Time-dependent prognostic score matching for recurrent event analysis to evaluate a treatment assigned during follow-up

Matching by OS Prognostic Score to Construct External Controls in Lung Cancer Clinical Trials

External controls (eControls) leverage historical data to create non-randomized control arms. The lack of randomization can result in confounding between the experimental and eControl cohorts. To balance potentially confounding variables between the cohorts, one of the proposed methods is to match on prognostic scores. Still, the performance of prognostic scores to construct eControls in oncology has not been analyzed yet. Using an electronic health record-derived de-identified database, we constructed eControls using one of three methods: ROPRO, a state-of-the-art prognostic score, or either a propensity score composed of five (5Vars) or 27 covariates (ROPROvars). We compared the performance of these methods in estimating the overall survival (OS) hazard ratio (HR) of 11 recent advanced non-small cell lung cancer. The ROPRO eControls had a lower OS HR error (median absolute deviation (MAD), 0.072, confidence interval (CI): 0.036-0.185), than the 5Vars (MAD 0.081, CI: 0.025-0.283) and ROPROvars eControls (MAD 0.087, CI: 0.054-0.383). Notably, the OS HR errors for all methods were even lower in the phase III studies. Moreover, the ROPRO eControl cohorts included, on average, more patients than the 5Vars (6.54%) and ROPROvars cohorts (11.7%). The eControls matched with the prognostic score reproduced the controls more reliably than propensity scores composed of the underlying variables. Additionally, prognostic scores could allow eControls to be built on many prognostic variables without a significant increase in the variability of the propensity score, which would decrease the number of matched patients.

Association of COVID-19 Versus COVID-19 Vaccination With Kidney Function and Disease Activity in Primary Glomerular Disease: A Report of the Cure Glomerulonephropathy Study

RATIONALE & OBJECTIVE

Patients with glomerular disease (GN) may be at increased risk of severe COVID-19, yet concerns over vaccines causing disease relapse may lead to vaccine hesitancy. We examined the associations of COVID-19 with longitudinal kidney function and proteinuria and compared these with similar associations with COVID-19 vaccination.

STUDY DESIGN

Observational cohort study from July 1, 2021, to January 1, 2023.

SETTING & PARTICIPANTS

A prospective observational study network of 71 centers from North America and Europe (CureGN) with children and adults with primary minimal change disease, focal segmental glomerulosclerosis, membranous nephropathy, or IgA nephropathy.

EXPOSURE

COVID-19 and COVID-19 vaccination.

OUTCOME

Repeated measure of estimated glomerular filtration rate (eGFR); recurrent time-to-event outcome of GN disease worsening as defined by doubling of the urinary protein-creatinine ratio (UPCR) to at least 1.5g/g or increase in dipstick urine protein by 2 ordinal levels to 3+(300mg/dL) or above.

ANALYTICAL APPROACH

Interrupted time series analysis for eGFR. Prognostic matched sequential stratification recurrent event analysis for GN disease worsening.

RESULTS

Among 2,055 participants, 722 (35%) reported COVID-19 infection; of these, 92 (13%) were hospitalized, and 3 died (<1%). The eGFR slope before COVID-19 infection was-1.40mL/min/1.73m2 (± 0.29 SD); within 6 months after COVID-19 infection, the eGFR slope was-4.26mL/min/1.73m2 (± 3.02 SD), which was not significantly different (P=0.34). COVID-19 was associated with increased risk of worsening GN disease activity (HR, 1.35 [95% CI, 1.01-1.80]). Vaccination was not associated with a change in eGFR (-1.34mL/min/1.73m2±0.15 SD vs-2.16mL/min/1.73m2±1.74 SD; P=0.6) or subsequent GN disease worsening (HR 1.02 [95% CI, 0.79-1.33]) in this cohort.

LIMITATIONS

Infrequent or short follow-up.

CONCLUSIONS

Among patients with primary GN, COVID-19 infection was severe for 1 in 8 cases and was associated with subsequent worsening of GN disease activity, as defined by proteinuria. By contrast, vaccination against COVID-19 was not associated with change in disease activity or kidney function decline. These results support COVID-19 vaccination for patients with GN.

PLAIN-LANGUAGE SUMMARY

In this cohort study of 2,055 patients with minimal change disease, focal segmental glomerulosclerosis, membranous nephropathy, or IgA nephropathy, COVID-19 resulted in hospitalization or death for 1 in 8 cases and was associated with a 35% increase in risk for worsening proteinuria. By contrast, vaccination did not appear to adversely affect kidney function or proteinuria. Our data support vaccination for COVID-19 in patients with glomerular disease.

Black Race Is Associated With Higher Rates of Early-Onset End-Stage Renal Disease and Increased Mortality Following Liver Transplantation

Black race is a risk factor for end-stage renal disease (ESRD). Racial disparities in the risks of early and long-term renal complications after liver transplantation (LT) have not been systematically studied. This study evaluated racial differences in the natural history of acute and chronic renal insufficiency after LT. This was a retrospective single-center cohort study of 763 non-Hispanic White and 181 Black LT recipients between 2008 and 2017. Black race was investigated as an independent predictor of the following outcomes: (1) receipt and duration of early post-LT hemodialysis and (2) time to post-LT ESRD. The interaction of race and post-LT ESRD on survival was also studied. Black recipients had higher rates of pre-LT hypertension (P < 0.001), but diabetes mellitus and renal function before LT were not different by race (all P > 0.05). Overall, 15.2% of patients required early hemodialysis immediately after LT with no difference by race (covariate-adjusted odds ratio, 0.89; P = 0.71). Early dialysis discontinuation was lower among Black recipients (covariate-adjusted hazard ratio [aHR], 0.47; P = 0.02), whereas their rate of post-LT ESRD was higher (aHR, 1.91; P = 0.005). Post-LT survival after ESRD was markedly worse for Black (aHR, 11.18; P < 0.001) versus White recipients (aHR, 5.83; P < 0.001; interaction P = 0.08). Although Black and White LT recipients had comparable pretransplant renal function, post-LT renal outcomes differed considerably, and the impact of ESRD on post-LT survival was greater for Black recipients. This study highlights the need for an individualized approach to post-LT management to improve outcomes for all patients.

Matching with time-dependent treatments: A review and look forward

Observational studies of treatment effects attempt to mimic a randomized experiment by balancing the covariate distribution in treated and control groups, thus removing biases related to measured confounders. Methods such as weighting, matching, and stratification, with or without a propensity score, are common in cross‐sectional data. When treatments are initiated over longitudinal follow‐up, a target pragmatic trial can be emulated using appropriate matching methods. The ideal experiment of interest is simple; patients would be enrolled sequentially, randomized to one or more treatments and followed subsequently. This tutorial defines a class of longitudinal matching methods that emulate this experiment and provides a review of existing variations, with guidance regarding study design, execution, and analysis. These principles are illustrated in application to the study of statins on cardiovascular outcomes in the Framingham Offspring cohort. We identify avenues for future research and highlight the relevance of this methodology to high‐quality comparative effectiveness studies in the era of big data.

Estimating the effect of a rare time-dependent treatment on the recurrent event rate

In many observational studies, the objective is to estimate the effect of treatment or state-change on the recurrent event rate. If treatment is assigned after the start of follow-up, traditional methods (eg, adjustment for baseline-only covariates or fully conditional adjustment for time-dependent covariates) may give biased results. We propose a two-stage modeling approach using the method of sequential stratification to accurately estimate the effect of a time-dependent treatment on the recurrent event rate. At the first stage, we estimate the pretreatment recurrent event trajectory using a proportional rates model censored at the time of treatment. Prognostic scores are estimated from the linear predictor of this model and used to match treated patients to as yet untreated controls based on prognostic score at the time of treatment for the index patient. The final model is stratified on matched sets and compares the posttreatment recurrent event rate to the recurrent event rate of the matched controls. We demonstrate through simulation that bias due to dependent censoring is negligible, provided the treatment frequency is low, and we investigate a threshold at which correction for dependent censoring is needed. The method is applied to liver transplant (LT), where we estimate the effect of development of post-LT End Stage Renal Disease (ESRD) on rate of days hospitalized.

National Assessment of Hospitalization Rates for Incident End-Stage Renal Disease After Liver Transplantation

BACKGROUND

We examined the association of incident end-stage renal disease (ESRD) after liver transplantation (LT) and resource utilization using a data linkage between the Scientific Registry of Transplant Recipients and claims data from the Centers for Medicare and Medicaid Services.

METHODS

The study cohort consisted of patients aged ≥18 years who underwent deceased donor LT between January 1, 2003, and December 31, 2010, with Medicare as primary or secondary insurance and were discharged alive from the index LT hospitalization (n = 7019). The association of ESRD and post-LT hospitalization was assessed by sequential stratification, which entailed prognostic score matching of ESRD-free patients to each LT recipient at ESRD onset. The prognostic score was developed from a model of time to hospitalization and included baseline factors and hospitalization history as predictors.

RESULTS

The overall hospitalization rates for LT recipients with and without ESRD were 2.7 and 1.1 per patient-year at risk, respectively. The total number of days hospitalized patient per year was 23 in ESRD and 7 in non-ESRD LT recipients. The adjusted post-LT hospitalization rate was 97% higher after reaching ESRD compared to non-ESRD (hazard ratio, 1.97; P < 0.0001).

CONCLUSIONS

Hospitalization rates increased significantly for LT recipients after ESRD onset. Early risk factor modification efforts targeting patients who are at high ESRD risk may reduce post-LT ESRD incidence and hence decrease morbidity and cost among LT recipients.