controlling for confounding bias is crucial in causal inference. Distinct methods are currently employed to mitigate the effects of confounding bias. Each requires the introduction of a set of covariates, which remains difficult to choose, especially regarding the different methods. We conduct a simulation study to compare the relative performance results obtained by using four different sets of covariates (those causing the outcome, those causing the treatment allocation, those causing both the outcome and the treatment allocation, and all the covariates) and four methods: g-computation, inverse probability of treatment weighting, full matching and targeted maximum likelihood estimator. our simulations are in the context of a binary treatment, a binary outcome and baseline confounders. the simulations suggest that considering all the covariates causing the outcome led to the lowest bias and variance, particularly for g-computation. The consideration of all the covariates did not decrease the bias but significantly reduced the power. We apply these methods to two real-world examples that have clinical relevance, thereby illustrating the real-world importance of using these methods. We propose an R package RISCA to encourage the use of g-computation in causal inference. The randomised controlled trial (RCT) remains the primary design for evaluating the marginal (population average) causal effect of a treatment, i.e., the average treatment effect between two hypothetical worlds where: i) everyone is treated and ii) everyone is untreated 1. Indeed, a well-designed RCT with a sufficient sample size ensures the baseline comparability between groups, thus allowing the estimation of a marginal causal effect. Nevertheless, it is well established that RCT is performed under optimal circumstances (e.g., over-representation of treatment-adherent patients, low frequency of morbidity), which may be different from real-life practices 2. Observational studies have the advantage of limiting the issue of external validity, but treated and untreated patients are often non-comparable, leading to a high risk of confounding bias. To reduce such confounding bias, the vast majority of observational studies have been based on multivariable models (mainly linear, logistic, or Cox models), allowing for the direct estimation of conditional (subject-specific) effects, i.e., the average effect across sub-populations of subjects who share the same characteristics. Several

Objective:To compare natalizumab and fingolimod on both clinical and MRI outcomes in patients with relapsing-remitting multiple sclerosis (RRMS) from 27 multiple sclerosis centers participating in the French follow-up cohort Observatoire of Multiple Sclerosis.Methods:Patients with RRMS included in the study were aged from 18 to 65 years with an Expanded Disability Status Scale score of 0–5.5 and an available brain MRI performed within the year before treatment initiation. The data were collected for 326 patients treated with natalizumab and 303 with fingolimod. The statistical analysis was performed using 2 different methods: logistic regression and propensity scores (inverse probability treatment weighting).Results:The confounder-adjusted proportion of patients with at least one relapse within the first and second year of treatment was lower in natalizumab-treated patients compared to the fingolimod group (21.1% vs 30.4% at first year, p = 0.0092; and 30.9% vs 41.7% at second year, p = 0.0059) and supported the trend observed in nonadjusted analysis (21.2% vs 27.1% at 1 year, p = 0.0775). Such statistically significant associations were also observed for gadolinium (Gd)-enhancing lesions and new T2 lesions at both 1 year (Gd-enhancing lesions: 9.3% vs 29.8%, p < 0.0001; new T2 lesions: 10.6% vs 29.6%, p < 0.0001) and 2 years (Gd-enhancing lesions: 9.1% vs 22.1%, p = 0.0025; new T2 lesions: 16.9% vs 34.1%, p = 0.0010) post treatment initiation.Conclusion:Taken together, these results suggest the superiority of natalizumab over fingolimod to prevent relapses and new T2 and Gd-enhancing lesions at 1 and 2 years.Classification of evidence:This study provides Class IV evidence that for patients with RRMS, natalizumab decreases the proportion of patients with at least one relapse within the first year of treatment compared to fingolimod.

ObjectiveIn this study, we compared the effectiveness of teriflunomide (TRF) and dimethyl fumarate (DMF) on both clinical and MRI outcomes in patients followed prospectively in the Observatoire Français de la Sclérose en Plaques.MethodsA total of 1,770 patients with relapsing-remitting multiple sclerosis (RRMS) (713 on TRF and 1,057 on DMF) with an available baseline brain MRI were included in intention to treat. The 1- and 2-year postinitiation outcomes were relapses, increase of T2 lesions, increase in Expanded Disability Status Scale score, and reason for treatment discontinuation. Propensity scores (inverse probability weighting) and logistic regressions were estimated.ResultsThe confounder-adjusted proportions of patients were similar in TRF- compared to DMF-treated patients for relapses and disability progression after 1 and 2 years. However, the adjusted proportion of patients with at least one new T2 lesion after 2 years was lower in DMF compared to TRF (60.8% vs 72.2%, odds ratio [OR] 0.60, p < 0.001). Analyses of reasons for treatment withdrawal showed that lack of effectiveness was reported for 8.5% of DMF-treated patients vs 14.5% of TRF-treated patients (OR 0.54, p < 0.001), while adverse events accounted for 16% of TRF-treated patients and 21% of DMF-treated patients after 2 years (OR 1.39, p < 0.001).ConclusionsAfter 2 years of treatment, we found similar effectiveness of DMF and TRF in terms of clinical outcomes, but with better MRI-based outcomes for DMF-treated patients, resulting in a lower rate of treatment discontinuation due to lack of effectiveness.Classification of evidenceThis study provides Class III evidence that for patients with RRMS, TRF and DMF have similar clinical effectiveness after 2 years of treatment.

No abstract