A conditional likelihood approach for regression analysis using biomarkers measured with batch-specific error

The Role of Conditional Likelihoods in Latent Variable Modeling

In psychometrics, the canonical use of conditional likelihoods is for the Rasch model in measurement. Whilst not disputing the utility of conditional likelihoods in measurement, we examine a broader class of problems in psychometrics that can be addressed via conditional likelihoods. Specifically, we consider cluster-level endogeneity where the standard assumption that observed explanatory variables are independent from latent variables is violated. Here, "cluster" refers to the entity characterized by latent variables or random effects, such as individuals in measurement models or schools in multilevel models and "unit" refers to the elementary entity such as an item in measurement. Cluster-level endogeneity problems can arise in a number of settings, including unobserved confounding of causal effects, measurement error, retrospective sampling, informative cluster sizes, missing data, and heteroskedasticity. Severely inconsistent estimation can result if these challenges are ignored.

Combining Biomarker Calibration Data to Reduce Measurement Error

Biomarker assay measurement often consists of a two-stage process where laboratory equipment yields a relative measure which is subsequently transformed to the unit of interest using a calibration curve. The calibration curve establishes the relation between the measured relative units and sample biomarker concentrations using stepped samples of known biomarker concentrations. Samples from epidemiologic studies are often measured in multiple batches or plates, each with independent calibration experiments. Collapsing calibration information across batches before statistical analysis has been shown to reduce measurement error and improves estimation. Additionally, collapsing in practice can also create an additional layer of quality control (QC) and optimization in a part of the laboratory measurement process that is often highly automated. Principled recalibration is demonstrated via. a three-step process of identifying batches where recalibration might be beneficial, forming a collapsed calibration curve and recalibrating identified batches, and using QC data to assess the appropriateness of recalibration. Here, we use inhibin B measured in biospecimens from the BioCycle study using 50 enzyme-linked immunosorbent assay (ELISA) batches (3875 samples) to motivate and display the benefits of collapsing calibration experiments, such as detecting and overcoming faulty calibration experiments, and thus improving assay coefficients of variation from reducing unwanted measurement error variability. Differences in the analysis of inhibin B by testosterone quartile are also demonstrated before and after recalibration. These simple and practical procedures are minor adjustments implemented by study personnel without altering laboratory protocols which could have positive estimation and cost-saving implications especially for population-based studies.

Transforming growth factors and receptor as potential modifiable pre-neoplastic biomarkers of risk for colorectal neoplasms

Increased colorectal epithelial cell proliferation is an early, common event in colorectal carcinogenesis. We conducted a pilot, colonoscopy-based case-control study (n = 49 cases, 154 controls) of incident, sporadic colorectal adenoma to investigate endogenous cell growth factors and receptor, as well as the balance of growth factors, as potential modifiable pre-neoplastic biomarkers of risk for colorectal neoplasms. We measured transforming growth factor alpha (TGFα), TGFβ(1), and TGFβ receptor II (TGFβRII) expression in normal-appearing mucosa from the rectum, sigmoid colon, and ascending colon using automated immunohistochemistry and quantitative image analysis. Diet and lifestyle were assessed via questionnaires. The mean ratio of rectal TGFα to TGFβ(1) expression and mean rectal TGFα expression were, respectively, 110% (P = 0.02) and 49% (P = 0.04) higher in cases than in controls, and associated with a more than two-fold (OR 2.42, 95% CI 0.85-6.87) and a 62% (OR 1.62, 95% CI 0.63-4.19) higher risk of colorectal adenoma. TGFβ(1) and TGFβRII expression were 6.7% (P = 0.75) and 7.2% (P = 0.49), respectively, lower in cases than in controls. The TGFα/TGFβ(1) expression ratio was 105% higher among smokers than among non-smokers (P = 0.03). These preliminary data suggest that the balance of TGFα and TGFβ(1) expression, and to a lesser extend TGFα alone, in the normal-appearing rectal mucosa may be directly associated with risk for incident, sporadic colorectal neoplasms, as well as with modifiable risk factors for colorectal neoplasms.

Association of hepatitis C with markers of hemostasis in HIV-infected and uninfected women in the women's interagency HIV study (WIHS)

BACKGROUND

Coinfection with HIV and hepatitis C virus (HCV) is common. HIV infection and treatment are associated with hypercoagulability; thrombosis in HCV is underinvestigated. Proposed markers of hemostasis in HIV include higher D-dimer, Factor VIII%, and plasminogen activator inhibitor-1 (PAI-1) antigen and lower total Protein S% (TPS) but have not been examined in HCV. We assessed the independent association of HCV with these 4 measures of hemostasis in a multicenter, prospective study of HIV: the Women's Interagency HIV Study.

METHODS

We randomly selected 450 HCV-infected (anti-HCV+ with detectable plasma HCV RNA) and 450 HCV-uninfected (anti-HCV-) women. HCV was the main exposure of interest in regression models.

RESULTS

Four hundred forty-three HCV+ and 425 HCV- women were included. HCV+ women had higher Factor VIII% (124.4% ± 3.9% vs. 101.8% ± 3.7%, P < 0.001) and lower TPS (75.7% ± 1.1% vs. 84.3% ± 1.1%, <0.001) than HCV- women, independent of HIV infection and viral load; there was little difference in PAI-1 or log10 D-dimer. After adjustment for confounders, these inferences remained. HIV infection was independently associated with higher Factor VIII% and log10 D-dimer and lower TPS.

CONCLUSIONS

HCV was independently associated with higher Factor VIII% and lower TPS consistent with hypercoagulability. Higher Factor VIII% and D-dimer and lower TPS were also strongly associated with HIV infection and levels of HIV viremia, independent of HCV infection. Further investigation is needed to determine if there is increased thrombotic risk from HCV. Studies examining hemostasis markers in HIV infection must also assess the contribution of HCV infection.