may adversely affect masculinization and result in genital anomalies.51 In early gestation, the fetal testosterone production is stimulated by human chorionic gonadotropin (hCG) from the placenta.52 Fluctuations in glucose, insulin, or other hormonal levels could play a role, altering the intrauterine milieu and thus affecting placental and fetal growth. The placentas are on average larger in diabetic pregnancies53 as are the fetuses, ultimately leading to macrosomia. Yet large placentas could be a compensatory mechanism in terms of the placenta aiming at balancing the growing fetal needs. Thus, placental dysfunction and hormonal alterations in diabetic women could influence the delicate morphogenesis of the developing fetus or by disturbing the fetal hormonal balance as also previously suggested.24,54–58 Further, our results could also support a role of poor glycemic control and findings from animal studies of detrimental effect on fetal testicular function and reproductive hormones further support this hypothesis.7
Previous studies on this subject are sparse and inconsistent. Some found an association between diabetes and hypospadias,2,8–12 while others did not.13–25 Aberg et al.8 used Swedish register-based data from 1987 to 1997 and found a higher frequency of hypospadias among boys of mothers with pre-existing diabetes. Mavrogenis et al.12 used data from the Hungarian Case-Control Surveillance of Congenital Abnormalities from 1980 to 1996 and showed a higher incidence of gestational diabetes among mothers of boys with hypospadias. In contrast, Trabert et al.14 found no association in a large population of women with gestational diabetes verified by glucose tolerance tests. Consistent with our results, a large study performed by Porter et al.9 reported an odds ratio (OR) of 2.18 (95% CI, 1.03, 4.60) for boys of mothers with prepregnancy diabetes and no association with gestational diabetes. Agopian et al.10 found a slightly increased risk of hypospadias, but did not assess diabetes type. Thus, previous results for the association between maternal diabetes and hypospadias are conflicting, but the largest studies point toward an association.
On the other hand, apart from 1 finding of higher risk of cryptorchidism among boys of mothers with gestational diabetes,26 the majority of studies found no association.10,14,27–36 Recently, two large studies performed by Trabert et al.14 and Agopian et al.10 reported no association, but they were unable to distinguish between types of diabetes. A recent meta-analysis found marginally higher risk of cryptorchidism (OR = 1.21 [95% CI, 1.00, 1.46]), however, without considering type of diabetes.37 Taken together, comparison across the existing literature is challenging and lacks solid conclusions.
In this study, the findings from each of the two countries were in the same direction, thus providing some immediate replication. As the coverage of Nordic birth registers is close to complete,59 risk of selection bias is practically absent. The quality of Danish and Swedish registers are considered high, including information on pregestational diabetes,60 hypospadias,61,62 and cryptorchidism.63 Cryptorchidism registration in the Danish National Patient Register has a positive predictive value of 99% for diagnoses verified by corrective surgery63 and hypospadias registration was recently validated yielding an overall positive predictive value of 97.6%.62 Nonetheless, some misclassification is present and the National Patient Registers are unlikely to capture all cases, as the mildest may not be recognized clinically. Information on gestational diabetes is expected to be underestimated but we consider that any exposure misclassification is nondifferential and most likely will bias the results toward the null, because diabetes registration was made before outcome ascertainment.
We adjusted for several potentially confounding factors, perhaps most importantly; the results on pregestational diabetes were robust to adjustment for maternal BMI and cigarette smoking. Although some residual confounding probably exists, our results are consistent with the hypothesized link between pregestational diabetes and genital anomalies in sons.
During pregnancy, a relatively large proportion of embryos are lost by miscarriage or stillbirth, and these have more congenital malformations than live-born births.64 As our population consisted of live-born boys, we left out some of the population at risk, which may be a source of selection bias. If more fetuses with genital anomalies are lost in utero among diabetic women, the inclusion of only live births could have underestimated the association. However, neither hypospadias nor cryptorchidism is expected to increase fetal death and, if they do, only when in combination with severe malformations, and we consider live-birth bias to be a limited problem.
Another potential source of bias important to consider is that children of women with diabetes may have more thorough clinical examinations, thus that their genital anomalies are detected earlier in the postpartum period. With the large data and long follow-up available (up to 35 and 26 years for Denmark and Sweden, respectively), we were able to explore this potential bias further. During the first years of follow-up, especially in Sweden, boys of mothers with prepregnancy diabetes seemed to be diagnosed earlier than those of nondiabetic women as indicated by the high RRs at birth (eFigure 1A and 1B; https://links.lww.com/EDE/B295). This can be ascribed to detection bias. However, because of the long follow-up, the validity of our study is not compromised (eFigure 1A and 1B; https://links.lww.com/EDE/B295).
We hypothesized that boys of mothers with pregestational diabetes, and especially those with poor glycemic control, as indicated by diabetic complications, had a higher risk of genital anomalies. Furthermore, we expected no association between gestational diabetes and hypospadias, as gestational diabetes develops after the critical time window. Our results supported the hypothesis that poor glycemic control increases the risk. This interpretation was supported by the lack of association with paternal diabetes or maternal diabetes diagnosed later in life, and by the robustness when adjusting for maternal BMI and cigarette smoking. We acknowledge that the higher risk related to poor glycemic control could be explained by unmeasured maternal factors associated with the ability to control diabetes, for example, nutrition, alcohol intake, and other risk behaviors.
In conclusion, type 1 diabetes was associated with increased risks of genital anomalies, and the highest occurrence was among boys of mothers with diabetic complications. These findings are consistent with a potential influence of poor glycemic control on fetal genital development in the critical early period of organogenesis. Considering the increasing prevalence of pregestational diabetes in women of childbearing age and the challenges of poor disease control despite careful prepregnancy counseling, our findings have public health importance and provide further evidence in support of the importance of optimizing glycemic control in the periconceptional period.
REFERENCES
1. Getahun D, Nath C, Ananth CV, Chavez MR, Smulian JCGestational diabetes in the United States: temporal trends 1989 through 2004. Am J Obstet Gynecol. 2008;198:525.e1–525.e5.
2. Correa A, Gilboa SM, Besser LM, et alDiabetes mellitus and birth defects. Am J Obstet Gynecol. 2008;199:237.e1–237.e9.
3. Huynh J, Dawson D, Roberts D, Bentley-Lewis RA systematic review of placental pathology in maternal diabetes mellitus. Placenta. 2015;36:101–114.
4. Morisset AS, Dubé MC, Drolet R, et alAndrogens in the maternal and fetal circulation: association with insulin resistance. J Matern Fetal Neonatal Med. 2013;26:513–519.
5. Magee MS, Walden CE, Benedetti TJ, Knopp RHInfluence of diagnostic criteria on the incidence of gestational diabetes and perinatal morbidity. JAMA. 1993;269:609–615.
6. Berglund SK, García-Valdés L, Torres-Espinola FJ, et alPREOBE team. Maternal, fetal and perinatal alterations associated with obesity, overweight and gestational diabetes: an observational cohort study (PREOBE). BMC Public Health. 2016;16:207.
7. Jelodar G, Khaksar Z, Pourahmadi MEndocrine profile and testicular histomorphometry in adult rat offspring of diabetic mothers. J Physiol Sci. 2009;59:377–382.
8. Aberg A, Westbom L, Källén BCongenital malformations among infants whose mothers had gestational diabetes or preexisting diabetes. Early Hum Dev. 2001;61:85–95.
9. Porter MP, Faizan MK, Grady RW, Mueller BAHypospadias in Washington State: maternal risk factors and prevalence trends. Pediatrics. 2005;115:e495–e499.
10. Agopian AJ, Langlois PH, Ramakrishnan A, Canfield MAEpidemiologic features of male genital malformations and subtypes in Texas. Am J Med Genet A. 2014;164A:943–949.
11. Woud SG, van Rooij IA, van Gelder MM, et alNational Birth Defects Prevention Study. Differences in risk factors for second and third degree hypospadias in the national birth defects prevention study. Birth Defects Res A Clin Mol Teratol. 2014;100:703–711.
12. Mavrogenis S, Urban R, Czeizel AEPregnancy complications in the mothers who delivered boys with isolated hypospadias - a population-based case-control study. J Matern Fetal Neonatal Med. 2015;28:489–493.
13. Hussain N, Chaghtai A, Herndon CD, Herson VC, Rosenkrantz TS, McKenna PHHypospadias and early gestation growth restriction in infants. Pediatrics. 2002;109:473–478.
14. Trabert B, Chodick G, Shalev V, Sella T, Longnecker MP, McGlynn KAGestational diabetes and the risk of cryptorchidism and hypospadias. Epidemiology. 2014;25:152–153.
15. Källén BCase control study of hypospadias, based on registry information. Teratology. 1988;38:45–50.
16. Aschim EL, Haugen TB, Tretli S, Daltveit AK, Grotmol TRisk factors for hypospadias in Norwegian boys - association with testicular dysgenesis syndrome? Int J Androl. 2004;27:213–221.
17. Becerra JE, Khoury MJ, Cordero JF, Erickson JDDiabetes mellitus during pregnancy and the risks for specific birth defects: a population-based case-control study. Pediatrics. 1990;85:1–9.
18. Brouwers MM, van der Zanden LF, de Gier RP, et alHypospadias: risk factor patterns and different phenotypes. BJU Int. 2010;105:254–262.
19. Morera AM, Valmalle AF, Asensio MJ, et alA study of risk factors for hypospadias in the Rhône-Alpes region (France). J Pediatr Urol. 2006;2:169–177.
20. Polednak AP, Janerich DTMaternal characteristics and hypospadias: a case-control study. Teratology. 1983;28:67–73.
21. Ramos-Arroyo MA, Rodriguez-Pinilla E, Cordero JFMaternal diabetes: the risk for specific birth defects. Eur J Epidemiol. 1992;8:503–508.
22. Sørensen HT, Pedersen L, Nørgaard M, Wogelius P, Rothman KJMaternal asthma, preeclampsia and risk of hypospadias. Epidemiology. 2005;16:806–807.
23. Sun G, Tang D, Liang J, Wu MIncreasing prevalence of hypospadias associated with various perinatal risk factors in Chinese newborns. Urology. 2009;73:1241–1245.
24. Stoll C, Alembik Y, Roth MP, Dott BGenetic and environmental factors in hypospadias. J Med Genet. 1990;27:559–563.
25. Winston JJ, Meyer RE, Emch MEGeographic analysis of individual and environmental risk factors for hypospadias births. Birth Defects Res A Clin Mol Teratol. 2014;100:887–894.
26. Virtanen HE, Tapanainen AE, Kaleva MM, et alMild gestational diabetes as a risk factor for congenital cryptorchidism. J Clin Endocrinol Metab. 2006;91:4862–4865.
27. Davies TW, Williams DR, Whitaker RHRisk factors for undescended testis. Int J Epidemiol. 1986;15:197–201.
28. Berkowitz GS, Lapinski RH, Godbold JH, Dolgin SE, Holzman IRMaternal and neonatal risk factors for cryptorchidism. Epidemiology. 1995;6:127–131.
29. Berkowitz GS, Lapinski RHRisk factors for cryptorchidism: a nested case-control study. Paediatr Perinat Epidemiol. 1996;10:39–51.
30. Jones ME, Swerdlow AJ, Griffith M, Goldacre MJPrenatal risk factors for cryptorchidism: a record linkage study. Paediatr Perinat Epidemiol. 1998;12:383–396.
31. Biggs ML, Baer A, Critchlow CWMaternal, delivery, and perinatal characteristics associated with cryptorchidism: a population-based case-control study among births in Washington State. Epidemiology. 2002;13:197–204.
32. McGlynn KA, Graubard BI, Klebanoff MA, Longnecker MPRisk factors for cryptorchism among populations at differing risks of testicular cancer. Int J Epidemiol. 2006;35:787–795.
33. Main KM, Kiviranta H, Virtanen HE, et alFlame retardants in placenta and breast milk and cryptorchidism in newborn boys. Environ Health Perspect. 2007;115:1519–1526.
34. Zakaria M, Azab S, El baz M, Fawaz L, Bahagat ACryptorchidism in Egyptian neonates. J Pediatr Urol. 2013;9(6 Pt A):815–819.
35. Adams SV, Hastert TA, Huang Y, Starr JRNo association between maternal pre-pregnancy obesity and risk of hypospadias or cryptorchidism in male newborns. Birth Defects Res A Clin Mol Teratol. 2011;91:241–248.
36. Brouwers MM, de Bruijne LM, de Gier RP, Zielhuis GA, Feitz WF, Roeleveld NRisk factors for undescended testis. J Pediatr Urol. 2012;8:59–66.
37. Zhang L, Wang XH, Zheng XM, et alMaternal gestational smoking, diabetes, alcohol drinking, pre-pregnancy obesity and the risk of cryptorchidism: a systematic review and meta-analysis of observational studies. PLoS One. 2015;10:e0119006.
38. Knudsen LB, Olsen JThe Danish Medical Birth Registry. Dan Med Bull. 1998;45:320–323.
39. Odlind V, Haglund B, Pakkanen M, Otterblad Olausson PDeliveries, mothers and newborn infants in Sweden, 1973-2000. Trends in obstetrics as reported to the Swedish Medical Birth Register. Acta Obstet Gynecol Scand. 2003;82:516–528.
40. Lynge E, Sandegaard JL, Rebolj MThe Danish National Patient Register. Scand J Public Health. 2011;39(7 Suppl):30–33.
41. Carstensen B, Kristensen JK, Ottosen P, Borch-Johnsen KSteering Group of the National Diabetes Register. The Danish National Diabetes Register: trends in incidence, prevalence and mortality. Diabetologia. 2008;51:2187–2196.
42. Green A, Sortsø C, Jensen PB, Emneus MValidation of the Danish national diabetes register. Clin Epidemiol. 2015;7:5–15.
43. Socialstyrelsen. The Swedish Patient Registry. Available at:
http://www.socialstyrelsen.se/register/halsodataregister/patientregister/. Assessed 11 August 2016.
44. Robins JM, Hernán MA, Brumback BMarginal structural models and causal inference in epidemiology. Epidemiology. 2000;11:550–560.
45. Hernán MA, Hernández-Díaz S, Werler MM, Mitchell AACausal knowledge as a prerequisite for confounding evaluation: an application to birth defects epidemiology. Am J Epidemiol. 2002;155:176–184.
46. Rubin DBMultiple Imputation for Nonresponse in Surveys. 1987.New York, NY: John Wiley & Sons;
47. Sterne JA, White IR, Carlin JB, et alMultiple imputation for missing data in epidemiological and clinical research: potential and pitfalls. BMJ. 2009;338:b2393.
48. Xu R, O’Quigley JEstimating average regression effect under non-proportional hazards. Biostatistics. 2000;1:423–439.
49. Overgaard M, Andersen PK, Parner ETRegression analysis of censored data using pseudo-observations. Stata J. 2010;10:408–422.
50. Symons MJ, Moore DTHazard rate ratio and prospective epidemiological studies. J Clin Epidemiol. 2002;55:893–899.
51. Macleod DJ, Sharpe RM, Welsh M, et alAndrogen action in the masculinization programming window and development of male reproductive organs. Int J Androl. 2010;33:279–287.
52. Huhtaniemi IT, Korenbrot CC, Jaffe RBHCG binding and stimulation of testosterone biosynthesis in the human fetal testis. J Clin Endocrinol Metab. 1977;44:963–967.
53. Strøm-Roum EM, Haavaldsen C, Tanbo TG, Eskild APlacental weight relative to birthweight in pregnancies with maternal diabetes mellitus. Acta Obstet Gynecol Scand. 2013;92:783–789.
54. Fujimoto T, Suwa T, Kabe K, Adachi T, Nakabayashi M, Amamiya TPlacental insufficiency in early gestation is associated with hypospadias. J Pediatr Surg. 2008;43:358–361.
55. Boisen KA, Chellakooty M, Schmidt IM, et alHypospadias in a cohort of 1072 Danish newborn boys: prevalence and relationship to placental weight, anthropometrical measurements at birth, and reproductive hormone levels at three months of age. J Clin Endocrinol Metab. 2005;90:4041–4046.
56. Yinon Y, Kingdom JC, Proctor LK, et alHypospadias in males with intrauterine growth restriction due to placental insufficiency: the placental role in the embryogenesis of male external genitalia. Am J Med Genet A. 2010;152A:75–83.
57. Jensen MS, Wilcox AJ, Olsen J, et alCryptorchidism and hypospadias in a cohort of 934,538 Danish boys: the role of birth weight, gestational age, body dimensions, and fetal growth. Am J Epidemiol. 2012;175:917–925.
58. Arendt LH, Ramlau-Hansen CH, Wilcox AJ, Henriksen TB, Olsen J, Lindhard MSPlacental weight and male genital anomalies: a Nationwide Danish Cohort Study. Am J Epidemiol. 2016;183:1122–1128.
59. The Danish Health and Medicines Authority. The Birth Registry 2004 [in Danish]. Available at:
http://sundhedsstyrelsen.dk/publ/tidsskrifter/nyetal/pdf/2005/04_05.pdf. Assessed 12 October 2016.
60. Langhoff-Roos J, Rasmussen SValidering af Landspatientregisteret (LPR) med henblik på obstetrisk forskning og kvalitetssikring – et kvalitetsudviklingsprojekt. 2003. Available at:
http://www.statensnet.dk/pligtarkiv/fremvis.pl?vaerkid=23080&reprid=0&iarkiv=1. Accessed 27 November 2017.
61. Pedersen L, Nørgaard M, Skriver MV, Olsen J, Sørensen HTPrenatal exposure to loratadine in children with hypospadias: a nested case-control study within the Danish National Birth Cohort. Am J Ther. 2006;13:320–324.
62. Arendt LH, Ernst A, Lindhard MS, et alAccuracy of the hypospadias diagnoses and surgical treatment registrations in the Danish National Patient Register. Clin Epidemiol. 2017;9:483–489.
63. Jensen MS, Snerum TM, Olsen LH, et alAccuracy of cryptorchidism diagnoses and corrective surgical treatment registration in the Danish National Patient Registry. J Urol. 2012;188:1324–1329.
64. EUROCAT Prevalence Tables. Available at:
http://www.eurocat-network.eu/accessprevalencedata/prevalencetables. Accessed 28 June 2016.
