Vascular Limb Defects and Maternal Age Hanah Z. Nasri*1, Marie-Noel Westgate1, Eric A. Macklin2, and Lewis B. Holmes1
Background: The prenatal diagnosis procedure chorionic villus sampling is associated with increased risk of vascular disruption limb defects. Some studies have suggested that these defects are more common among infants born to women 35 years and older while other studies have shown a correlation with younger mothers. Methods: All infants with vascular disruption defects were identified in the Active Malformations Surveillance Program at Brigham and Women’s Hospital in the years 1972–1974, 1979–2011. We compared the rate of occurrence of infants with vascular limb defects among women in theses age groups: 19, 20 to 34, and 35 years to the rate of occurrence of infants with preaxial polydactyly, adjusting for race. Infants with an identifiable cause of their defects were excluded. Results: 106 infants with vascular disruption defects and 67 with preaxial polydactyly were identified. Seventeen percent of the infants with vascular disruption defects and 25% of the infants with preaxial polydactyly were born
to women 35 and older (p 5 0.23). In contrast, 16% of the infants with vascular disruption defects were born to young mothers (19 years) compared with 6.0% of the mothers of infants with preaxial polydactyly (adjusted odds ratio vs. 351 years 5 5.3, 95% confidence interval 1.4–21, p 5 0.017). Conclusion: Women 35 years old or older did not have increased risk for having a child with vascular disruption defects, but these defects were more common among infants of young (19) mothers, compared with the preaxial polydactyly group.
Introduction
including those in pregnancies which were terminated electively because of the presence of fetal anomalies. They also noted the heterogeneity of limb deficiency phenotypes. However, others have proposed that birth defects attributed to vascular disruption are more likely among infants born to younger mothers (Lubinsky et al., 1997). Werler et al. (2009) found that infants with amniotic band defects were more likely to be born to mothers less than 25 years old and African American mothers. This report focuses on limb defects attributed to the process of vascular disruption. This term is used to describe the changes in limb structures that had formed normally, but were damaged subsequently by a process of hypoperfusion, hypoxia and hemorrhage (Bouwes Bavinck and Weaver, 1986; Los et al., 1999; Olney et al., 1995; Van Allen, 1981; Webster and Abela, 2007; Werler et al., 2009). These defects include those attributed to the: (a) amniotic band syndrome, which involves more often fingers 2, 3 and 4 (Light and Ogden, 1993); (b) terminal transverse defects, involving loss of most or all of the five digits on one or more limbs (Olney et al., 1995); (c) tapering, stiffness and loss of the distal portion of the third finger (Golden et al., 2003); (d) specific conditions, such as Poland anomaly, which have been attributed to the process of vascular disruption. In the analysis of all limb deficiencies identified among 206,244 live-born and stillborn infants and elective terminations because of fetal anomalies at this hospital in the years 1972 to 1974 and 1979 to 2011, 22% of the defects were attributed to vascular disruption (Gold et al., 2011). 21% of the infants with limb malformations were among the elective terminations of pregnancies. We have used a consecutive sample of newborn infants (including stillbirths and elective terminations) with vascular disruption limb defects identified in the Active Malformations Surveillance Program to address the hypothesis
The prenatal diagnostic procedure chorionic villus sampling (CVS) has been used extensively throughout the world as a common method of prenatal diagnosis. Beginning in 1991 (Firth et al., 1991), several clinical investigators have reported an increased frequency of several types of abnormalities in CVS-exposed infants: terminal transverse limb defects, vascular disruption limb defects, hypoplasia of cranial nerves, intestinal atresia, club foot deformity and hemangiomas (Burton et al., 1993; Golden et al., 2003; Holmes, 2009; Olney et al., 1995; Stoler et al., 1999). These defects were more common the earlier in pregnancy the procedure had been performed, particularly when performed at 9.5 weeks of gestation or earlier (Firth et al., 1991; Rodeck, 1993). In the assessment of the harmful fetal effects of CVS, some clinical investigators have raised the question as to whether advanced maternal age is a risk factor for terminal transverse limb defects and limb deficiencies (Halliday et al., 1993; Mastroiacovo and Lorenzo, 1994), unrelated to the CVS procedure. These investigators noted the importance of identifying all infants with limb deficiencies,
1 Medical Genetics Unit, MassGeneral Hospital for Children, Boston, Massachusetts 2 The Biostatistics Center, Massachusetts General Hospital; Harvard Medical School, Boston, Massachusetts.
Salary support was provided through a sub-grant from the Center for Birth Defects Research and Prevention in the Massachusetts Department of Public Health. The funding agency had no role in the conduction of the study nor was it involved in the analysis of the findings. *Correspondence to: Hanah Nasri, Medical Genetics Unit, MassGeneral Hospital for Children, 175 Cambridge Street, Room 502, Boston, MA 02114. E-mail: [email protected] Published online 2 September 2014 in Wiley Online Library (wileyonlinelibrary. com). Doi: 10.1002/bdra.23294
C 2014 Wiley Periodicals, Inc. V
Birth Defects Research (Part A) 100:760–763, 2014. C 2014 Wiley Periodicals, Inc. V
Key words: vascular disruption limb defects; maternal age
BIRTH DEFECTS RESEARCH (PART A) 100:760–763 (2014)
761
TABLE 1. Type of Defect and Age Group Cross-Tabulation
Age group 19 Type of defect
Vascular limb defect
Count Percent within “Type of defect”
Pre-axial polydactyly
Total
Count
20–34
35
Total
17
71
18
106
16.0%
67.0%
17.0%
100%
4
46
17
67
Percent within “Type of defect”
6.0%
68.7%
25.4%
100%
Count
21
117
35
173
A cross-tabulation of the two types of limb defects, vascular disruption limb defects and the control, pre-axial polydactyl along with the row percentages to highlight the frequency of occurrence of a defect based on maternal age group.
that these defects were more common among the infants of older mothers.
Materials and Methods The diagnoses recorded for all elective terminations, stillbirths, and infants identified by the Active Malformations Surveillance Program at Brigham and Women’s Hospital (1972–1974, 1979–2011) in Boston were reviewed to identify those with vascular disruption defects. These defects were defined as limb deficiencies that included: terminal transverse limb defects with and without small digit-like “nubbins,” absence of one or more fingers or toes, and a limb deficiency associated with constriction rings, loss of distal structures, syndactyly and denuded or crusted tips of affected digits with or without attached fibrous strands (amniotic band limb deformity), and limb deficiency attributed to the disruption of specific arteries, such as the subclavian artery and the internal mammary artery, as has been postulated for the Poland anomaly (Bouwes Bavinck and Weaver, 1986). Nineteen infants with vascular disruption limb defects were excluded because of a recognizable cause: prenatal exposure to misoprostol (three infants; Genest et al., 1999); substance abuse with cocaine during pregnancy, confirmed by urine toxicology (one infant); CVS (two infants); a failed dilation and curettage procedure, (one infant; Holmes, 1995); homozygous alpha thalassemia (one infant; Harmon et al., 1995); Adams-Oliver Syndrome (one infant; Lin et al., 1998); chromosome abnormalities (four infants); infants with multiple anomalies, including omphalocele, anencephaly, renal agenesis, urethral atresia and esophageal atresia (six infants). After these exclusions, there were 106 infants with vascular disruption limb defects: amniotic band syndrome (52 infants), terminal transverse limb defects with and without tiny digit-like nubbins (41 infants), Poland anomaly (five infants) and acardia (eight infants). The comparison group, infants with preaxial polydactyly, included 54 infants with this defect as an isolated
unilateral deformity, and 13 infants with preaxial polydactyly in combination with other anomalies for a total of 67 infants. The infants with preaxial polydactyly who were excluded were those with: chromosome abnormalities (seven infants); hereditary disorders, specifically achondroplasia and crossed polysyndactyly (three infants); maternal insulin-dependent diabetes mellitus (four infants) and infants with multiple anomalies, including anencephaly (one infant), oculo-auriculo-vertebral (OAV) syndrome (one infant) and VACTERL association (one infant) and one infant with an affected mother. The race of each affected infant was established in the postpartum interview with the mother. She was given a list of racial groups and asked to identify the one that was most appropriate for her and for the father of her infant. Racial designations were available for 95 of the 106 infants with vascular disruption defects (58 white, 61.1%, 14 black, 14.7%, and 23 other, 24.2%) and 52 of the 67 infants with preaxial polydactyly (34 white, 65.4%, 6 black, 11.5%, and 12 other, 23.1%). Mothers were also asked about their medications, alcohol use, cigarette smoking and illicit drug use. In the vascular disruption group 10 mothers reported smoking during pregnancy. The most these mothers reported smoking was half a pack of cigarettes per day. In the preaxial polydactyl group, eight mothers reported smoking during pregnancy; one pack per day was the highest amounted reported. For statistical analysis we compared the rate of occurrence of vascular disruption limb defects in infants born to women age 19 years or less, 20 to 34 years, and 35 years old or older by logistic regression, adjusting for selfreported race (four levels: Caucasian, Black, other known race, not reported). We used linear regression to estimate the relative odds of vascular disruption limb defects versus preaxial polydactyly between age groups. Wald estimates were used to construct 95% confidence intervals. Comparisons were not adjusted for multiple comparisons. Analyses were conducted in SAS (version 9.3, SAS Institute, Cary, NC).
762
Results As is indicated in the cross-tabulation (Table 1), 16% of the mothers of the 106 infants with vascular disruption limb defects were younger 19 years old or younger, in comparison to 6.0% of the mothers of the 67 infants with preaxial polydactyly. The adjusted odds ratio for vascular limb defect 19 compared with the 20 to 34 group was 0.62, (95%CI 0.28–1.4; p 5 0.23). 17% of the mothers with vascular disruption limb defects were 35 years old or older, in comparison to 25% of the mothers of infants with preaxial polydactyly. The adjusted odds ratio for vascular limb defect 35 compared with the 20 to 34 group was 3.3 (95%CI 0.98–11; p 5 0.053). However, the adjusted odds ratio for vascular limb defect 19 compared the 35 group was 5.3 (95%CI 1.4–21; p 5 0.017).
Discussion The findings in this study show that advanced maternal age (35 years old or older) was not associated with increased odds of vascular limb defects relative to the ages of mothers of infants with preaxial polydactyly. This means that advanced maternal age is not a factor in the association between CVS and the occurrence of the vascular disruption type of limb defects. Instead, vascular limb deficiencies were 5 times more common among the infants of younger mothers, 19 years old or younger as compared to older mothers, 35 years old or older. A strength of this analysis is the accuracy of the identification of the phenotypes of the affected infants by the Active Malformations Surveillance Program (Holmes, 2012; Nelson and Holmes, 1989). Several specific causes of vascular disruption limb defects and preaxial polydactyly had been identified and could be excluded. In addition, the procedure CVS was not performed routinely at this hospital until 1995 (Wilkins-Haug, personal communication). After 1995, the interview of each postpartum mother included the question as to whether or not she had had the CVS procedure early in her pregnancy. The preaxial location of the polydactyly had been recorded consistently in the initial examinations of the affected infants and was available to the Research Assistant surveying the medical records of all of the births. The dates of birth of all mothers were recorded as a routine in each medical record. One potential weakness of this study is that the interview and the medical records of the infants surveyed might have not identified all causes of the vascular disruption limb defect or preaxial polydactyly. For example, some affected infants with a chromosomal abnormality might not have had chromosome analysis. In addition, the mother might not have reported to the interviewer that she used misoprostol or that she had had CVS. Theoretically, the accurate identification of a pregnant woman who was using a vascoconstrictive drug, like cocaine, could have not been identified. Another limitation to this study is that the sample is not representative of the general
VASCULAR LIMB DEFECTS AND MATERNAL AGE
population; the participants used in this study were mostly from Massachusetts, and were not randomly selected. Likewise, the control group sample size was small and not chosen randomly. Our findings are not consistent with the previous observations of the studies by Halliday et al. (1993) but are consistent with Mastroiacovo and Botto (1994). Halliday et al. (1993) identified all infants with limb deficiency born in 1990 and 1991 in the state of Victoria, Australia. The exposures to CVS and amniocentesis during each pregnancy had been recorded for each infant. Infants with a chromosome abnormality and a recognized syndrome were excluded. In addition, the authors excluded infants with “recognized amniotic bands.” The reason for this exclusion was not provided. Although the studies focused on transverse limb defects, the findings showed a significant increased risk in the advanced maternal age group (>35) as compared to those without advanced maternal age (
Background: The prenatal diagnosis procedure chorionic villus sampling is associated with increased risk of vascular disruption limb defects. Some studies have suggested that these defects are more common among infants born to women 35 years and older while other studies have shown a correlation with younger mothers. Methods: All infants with vascular disruption defects were identified in the Active Malformations Surveillance Program at Brigham and Women’s Hospital in the years 1972–1974, 1979–2011. We compared the rate of occurrence of infants with vascular limb defects among women in theses age groups: 19, 20 to 34, and 35 years to the rate of occurrence of infants with preaxial polydactyly, adjusting for race. Infants with an identifiable cause of their defects were excluded. Results: 106 infants with vascular disruption defects and 67 with preaxial polydactyly were identified. Seventeen percent of the infants with vascular disruption defects and 25% of the infants with preaxial polydactyly were born
to women 35 and older (p 5 0.23). In contrast, 16% of the infants with vascular disruption defects were born to young mothers (19 years) compared with 6.0% of the mothers of infants with preaxial polydactyly (adjusted odds ratio vs. 351 years 5 5.3, 95% confidence interval 1.4–21, p 5 0.017). Conclusion: Women 35 years old or older did not have increased risk for having a child with vascular disruption defects, but these defects were more common among infants of young (19) mothers, compared with the preaxial polydactyly group.
Introduction
including those in pregnancies which were terminated electively because of the presence of fetal anomalies. They also noted the heterogeneity of limb deficiency phenotypes. However, others have proposed that birth defects attributed to vascular disruption are more likely among infants born to younger mothers (Lubinsky et al., 1997). Werler et al. (2009) found that infants with amniotic band defects were more likely to be born to mothers less than 25 years old and African American mothers. This report focuses on limb defects attributed to the process of vascular disruption. This term is used to describe the changes in limb structures that had formed normally, but were damaged subsequently by a process of hypoperfusion, hypoxia and hemorrhage (Bouwes Bavinck and Weaver, 1986; Los et al., 1999; Olney et al., 1995; Van Allen, 1981; Webster and Abela, 2007; Werler et al., 2009). These defects include those attributed to the: (a) amniotic band syndrome, which involves more often fingers 2, 3 and 4 (Light and Ogden, 1993); (b) terminal transverse defects, involving loss of most or all of the five digits on one or more limbs (Olney et al., 1995); (c) tapering, stiffness and loss of the distal portion of the third finger (Golden et al., 2003); (d) specific conditions, such as Poland anomaly, which have been attributed to the process of vascular disruption. In the analysis of all limb deficiencies identified among 206,244 live-born and stillborn infants and elective terminations because of fetal anomalies at this hospital in the years 1972 to 1974 and 1979 to 2011, 22% of the defects were attributed to vascular disruption (Gold et al., 2011). 21% of the infants with limb malformations were among the elective terminations of pregnancies. We have used a consecutive sample of newborn infants (including stillbirths and elective terminations) with vascular disruption limb defects identified in the Active Malformations Surveillance Program to address the hypothesis
The prenatal diagnostic procedure chorionic villus sampling (CVS) has been used extensively throughout the world as a common method of prenatal diagnosis. Beginning in 1991 (Firth et al., 1991), several clinical investigators have reported an increased frequency of several types of abnormalities in CVS-exposed infants: terminal transverse limb defects, vascular disruption limb defects, hypoplasia of cranial nerves, intestinal atresia, club foot deformity and hemangiomas (Burton et al., 1993; Golden et al., 2003; Holmes, 2009; Olney et al., 1995; Stoler et al., 1999). These defects were more common the earlier in pregnancy the procedure had been performed, particularly when performed at 9.5 weeks of gestation or earlier (Firth et al., 1991; Rodeck, 1993). In the assessment of the harmful fetal effects of CVS, some clinical investigators have raised the question as to whether advanced maternal age is a risk factor for terminal transverse limb defects and limb deficiencies (Halliday et al., 1993; Mastroiacovo and Lorenzo, 1994), unrelated to the CVS procedure. These investigators noted the importance of identifying all infants with limb deficiencies,
1 Medical Genetics Unit, MassGeneral Hospital for Children, Boston, Massachusetts 2 The Biostatistics Center, Massachusetts General Hospital; Harvard Medical School, Boston, Massachusetts.
Salary support was provided through a sub-grant from the Center for Birth Defects Research and Prevention in the Massachusetts Department of Public Health. The funding agency had no role in the conduction of the study nor was it involved in the analysis of the findings. *Correspondence to: Hanah Nasri, Medical Genetics Unit, MassGeneral Hospital for Children, 175 Cambridge Street, Room 502, Boston, MA 02114. E-mail: [email protected] Published online 2 September 2014 in Wiley Online Library (wileyonlinelibrary. com). Doi: 10.1002/bdra.23294
C 2014 Wiley Periodicals, Inc. V
Birth Defects Research (Part A) 100:760–763, 2014. C 2014 Wiley Periodicals, Inc. V
Key words: vascular disruption limb defects; maternal age
BIRTH DEFECTS RESEARCH (PART A) 100:760–763 (2014)
761
TABLE 1. Type of Defect and Age Group Cross-Tabulation
Age group 19 Type of defect
Vascular limb defect
Count Percent within “Type of defect”
Pre-axial polydactyly
Total
Count
20–34
35
Total
17
71
18
106
16.0%
67.0%
17.0%
100%
4
46
17
67
Percent within “Type of defect”
6.0%
68.7%
25.4%
100%
Count
21
117
35
173
A cross-tabulation of the two types of limb defects, vascular disruption limb defects and the control, pre-axial polydactyl along with the row percentages to highlight the frequency of occurrence of a defect based on maternal age group.
that these defects were more common among the infants of older mothers.
Materials and Methods The diagnoses recorded for all elective terminations, stillbirths, and infants identified by the Active Malformations Surveillance Program at Brigham and Women’s Hospital (1972–1974, 1979–2011) in Boston were reviewed to identify those with vascular disruption defects. These defects were defined as limb deficiencies that included: terminal transverse limb defects with and without small digit-like “nubbins,” absence of one or more fingers or toes, and a limb deficiency associated with constriction rings, loss of distal structures, syndactyly and denuded or crusted tips of affected digits with or without attached fibrous strands (amniotic band limb deformity), and limb deficiency attributed to the disruption of specific arteries, such as the subclavian artery and the internal mammary artery, as has been postulated for the Poland anomaly (Bouwes Bavinck and Weaver, 1986). Nineteen infants with vascular disruption limb defects were excluded because of a recognizable cause: prenatal exposure to misoprostol (three infants; Genest et al., 1999); substance abuse with cocaine during pregnancy, confirmed by urine toxicology (one infant); CVS (two infants); a failed dilation and curettage procedure, (one infant; Holmes, 1995); homozygous alpha thalassemia (one infant; Harmon et al., 1995); Adams-Oliver Syndrome (one infant; Lin et al., 1998); chromosome abnormalities (four infants); infants with multiple anomalies, including omphalocele, anencephaly, renal agenesis, urethral atresia and esophageal atresia (six infants). After these exclusions, there were 106 infants with vascular disruption limb defects: amniotic band syndrome (52 infants), terminal transverse limb defects with and without tiny digit-like nubbins (41 infants), Poland anomaly (five infants) and acardia (eight infants). The comparison group, infants with preaxial polydactyly, included 54 infants with this defect as an isolated
unilateral deformity, and 13 infants with preaxial polydactyly in combination with other anomalies for a total of 67 infants. The infants with preaxial polydactyly who were excluded were those with: chromosome abnormalities (seven infants); hereditary disorders, specifically achondroplasia and crossed polysyndactyly (three infants); maternal insulin-dependent diabetes mellitus (four infants) and infants with multiple anomalies, including anencephaly (one infant), oculo-auriculo-vertebral (OAV) syndrome (one infant) and VACTERL association (one infant) and one infant with an affected mother. The race of each affected infant was established in the postpartum interview with the mother. She was given a list of racial groups and asked to identify the one that was most appropriate for her and for the father of her infant. Racial designations were available for 95 of the 106 infants with vascular disruption defects (58 white, 61.1%, 14 black, 14.7%, and 23 other, 24.2%) and 52 of the 67 infants with preaxial polydactyly (34 white, 65.4%, 6 black, 11.5%, and 12 other, 23.1%). Mothers were also asked about their medications, alcohol use, cigarette smoking and illicit drug use. In the vascular disruption group 10 mothers reported smoking during pregnancy. The most these mothers reported smoking was half a pack of cigarettes per day. In the preaxial polydactyl group, eight mothers reported smoking during pregnancy; one pack per day was the highest amounted reported. For statistical analysis we compared the rate of occurrence of vascular disruption limb defects in infants born to women age 19 years or less, 20 to 34 years, and 35 years old or older by logistic regression, adjusting for selfreported race (four levels: Caucasian, Black, other known race, not reported). We used linear regression to estimate the relative odds of vascular disruption limb defects versus preaxial polydactyly between age groups. Wald estimates were used to construct 95% confidence intervals. Comparisons were not adjusted for multiple comparisons. Analyses were conducted in SAS (version 9.3, SAS Institute, Cary, NC).
762
Results As is indicated in the cross-tabulation (Table 1), 16% of the mothers of the 106 infants with vascular disruption limb defects were younger 19 years old or younger, in comparison to 6.0% of the mothers of the 67 infants with preaxial polydactyly. The adjusted odds ratio for vascular limb defect 19 compared with the 20 to 34 group was 0.62, (95%CI 0.28–1.4; p 5 0.23). 17% of the mothers with vascular disruption limb defects were 35 years old or older, in comparison to 25% of the mothers of infants with preaxial polydactyly. The adjusted odds ratio for vascular limb defect 35 compared with the 20 to 34 group was 3.3 (95%CI 0.98–11; p 5 0.053). However, the adjusted odds ratio for vascular limb defect 19 compared the 35 group was 5.3 (95%CI 1.4–21; p 5 0.017).
Discussion The findings in this study show that advanced maternal age (35 years old or older) was not associated with increased odds of vascular limb defects relative to the ages of mothers of infants with preaxial polydactyly. This means that advanced maternal age is not a factor in the association between CVS and the occurrence of the vascular disruption type of limb defects. Instead, vascular limb deficiencies were 5 times more common among the infants of younger mothers, 19 years old or younger as compared to older mothers, 35 years old or older. A strength of this analysis is the accuracy of the identification of the phenotypes of the affected infants by the Active Malformations Surveillance Program (Holmes, 2012; Nelson and Holmes, 1989). Several specific causes of vascular disruption limb defects and preaxial polydactyly had been identified and could be excluded. In addition, the procedure CVS was not performed routinely at this hospital until 1995 (Wilkins-Haug, personal communication). After 1995, the interview of each postpartum mother included the question as to whether or not she had had the CVS procedure early in her pregnancy. The preaxial location of the polydactyly had been recorded consistently in the initial examinations of the affected infants and was available to the Research Assistant surveying the medical records of all of the births. The dates of birth of all mothers were recorded as a routine in each medical record. One potential weakness of this study is that the interview and the medical records of the infants surveyed might have not identified all causes of the vascular disruption limb defect or preaxial polydactyly. For example, some affected infants with a chromosomal abnormality might not have had chromosome analysis. In addition, the mother might not have reported to the interviewer that she used misoprostol or that she had had CVS. Theoretically, the accurate identification of a pregnant woman who was using a vascoconstrictive drug, like cocaine, could have not been identified. Another limitation to this study is that the sample is not representative of the general
VASCULAR LIMB DEFECTS AND MATERNAL AGE
population; the participants used in this study were mostly from Massachusetts, and were not randomly selected. Likewise, the control group sample size was small and not chosen randomly. Our findings are not consistent with the previous observations of the studies by Halliday et al. (1993) but are consistent with Mastroiacovo and Botto (1994). Halliday et al. (1993) identified all infants with limb deficiency born in 1990 and 1991 in the state of Victoria, Australia. The exposures to CVS and amniocentesis during each pregnancy had been recorded for each infant. Infants with a chromosome abnormality and a recognized syndrome were excluded. In addition, the authors excluded infants with “recognized amniotic bands.” The reason for this exclusion was not provided. Although the studies focused on transverse limb defects, the findings showed a significant increased risk in the advanced maternal age group (>35) as compared to those without advanced maternal age (
