Study population

The NBDPS is an ongoing study that is based on birth defects surveillance systems in the following states: Arkansas, California, Georgia/Centers for Disease Control and Prevention, Iowa, Massachusetts, New Jersey (through 2002), New York, North Carolina (beginning 2003), Texas, and Utah (beginning 2003). Case infants who were selected for the study had at least 1 of >30 eligible birth defects and were liveborn, stillborn, or electively terminated. Case records were reviewed systematically by clinical geneticists to exclude case infants with recognized or strongly suspected single-gene conditions or chromosomal abnormalities. Control infants were liveborn infants without birth defects who were selected randomly either from birth certificates or hospital birth records. Mothers were interviewed in either English or Spanish by telephone 6 weeks to 24 months after the estimated date of delivery with the use of a computer-based questionnaire. Interviewers obtained information on maternal demographic characteristics, exposures (eg, nutritional, behavioral, or occupational), and medication use both before and during pregnancy. Interview participation rates were 70% among mothers of case infants and 67% among mothers of control infants. The NBDPS was approved by the institutional review boards of the Centers for Disease Control and Prevention and the participating study centers.

Clinical information on case infants was reviewed by a team of clinical geneticists and clinicians with expertise in pediatric cardiology. Case infants were classified as having an isolated birth defect if they had (a) 1 major birth defect only; (b) 1 major birth defect and ≥1 minor birth defects; (c) ≥1 major birth defects that affect 1 organ system only; or (d) 1 major birth defect with a well-described sequence of related defects and no major unrelated birth defects. Case infants were classified as having multiple birth defects if they had ≥2 major unrelated defects in different organ systems. For case infants with a congenital heart defect (CHD), an additional layer of classification was used to denote “simple” cases as anatomically discrete or having a well-recognized single malformation (eg, hypoplastic left heart syndrome or tetralogy of Fallot).

Definitions of exposures and covariates

All information was self-reported during the maternal telephone interviews. Mothers reported whether a physician had diagnosed them previously with preexisting diabetes mellitus or gestational diabetes mellitus. Based on such reports, we classified case and control infants into 1 of 4 mutually exclusive categories: (1) infant of a mother with preexisting diabetes mellitus, if the mother reported having been diagnosed with type 1 or type 2 diabetes mellitus before the estimated date of conception of the index infant; (2) infant of a mother with gestational diabetes mellitus, if the mother reported having been diagnosed with gestational diabetes mellitus during the index pregnancy; (3) infant of a nondiabetic mother, if the mother reported never having been diagnosed with any type of diabetes mellitus; and (4) unknown, if the response was missing or inconstant (maternal report of preexisting diabetes mellitus diagnosed during the index pregnancy). The current analyses covered only infants of mothers who had been classified into categories 1 and 3.

Mothers were asked about their use of a multivitamin, prenatal vitamin, or single-component vitamin, including information on the product brand used, start and stop dates (and/or duration of use), and frequency of use. If exact dates of use were unknown, mothers could report less specific information, such as a pregnancy month (eg, first month of pregnancy) or time of year (eg, beginning of the year), which was converted into dates to determine the timing of the use in relation to the pregnancy. NBDPS investigators determined whether the specific product that was reported contained folic acid. Periconceptional users of vitamins or supplements that contain folic acid were identified as mothers who reported any use during the month before conception or during the first 3 months of pregnancy. Those who reported no use during the entire time period from 1 month before conception through the end of the first trimester were considered nonusers. Mothers with an unknown intake, those who began intake after the end of the first trimester, or those who began (and ended) intake before the start of the month before conception were excluded from these analyses.

Several covariates were considered potential confounders. Self-reported prepregnancy height and weight were converted to metric units and maternal body mass index was calculated as weight in kilograms divided by height in square meters (kg/m ² ). Four body mass index groups were formed: (1) underweight (<18.5 kg/m ² ), (2) normal weight (18.5-24.9 kg/m ² ), (3) overweight (25.0-29.9 kg/m ² ), and (4) obese (≥30.0 kg/m ² ). Additional variables included maternal age (<20, 20-24, 25-29, 30-34, and ≥35 years), maternal race or ethnicity (non-Hispanic white, non-Hispanic black or African American, Hispanic, and other race or ethnicity), timing of entry into prenatal care (≤10 weeks gestation vs later), annual household income (≥$40,000 vs less), and parity (first vs subsequent livebirth).

Exclusions

Case and control infants who were delivered during the period from October 1, 1997, through December 31, 2004, were eligible for this study. We restricted the analysis to case and control mothers with preexisting diabetes mellitus (type 1 or type 2) with a known date of diagnosis (month and year) before the index pregnancy and mothers with no diabetes mellitus of any type. Of the 16,419 case mothers and 5958 control mothers who participated in the NBDPS, 1313 case and 386 control mothers with gestational diabetes mellitus or with unknown or inconsistent diabetes mellitus status were excluded, as were 441 case and 147 control mothers who were neither definitive users nor nonusers of vitamins or supplements that contain folic acid during the period of 1 month before pregnancy through the third month of pregnancy. Because 56 case and 12 control mothers met both exclusion criteria, the final analyses comprised 14,721 case mothers and 5437 control mothers. Control infants who were included in the analyses of hypospadias were restricted to male infants only.

Statistical analysis

We conducted multiple logistic regressions using the covariates that were described previously to estimate relative risks with adjusted odds ratios and 95% confidence intervals. We evaluated the independent and joint effects of preexisting diabetes mellitus and the absence of the periconceptional intake of vitamins or supplements that contain folic acid by comparing the risk for birth defects among 4 mutually exclusive groupings of mothers: (1) mothers without diabetes mellitus with periconceptional intake of vitamins or supplements that contain folic acid (reference group), (2) mothers without diabetes mellitus with no periconceptional intake of vitamins (the “independent” effect of no periconceptional intake), (3) mothers with preexisting diabetes mellitus with periconceptional intake of vitamins that contain folic acid (the “independent” effect of preexisting diabetes mellitus), and (4) mothers with preexisting diabetes mellitus with no periconceptional intake of vitamins (joint effect). To assess whether there was an interaction between diabetes mellitus and a lack of intake of vitamins that contain folic acid that departed from additivity of effects, we calculated the relative excess risk due to interaction (RERI) and its 95% confidence interval. The confidence intervals were calculated based on Taylor expansions of the variances and covariances from the multiple logistic regression models. A spreadsheet (Microsoft Excel, version 1997-2003; Microsoft Corporation, Redmond, WA) that was developed by Andersson et al and available at www.epinet.se facilitated these calculations. RERI estimates greater than zero suggested superadditive effects, although estimates equal to zero suggested additive effects only.

We assessed the sensitivity of our results to certain exclusions and definitions by looking at changes in estimates associated with (1) the exclusion of multiple gestations and a first-degree family history of birth defects (2 strong, but uncommon, risk factors), (2) changes in the definition of periconceptional intake of vitamins or supplements that contain folic acid (ie, changes in the gestational months of periconceptional use), and (3) the restriction of the analyses to cases of isolated birth defects only. All analyses were conducted with SAS software (version 9.2; SAS Institute, Inc, Cary, NC).