Final FDA Guidance "Adjusting for Covariates in Randomized Clinical Trials for Drugs and Biological Products" - what we learned?

In May 2023, FDA published the final guidance for industry "Adjusting for Covariates in Randomized Clinical Trials for Drugs and Biological Products". This final version was based on the draft guidance with the same title that was released 4 years ago in April 2019 (see a previous post "FDA and EMA Guidance on Adjusting for Covariates in Randomized Clinical Trials".

FDA created guidance snapshot below:

The final guidance provided the general guidelines on several issues related to the covariates or baseline covariates. 

Both unadjusted analysis and analysis adjusted for baseline covariates are acceptable. However, if analysis is adjusted for baseline covariates, the details about the covariates need to be pre-specified in the statistical analysis plan before the study unblinding. Our experience is that the details about which baseline covariates to be included and whether the baseline covariates are continuous or categorized need to be pre-specified. 

Usually, the analysis adjusted for baseline covariates leads to efficiency gain and is more powerful than the unadjusted analysis. 

It is acceptable to calculate the sample size based on adjusted analysis, but perform the final analysis based on analysis adjusted for baseline covariates. In practice, the sample size calculation is commonly based on adjusted analysis regardless of the final analysis. For example, for a study to compare two group means, the sample size may be calculated based on t-test approach, but the analysis may be based on analysis of covariates where the adjustment for baseline covariates are used. 

For studies with stratified randomization where the randomization is stratified by one or more baseline covariates (categorical), the stratification factors are usually included in the analysis model even though the treatment assignments are generally balanced within each stratum. 

For studies with stratified randomization, it is not uncommon that incorrect stratification may occur where the treatment assignment is picked from the incorrect stratum. When this occurs, there will be two sets of the randomization stratification information (two different strata variables): strata as randomized versus actual strata. It is acceptable to use either strata variable as randomized (intention-to-treat principle) or actual strata variable (correct strata information for all patients). When mis-stratification occurs, there should be any attempt to go back to the randomization systems (such as IRT, IVR, IWR) to correct the stratification allocation. Once randomized, it is randomized. While the incorrect stratification is used for randomization, the correct stratification can be recorded on the case report form or EDC (electronic data capture). See a previous post "Handling Randomization Errors in Clinical Trials with Stratified Randomization".

For studies with continuous outcome measures, the endpoint is usually the change from baseline to a specific visit. Baseline covariate is used in the change from baseline calculation. In the analysis adjusted for baseline covariate, the baseline covariate can still be included in the model even though it gives an impression that the baseline measure is used twice. 

The guidance contains additional guidelines on linear models and non-linear models. For example, for linear models, the issue related to treatment group by covariate interactions is discussed: 

For non-linear models, binary outcome (logistic regression), ordinal outcome (generalized linear model), count outcome (Poisson regression), or time-to-event outcome (Cox regression) are analyzed. The estimators like odds ratio and hazard ratio are called are non-collapsible effect measures. Non-collapsibility implies that the effect parameter is not the same for different sets of covariates that are conditioned on, even if these covariates are independent of the exposure. Even when all subgroup treatment effects are identical, this subgroup specific conditional treatment effect can differ from the unconditional treatment effect. 

References: raps.org: FDA finalizes guidance on adjusting for covariates in randomized trials

Change from Baseline versus Percent Change from Baseline

The US Food and Drug Administration (FDA) has published its fourth and final guidance in a series of patient-focused drug development (PFDD) guidances meant to help sponsors collect and incorporate patient experience information that can factor into regulatory decision-making. The latest guidance "Patient-Focused Drug Development: Incorporating Clinical Outcome Assessments Into Endpoints For Regulatory Decision-Making" focuses on how clinical outcomes assessments (COA) can be used as endpoints to support a product. It is interesting to see that there is an entire section to discuss the difference between using change from baseline versus percent (or percentage) change from baseline as the endpoint. While the discussion is specifically for COA (clinical outcomes assessments) endpoints, the same discussion points are applicable to other endpoints where the outcome measures are continuous variables. 

Clinical trials are usually designed as longitudinal studies where the baseline measures are performed before the randomization or the first dose of the study drug and then there will be periodic measures (or repeated measures) for the post-baseline visits. 

The statistical analyses may be performed on the original measures, but are more often performed using change from baseline values. For each post-baseline visit, the change from baseline values will be calculated and statistical analyses will include the change from baseline values as the dependent variable and the baseline values will be used as a covariate in the model. Both FDA and EMA have guidelines related to the adjustment for baseline values. 

• FDA guidance (2021) "Adjusting for Covariates in Randomized ClinicalTrials for Drugs and Biological Products"

"Clinical trials often record a baseline measurement of a defined characteristic and record a later measurement of the characteristic to be used as an outcome. When using this approach, adjusting for the baseline value rather than (or in addition to) defining the primary endpoint as a change from baseline is generally acceptable."

• EMA guidelines (2015) "Guideline on adjustment for baseline covariates in clinical trials"

"5.6. Change from baseline’ analyses

When the primary analysis is based on a continuous outcome there is commonly the choice of whether to use the raw outcome variable or the change from baseline as the primary endpoint. Whichever of these endpoints is chosen, the baseline value should be included as a covariate in the primary analysis. The use of change from baseline with adjustment for baseline is generally more precise than change of baseline without adjustment. Note that when the baseline is included as a covariate in a standard linear model, the estimated treatment effects are identical for both ‘change from baseline’ (on an additive scale) and the ‘raw outcome’ analysis. Consequently if the appropriate adjustment is done, then the choice of endpoint becomes solely an issue of interpretability."

Percent change from baseline may also be used as the endpoint even though it is less commonly used in the analysis of clinical trial data. FDA's COA guidance clearly indicated the potential issues for using percent change from baseline in the analysis. 

For statistical modeling, Percent change from baseline has some undesirable properties: 

• It is asymmetric, e.g. a change from 4 to 5 is 25%, but a change from 5 to 4 is -20%. While this is asymmetric, the percent change has been commonly used in measuring the fluctuation of stock prices, a change from $50 to $100 is a 100% increase, but a change from $100 to $50 is a 50% decrease. 
• The variance is not easily defined. If we rewrite the percent change from baseline, it will be 1−Post-baseline measures/Baseline measure, so the variance is only defined by the ratio of the Post baseline values/baseline value. In statistical modeling, people don't like to deal with the ratio unless there is no better way. When we handle log-normal data (such as pharmacokinetic data), the geometric mean ratio is commonly used. 
• It is undefined if the baseline is zero

In a paper by Vickers "The use of percentage change from baseline as an outcome in a controlled trial is statistically inefficient: a simulation study", the following conclusions were made:

"Percentage change from baseline has the lowest statistical power and was highly sensitive to changes in variance. Theoretical considerations suggest that percentage change from baseline will also fail to protect from bias in the case of baseline imbalance and will lead to an excess of trials with non-normally distributed outcome data."

Nevertheless, the percent change from the baseline may still be used as the primary efficacy endpoint in some clinical trials. Recently, therapeutic trials for weight loss in non-diabetic patients are hot topics. In clinical trials for weight loss, the primary efficacy endpoint is the Percent Change from Baseline to Week x in weight or BMI (body mass index). 

In a paper by Weghuber, et al "Once-Weekly Semaglutide in Adolescents with Obesity", the primary efficacy endpoint was percent change from baseline to week 68 in BMI. 

Efficacy end points were assessed from baseline (the time of randomization [week 0]) to week 68, unless otherwise stated. The primary end point was the percentage change in BMI, and the secondary confirmatory end point was a reduction in body weight of

at least 5%.

In a paper by Rubino et al "Effect of Weekly Subcutaneous Semaglutide vs Daily Liraglutide on BodyWeight in Adults With Overweight or Obesity Without Diabetes The STEP 8 Randomized Clinical Trial", the primary efficacy endpoint was percentage change from baseline in body weight at week 68. Percent change from baseline was analyzed using analysis of covariance, with randomized treatment as a factor and baseline value of the outcome measure of interest (eg, baseline body weight in kilograms for analysis of percentage change in body weight) as a covariate. Multiple imputation approach was used to handle the missing values. 

In SURMOUNT-1 study by Eli Lilly (Jastreboff, et al "Tirzepatide Once Weekly for the Treatment of Obesity"), co-primary efficacy endpoints were specified: 

Percent change in body weight from baseline to Week 72

AND

Percentage of participants with ≥5% body weight reduction at Week 72