Adaptive Clinical Trial Designs with Surrogates: When Should We Bother?

The success of a new drug is assessed within a clinical trial using a primary endpoint, which is typically the true outcome of interest—for example, overall survival. However, regulators sometimes approve drugs using a surrogate outcome—an intermediate indicator that is faster or easier to measure than the true outcome of interest—for example, progression-free survival—as the primary endpoint when there is demonstrable medical need. Although using a surrogate outcome (instead of the true outcome) as the primary endpoint can substantially speed up clinical trials and lower costs, it can also result in poor drug-approval decisions because the surrogate is not a perfect predictor of the true outcome. In this paper, we propose combining data from both surrogate and true outcomes to improve decision making within a late-phase clinical trial. In contrast to broadly used clinical trial designs that rely on a single primary endpoint, we propose a Bayesian adaptive clinical trial design that simultaneously leverages both observed outcomes to inform trial decisions. We perform comparative statics on the relative benefit of our approach, illustrating the types of diseases and surrogates for which our proposed design is particularly advantageous. Finally, we illustrate our proposed design on metastatic breast cancer. We use a large-scale clinical trial database to construct a Bayesian prior and simulate our design on a subset of clinical trials. We estimate that our design would yield a 16% decrease in trial costs relative to existing clinical trial designs, while maintaining the same Type I/II error rates. This paper was accepted by J. George Shanthikumar for the Special Issue on Data-Driven Prescriptive Analytics.

383 ECG in biopsy-proven and clinically suspected myocarditis: morpho-functional correlates and prognostic implications

Aims A distinction exists between biopsy-proven (BP) and clinically suspected (CS) myocarditis, the latter being an exclusion diagnosis based on clinical and instrumental findings. A clear diagnostic and prognostic role of the ECG in these two groups of patients has not yet been defined. (i) To describe frequency and characteristics of ECG findings in myocarditis, and to assess any difference between CS and BP myocarditis; (2) to identify morpho-functional correlates between ECG and cardiac magnetic resonance (CMR); and (iii) to evaluate the prognostic value of ECG findings. Methods and results 162 patients were included (median age 36 years, 70% male, median follow-up 32.9 months), 36 with BP and 126 with CS myocarditis. All patients underwent CMR; for ECG-CMR correlates, the ECG nearest in time to CMR was assessed. Surrogate outcome was defined as left ventricular (LV) ejection fraction (EF) <50% and/or NYHA class >I during follow-up. In the entire cohort ECG alterations were numerous: T-wave inversion (TWI) (82%), fragmented QRS (34%), low voltages (14%), ST elevation (STE) (13%). Compared to CS myocarditis, BP myocarditis patients showed higher frequency of non-sinus rhythm (17% vs. 2%, P < 0.001), long QT (28% vs. 0%, P < 0.001), lateral TWI (36% vs. 19%, P = 0.031) and bundle branch block (19% vs. 2%, P < 0.001). BP myocarditis patients had worse clinical features at diagnosis: heart failure (64% vs. 6%, P < 0.001), arrhythmic (14% vs. 4%, P = 0.029) and fulminant presentation (14% vs. 0%, P < 0.001), and presented higher LV end-diastolic volume and lower LVEF by echocardiography and CMR (109.5 mL/m2 vs. 85.50 ml/m2, P < 0.001; 31.5% vs. 59%, P < 0.001). A correlation was observed between number of myocardial segments with oedema at CMR and low ECG voltages (P = 0.010) and between late gadolinium enhancement (LGE) mass at CMR and lateral STE (P = 0.004 and P = 0.049, respectively). Several ECG alterations correlated with the surrogate outcome: long QT (P = 0.029), lateral TWI (P = 0.006), left bundle branch block (P < 0.001), ventricular ectopic beats (P = 0.020), and atrial fibrillation (P < 0.001). Conclusions A significant difference in ECG findings between CS and BP myocarditis has been demonstrated: ECG alterations are more frequent and more severe in BP myocarditis and correlate with prognosis. Moreover, ECG alterations identified patients with pathologic morpho-functional correlates.

Evidence from Studies of Patient-Reported Outcomes Supports a Policy of Using a Dialysate Sodium Concentration of 140 mEq/L for the Majority of Patients

The best evidence available to guide a policy for prescribing the dialysate sodium concentration, [DNa], comes from large randomly selected observational studies, such as the Dialysis Outcomes and Practice Patterns Study (DOPPS). These show that, after adjustment for differences in demographics and comorbidity, using a [DNa] lower than 140 mEq/L is associated with patients taking longer to recover after a dialysis treatment, worse symptoms of kidney failure, a higher score for the burden of kidney disease and worse mental and physical health-related quality of life. It is also associated with greater risks of being admitted to hospital and dying. These outcomes are more important than any medically determined surrogate outcome, such as the control of blood pressure or interdialytic weight gain. The most appropriate policy for prescribing the dialysate sodium concentration is to use a [DNa] of 140 mEq/L for the majority of patients.

Role of S100B Serum Concentration as a Surrogate Outcome Parameter After Mechanical Thrombectomy

Objective:To establish serum concentration of protein S100B as an objective biomarker surrogate for astroglial tissue damage after mechanical thrombectomy in patients with acute ischemic stroke.Methods:This prospective two-center study recruited patients with acute middle cerebral artery infarctions caused by large vessel occlusion treated with mechanical thrombectomy. Blood samples were collected at day 2 after intervention and analyzed for S100B serum concentrations using ELISA techniques. Infarct size was determined on follow-up brain imaging, and functional outcome according to modified Rankin scale (mRS) was assessed at 90 days.Results:171 patients were included (mean age ±SD: 70±14 years, 42% female). S100B levels correlated with infarct size. Median S100B concentrations at day 2 after intervention were lower in patients with favorable outcome (mRS score 0-1) at 90 days compared to patients with unfavorable outcome (mRS score 2-6) (median 0.10 µg/L [IQR 0.07-0.14] vs. 0.20 µg/L [0.11-0.48], p<0.001). Younger age (OR 1.120 [CI 1.068-1.174; p<0.001), lower NIHSS 24h after symptom onset (OR 1.232 [CI 1.106-1.372; p<0.001) and lower S100B serum concentrations (OR 1.364 [CI 1.105-1.683]; p=0.004) were independently associated with a favorable outcome. S100B was able to eliminate the lateralization bias associated with the use of mRS for functional outcome assessment at 90 days after stroke.Conclusion:S100B serum concentrations after mechanical thrombectomy indicate the extent of ischemic tissue damage. It can be rapidly assessed, independent of brain imaging and clinical outcome scales. Following prospective validation in further studies, it may provide an objective surrogate outcome parameter both in clinical routine and interventional trials.Classification of Evidence:This study provides Class I evidence that S100B 2 days following mechanical thrombectomy for acute ischemic stroke accurately distinguishes favorable from unfavorable functional outcome.

Sweat Chloride Testing and Nasal Potential Difference (NPD) Are Primary Outcome Parameters in Treatment with Cystic Fibrosis Transmembrane Conductance Regulator (CFTR) Modulators

: With the advent of CFTR modulators, surrogate outcome parameters that accurately quantify the improvement in CFTR activity are needed. In vivo biomarkers that reflect CFTR ion transport and can serve as outcomes in the treatment of CFTR modulators are the sweat Cl− test (SCT), the nasal potential difference (NPD) measurement or the intestinal current measurement (ICM). This review focus on the SCT and NPD. The SCT displays a low intra-patient variability in contrast to the NPD. It has been used extensively as a biomarker of CFTR function in clinical trials of CFTR modulator therapies and provides evidence for change in the short term. The level of functional rescue in the NPD increases up to 40% of normal CFTR in patients with a Gly551Asp treated with ivacaftor monotherapy, while in F508del homozygous patients treated with ivacaftor-lumacaftor, activity increased on average up to ~20% of normal activity. While both tests provide evidence of the effect on CFTR activity, they cannot be used at an individual level to predict the response to any CFTR modulators. Nevertheless, their rapid modification, reflecting electrophysiological properties, highlight their potential use in proof-of-concept studies for CFTR modulators.

Meta-regression of randomized control trials with antithrombotics: weak correlation between net clinical benefit and all cause-mortality

This study aimed to explore the validity of the use of the net clinical benefit (NCB), i.e. the sum of major bleeding and thrombotic events, as a potential surrogate for all-cause mortality in clinical trials assessing antithrombotics. Published randomized controlled trials testing anticoagulants in the prevention or treatment of venous thromboembolism (VTE) and non-valvular atrial fibrillation (NVAF) were systematically reviewed. The validity of NCB as a surrogate endpoint was estimated by calculating the strength of correlation of determination (R2) and its 95% confidence interval (CI) between the relative risks of NCB and all-cause mortality. Amongst the 125 trials retrieved, the highest R2trial values were estimated for NVAF (R2trial = 0.41, 95% CI [0.03; 0.48]), and acute VTE (R2trial = 0.30, 95% CI [0.04; 0.84]). Conversely, the NCB did not correlate with all-cause mortality in prevention studies with medical (R2trial = 0.12, 95% CI [0.00; 0.36]), surgical (R2trial = 0.05, 95% CI [0.00; 0.23]), and cancer patients (R2trial = 0.006, 95% CI [0.00; 1.00]). A weak correlation between NCB and all cause-mortality was found in NVAF and acute VTE, whereas no correlation was observed in clinical situations where the mortality rate was low. Consequently, NCB should not be considered a surrogate outcome for all cause-mortality in anticoagulation trials.

Quantifying Duration of Proteinuria Remission and Association with Clinical Outcome in IgA Nephropathy

BackgroundOn the basis of findings of observational studies and a meta-analysis, proteinuria reduction has been proposed as a surrogate outcome in IgA nephropathy. How long a reduction in proteinuria needs to be maintained to mitigate the long-term risk of disease progression is unknown.MethodsIn this retrospective multiethnic cohort of adult patients with IgA nephropathy, we defined proteinuria remission as a ≥25% reduction in proteinuria from the peak value after biopsy, and an absolute reduction in proteinuria to <1 g/d. The exposure of interest was the total duration of first remission, treated as a time-varying covariate using longitudinal proteinuria measurements. We used time-dependent Cox proportional hazards regression models to quantify the association between the duration of remission and the primary outcome (ESKD or a 50% reduction in eGFR).ResultsDuring a median follow-up of 3.9 years, 274 of 1864 patients (14.7%) experienced the primary outcome. The relationship between duration of proteinuria remission and outcome was nonlinear. Each 3 months in sustained remission up to approximately 4 years was associated with an additional 9% reduction in the risk of disease progression (hazard ratio [HR], 0.91; 95% confidence interval [95% CI], 0.89 to 0.93). Thereafter, each additional 3 months in remission was associated with a smaller, nonsignificant risk reduction (HR, 0.99; 95% CI, 0.96 to 1.03). These findings were robust to multivariable adjustment and consistent across clinical and histologic subgroups.ConclusionsOur findings support the use of proteinuria as a surrogate outcome in IgA nephropathy, but additionally demonstrate the value of quantifying the duration of proteinuria remission when estimating the risk of hard clinical endpoints.