The higher risk of cardiovascular disease (CVD) among patients with chronic kidney disease (CKD) is not reflected in their poor level of inclusion in cardiovascular trials.
The result has been the inability to ground treatment and prevention recommendations for such patients in evidence from randomized controlled trials (RCTs), considered the gold standard in evidence-based medicine.
A new preprint posted to the medRxiv* server looks at how and where patients with CKD have been excluded from CVD RCTs over the last two decades and the resulting gaps in evidence for medications for cardiovascular risk management (CVRM).
Study: The persistent underrepresentation of patients with chronic kidney disease in cardiovascular trials: a systematic review and evidence map of exclusion and outcomes. Image Credit: PeopleImages.com-YuriA/Shutterstock.com
*Important notice: medRxiv publishes preliminary scientific reports that are not peer-reviewed and, therefore, should not be regarded as conclusive, guide clinical practice/health-related behavior, or treated as established information.
Introduction
CKD affects around 700 million patients around the world, being the cause of almost two million deaths. Of these, about 60% are patients with CVD, which causes most of these deaths rather than renal failure.
The increased CVD risk among CKD patients increases with the disease stage, irrespective of other risk factors such as diabetes mellitus or hypertension. This emphasizes the need for proper CVRM; indeed, nine out of every ten patients with CKD are on CVRM medication.
However, little is known about how effective these medications are in CKD patients since they are routinely left out of RCTs assessing cardiovascular interventions. Such exclusion criteria often cite concerns about the safety and efficacy of the proposed measure in CKD.
Interestingly, trials that do not specifically exclude such patients still often fail to have them in representative numbers or to analyze treatment effects in this subgroup.
As CKD causes profound shifts in the underlying CVD process and increases the CVD risk, the efficacy of CVRM medications should be carefully assessed in this group rather than based on their use in non-CKD patients.
It is possible, for instance, that the shift in pathophysiology observed in CKD patients could make standard CVRM therapy useless.
Prior research has pointed out the underrepresentation of CKD patients in most cardiovascular RCTs. In the current study, the researchers looked at cardiovascular RCTs evaluating the impact of CVRM medications in individuals who have either CVD or risk factors for CVD.
Their interest was to identify CKD exclusion and the separate reporting of results for CKD patients.
What did the study show?
The study included almost 1,200 RCTs with over two million participants. A quarter of them reported the estimated glomerular filtration rate (eGFR), and half reported the median serum creatinine levels, which are important in classifying CKD severity.
The researchers found that almost 80% of cardiovascular RCTs conducted from the year 2000 onwards did not include CKD patients, up from 66% earlier. CKD stages 1-3 were excluded in half of all RCTs that excluded CKD, and about 40% of all included RCTs.
CKD stages 4-5 were excluded more frequently over the last two decades,
In three out of four RCTs, a higher-than-expected number of patients was excluded pleading safety. When examined with greater granularity, almost two-thirds of RCTs that did not require dose adjustment of the tested medication with renal function still excluded CKD patients.
For medications that needed dosage adjustment or were contraindicated in impaired renal function, over 80% of RCTs excluded these patients. About 80% of these trials excluded an unreasonably high number of CKD patients based on safety concerns.
Almost one in seven RCTs did report subgroup analyses for CKD patients, but the mean eGFR was 71 ml/min/1.73m2. However, only 40% of RCTs reported eGFR in the participants.
Less than 5% of RCTs restricted their population to CKD patients. The reporting of results for CKD patients from RCTs remains poor, as it has been over 20 years.
Reporting endpoints also remained nonspecific, covering composite outcomes rather than individual cardiovascular endpoints or other separate reports for CKD patients. Thus, there were large gaps in the evidence for various CVRM measures, especially for those in stage 4 or 5 CKD.
Among patients with a low eGFR of below 30 ml/min/1.73m2, only 23 trials reported results. This number dropped to 15 for dialysis patients, with only one trial involving renal transplant patients.
A single RCT has not yet evaluated some anti-hypertensive drugs. Others have been assessed in different phases of CKD, including post-transplant and dialysis patients. Lipid-lowering drugs have been evaluated in CKD patients, but mostly only statins, with or without ezetimibe.
Anticoagulant and antiplatelet therapies have been assessed only partly and mostly in non-severe CKD. Newer glucose-lowering drugs have been tested in CKD patients with eGFR <60 ml/min/1.73m2, but less so with older drugs, including insulin, metformin, or sulfonylureas.
Some trials have compared the efficacy of older and novel glucose-lowering agents in CKD patients with eGFR less than 30 ml/min/1.73m2.
What are the implications?
“The underrepresentation of patients with CKD in cardiovascular RCTs has not improved in the past two decades.”
The prevalence of CKD is increasing, and CVRM medications are being prescribed to this group.
Yet, the lack of representation of CKD patients in most CVRM RCTs has led to poor evidence for many of these medications in this patient set, especially with advanced CKD, namely, stages 3 and 4.
The reasons for such exclusion are likely to be pragmatic, relating to cost, lower life expectancy, or differences expected in treatment effects for CKD patients compared to others. This has resulted in stricter exclusion criteria compared to standard practice during RCT recruitment.
As a result of such exclusion...
“in practice, practitioners must resort to extrapolating results from RCTs conducted in other populations, assuming that the treatment effects are comparable.”
This necessarily ignores the fact that increasing CKD severity, other risk factors emerge, such as high serum urea levels, chronic inflammation, arterial medial calcification, and left ventricular hypertrophy.
These modify the risk profile, making it impossible to know, in the absence of RCT-generated evidence, whether the patients are being benefited or harmed by the intervention.
Small sample size, subgroup analysis, wide variations in exclusion criteria, and poor reporting of eGFR and serum creatinine levels cause a loss of comparability between RCTs and imprecise effect estimates.
Similarly, when individual cardiovascular and kidney endpoints are not assessed for the effects of CVRM medications, these cannot be considered beneficial. Finally, subgroup analysis confounds randomization.
“These limitations are likely to amount in a GRADE recommendation of low or very low certainty of evidence for most CVRM medications in patients with CKD.”
The need to increase their representation should be urgently taken up by all stakeholders, perhaps adapting the current RCT design to reduce the cost while improving the speed at which results can be generated on the effects of various treatments on different CKD groups.
*Important notice: medRxiv publishes preliminary scientific reports that are not peer-reviewed and, therefore, should not be regarded as conclusive, guide clinical practice/health-related behavior, or treated as established information.
Novel multi-biomarker approach enhances chronic kidney disease risk assessment