Hwan Chung

A longitudinal study of angiogenic (placental growth factor) and anti-angiogenic (soluble endoglin and soluble vascular endothelial growth factor receptor-1) factors in normal pregnancy and patients destined to develop preeclampsia and deliver a small for gestational age neonate

Introduction. Accumulating evidence suggests that an imbalance between pro-angiogenic (i.e., vascular endothelial growth factor (VEGF) and placental growth factor (PlGF)) and anti-angiogenic factors (i.e., soluble VEGF receptor-1 (sVEGFR-1, also referred to as sFlt1)) is involved in the pathophysiology of preeclampsia (PE). Endoglin is a protein that regulates the pro-angiogenic effects of transforming growth factor β, and its soluble form has recently been implicated in the pathophysiology of PE. The objective of this study was to determine if changes in maternal plasma concentration of these angiogenic and anti-angiogenic factors differ prior to development of disease among patients with normal pregnancies and those destined to develop PE (preterm and term) or to deliver a small for gestational age (SGA) neonate. Methods. This longitudinal nested case–control study included 144 singleton pregnancies in the following groups: (1) patients with uncomplicated pregnancies who delivered appropriate for gestational age (AGA) neonates (n = 46); (2) patients who delivered an SGA neonate but did not develop PE (n = 56); and (3) patients who developed PE (n = 42). Longitudinal samples were collected at each prenatal visit, scheduled at 4-week intervals from the first or early second trimester until delivery. Plasma concentrations of soluble endoglin (s-Eng), sVEGFR-1, and PlGF were determined by specific and sensitive ELISA. Results. (1) Patients destined to deliver an SGA neonate had higher plasma concentrations of s-Eng throughout gestation than those with normal pregnancies; (2) patients destined to develop preterm PE and term PE had significantly higher concentrations of s-Eng than those with normal pregnancies at 23 and 30 weeks, respectively (for preterm PE: p < 0.036 and for term PE: p = 0.002); (3) patients destined to develop PE (term or preterm) and those who delivered an SGA neonate had lower plasma concentrations of PlGF than those with a normal pregnancy throughout gestation, and the maternal plasma concentration of this analyte became detectable later among patients with pregnancy complications, compared to normal pregnant women; (4) there were no significant differences in the plasma concentrations of sVEGFR-1 between patients destined to deliver an SGA neonate and those with normal pregnancies; (5) patients destined to develop preterm and term PE had a significantly higher plasma concentration of sVEGFR-1 at 26 and 29 weeks of gestation than controls (p = 0.009 and p = 0.0199, respectively); and (6) there was no significant difference in the increment of sVEGFR-1 between control patients and those who delivered an SGA neonate (p = 0.147 at 25 weeks and p = 0.8285 at 40 weeks). Conclusions. (1) Changes in the maternal plasma concentration of s-Eng, sVEGFR-1, and PlGF precede the clinical presentation of PE, but only changes in s-Eng and PlGF precede the delivery of an SGA neonate; and (2) differences in the profile of angiogenic and anti-angiogenic response to intrauterine insults may determine whether a patient will deliver an SGA neonate, develop PE, or both.

Mid-pregnancy circulating cytokine levels, histologic chorioamnionitis and spontaneous preterm birth

Some spontaneous preterm deliveries (PTD) are caused by occult infections of the fetal membranes (histologic chorioamnionitis [HCA]). High levels of infection-related markers, including some cytokines, sampled from maternal circulation in mid-pregnancy have been linked to PTD, but whether these specifically identify HCA has not been established. We have tested associations between 13 Th1, Th2 and Th17 cytokines and PTD with and without HCA in a prospective cohort study. The study sample included 926 Pregnancy Outcomes and Community Health Study subcohort women; women with medically indicated PTD or incomplete data excluded. A panel of cytokines was assessed using a multiplex assay in maternal plasma collected at 15-27 weeks of gestation. Severe HCA was scored by a placental pathologist blinded to clinical variables. Multivariable polytomous logistic regression was used to estimate adjusted odds ratios (OR) per 1 standard deviation (S.D.) increase in cytokine levels using a 5 level outcome variable: PTD <35 weeks with HCA, PTD <35 weeks without HCA, PTD 35-36 weeks with HCA, PTD 35-36 weeks without HCA, and term (referent). Interleukin (IL)-1beta, IL-2, IL-12, interferon-gamma, IL-4, IL-6 and transforming growth factor-beta were all significantly associated with PTD <35 weeks with HCA, with ORs of 1.6-2.3 per S.D. increase. None of these were associated with PTD <35 weeks without HCA or PTD 35-36 weeks with HCA. Although the tissues of origin of circulating cytokines are unclear, the observed elevations across many cytokines among women who later delivered <35 weeks with HCA may represent a robust immune response to infection within gestational tissues. These results suggest that women with HCA could be identified using relatively non-invasive means.

The latent structure of post-traumatic stress disorder : tests of invariance by gender and trauma type

BACKGROUND Measurement invariance of Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV) post-traumatic stress disorder (PTSD) criterion symptoms was tested by gender and trauma type, assaultive and non-assaultive. METHOD Analysis was conducted using latent class analysis (LCA), based on findings that the three-class LCA model from Breslau et al. (Archives of General Psychiatry 2005, 62, 1343-1351) fits the data across the four groups best. The classes represent three levels of PTSD-related disturbance: no disturbance, intermediate disturbance and pervasive disturbance, with the last one approximating the DSM-IV PTSD diagnosis. RESULTS Analysis of measurement invariance showed that, with respect to gender, there was no evidence of differential symptom reporting within the same disturbance class. DSM-IV symptom indicators represent the latent structure of PTSD equally in males and females. We found that more female than male victims of assaultive violence experienced pervasive disturbance. In the absence of measurement variability associated with gender, the finding is likely to reflect a gender difference in susceptibility. The analysis of measurement invariance detected evidence of variability associated with trauma type. Victims of assaultive violence in the pervasive disturbance class report more severe distress (especially emotional numbing) than do victims of non-assaultive violence in the same class. CONCLUSIONS The finding of measurement bias associated with type of trauma raises questions about the applicability of a single definition for PTSD associated with assaultive violence and PTSD associated with traumatic events of lesser magnitude.

Difficulties in Drawing Inferences With Finite-Mixture Models: A Simple Example With a Simple Solution

Likelihood functions from finite mixture models have many unusual features. Maximum likelihood (ML) estimates may behave poorly over repeated samples, and the abnormal shape of the likelihood often makes it difficult to assess the uncertainty in parameter estimates. Bayesian inference via Markov chain Monte Carlo (MCMC) can be a useful alternative to ML, but the component labels may switch during the MCMC run, making the output difficult to interpret. Two basic methods for handling the label-switching problem have been proposed: imposing constraints on the parameter space and cluster-based relabeling of the simulated parameters. We have found that label switching may also be reduced by supplying small amounts of prior information that are asymmetric with respect to the mixture components. Simply assigning one observation to each component a priori may effectively eliminate the problem. Using a very simple example—a univariate sample from a mixture of two exponentials—we evaluate the performance of likelihood and MCMC-based estimates and intervals over repeated sampling. Our simulations show that MCMC performs much better than ML if the label-switching problem is adequately addressed, and that asymmetric prior information performs as well as or better than the other proposed methods.

Interplay of cytokine polymorphisms and bacterial vaginosis in the etiology of preterm delivery.

Recent findings suggest that the association between inflammation-related genes and preterm delivery may be stronger in the presence of bacterial vaginosis (BV). Tumor necrosis factor-alpha (TNFα) and interleukin 1-beta (IL-1β) are pro-inflammatory cytokines capable of inducing preterm labor in non-human primates. In this study the authors tested associations among two TNFα promoter polymorphisms (-G308A and -G238A), a single IL-1β polymorphism (+C3954T), vaginal microbial findings, and risk of preterm delivery. Data were from the Pregnancy Outcomes and Community Health (POUCH) Study (n=777 term and n=230 preterm deliveries). Vaginal smears collected at mid-pregnancy (15-27 weeks gestation) were scored according to Nugents criteria. A Nugent score of ≥ 4 was modeled as the cut-point for intermediate and positive BV. Logistic regression was used to estimate odds ratios for associations among independent covariates (vaginal flora, genotype) and preterm delivery. Results showed that women with a Nugent score of≥ 4 and the TNFα -238 A/G or A/A were at increased risk of delivering preterm (race/ethnicity adjusted OR 2.6, 95% CI 1.2, 5.8). The p-value for the genotype and Nugent score interaction=0.02. This study points to one more example of a potential gene-environment interaction in a preterm delivery pathway. Future tests of this finding will determine the robustness of these results.

Midpregnancy Vaginal Fluid Defensins, Bacterial Vaginosis, and Risk of Preterm Delivery

OBJECTIVE: To assess relations among midpregnancy vaginal defensin levels, a component of the host innate immune response, bacterial vaginosis, and risk of preterm delivery. These relations are compared across race groups because previous studies have repeatedly shown that the prevalence of bacterial vaginosis and the risk of preterm delivery are greater in African-American women compared with that in white women. METHODS: Data are from a prospective study that enrolled pregnant women from 52 clinics in five Michigan communities. In the study subcohort, defensins (human neutrophil peptides 1, 2 and 3) and bacterial vaginosis (Nugent criteria) were measured in vaginal fluid collected at enrollment (15th through 27th week of pregnancy) from 1,031 non-Hispanic white and African-American women (787 term, 244 preterm). Preterm deliveries were categorized by clinical circumstances, ie, spontaneous and medically indicated. RESULTS: Among African Americans, vaginal human neutrophil peptides 1–3 levels greater than or equal to the median were associated with bacterial vaginosis and specifically with spontaneous preterm delivery only (adjusted odds ratio 2.3, 95% confidence interval 1.2–4.3). Once African-American women were stratified by human neutrophil peptide 1–3 levels, bacterial vaginosis added nothing to the prediction of spontaneous preterm delivery risk. None of the above associations were observed in non-Hispanic whites. CONCLUSION: The relations among human neutrophil peptide 1–3 levels, bacterial vaginosis, and preterm delivery vary by race group. In African Americans, midpregnancy human neutrophil peptide 1–3 levels were more informative to preterm delivery risk than was bacterial vaginosis, suggesting an important role for host response. In addition, elevated human neutrophil peptide 1–3 levels may be a marker for particular high-risk vaginal milieus that are not distinguished by the current bacterial vaginosis Nugent scoring system. LEVEL OF EVIDENCE: II

On the performance of bias-reduction techniques for variance estimation in approximate Bayesian bootstrap imputation

Multiply imputed data sets can be created with the approximate Bayesian bootstrap (ABB) approach under the assumption of ignorable nonresponse. The theoretical development and inferential validity are predicated upon asymptotic properties; and biases are known to occur in small-to-moderate samples. There have been attempts to reduce the finite-sample bias for the multiple imputation variance estimator. In this note, we present an empirical study for evaluating the comparative performance of the two proposed bias-correction techniques and their impact on precision. The results suggest that to varying degrees, bias improvements are outweighed by efficiency losses for the variance estimator. We argue that the original ABB has better small-sample properties than the modified versions in terms of the integrated behavior of accuracy and precision, as measured by the root mean-square error.

Innate immune system gene polymorphisms in maternal and child genotype and risk of preterm delivery

Abstract Objective. There is little information about the combination of genetic variability in pregnant women and their children in relation to the risk of preterm delivery (PTD). In a sub-cohort of 487 non-Hispanic white and 288 African-American mother/child pairs, the Pregnancy Outcomes and Community Health Study assessed 10 functional polymorphisms in 9 genes involved in innate immune function. Methods. Race-stratified weighted logistic regression models were used to calculate odds ratios for genotype and PTD/PTD subtypes. Polymorphisms significantly associated with PTD/PTD subtypes were tested for mother/child genotype interactions. Results. Three maternal polymorphisms (IL-1 receptor antagonist intron two repeat (IL-1RN), matrix metalloproteinase- −C1562T, and TNF receptor two M196R (TNFR2)) and three child polymorphisms (IL1-RN, tumor necrosis factor-alpha −G308A, and TNFR2) were associated with PTD, but associations varied by PTD subtype and race. Two interactions were detected for maternal and child genotype. Among non-Hispanic white women, the odds of PTD was higher when both mother and child carried the IL-1RN allele two (additive interaction p < 0.05). Among African-American women, the odds of PTD were higher when both mother and child carried the TNFR2 R allele (multiplicative interaction p < 0.05). Conclusion. These results highlight the importance of assessing both maternal and child genotype in relation to PTD risk.

Latent class regression: inference and estimation with two-stage multiple imputation.

Latent class regression (LCR) is a popular method for analyzing multiple categorical outcomes. While nonresponse to the manifest items is a common complication, inferences of LCR can be evaluated using maximum likelihood, multiple imputation, and two-stage multiple imputation. Under similar missing data assumptions, the estimates and variances from all three procedures are quite close. However, multiple imputation and two-stage multiple imputation can provide additional information: estimates for the rates of missing information. The methodology is illustrated using an example from a study on racial and ethnic disparities in breast cancer severity.

Latent class regression

Latent class regression (LCR) is a popular method for analyzing multiple categorical outcomes. While nonresponse to the manifest items is a common complication, inferences of LCR can be evaluated using maximum likelihood, multiple imputation, and two-stage multiple imputation. Under similar missing data assumptions, the estimates and variances from all three procedures are quite close. However, multiple imputation and two-stage multiple imputation can provide additional information: estimates for the rates of missing information. The methodology is illustrated using an example from a study on racial and ethnic disparities in breast cancer severity.