Accounting for informative observation process in transition models of binary longitudinal outcome: Application to medical record data.

When extracting medical record data to form a retrospective cohort, investigators typically focus on a pre-specified study window, and select subjects who had hospital visits during that study window. However, such data extraction may suffer from an informative observation process, since sicker patients may have hospital visits more frequently. For example, Consecutive Pregnancy Study is a retrospective cohort study of women with multiple pregnancies in 23 Utah hospitals from 2003 to 2010, where the interest is to understand the risk factors of recurrent pregnancy outcomes, such as preterm birth. The observation process is informative in the sense that, women with adverse pregnancy outcomes may be less likely/willing/able to endure subsequent pregnancies. We proposed a three-part joint model with shared random effects structure to address this analytic complication. Particularly, a first-order transition model is used to model the longitudinal binary outcome; a gamma regression model is assumed for the inter-pregnancy intervals; a continuation ratio model specifies the probability of continuing with more births in the future. We note that the latter two parts give rise to a parametric cure-rate survival model. The performance of the proposed method was examined in extensive simulation studies, with both correctly and mis-specified models. The analyses of Consecutive Pregnancy Study data further demonstrate the inadequacies of fitting the transition model alone ignoring the informative observation process.

An Updated Evaluation of Atrazine-Cancer Incidence Associations among Pesticide Applicators in the Agricultural Health Study Cohort.

BACKGROUND: Atrazine is a common agricultural herbicide in the United States. Few epidemiologic studies have evaluated cancer risks. Previous analyses within the Agricultural Health Study (AHS) have found some evidence of associations with cancer at some sites. OBJECTIVE: We updated exposure information, incident cases, and follow-up time to assess the associations between atrazine use and cancer at specific sites in the AHS. METHODS: Information about lifetime pesticide use was reported at enrollment (1993-1997) and follow-up (1999-2005). Among 53,562 pesticide applicators in North Carolina and Iowa, we identified 8,915 incident cases through cancer registry linkages through 2014 (North Carolina)/2017 (Iowa). We used Poisson regression to evaluate the association between ever/never and intensity-weighted lifetime days of atrazine use and incident cancer risk controlling for several confounders. We also evaluated lagged exposures and age-stratified risk. RESULTS: Approximately 71.2% of applicators reported ever using atrazine, which was associated with lung cancer [rate ratios (RR)=1.24; 95% confidence interval (CI): 1.04, 1.46]. Aggressive prostate cancer risk was increased in the highest quartile (RRQ4=1.20; 95% CI: 0.95, 1.52; p-trend=0.19), particularly among those <60 years old (RRQ4=3.04; 95% CI: 1.61, 5.75; p-trend<0.001; p-interaction=0.04). Among applicators <50 years of age, ever-atrazine use was associated with non-Hodgkin lymphoma (NHL) (RR=2.43; 95% CI: 1.10, 5.38; p-interaction=0.60). For soft tissue sarcoma, there was an elevated risk in the highest tertile of exposure (RRT3: 2.54; 95% CI: 0.97, 6.62; p-trend=0.31). In analyses with exposure lagged by 25 years, there was an elevated risk of pharyngeal (RRT3=3.04; 95% CI: 1.45, 6.36; p-trend=0.07) and kidney (RRQ4=1.62; 95% CI: 1.15, 2.29; p-trend<0.005) cancers. DISCUSSION: We observed suggestive associations with some malignancies in overall, age-specific, and lagged analyses. Associations with aggressive prostate cancer and NHL were apparent among those diagnosed at younger ages and with cancers of the pharynx and kidney, and soft tissue sarcomas were observed in lagged analyses. Further work is needed to confirm these observed associations and elucidate potential underlying mechanisms. https://doi.org/10.1289/EHP13684.

Changes in smoking use and subsequent lung cancer risk in the ATBC study.

BACKGROUND: Reducing cigarettes/day may lower the risk of lung cancer compared with continuing to smoke at the same intensity. Other changes in smoking behaviors, such as increasing cigarette consumption or quitting for a period and relapsing, may also affect lung cancer risk. METHODS: We examined changes in smoking status and cigarettes/day among 24,613 Finnish male smokers aged 50-69 years who participated in the Alpha-Tocopherol, Beta-Carotene Cancer Prevention (ATBC) Study. Longitudinal data on smoking were collected during study follow-up visits three times a year (approximately every 4 months) between 1985 and 1993. Incident lung cancer cases through 2012 were identified by the Finnish Cancer Registry. Risk ratios (RRs) and 95% confidence intervals (95%CI) were estimated using Cox proportional hazards regression. RESULTS: Compared with smoking 20 cigarettes/day continuously across the intervention period, reducing an average of 5 cigarettes/day per year while smoking was associated with a 20% lower risk of lung cancer (95%CI : 0.71 to 0.90). A substantially lower risk of lung cancer was also observed when participants smoked at 50% (RR = 0.72; 95%CI : 0.57-0.90) and 10% (RR = 0.55; 95%CI : 0.36-0.83) of study visits, relative to smoked at 100% of study visits. CONCLUSIONS: Smokers may lower their risk of lung cancer by reducing smoking intensity (cigarettes per day while smoking) and the time they smoke. However, quitting smoking completely is the most effective way for smokers to reduce their risk of lung cancer.

A hidden Markov modeling approach combining objective measure of activity and subjective measure of self-reported sleep to estimate the sleep-wake cycle.

Characterizing the sleep-wake cycle in adolescents is an important prerequisite to better understand the association of abnormal sleep patterns with subsequent clinical and behavioral outcomes. The aim of this research was to develop hidden Markov models (HMM) that incorporate both objective (actigraphy) and subjective (sleep log) measures to estimate the sleep-wake cycle using data from the NEXT longitudinal study, a large population-based cohort study. The model was estimated with a negative binomial distribution for the activity counts (1-minute epochs) to account for overdispersion relative to a Poisson process. Furthermore, self-reported measures were dichotomized (for each one-minute interval) and subject to misclassification. We assumed that the unobserved sleep-wake cycle follows a two-state Markov chain with transitional probabilities varying according to a circadian rhythm. Maximum-likelihood estimation using a backward-forward algorithm was applied to fit the longitudinal data on a subject by subject basis. The algorithm was used to reconstruct the sleep-wake cycle from sequences of self-reported sleep and activity data. Furthermore, we conduct simulations to examine the properties of this approach under different observational patterns including both complete and partially observed measurements on each individual.

Nutritional Intake in Dichorionic Twin Pregnancies: A Descriptive Analysis of a Multisite United States Cohort.

INTRODUCTION: Twin gestations have greater nutritional demands than singleton gestations, yet dietary intakes of women with twin gestations have not been well described. METHODS: In a prospective, multi-site US study of 148 women with dichorionic twin gestations (2012-2013), we examined longitudinal changes in diet across pregnancy. Women completed a food frequency questionnaire during each trimester of pregnancy. We examined changes in means of total energy and energy-adjusted dietary components using linear mixed effects models. RESULTS: Mean energy intake (95% CI) across the three trimesters was 2010 kcal/day (1846, 2175), 2177 kcal/day (2005, 2349), 2253 kcal/day (2056, 2450), respectively (P = 0.01), whereas the Healthy Eating Index-2010 was 63.9 (62.1, 65.6), 64.5 (62.6, 66.3), 63.2 (61.1, 65.3), respectively (P = 0.53). DISCUSSION: Women with twin gestations moderately increased total energy as pregnancy progressed, though dietary composition and quality remained unchanged. These findings highlight aspects of nutritional intake that may need to be improved among women carrying twins.

Prediagnostic serum concentrations of per- and polyfluoroalkyl substances and risk of papillary thyroid cancer in the Finnish Maternity Cohort.

Human exposure to per- and polyfluoroalkyl substances (PFAS) occurs globally through contaminated food, dust, and drinking water. Studies of PFAS and thyroid cancer have been limited. We conducted a nested case-control study of prediagnostic serum levels of 19 PFAS and papillary thyroid cancer (400 cases, 400 controls) in the Finnish Maternity Cohort (pregnancies 1986-2010; follow-up through 2016), individually matched on sample year and age. We used conditional logistic regression to estimate odds ratios (OR) and 95% confidence intervals (CI) for log2 transformed and categorical exposures, overall and stratified by calendar period, birth cohort, and median age at diagnosis. We adjusted for other PFAS with Spearman correlation rho = 0.3-0.6. Seven PFAS, including perfluoroctanoic acid (PFOA), perfluorooctanesulfonic acid (PFOS), N-ethyl-perfluorooctane sulfonamidoacetic acid (EtFOSAA), perfluorononanoic acid (PFNA), perfluorodecanoic acid (PFDA), and perfluorohexane sulfonic acid (PFHxS) were detected in >50% of women. These PFAS were not associated with risk of thyroid cancer, except for PFHxS, which was inversely associated (OR log2 = 0.82, 95% CI: 0.70-0.97). We observed suggestive but imprecise increased risks associated with PFOA, PFOS, and EtFOSAA for those diagnosed at ages <40 years, whereas associations were null or inverse among those diagnosed at 40+ years (P-interaction: .02, .08, .13, respectively). There was little evidence of other interactions. These results show no clear association between PFAS and papillary thyroid cancer risk. Future work would benefit from evaluation of these relationships among those with higher exposure levels and during periods of early development when the thyroid gland may be more susceptible to environmental harms.

Comparison of Vital Status, Cause of Death, and Follow-Up after Hematopoietic Cell Transplantation in Linked Center for International Blood and Marrow Transplant Research and California Cancer Registry Data, 1991 to 2018.

Assessing outcomes following hematopoietic cell transplantation (HCT) poses challenges due to the necessity for systematic and often prolonged patient follow-up. Linking the HCT database of the Center for International Blood and Marrow Transplant Research (CIBMTR) with cancer registry data may improve long-term outcome ascertainment, but the reliability of mortality data in death certificates from cancer registries among HCT recipients remains unknown. We compared the classification of vital status and primary cause of death (COD), as well as the length of follow-up between the CIBMTR and California Cancer Registry (CCR) to assess the possibility of supplementing the CIBMTR with cancer registry data. This retrospective study leveraged a linked CIBMTR-CCR dataset. We included patients who were California residents at the time of HCT and received a first allogeneic (allo) or autologous (auto) HCT for a hematologic malignancy diagnosed during 1991-2016. Follow-up was through 2018. We analyzed 18,450 patients (alloHCT, n = 8232; autoHCT, n = 10,218). The Vital status agreement was 97.7% for alloHCT and 97.2% for autoHCT. Unknown COD was higher in CIBMTR (12.9%) than in CCR (1.6%). After excluding patients with unknown COD information, the overall agreement of primary COD (cancer versus noncancer) was 53.7% for alloHCT and 83.2% for autoHCT. This agreement was lower within the first 100 days post-HCT (alloHCT, 31.0%; autoHCT, 54.6%). Compared with CIBMTR, deaths due to cancer were higher in CCR (alloHCT, 90.0%; autoHCT, 90.1% versus alloHCT, 47.3%; autoHCT, 82.5% in CIBMTR). CIBMTR reports more frequently noncancer-related deaths, including graft-versus-host disease and infections. The cumulative incidence of cancer-specific mortality at 20 years differed, particularly for alloHCT (CCR, 53.7%; CIBMTR, 27.6%). The median follow-up among alive patients was longer in CCR (alloHCT, 6.0 years; autoHCT, 4.7 years) than in CIBMTR (alloHCT, 5.0 years; autoHCT, 3.8 years). Our findings highlight the completeness of vital status data in CIBMTR but reveal substantial disagreement in primary COD. Consequently, caution is required when interpreting HCT studies that use only death certificates to estimate cause-specific mortality outcomes. Improving the accuracy of COD registration and follow-up completeness by developing communication pathways between cancer registries and hospital-based cohorts may enhance our understanding of late effects and long-term outcomes among HCT survivors.

Maternal serum concentrations of per- and polyfluoroalkyl substances and childhood acute lymphoblastic leukemia.

BACKGROUND: Per- and polyfluoroalkyl substances (PFAS) are widespread and environmentally persistent chemicals with immunotoxic properties. Children are prenatally exposed through maternal transfer of PFAS to cord blood, but no studies have investigated the relationship with childhood leukemia. METHODS: We measured maternal serum levels of 19 PFAS in first-trimester samples collected in 1986-2010 and evaluated associations with acute lymphoblastic leukemia (ALL) in full-term offspring (<15 years) for 400 cases and 400 controls in the Finnish Maternity Cohort, matched on sample year, mother's age, gestational age, birth order, and child's sex. We analyzed continuous and categorical exposures, estimating odds ratios (OR) and 95% confidence intervals (CI) via conditional logistic regression adjusted for maternal smoking and correlated PFAS (ρ ≥ ±0.3). We also stratified by calendar period, mean diagnosis age, and the child's sex. RESULTS: N­methyl­perfluorooctane sulfonamidoacetic acid (MeFOSAA) was associated with ALL in continuous models (per each doubling in levels: ORperlog2=1.22, CI = 1.07-1.39), with a positive exposure-response across categories (OR>90th percentile=2.52, CI = 1.33-4.78; p-trend = 0.01). While we found no relationship with perfluorooctane sulfonic acid (PFOS) overall, an association was observed in samples collected 1986-1995, when levels were highest (median = 17.9 µg/L; ORperlog2=4.01, CI = 1.62-9.93). A positive association with perfluorononanoic acid was suggested among first births (p-interaction = 0.06). The MeFOSAA association was mainly limited to children diagnosed before age 5 (p-interaction = 0.02). We found no consistent patterns of association with other PFAS, nor differences by sex. CONCLUSIONS: These novel data offer evidence of a relationship between some PFAS and risk of the most common childhood cancer worldwide, including associations with the highest levels of PFOS and with a precursor, MeFOSAA.

The Pediatric Proton and Photon Therapy Comparison Cohort: Study Design for a Multicenter Retrospective Cohort to Investigate Subsequent Cancers After Pediatric Radiation Therapy.

Purpose: The physical properties of protons lower doses to surrounding normal tissues compared with photons, potentially reducing acute and long-term adverse effects, including subsequent cancers. The magnitude of benefit is uncertain, however, and currently based largely on modeling studies. Despite the paucity of directly comparative data, the number of proton centers and patients are expanding exponentially. Direct studies of the potential risks and benefits are needed in children, who have the highest risk of radiation-related subsequent cancers. The Pediatric Proton and Photon Therapy Comparison Cohort aims to meet this need. Methods and Materials: We are developing a record-linkage cohort of 10,000 proton and 10,000 photon therapy patients treated from 2007 to 2022 in the United States and Canada for pediatric central nervous system tumors, sarcomas, Hodgkin lymphoma, or neuroblastoma, the pediatric tumors most frequently treated with protons. Exposure assessment will be based on state-of-the-art dosimetry facilitated by collection of electronic radiation records for all eligible patients. Subsequent cancers and mortality will be ascertained by linkage to state and provincial cancer registries in the United States and Canada, respectively. The primary analysis will examine subsequent cancer risk after proton therapy compared with photon therapy, adjusting for potential confounders and accounting for competing risks. Results: For the primary aim comparing overall subsequent cancer rates between proton and photon therapy, we estimated that with 10,000 patients in each treatment group there would be 80% power to detect a relative risk of 0.8 assuming a cumulative incidence of subsequent cancers of 2.5% by 15 years after diagnosis. To date, 9 institutions have joined the cohort and initiated data collection; additional centers will be added in the coming year(s). Conclusions: Our findings will affect clinical practice for pediatric patients with cancer by providing the first large-scale systematic comparison of the risk of subsequent cancers from proton compared with photon therapy.

Early Pregnancy Maternal Plasma Phospholipid Saturated Fatty Acids and Fetal Growth: Findings from a Multi-Racial/Ethnic Birth Cohort in US.

Saturated fatty acids (SFAs) during pregnancy are associated with disrupted metabolic programming among offspring at birth and later growth. We examined plasma phospholipid SFAs in early pregnancy and fetal growth throughout pregnancy. We enrolled 321 pregnant women from the NICHD Fetal Growth Studies-Singleton Cohort at gestational weeks 8-13. Ultrasonogram schedules were randomly assigned to capture weekly fetal growth. We measured plasma phospholipid SFAs at early pregnancy using blood samples and modeled fetal growth trajectories across tertiles of SFAs with cubic splines using linear mixed models after full adjustment. We then compared pairwise weekly fetal growth biometrics referencing the lowest tertile in each SFA using the Wald test. We found that even-chain and very long even-chain SFAs were inversely associated, whereas odd-chain SFAs were positively associated with fetal weight and size. Compared with the lowest tertile, the highest tertile of pentadecanoic acid (15:0) had a greater fetal weight and size, starting from week 13 until late pregnancy (at week 39: 3429.89 vs. 3269.08 g for estimated fetal weight; 328.14 vs. 323.00 mm for head circumference). Our findings could inspire future interventions using an alternative high-fat diet rich in odd-chain SFAs for optimal fetal growth.

A latent functional approach for modeling the effects of multidimensional exposures on disease risk.

Understanding the relationships between exposure and disease incidence is an important problem in environmental epidemiology. Typically, a large number of these exposures are measured, and it is found either that a few exposures transmit risk or that each exposure transmits a small amount of risk, but, taken together, these may pose a substantial disease risk. Further, these exposure effects can be nonlinear. We develop a latent functional approach, which assumes that the individual effect of each exposure can be characterized as one of a series of unobserved functions, where the number of latent functions is less than or equal to the number of exposures. We propose Bayesian methodology to fit models with a large number of exposures and show that existing Bayesian group LASSO approaches are a special case of the proposed model. An efficient Markov chain Monte Carlo sampling algorithm is developed for carrying out Bayesian inference. The deviance information criterion is used to choose an appropriate number of nonlinear latent functions. We demonstrate the good properties of the approach using simulation studies. Further, we show that complex exposure relationships can be represented with only a few latent functional curves. The proposed methodology is illustrated with an analysis of the effect of cumulative pesticide exposure on cancer risk in a large cohort of farmers.

The efficient design of Nested Group Testing algorithms for disease identification in clustered data.

Group testing study designs have been used since the 1940s to reduce screening costs for uncommon diseases; for rare diseases, all cases are identifiable with substantially fewer tests than the population size. Substantial research has identified efficient designs under this paradigm. However, little work has focused on the important problem of disease screening among clustered data, such as geographic heterogeneity in HIV prevalence. We evaluated designs where we first estimate disease prevalence and then apply efficient group testing algorithms using these estimates. Specifically, we evaluate prevalence using individual testing on a fixed-size subset of each cluster and use these prevalence estimates to choose group sizes that minimize the corresponding estimated average number of tests per subject. We compare designs where we estimate cluster-specific prevalences as well as a common prevalence across clusters, use different group testing algorithms, construct groups from individuals within and in different clusters, and consider misclassification. For diseases with low prevalence, our results suggest that accounting for clustering is unnecessary. However, for diseases with higher prevalence and sizeable between-cluster heterogeneity, accounting for clustering in study design and implementation improves efficiency. We consider the practical aspects of our design recommendations with two examples with strong clustering effects: (1) Identification of HIV carriers in the US population and (2) Laboratory screening of anti-cancer compounds using cell lines.

A Bayesian hierarchical sparse factor model for estimating simultaneous covariance matrices for gestational outcomes in consecutive pregnancies.

Covariance estimation for multiple groups is a key feature for drawing inference from a heterogeneous population. One should seek to share information about common features in the dependence structures across the various groups. In this paper, we introduce a novel approach for estimating the covariance matrices for multiple groups using a hierarchical latent factor model that shrinks the factor loadings across groups toward a global value. Using a sparse spike and slab model on these loading coefficients allows for a sparse formulation of our model. Parameter estimation is accomplished through a Markov chain Monte Carlo scheme, and a model selection approach is used to select the number of factors to use. We validate our model through extensive simulation studies. Finally, we apply our methodology to the NICHD Consecutive Pregnancies Study to estimate the correlations between birth weights and gestational ages of three consecutive birth within four different subgroups (underweight, normal, overweight, and obese) of women.

Metabolomic Profiling of an Ultraprocessed Dietary Pattern in a Domiciled Randomized Controlled Crossover Feeding Trial.

BACKGROUND: Objective markers of ultraprocessed foods (UPF) may improve the assessment of UPF intake and provide insight into how UPF influences health. OBJECTIVES: To identify metabolites that differed between dietary patterns (DPs) high in or void of UPF according to Nova classification. METHODS: In a randomized, crossover, controlled-feeding trial (clinicaltrials.govNCT03407053), 20 domiciled healthy participants (mean ± standard deviation: age 31 ± 7 y, body mass index [kg/m2] 22 ± 11.6) consumed ad libitum a UPF-DP (80% UPF) and an unprocessed DP (UN-DP; 0% UPF) for 2 wk each. Metabolites were measured using liquid chromatography with tandem mass spectrometry in ethylenediaminetetraacetic acid plasma, collected at week 2 and 24-h, and spot urine, collected at weeks 1 and 2, of each DP. Linear mixed models, adjusted for energy intake, were used to identify metabolites that differed between DPs. RESULTS: After multiple comparisons correction, 257 out of 993 plasma and 606 out of 1279 24-h urine metabolites differed between UPF-DP and UN-DP. Overall, 21 known and 9 unknown metabolites differed between DPs across all time points and biospecimen types. Six metabolites were higher (4-hydroxy-L-glutamic acid, N-acetylaminooctanoic acid, 2-methoxyhydroquinone sulfate, 4-ethylphenylsulfate, 4-vinylphenol sulfate, and acesulfame) and 14 were lower following the UPF-DP; pimelic acid, was lower in plasma but higher in urine following the UPF-DP. CONCLUSIONS: Consuming a DP high in, compared with 1 void of, UPF has a measurable impact on the short-term human metabolome. Observed differential metabolites could serve as candidate biomarkers of UPF intake or metabolic response in larger samples with varying UPF-DPs. This trial was registered at clinicaltrials.gov as NCT03407053 and NCT03878108.

Uncovering circadian rhythms in metabolic longitudinal data: A Bayesian latent class modeling approach.

Researchers in biology and medicine have increasingly focused on characterizing circadian rhythms and their potential impact on disease. Understanding circadian variation in metabolomics, the study of chemical processes involving metabolites may provide insight into important aspects of biological mechanism. Of scientific importance is developing a statistical rigorous approach for characterizing different types of 24-hour patterns among high dimensional longitudinal metabolites. We develop a latent class approach to incorporate variation in 24-hour patterns across metabolites where profiles are modeled with finite mixtures of distinct shape-invariant circadian curves that themselves incorporate variation in amplitude and phase across metabolites. An efficient Markov chain Monte Carlo sampling is used to carry out Bayesian posterior computation. When the model was fit separately by individual to the data from a small group of participants, two distinct 24-hour rhythms were identified, with one being sinusoidal and the other being more complex with multiple peaks. Interestingly, the latent pattern associated with circadian variation (simple sinusoidal curve) had a similar phase across the three participants, while the more complex latent pattern reflecting diurnal variation differed across individual. The results suggested that this modeling framework can be used to separate 24-hour rhythms into an endogenous circadian and one or more exogenous diurnal patterns in describing human metabolism.

Evaluation of alcohol-free mouthwash for studies of the oral microbiome.

Oral bacteria play important roles in human health and disease. Oral samples collected using ethanol-containing mouthwash are widely used for oral microbiome studies. However, ethanol is flammable and not ideal for transportation/storage in large quantities, and some individuals may avoid ethanol due to the burning sensation or due to various personal, medical, religious, and/or cultural factors. Here, we compared ethanol-free and ethanol-containing mouthwashes using multiple microbiome metrics and assessed the stability of the mouthwash samples stored up to 10 days before processing. Forty volunteers provided oral wash samples collected using ethanol-free and ethanol-containing mouthwashes. From each sample, one aliquot was immediately frozen, one was stored at 4°C for 5 days and frozen, while the third aliquot was stored for 5 days at 4°C and 5 days at ambient temperature to mimic shipping delays and then frozen. DNA was extracted, the 16S rRNA gene V4 region was amplified and sequenced, and bioinformatic processing was performed using QIIME 2. Microbiome metrics measured in the two mouthwash types were very similar, with intraclass correlation coefficients (ICCs) for alpha and beta diversity metrics greater than 0.85. Relative abundances of some taxa were significantly different, but ICCs of the top four most abundant phyla and genera were high (> 0.75) for the comparability of the mouthwashes. Stability during delayed processing was also high for both mouthwashes based on alpha and beta diversity measures and relative abundances of the top four phyla and genera (ICCs ≥ 0.90). These results demonstrate ethanol-free mouthwash performs similarly to ethanol-containing mouthwash for microbial analyses, and both mouthwashes are stable for at least 10 days without freezing prior to laboratory processing. Ethanol-free mouthwash is suitable for collecting and shipping oral wash samples, and these results have important implications for planning future epidemiologic studies of the oral microbiome.

The Influence of Preanalytical Biospecimen Handling on the Measurement of B Vitamers, Amino Acids, and Other Metabolites in Blood.

Introduction: Sample handling can influence biomarker measurement and introduce variability when combining data from multiple studies or study sites. To inform the development of blood collection protocols within a multisite cohort study, we directly quantified concentrations of 54 biomarkers in blood samples subjected to different handling conditions. Materials and Methods: We obtained serum, lithium heparin plasma, and EDTA plasma from 20 adult volunteers. Tubes of chilled whole blood were either centrifuged and processed within 2 hours of collection (the "reference standard") or were stored with cool packs for 24 or 48 hours; centrifuged before and/or after this delay; or collected in tubes with/without gel separators. We used linear mixed models with random intercepts to estimate geometric mean concentrations and relative percent differences across the conditions. Results: Compared to the reference standard tubes, concentrations of many biomarkers changed after processing delays, but changes were often small. In serum, we observed large differences for B vitamers, glutamic acid (37% and 73% increases with 24- and 48-hour delays, respectively), glycine (12% and 23% increases), serine (16% and 27% increases), and acetoacetate (-19% and -26% decreases). Centrifugation timing and separator tube use did not affect concentrations of most biomarkers. Conclusion: Sample handling should be consistent across samples within an analysis. The length of processing delays should be recorded and accounted for when this is not feasible.

Prospective and Cross-sectional Associations of the Rectal Tissue Microbiome with Colorectal Adenoma Recurrence.

BACKGROUND: The gut microbiome is plausibly associated with colorectal cancer risk; however, previous studies mostly investigated this association cross-sectionally. We investigated cross-sectional and prospective associations of the rectal tissue microbiome with adenoma recurrence in the Polyp Prevention Trial (PPT). METHODS: PPT is a 4-year randomized clinical trial of the effect of a dietary intervention on adenoma recurrence among community members. We extracted DNA from rectal biopsies at baseline, end of year 1, and end of year 4 among 455 individuals and sequenced the V4 region of the 16S rRNA gene. At each timepoint, we investigated associations of alpha diversity, beta diversity, and presence and relative abundance of select taxa with adenoma recurrence using multivariable logistic regression. RESULTS: Variation in beta diversity was primarily explained by subject and minimally by year of collection or time between biopsy and colonoscopy. Cross-sectionally, year 4 alpha diversity was strongly, inversely associated with adenoma prevalence [ORQ3 vs. Q1 Shannon index = 0.40 (95% confidence interval, CI: 0.21-0.76)]. Prospective alpha diversity associations (i.e., baseline/year 1 alpha diversity with adenoma recurrence 3-4 years later) were weak or null, as were cross-sectional and prospective beta diversity-adenoma associations. Bacteroides abundance was more strongly, positively associated with adenoma prevalence cross-sectionally than prospectively. CONCLUSIONS: Rectal tissue microbiome profiles may be associated with prevalent adenomas, with little evidence supporting prospective associations. IMPACT: Additional prospective studies, with serial fecal and tissue samples, to explore microbiome-colorectal cancer associations are needed. Eventually, it may be possible to use microbiome characteristics as intervenable risk factors or screening tools.

Combination of Fundal Height and Ultrasound to Predict Small for Gestational Age at Birth.

OBJECTIVE: The objective of the study was to determine whether adding longitudinal measures of fundal height (FH) to the standard cross-sectional FH to trigger third trimester ultrasound estimated fetal weight (EFW) would improve small for gestational age (SGA) prediction. STUDY DESIGN: We developed a longitudinal FH calculator in a secondary analysis of a prospective cohort study of 1,939 nonobese pregnant women who underwent serial FH evaluations at 12 U.S. clinical sites. We evaluated cross-sectional FH measurement ≤ -3 cm at visit 3 (mean: 32.0 ± 1.6 weeks) versus the addition of longitudinal FH up to and including visit 3 to trigger an ultrasound to diagnose SGA defined as birth weight <10th percentile. If the FH cut points were not met, the SGA screen was classified as negative. If FH cut points were met and EFW was <10th percentile, the SGA screen was considered positive. If EFW was ≥10th percentile, the SGA screen was also considered negative. Sensitivity, specificity, and positive predictive value (PPV) and negative predictive value (NPV) were computed. RESULTS: In a comparison of methods, 5.8% of women were classified as at risk of SGA by both cross-sectional and longitudinal classification methods; cross-sectional FH identified an additional 4.0%, and longitudinal fundal height identified a separate, additional 4.5%.Using cross-sectional FH as an ultrasound trigger, EFW had a PPV and NPV for SGA of 69 and 92%, respectively. After adding longitudinal FH, PPV increased to 74%, whereas NPV of 92% remained unchanged; however, the number of women who underwent triggered EFW decreased from 9.7 to 5.7%. CONCLUSION: An innovative approach for calculating longitudinal FH to the standard cross-sectional FH improved identification of SGA birth weight, while simultaneously reducing the number of triggered ultrasounds. As an essentially free-of-charge screening test, our novel method has potential to decrease costs as well as perinatal morbidity and mortality (through better prediction of SGA). KEY POINTS: · We have developed an innovative calculator for fundal height trajectory.. · Longitudinal fundal height improves detection of SGA.. · As a low cost screening test, the fundal height calculator may decrease costs and morbidity through better prediction of SGA..

Using group testing in a two-phase epidemiologic design to identify the effects of a large number of antibody reactions on disease risk.

BACKGROUND: The role of immunological responses to exposed bacteria on disease incidence is increasingly under investigation. With many bacterial species, and many potential antibody reactions to a particular species, the large number of assays required for this type of discovery can make it prohibitively expensive. We propose a two-phase group testing design to more efficiently screen numerous antibody effects in a case-control setting. METHODS: Phase 1 uses group testing to select antibodies that are differentially expressed between cases and controls. The selected antibodies go on to Phase 2 individual testing. RESULTS: We evaluate the two-phase group testing design through simulations and example data and find that it substantially reduces the number of assays required relative to standard case-control and group testing designs, while maintaining similar statistical properties. CONCLUSION: The proposed two-phase group testing design can dramatically reduce the number of assays required, while providing comparable results to a case-control design.