Hurdle models for multilevel zero-inflated data via h-likelihood.

Count data often exhibit overdispersion. One type of overdispersion arises when there is an excess of zeros in comparison with the standard Poisson distribution. Zero-inflated Poisson and hurdle models have been proposed to perform a valid likelihood-based analysis to account for the surplus of zeros. Further, data often arise in clustered, longitudinal or multiple-membership settings. The proper analysis needs to reflect the design of a study. Typically random effects are used to account for dependencies in the data. We examine the h-likelihood estimation and inference framework for hurdle models with random effects for complex designs. We extend the h-likelihood procedures to fit hurdle models, thereby extending h-likelihood to truncated distributions. Two applications of the methodology are presented.

Resolution of hypertension and proteinuria after preeclampsia.

OBJECTIVE: To estimate the time required for hypertension and proteinuria to resolve after preeclampsia, and to estimate how this time to resolution correlates with the levels of blood pressure and proteinuria during preeclampsia and prolonging pregnancy after the development of preeclampsia. METHODS: This is a historic prospective cohort study of 205 preeclamptic women who were admitted between 1990 and 1992 at the Erasmus MC Medical Centre, Rotterdam, The Netherlands. Data were collected at 1.5, 3, 6, 12, 18, and 24 months after delivery. Hypertension was defined as a blood pressure 140/90 mm Hg or higher or use of antihypertensive drugs. Proteinuria was defined as 0.3 g/d or more. Resolution of hypertension and proteinuria were analyzed with the Turnbull extension to the Kaplan-Meier procedure. Correlations were calculated with an accelerated failure time model. RESULTS: At 3 months postpartum, 39% of women still had hypertension, which decreased to 18% at 2 years postpartum. Resolution time increased by 60% (P<.001) for every 10-mm Hg increase in maximal systolic blood pressure, 40% (P=.044) for every 10-mm Hg increase in maximal diastolic blood pressure, and 3.6% (P=.001) for every 1-day increase in the diagnosis-to-delivery interval. At 3 months postpartum, 14% still had proteinuria, which decreased to 2% at 2 years postpartum. Resolution time increased by 16% (P=.001) for every 1-g/d increase in maximal proteinuria. Gestational age at onset of preeclampsia was not correlated with resolution time of hypertension and proteinuria. CONCLUSION: The severity of preeclampsia and the time interval between diagnosis and delivery are associated with postpartum time to resolution of hypertension and proteinuria. After preeclampsia, it can take up to 2 years for hypertension and proteinuria to resolve. Therefore, the authors suggest that further invasive diagnostic tests for underlying renal disease may be postponed until 2 years postpartum. LEVEL OF EVIDENCE: III.

Proteolytic roles of matrix metalloproteinase (MMP)-13 during progression of chronic periodontitis: initial evidence for MMP-13/MMP-9 activation cascade.

AIM: Matrix metalloproteinases (MMP)-13 can initiate bone resorption and activate proMMP-9 in vitro, and both these MMPs have been widely implicated in tissue destruction associated with chronic periodontitis. We studied whether MMP-13 activity and TIMP-1 levels in gingival crevicular fluid (GCF) associated with progression of chronic periodontitis assessed clinically and by measuring carboxy-terminal telopeptide of collagen I (ICTP) levels. We additionally addressed whether MMP-13 could potentiate gelatinase activation in diseased gingival tissue. MATERIALS AND METHODS: In this prospective study, GCF samples from subjects undergoing clinical progression of chronic periodontitis and healthy controls were screened for ICTP levels, MMP-13 activity and TIMP-1. Diseased gingival explants were cultured, treated or not with MMP-13 with or without adding CL-82198, a synthetic MMP-13 selective inhibitor, and assayed by gelatin zymography and densitometric analysis. RESULTS: Active sites demonstrated increased ICTP levels and MMP-13 activity (p<0.05) in progression subjects. The MMP-9 activation rate was elevated in MMP-13-treated explants (p<0.05) and MMP-13 inhibitor prevented MMP-9 activation. CONCLUSIONS: MMP-13 could be implicated in the degradation of soft and hard supporting tissues and proMMP-9 activation during progression of chronic periodontitis. MMP-13 and -9 can potentially form an activation cascade overcoming the protective TIMP-1 shield, which may become useful for diagnostic aims and a target for drug development.

The use of a biplot in studying outcomes after stroke.

BACKGROUND AND PURPOSE: This study aimed to unravel the multidimensional profile of stroke outcomes by investigating the global correlation structure of motor, functional, and emotional problems of patients, as well as their caregivers' strain, at 6 months after stroke. Potential differential associations based on patients' level of functioning on admission to the rehabilitation center were analyzed. METHODS: Data were collected within the CERISE-study (Collaborative Evaluation of Rehabilitation in Stroke across Europe). Six months after stroke, the Rivermead Motor Assessment (RMA), Extended Activities of Daily Living (EADL), Hospital Anxiety and Depression Scale-Anxiety (HADS-A) and Hospital Anxiety and Depression Scale-Depression (HADS-D), EuroQol-Health State (EQ-HS), EuroQol-Visual Analogue Scale (EQ-VAS), and Caregiver Strain Index (CSI) were administered. Patients were classified into 3 categories according to their Barthel Index (BI) score on admission to the rehabilitation center. Principal component analysis was carried out, and a biplot was constructed. RESULTS: Data were available on 510 patients. One cluster was formed by RMA and EADL, and a second one by HADS-A, HADS-D, and EQ-VAS. EQ-HS was situated between these two. CSI formed a third dimension. Patients with low BI scores on admission to the rehabilitation center had higher HADS-A and HADS-D scores 6 months after stroke. High BI scores were associated with large variations in HADS-A and HADS-D scores. CONCLUSIONS: This novel biplot strategy for rehabilitation studies revealed 2 clusters: one of motor/functional problems and one of emotional problems. Patients with mild functional deficit measured on admission to the rehabilitation center can suffer from mild to severe anxiety and depression at 6 months poststroke. Screening for emotional disorders in all patients is recommended.

A comparison of three random effects approaches to analyze repeated bounded outcome scores with an application in a stroke revalidation study.

Discrete bounded outcome scores (BOS), i.e. discrete measurements that are restricted on a finite interval, often occur in practice. Examples are compliance measures, quality of life measures, etc. In this paper we examine three related random effects approaches to analyze longitudinal studies with a BOS as response: (1) a linear mixed effects (LM) model applied to a logistic transformed modified BOS; (2) a model assuming that the discrete BOS is a coarsened version of a latent random variable, which after a logistic-normal transformation, satisfies an LM model; and (3) a random effects probit model. We consider also the extension whereby the variability of the BOS is allowed to depend on covariates. The methods are contrasted using a simulation study and on a longitudinal project, which documents stroke rehabilitation in four European countries using measures of motor and functional recovery.