Materials and Methods
The Eunice Kennedy Shriver National Institute of Child Health and Human Development Stillbirth Collaborative Research Network (SCRN) conducted a population-based case-control study of stillbirth, with participant enrollment at the time of delivery from March 2006 through September 2008. The study design and methods have been described in detail. The study was approved by the institutional review board of each clinical site and the data-coordinating center. An advisory board reviewed the progress and safety of the study. All participants gave written informed consent.
A stillborn fetus was defined by Apgar scores of 0 at 1 and 5 minutes, and no signs of life by direct observation. Deliveries resulting from the termination of a live fetus were excluded. Residents of a SCRN catchment area were eligible for participation. Catchment areas were defined by state and county boundaries and included portions of 5 states: Rhode Island, Massachusetts, Georgia, Texas, and Utah. Daily surveillance and enrollment occurred at 59 tertiary care and community hospitals servicing the catchment areas. These hospitals deliver >80,000 infants per year. Attempts were made to enroll all stillbirths and a representative sample of live births with oversampling of preterm births and non-Hispanic black women to ensure adequate numbers for stratified analyses.
Study components included a comprehensive standardized fetal postmortem examination and uniform placental pathology evaluation performed by a perinatal pathologist in both stillbirths and live births. A standardized maternal interview during the delivery hospitalization and detailed chart abstraction of prenatal office visits, antepartum hospitalizations, and the delivery hospitalization was performed. Biospecimens collected included maternal blood for serum and DNA, fetal blood from the umbilical cord (when available), placental samples, and fetal tissue (in cases). Maternal serum was stored at –80°C for 2-5 years prior to assay.
Gestational age was determined by the best clinical estimate. Multiple sources were utilized including assisted reproduction with documentation of the day of ovulation or embryo transfer, first day of the last menstrual period, and obstetrical sonograms. Fetal deaths at 18 or 19 weeks and without good dating were included in the study to ensure inclusion of all possible eligible stillbirths.
Maternal serum in cases and controls was assessed for total bile acids at ARUP Laboratories, Salt Lake City, UT. Deuterated internal standards were added to the serum and proteins were precipitated. Two-dimensional high-pressure liquid chromatography was performed, followed by electrospray ionization tandem mass spectrometry with negative ion monitoring. Bile acid levels ≥10 were considered to be abnormal by ARUP Laboratories. This is a commonly used threshold to define ICP. Bile acids ≥40 μmol/L also were assessed since this level has been associated with an increase in the risk for stillbirth in some studies. Finally, receiver operating characteristic (ROC) analysis was done to determine if there is another threshold that might optimally differentiate between stillbirth and live birth.
The delivery, defined as a case if there were any stillbirths delivered and as a control if all live births were delivered, was the unit of analysis. The analyses were weighted for oversampling of live births and other aspects of the study design as well as for differential consent among the women with stillbirth and among the women with live birth using SUDAAN software, version 11.0. Construction of the weights has been previously described. Weighted samples of live births were used in the analyses to approximate random selection of live births in the catchment areas over the enrollment period since all live births could not be enrolled.
Crude and adjusted stillbirth odds ratios and 95% confidence intervals (CIs) were calculated from univariate and multivariable logistic regression models using thresholds for bile acids. Additionally, analysis of variance was used to compare mean levels of bile acids for stillbirth to live birth groups with and without covariate adjustment. The distribution of bile acids was found to be highly skewed to the right. Thus, analysis of variance was conducted on the original scale (x), for completeness, and on a natural logarithmic scale (log e [x+1]). Results on the logarithmic scale were back transformed to geometric means and 95% CI to reflect the central tendency of the data. Conclusions from the analysis of variance are appropriately drawn from the results on the logarithmic scale and the corresponding geometric means. All tests were performed at a nominal significance level of α = 0.05. All single degree of freedom tests were 2-sided.
The adjusted analyses accounted for stillbirth risk factors known at pregnancy confirmation (baseline) using a risk factor score for stillbirth that was developed on the logit scale using the coefficients from a logistic regression model. The model used data on all SCRN deliveries where a maternal interview was conducted and a prenatal chart was abstracted. Variables contributing to the baseline risk factor score were those described previously and included the following maternal characteristics: age, race and ethnicity, marital status, education, pregnancy history, body mass index, smoking status, alcohol use, illicit drug use, hypertension, diabetes, seizure disorder, blood type, Rh factor, and multiple gestation in current pregnancy, as well as paternal age, family income, insurance and method of payment, and clinical site. All variables included in the score were categorical and an “average” of the regression coefficients associated with the categories was used when a variable was missing for an observation. The average was based on the sample-weighted proportion of live births by category.
Subanalysis was also performed in stillbirths excluding those with major anomalies and obstetric complications as previously described. The Initial Causes of Fetal Death (INCODE) system developed by the SCRN was used to assign causes of death to each case. Some analyses were performed on subsets of cases and controls, including nonanomalous stillbirths and stillbirths without obstetric complications. Nonanomalous stillbirths excluded those stillbirths with possible or probable causes of stillbirth that included fetal genetic, structural, and karyotypic abnormalities. Cases with possible or probable causes of stillbirth including fetal maternal hemorrhage, cervical insufficiency, preterm labor, preterm premature rupture of membranes, clinical chorioamnionitis, intrapartum death, abruption, complications of multiple gestation, and uterine rupture were excluded in the subset of stillbirths without obstetric complications.
Results
The flow of patients included in the study is shown in the Figure . There were a total of 581 women with stillbirth and 1546 with live births who had adequate data and blood available for assessment of bile acid levels. Cases that did and did not enroll in the study were similar with regard to maternal age, maternal race/ethnicity, insurance/method of payment, and gestational age at delivery. Women with live births who did not enroll differed from those who enrolled by maternal race/ethnicity and gestational age at delivery.
The characteristics of women enrolled in the study with and without testing for bile acids are shown in Table 1 . Among women with stillbirth, those without testing (eg, refusing blood draw) were more likely to be non-Hispanic black and nulliparous than those with testing. Women with live births who were non-Hispanic black, <20 or ≥40 years of age, and did not receive early prenatal care were less likely to have test results.
| Characteristic–weighted % | Stillbirths | Live births | ||||
|---|---|---|---|---|---|---|
| Testing | Testing | |||||
| Yes | No | P value | Yes | No | P value | |
| Unweighted sample size, N | 581 | 82 | 1546 | 386 | ||
| Weighted sample size, N w | 579 | 84 | 1183 | 256 | ||
| Maternal age at delivery, y | ||||||
| <20 | 13.0% | 14.3% | .6075 | 9.5% | 13.9% | .0284 |
| 20-34 | 70.2% | 66.1% | 76.6% | 71.4% | ||
| 35-39 | 11.8% | 16.3% | 12.2% | 10.6% | ||
| ≥40 | 5.0% | 3.2% | 1.7% | 4.1% | ||
| Maternal race/ethnicity | ||||||
| White, non-Hispanic | 35.8% | 17.5% | .0004 | 47.3% | 38.9% | <.0001 |
| African American, non-Hispanic | 20.7% | 41.2% | 9.4% | 22.5% | ||
| Hispanic | 36.9% | 32.3% | 36.1% | 29.2% | ||
| Other | 6.6% | 9.0% | 7.2% | 9.4% | ||
| Marital status/cohabitating | ||||||
| Not married or cohabitating | 25.1% | 27.3% | .5635 | 14.9% | 17.3% | .4900 |
| Cohabitating | 25.3% | 30.2% | 23.7% | 25.1% | ||
| Married | 49.6% | 42.6% | 61.4% | 57.7% | ||
| Maternal education | ||||||
| 0-11 (None/primary/some secondary) | 23.4% | 26.8% | .6022 | 18.3% | 18.4% | .8478 |
| 12 (Completed secondary) | 29.3% | 32.5% | 25.6% | 27.3% | ||
| ≥13 (College) | 47.4% | 40.6% | 56.1% | 54.3% | ||
| Insurance/method of payment | ||||||
| No insurance | 5.3% | 10.0% | .2482 | 3.6% | 3.1% | .4770 |
| Any public/private assistance | 53.7% | 52.3% | 47.9% | 52.1% | ||
| VA/commercial health insurance/HMO | 40.9% | 37.8% | 48.5% | 44.8% | ||
| Income | ||||||
| Only public/private assistance | 9.3% | 3.7% | .3708 | 5.9% | 5.9% | .7436 |
| Assistance and personal income | 37.9% | 37.7% | 37.0% | 39.6% | ||
| Only personal income | 52.8% | 58.7% | 57.1% | 54.4% | ||
| No first- or second-trimester prenatal care | 7.3% | 12.5% | .1292 | 2.5% | 4.7% | .0328 |
| Gestational age, mo | ||||||
| 18-19 | 2.3% | 3.7% | .1593 | 0.0% | 0.0% | .2648 |
| 20-23 | 32.0% | 46.2% | 0.4% | 0.3% | ||
| 24-27 | 15.9% | 15.4% | 0.7% | 1.0% | ||
| 28-31 | 12.9% | 11.3% | 1.0% | 1.2% | ||
| 32-26 | 19.8% | 11.6% | 8.2% | 10.6% | ||
| ≥37 | 17.0% | 11.8% | 89.8% | 86.9% | ||
| Nulliparous | 43.4% | 56.1% | .0373 | 34.6% | 38.7% | .2342 |
| Multifetal pregnancy | 6.5% | 4.2% | .4069 | 1.7% | 2.9% | .2440 |
a Results shown are weighted for study design and differential consent based on characteristics recorded in screened population. Overall sample sizes are given, unweighted and weighted. Sample sizes for weighted percentages vary slightly by characteristic.
Table 2 depicts levels of maternal bile acids in all stillbirths compared to all live births and to term live births. Bile acids tended to be slightly higher in women with stillbirth (geometric mean [95% CI] = 3.2 [3.0–3.5]) compared to live births (2.9 [2.7–3.1], P = .0327). However the difference was not significant after adjustment for baseline risk factors for stillbirth ( P = .23). The proportion of women with bile acid levels ≥10 μmol/L was similar in stillbirths and live births and the same was true for women with levels ≥40 μmol/L. Similar results were noted when all women with stillbirths were compared to women with term live births.
| Bile acid levels, μmol/L | Stillbirths | Live births | Term live births | Stillbirths vs live births d | Stillbirths vs term live births d | ||
|---|---|---|---|---|---|---|---|
| Unadjusted | Adjusted e | Unadjusted | Adjusted e | ||||
| Unweighted sample size, N | 581 | 1546 | 1163 | ||||
| Weighted sample size, N w | 579 | 1183 | 1062 | ||||
| Arithmetic mean (95% CI) | 5.3 (4.2–6.3) | 4.4 (3.9–4.8) | 4.2 (3.8–4.7) | P = .1211 | P = .2947 | P = .0801 | P = .2133 |
| Geometric mean (95% CI) c | 3.2 (3.0–3.5) | 2.9 (2.7–3.1) | 2.9 (2.7–3.0) | P = .0327 | P = .2271 | P = .0171 | P = .1387 |
| Median (minimum-maximum) | 2.5 (0-193) | 2.0 (0-108) | 2.0 (0-108) | ||||
| Cut points: | OR (95% CI) | OR (95% CI) | |||||
| ≥5 μmol/L | 32.3% | 25.0% | 24.6% | 1.43 (1.14–1.79) | 1.27 (1.00–1.61) | 1.46 (1.16–1.84) | 1.31 (1.02–1.68) |
| ≥10 μmol/L | 9.2% | 7.9% | 7.2% | 1.19 (0.83–1.69) | 1.01 (0.69–1.49) | 1.30 (0.90–1.89) | 1.10 (0.73–1.66) |
| ≥40 μmol/L | 1.0% | 0.8% | 0.7% | 1.25 (0.46–3.42) | 1.09 (0.42–2.82) | 1.43 (0.50–4.07) | 1.12 (0.41–3.09) |
a Term = gestational age ≥37 wk at birth
b Results shown are weighted for study design and differential consent based on characteristics recorded on all eligible pregnancies that were screened for study–overall sample sizes are given, unweighted and weighted
c Geometric mean (95% CI) is based on natural logarithmic transformation of bile acids values, log e (x+1), where bile acids value is denoted by x
d P value from weighted analysis of variance
e Adjusted P values and OR account for prepregnancy risk factors for stillbirth via score developed on logit scale and using data on all participants to Stillbirth Collaborative Research Network. Variables contributing to baseline risk factor score included following: maternal characteristics: age, race/ethnicity, marital status, education, pregnancy history, body mass index, smoking status, alcohol use, illicit drug use, hypertension, diabetes, seizure disorder, blood type, Rh factor, and multiple gestation in current pregnancy, as well as paternal age, family income, insurance/method of payment, and clinical site. Score is available on 552 of 581 stillbirth pregnancies, 1498 of 1546 live birth pregnancies, and 1131 of 1163 term live birth pregnancies.
Comparisons of nonanomalous stillbirths and term live births and nonanomalous stillbirths without obstetric complications and term live births are shown in Table 3 . When just the subset of nonanomalous stillbirths were compared to term live births, higher levels of bile acids were noted in mothers delivering stillbirths (geometric mean [95% CI] = 3.3 [3.0–3.6]) compared to live births (2.9 [2.7–3.0], P = .0202). Again, the difference was not significant after multivariable analysis ( P = .14) and groups were similar with regard to the proportion of women with bile acid levels ≥10 and ≥40 μmol/L. Finally, women with stillbirths not associated with either fetal anomalies or obstetric complications were compared to those with term live births. Mean levels of bile acid were higher in women with stillbirth compared to live birth (geometric mean [95% CI] = 3.4 [3.0–3.8] vs 2.9 [2.7–3.0], P = .0152, unadjusted; P = .06 adjusted). However, the percentage of women in both groups with bile acid levels ≥10 and ≥40 μmol/L were similar.
| Bile acid levels, μmol/L | Nonanomalous stillbirths | Nonanomalous stillbirths without complications | Term live births | Nonanomalous stillbirths vs term live births d | Nonanomalous stillbirths without complications vs term live births d | ||
|---|---|---|---|---|---|---|---|
| Unadjusted | Adjusted e | Unadjusted | Adjusted e | ||||
| Unweighted sample size, N | 390 | 273 | 1163 | ||||
| Weighted sample size, N w | 387 | 268 | 1062 | ||||
| Arithmetic mean (95% CI) | 5.7 (4.1–7.2) | 6.2 (4.0–8.4) | 4.2 (3.8–4.7) | P = .0846 | P = .1796 | P = .0829 | P = .1214 |
| Geometric mean (95% CI) c | 3.3 (3.0–3.6) | 3.4 (3.0–3.8) | 2.9 (2.7–3.0) | P = .0202 | P = .1443 | P = .0152 | P = .0614 |
| Median (minimum-maximum) | 2.5 (0-193) | 2.5 (0-193) | 2.0 (0-108) | ||||
| Cut points: | OR (95% CI) | OR (95% CI) | |||||
| ≥5 μmol/L | 32.2% | 33.5% | 24.6% | 1.45 (1.12–1.89) | 1.27 (0.96–1.69) | 1.54 (1.15–2.07) | 1.36 (0.99–1.86) |
| ≥10 μmol/L | 10.2% | 10.7% | 7.2% | 1.47 (0.98–2.22) | 1.20 (0.76–1.89) | 1.54 (0.97–2.44) | 1.29 (0.78–2.15) |
| ≥40 μmol/L | 1.4% | 1.7% | 0.7% | 2.15 (0.75–6.12) | 1.66 (0.60–4.58) | 2.58 (0.85–7.84) | 2.28 (0.79–6.56) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
