Is It Time to Re-Imagine Clinical Trials in Nephrology?

A recent article in The New York Times sparked my interest—a cheap, off-patent drug used successfully to restore hair loss. The drug in is question, minoxidil. Minoxidil, usually applied topically, was troublesome due to the stickiness of the preparation and had middling results for hair restoration. However, when administered orally at lower doses than that for hypertension, it was remarkably effective. The supporting data were a before and after scalp shot—a thinning hair pattern transforming into a lovely thick head of hair.¹ My main chagrin with the article was the expert stating that no proper randomized controlled trial (RCT) would ever be realized to prove its efficacy because there would be little profit for pharmaceutical companies in repeddling the agent. This is a pity and speaks to the structures we have in place to scientifically question our medical practices which center on RCTs and meta-analysis.

RCT practices have developed hand in hand with the scientific breakthroughs that brought new drugs and devices on stream. Recently, COVID-19 has shone a light on optimizations in how we design and implement studies. The metatrial is an example of adaptions in clinical trial methodologies used during the response to COVID-19.² A metatrial is a prospectively planned, pooled analysis of multiple clinical trials with the purpose of allowing faster accumulation of data and to enable a rapid determination of efficacy of relevant clinical end points. It engenders many of the requirements that should be ingrained in clinical trials: a collaborative culture, reduced exposure of trial participants to less effective treatments and unjustified research, and the avoidance of publication bias.³ A metatrial encompasses features of both RCT and meta-analysis that can overcome the expense required to achieve adequate power with the former and the heterogeneity problematic with the latter (Table 1). One such metatrial created in response to COVID-19 was the Awake Prone Positioning Metatrial study group.³ Awake prone positioning (APP), which involves nursing patients with severe COVID-19 on their stomach, results in a transient nonsustained improvement in oxygenation. Although encouraging observations had been made, there was a clear equipoise for a clinical trial to test whether APP was an effective strategy in managing patient with COVID-19. Owing to lockdown-related fall-off in COVID-19 cases, the principal investigators of five national randomized controlled open-label superiority trials of APP met online to align their studies as part of a metatrial on APP to improve recruitment. This brought about the creation of a metatrial protocol that harmonized eligibility, randomization procedures, and outcomes across the studies. Investigators identified a common set of core data for extraction from each trial, aligned primary and secondary outcomes, and planned a collaborative interim and final statistical analysis at the metatrial level. Finally, they agreed to report the findings jointly as a unified group of investigators. The study found that APP reduced the composite outcome of treatment failure defined by either intubation or death within 28 days.²

Table 1.

Comparison of the metatrial concept with alternative designs

Study Characteristic	Individual Trials Followed by a Retrospective Meta-Analysis	International Single Trial	Meta-Trial: Prospective International Meta-Analysis
Eligibility criteria for participants	Heterogeneous between trials	Uniform within the trial	Similar between trials (may have some heterogeneity within clinical relevance)
Baseline data	Heterogeneous between trials	Uniform within the trial	Common set of variables in data sharing agreement
Intervention details and how they were administered	Heterogeneous between trials	Uniform within the trial	Uniform between trials: agreement between individual investigators to deliver same intervention
Prespecified primary and secondary outcome measures	Heterogeneous between trials	Uniform within the trial	Uniform between trials (investigator agree on a common set of outcomes)
Sample size and interim analysis	Heterogeneous between trials, interim analysis impossible at the metalevel	One sample size calculation for the trial, interim analysis possible	Metatrial design transcends original sample size calculation, interim analysis possible at the metalevel
Randomization—sequence generation, stratification, allocation sequence, concealment, and blinding	Heterogeneous between trials	Centralized randomization	May differ for each site but fundamental randomization principles adhered to
Statistical methods	Heterogeneous original analyses, meta-analysis on effect sizes to compute a summary effect	Uniform within the trial, adjustments possible	Uniform within trials, meta-analysis on individual participant data, adjustment possible
Analysis populations: intention to treat, per-protocol, subgroups	Heterogeneous between trials	Uniform within the trial	Uniform between trials (agreement on uniform analysis population)
Data quality and safety monitoring	Each trialist is responsible for his or her trial	Centralized data monitoring	Each trialist is responsible for his or her trial
Funding	Multiple funding	Centralized funding	Multiple funding
Set-up time	Short	Long	Short
Time to completion	Long	Short	Short
Protocols	Multiple original protocols	Single protocol	Multiple original protocols followed by a metatrial protocol
Ethics	Each trialist is responsible for his or her trial	Centralized submission process	Each trialist is responsible for his or her trial

Adapted from ref. 3, with permission.

One strength of the metatrial approach is the ability to improve patient recruitment. Powering studies for meaningful clinical outcomes is frequently unrealistic for small-scale discovery studies in which biomarkers and surrogates are used to determine the effect of a treatment. In this scenario, participation in a metatrial cooperative allows investigators to pool their patients' outcomes to improve detection of such clinical outcomes not possible in an individual study.⁴ Another advantage of the metatrial approach is investigator independence. Although trial protocols are harmonized to reduce heterogeneity, involvement in the metatrial does not preclude reporting of additional physiological investigations or secondary outcomes for individual studies.⁴ A disadvantage may be the lack of a single principal investigator who takes charge of the study. Clear rules regarding presentation and publication need to be decided a priori to avoid conflict at later stages.

The requirement for regulatory rigor and with it, complexity in clinical trials is appropriate. Standards in ethics have been introduced in response to disregard to safety and well-being of patients as evidenced in Tuskagee and Guatamala.⁵ These unethical practices rightfully leave minorities and underserved patients suspicious of the clinical trial process.⁵ However, it only serves to deepen these wrongs if regulatory creep unrelated to trial safety further reduces access of the underserved and their practitioners to clinical trials. The metatrial is not a means to have a less-regulated study, and each individual as well as the overall meta-trial needs to be designed with the same rigor as any clinical trial. It should be noted that the metatrial is one of many adaptations in trial methodology applied during COVID-19 which includes platform studies, new approaches to Bayesian statistical analysis, as well as streamlined point-of-care studies that were conducted both efficiently and economically.^6,7 Progress in improving the accessibility and infrastructure that broadens the scope of clinical trial methodology needs to be encouraged and welcomed. It is interesting to contrast the higher recruitment of patients with chronic kidney disease in streamlined point-of-care COVID-19 studies compared with that of pharma-led cardiology-led studies.^8,9

Nephrology was an early adopter of the metatrial methodology as part of the 1998 International Society of Hypertension Blood Pressure-Lowering Treatment Trialists' Collaboration.¹⁰ This was an effort to provide more reliable information about the effects of newer blood pressure-lowering drugs than one study alone could achieve. It is useful to consider the barriers to clinical trials within our field where difficulty in recruitment and retention of patients with CKD and ESKD is recognized due to the age profile of patients and their other additional disease burdens. Disconnectedness of the primary nephrology team from the trial process may further reduce patients' interest and incentive to participate in studies.¹¹ The glacial pace and mammoth expense leads to studies that test patients in an unrealistic environment or, worse, select nonrepresentative patients who may be more likely to consent and go through the trial processes. It also precludes entry of ordinary practitioners into the trial industry, which centers predominantly on university centers which have the resources to handle the rigorous requirements on top of their own heavy clinical workload. Metatrials may assist here, where smaller more manageable studies can be conducted by investigators with protocols and outcome measures that harmonize and align with other studies in the field.

There clearly is an appetite for clinical trials—by May 2020, there were 1000 studies addressing various aspects of COVID-19 registered on ClinicalTrials.gov including more than 600 interventional studies and randomized controlled trials. Is this something we could match in Nephrology and how should this be facilitated? First, training in and technology available for clinical trials has markedly improved. This means answering important questions through clinical trials can be manageable, transparent, and performed to a high ethical and methodological standard, particularly for nonregulated studies where the requirements are less onerous. Educating practitioners and regulators about new trial designs is an important aspect of this. Workshops and online educational portals such as those by the Cochrane group have improved the reporting of meta-analysis and also include prospective meta-analysis resources.¹² Second, registration of meta-trials needs to be facilitated similar to ClinicalTrials.gov and Prospero. This will be important to ensure adherence to study protocols. Currently, there is no registry that records prospective meta-analysis although this had been planned in the past.¹³ Third, a mechanism to facilitate participation in a metatrial through reputable online portals is needed. This will allow for easier networking of researchers working on the same subject matter. Cochrane or NIH support of such an interface could permit easier identification of studies at early stages to participate in prospective meta-analysis.¹⁴ Finally, societies and funding bodies can insist on harmonization of clinically meaningful core outcome sets for studies so that every randomized event will not just answer a local investigators studies' question but be part of the bigger answer on what can improve the livelihood of our patients. More widespread application of the streamlined clinical trial approach such as platform studies and metatrials has the potential to empower the nephrology community to address important questions in areas of practice for which strong clinical trial evidence is often lacking such as chronic dialysis, glomerular disease, and long-term transplant outcomes.

Disclosures

B.A. McNicholas reports employer: Renalytix; consultancy: Teleflex; and advisory or leadership role: Deputy Chair Irish Critical Care Trials Group.

Funding

None.

Author Contributions

B.A. McNicholas conceptualized the study, wrote the original draft, and reviewed and edited the manuscript.