Correcting for Measurement Error in Time-Varying Covariates in Marginal Structural Models

Unbiased estimation of causal parameters from marginal structural models (MSMs) requires a fundamental assumption of no unmeasured confounding. Unfortunately, the time-varying covariates used to obtain inverse probability weights are often error-prone. Although substantial measurement error in important confounders is known to undermine control of confounders in conventional unweighted regression models, this issue has received comparatively limited attention in the MSM literature. Here we propose a novel application of the simulation-extrapolation (SIMEX) procedure to address measurement error in time-varying covariates, and we compare 2 approaches. The direct approach to SIMEX-based correction targets outcome model parameters, while the indirect approach corrects the weights estimated using the exposure model. We assess the performance of the proposed methods in simulations under different clinically plausible assumptions. The simulations demonstrate that measurement errors in time-dependent covariates may induce substantial bias in MSM estimators of causal effects of time-varying exposures, and that both proposed SIMEX approaches yield practically unbiased estimates in scenarios featuring low-to-moderate degrees of error. We illustrate the proposed approach in a simple analysis of the relationship between sustained virological response and liver fibrosis progression among persons infected with hepatitis C virus, while accounting for measurement error in γ-glutamyltransferase, using data collected in the Canadian Co-infection Cohort Study from 2003 to 2014.

CIHR Canadian HIV Trials Network Coinfection and Concurrent Diseases Core Research Group: 2016 Updated Canadian HIV/Hepatitis C Adult Guidelines for Management and Treatment

Background. Hepatitis C virus (HCV) coinfection occurs in 20-30% of Canadians living with HIV and is responsible for a heavy burden of morbidity and mortality. Purpose. To update national standards for management of HCV-HIV coinfected adults in the Canadian context with evolving evidence for and accessibility of effective and tolerable DAA therapies. The document addresses patient workup and treatment preparation, antiviral recommendations overall and in specific populations, and drug-drug interactions. Methods. A standing working group with HIV-HCV expertise was convened by The Canadian Institute of Health Research HIV Trials Network to review recently published HCV antiviral data and update Canadian HIV-HCV Coinfection Guidelines. Results. The gap in sustained virologic response between HCV monoinfection and HIV-HCV coinfection has been eliminated with newer HCV antiviral regimens. All coinfected individuals should be assessed for interferon-free, Direct Acting Antiviral HCV therapy. Regimens vary in content, duration, and success based largely on genotype. Reimbursement restrictions forcing the use of pegylated interferon is not acceptable if optimal patient care is to be provided. Discussion. Recommendations may not supersede individual clinical judgement. Treatment advances published since December 2015 are not considered in this document.

Sustained virological response in HIV/HCV co-infected patients treated with pegylated interferon/ribavirin can be predicted from the overall rate of viral load decline over the first 4 weeks of therapy

OBJECTIVE

It is not known whether the probability of achieving sustained virological response (SVR) can be determined on the basis of the magnitude of HCV viral decline over the first 4 weeks of Peg-IFN/RBV treatment of HIV/HCV co-infected patients who fail to achieve a rapid virological response (RVR).

METHODS

HIV patients co-infected with HCV genotype 1 naïve to Peg-IFN/RBV treatment were included. HCV viral decline from baseline to week 4 was graded. The positive predictive value (PPV) for SVR was evaluated according to the magnitude of HCV viral decline at week 4.

RESULTS

One hundred and fifty patients were included. Thirty-four (22.6%) patients achieved RVR, 33 of these (PPV [CI 95%]; 97.05% [86.34-99.85]) achieved SVR. In those patients who did not achieve RVR, the probability to achieving SVR was graded according to the magnitude of viral decline at week 4 (>2 log10 [55.5%], >2.5 log10 [73.3%] and >3 log10 [75%]). The combination of undetectable and magnitude of decline (>2.5 log10) had a PPV for SVR of 89.8% (CI 95%; 0.794-0.964).

CONCLUSIONS

The combination of undetectable HCV viral load and magnitude of decline at week 4 has a high PPV for SVR and identified a higher number of potential Peg-IFN/RBV responders.

Assessing long-term treatment efficacy in chronic hepatitis B and C: between evidence and common sense

Chronic infection with the hepatitis B and C virus represents a major health problem worldwide, as it is estimated that roughly 400 and 200 million people respectively, are infected by each virus. By definition, any antiviral therapy that claims to be effective should have as its ultimate efficacy end point an improvement in patients' survival, or at least a reduction in the development rates of liver-related complications. However, this is extremely complicated to prove as the natural course of both viral diseases is extremely slow, requiring decades to evolve in cirrhosis and even more years to lead to liver complications. For this reason, clinicians and health authorities have relied on so called surrogate end points to assess the efficacy of any therapeutic intervention for viral hepatitis. Obviously, this allows for standardization in study designs that ultimately translates into an accelerated time frame for therapeutic drugs as well as healthcare innovations to enter the viral hepatitis clinical practice. However, it also calls for demonstration that surrogate end points in the treatment of patients with chronic hepatitis B or C are good and reliable markers of long-term efficacy.

Twelve-week posttreatment follow-up to predict sustained virologic response for recurrent hepatitis C infection in liver recipients

BACKGROUND

The current standard for determining sustained virologic response (SVR) in patients treated for hepatitis C virus (HCV) infection is undetectable serum HCV-RNA 24 weeks after treatment. This study evaluates the value of HCV-RNA determination at 12 weeks posttreatment (W+12) to predict SVR in liver transplant (LT) patients treated with pegylated interferon and ribavirin for recurrent HCV infection.

METHODS

This study, performed in 2001 to 2010, included HCV-LT patients with an end-of-treatment response (undetectable serum HCV-RNA) and HCV-RNA testing at 12 and 24 weeks posttreatment (W+12/W+24). HCV-RNA was detected with a qualitative polymerase chain reaction assay (detection limit 50 IU/mL) and, when positive, measured by quantitative PCR (detection limit 600 IU/mL) up to 2006. Since 2007, a real-time PCR-based test (detection limit 15 IU/mL) has been used. The positive predictive value (PPV) was defined as the probability that SVR would occur in patients with undetectable HCV-RNA at W+12 and W+24.

RESULTS

Of 162 patients treated during the study period, 57 (35%) had end-of-treatment response and were included. Of these, 45 (79%) had SVR and 12 (21%) had virologic relapse. At W+12, HCV-RNA was undetectable in 45 (79%) patients, all of whom had SVR, yielding a PPV for SVR at W+12 of 100% (95% confidence interval, 75.8%-100%).

CONCLUSIONS

Undetectable HCV-RNA at W+12 posttreatment has a high PPV for predicting SVR. HCV-RNA testing to assess SVR at this time point seems as valid as W+24 testing and could be considered for predicting SVR in HCV-LT patients receiving treatment with pegylated interferon and ribavirin.

[Measurement of hepatitis C virus RNA 12 weeks after treatment predicts sustained virological response in HIV-HCV co-infected patients]

OBJECTIVE

The aim of this study was to assess whether measurement of hepatitis C virus RNA (HCV-RNA) at 12 weeks post-treatment could predict sustained virological response (SVR) to antiviral therapy for chronic hepatitis C (pegylated interferon alfa-2a and ribavirin) in HIV-co-infected patients.

PATIENTS AND METHODS

HIV-HCV co-infected patients were included if they completed a full course of anti-HCV therapy, achieved an end-of-treatment response and complied with the week +12 and +24 post-treatment follow-up schedule for serum HCV-RNA determination (Real-time HCV (Abbott, Wiesbaden, Germany) (lower limit of detection, 12 IU/ml).

RESULTS

Forty out of 66 patients (61%) showed an end-of-treatment response. They were assessed in a follow-up visit at +12 and at +24 weeks post-treatment. Serum HCV-RNA was undetectable in 28 of them at +12 weeks, and 100% of these remained undetectable at 24 weeks post-treatment (the gold standard of (SVR). The positive predictive value was 100% (95% confidence interval, 98.21-100%).

CONCLUSION

Post-treatment follow-up to identify virological relapse could be shortened to 12 weeks, providing a new definition of sustained virological response.