Percentage of trials excluding patients with kidney disease based on specific intervention categories (A) and specific diagnostic categories (B). The bar graphs represent the percentage of trials out of the total that evaluate a particular intervention or a particular diagnostic category. HFpEF indicates heart failure with preserved ejection fraction; HFrEF, heart failure with reduced ejection fraction; NSTEMI, non–ST-segment elevation myocardial infarction; PCI, percutaneous coronary intervention; RAAS, renin-angiotensin-aldosterone system; STEMI, ST-segment elevation myocardial infarction; UA, unstable angina.
aIncludes management, exercise, oxygen therapy, and systems-wide quality improvement initiatives.
Konstantinidis I, Nadkarni GN, Yacoub R, Saha A, Simoes P, Parikh CR, Coca SG. Representation of Patients With Kidney Disease in Trials of Cardiovascular InterventionsAn Updated Systematic Review. JAMA Intern Med. 2016;176(1):121-124. doi:10.1001/jamainternmed.2015.6102
The prevalence of kidney disease (KD) among the general population and in patients with cardiovascular disease (CVD) is high and growing.1 Also, standard therapeutic strategies may act differently in patients with KD, making extrapolation of trial data from patients without KD unreliable.2 In addition, trials in nephrology are more likely to be smaller and unblinded, leading to poorer-quality evidence.3 Despite the need for high-quality evidence for CVD interventions, 2 systematic reviews using data from 1985-2005 and 1998-2005 showed that patients with KD are underrepresented in randomized clinical trials of CVD interventions.4,5 We aimed to update the estimates of the representation of patients with KD in major randomized clinical trials of CVD interventions.
We performed a systematic search of MEDLINE between January 1, 2006, and December 31, 2014. Analysis was conducted from November 1, 2014, to August 31, 2015. We restricted the search to the top 10 medical journals ranked by impact factor from 2006 to 2013 to obtain 2350 citations and 2 of us (I.K. and G.N.N.) reviewed each full article. We included trials if they were randomized or controlled, treated heart failure or acute coronary syndrome, and randomized 100 participants or more. We excluded trials that did not report mortality or were subgroup analyses. We extracted the following data: trial characteristics, whether interventions were class I or II recommendations, exclusion of patients with KD, exclusion threshold of patients with KD (based on laboratory test results, renal replacement therapy, or nonspecific terms), reported indices of baseline renal function, proportion of patients with KD in each arm, and number of subgroup analyses by any nonrenal or renal characteristics. We evaluated differences in exclusion of patients with KD by each trial characteristic. Institutional review board approval was not sought because this is a systematic review of published data.
A total of 371 trials randomizing 590 040 participants were included for analysis. Overall, 212 trials (57.1%) excluded patients with KD. Patients with KD were more likely than patients without KD to be excluded from North American and Canadian trials vs European trials (84 of 129 [65.1%] vs 107 of 206 [51.9%]), were more likely to be included in trials that tested medications vs those that tested procedures (142 of 200 [71.0%] vs 32 of 94 [34.1%]), were more likely than patients without KD to be included in industry-funded trials vs government-funded trials (111 of 172 [64.5%] vs 49 of 101 [48.5%]), and were more likely than patients without KD to be included in trials of patients with heart failure vs trials of patients with acute coronary syndrome (91 of 144 [63.2%] vs 120 of 244 [53.6%]). The Figure shows the exclusion of patients with KD by specific categories of treatment and diagnosis. There were no significant differences in representation of patients with KD in trials testing an intervention that was a class I or II recommendation vs representation of patients with KD in all trials (Table).
Of 212 trials excluding patients with KD, 111 (52.4%) used serum creatinine levels as the exclusion threshold, 21 (9.9%) used estimated glomerular filtration rate, 22 (10.4%) used creatinine clearance thresholds, 60 (28.3%) used renal replacement therapy of any form, and 36 (17.0%) used nonspecific qualitative exclusion criteria. Only 156 (42.0%) and 84 (22.6%) trials, respectively, reported at least 1 measure of baseline renal function or proportion of patients with KD in each arm. Although 197 trials (53.1%) reported subgroup analyses by nonrenal baseline characteristics, only 60 (16.2%) reported subgroup analyses by renal characteristics (Table). In trials reporting renal subgroup analysis, 38 of 60 (63.3%) reported a test for interaction between renal function and outcome and, of those, 5 of 38 (13.2%) reported a significant interaction.
In this systematic review, we found continued underrepresentation of patients with KD in trials of CVD interventions despite previously highlighted underrepresentation of these patients in such trials. This finding is troubling since KD and CVD are independent risk factors leading to increased prevalence of KD in patients with CVD and vice versa.1 Exclusion of patients with KD in trials of CVD interventions may be owing to cautiousness in certain conditions (eg, anticoagulation) but not in lower-risk interventions or when KD is a major confounder. However, we found that only 13.2% of trials assessing effect by KD subgroup had any interaction and thus, concerns of harm may be overstated.
Most trials excluded patients by serum creatinine levels instead of estimated glomerular filtration rate. Given the availability of superior methods to estimate renal function and the inaccurateness of serum creatinine measurements,6 use of serum creatinine levels for exclusion is inappropriate. Finally, trials did not report baseline renal function or outcomes by renal function, leading to poor understanding of the response to interventions of patients with KD.
In summary, we urge trialists to include patients with KD in trials of CVD interventions and to both report and analyze outcomes by renal function to improve the understanding of the risks and benefits of interventions in this vulnerable population.
Corresponding Author: Girish N. Nadkarni, MD, MPH, CPH, Division of Nephrology, Department of Medicine, Icahn School of Medicine at Mount Sinai, One Gustave Levy Place, PO Box 1243, New York, NY 10029 (firstname.lastname@example.org).
Published Online: November 30, 2015. doi:10.1001/jamainternmed.2015.6102.
Author Contributions: Drs Konstantinidis and Nadkarni had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Drs Konstantinidis and Nadkarni contributed equally to this study.
Sudy concept and design: Konstantinidis, Nadkarni, Yacoub, Parikh, Coca.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Konstantinidis, Nadkarni, Yacoub, Simoes.
Critical revision of the manuscript for important intellectual content: Konstantinidis, Nadkarni, Yacoub, Saha, Parikh, Coca.
Statistical analysis: Konstantinidis, Yacoub, Parikh.
Administrative, technical, or material support: Konstantinidis, Saha, Simoes, Coca.
Study supervision: Nadkarni, Coca.
Conflict of Interest Disclosures: None reported.
Additional Contributions: Priti Poojary, MBBS, Department of Public Health, Icahn School of Medicine at Mount Sinai, assisted with data collection. She was not compensated for her contribution.