Evaluating the Role of Renal Function in Cardiovascular Disease Risk Prediction
GFR Found to be Weak Marker in an Apparently Healthy Female Population
By Genna Rollins
Cardiovascular disease (CVD) and chronic kidney disease (CKD) are related in that traditional CVD risk factors are increased in patients with impaired renal function, and have been demonstrated to predict the start of CKD. However, most studies exploring the link between CVD and renal function have involved individuals at high risk for or already diagnosed with CVD, or with already impaired renal function. New research examining the association between renal function and CVD and death in an apparently healthy cohort of women is the subject of this issue of Strategies.
Various studies have suggested that renal function is a good predictor of future CVD events, but for the most part, these analyses have involved subjects who already had or were at high risk for CVD or who already had impaired kidney function. Studies of this nature involving individuals not selected on the basis of either CVD risk or kidney function have produced inconsistent results. However, interest has been high in finding markers of CVD that can refine traditional risk models, such as the Framingham Risk Score. In a newly published study, the authors therefore set out to look at the association between kidney function, CVD, and death in a large cohort of apparently healthy women (BMJ 2009;338:b2392 doi:10.1136/bmj.b2392).
The study involved participants in the Women’s Health Study, a landmark randomized, controlled trial with nearly 40,000 female health professionals designed to assess the benefits and risks of low-dose aspirin and vitamin E in the primary prevention of CVD and cancer. The analysis of kidney function and CVD risk involved 27,939 Women’s Health Study participants from whom blood samples were collected prior to randomization. Creatinine levels were measured by a rate blanked method that is based on the Jaffe reaction. The researchers used the Modification of Diet in Renal Disease Study (MDRD) equation to estimate glomerular filtration rate (GFR) as the primary measure of kidney function and classified participants into four groups based on GFR, including <60, 60 to 74.9, 75 to 89.9, and
≥90 ml/min/1.73 m2. They also attempted a sensitivity analysis for GFR ≤40 ml/min/1.73 m2, but had too few participants to perform meaningful analysis. The primary outcome of the study was any first CVD event, which was defined as non-fatal stroke, non-fatal myocardial infarction, coronary revascularization procedures, or death due to CVD. The investigators also assessed the association between renal function and any first CVD event, all-cause mortality, and non-CVD-related deaths.
The subjects were all at least 45 years old at study entry, and they were followed for a mean of 12 years. The vast majority (94%) were Caucasian, and none had a history of CVD, cancer, or other major illnesses. A majority (53.5%) had GFR values ≥90 ml/min/1.73 m2, while 28.9% had values between 75 and 89 ml/min/1.73 m2, 12.8% had values between 60 and 74.9 ml/min/1.73 m2, and 4.7% had values <60 ml/min/1.73 m2. During follow-up there were 1,199 incident CVD events, and 856 deaths, of which 179 were from CVD. Overall the investigators found no association between GFR categories and risk of major CVD event or death from any cause, and no association between GFR and myocardial infarction, coronary revascularization procedures, stroke, or non-CVD death. Compared with women who had GFR
values ≥90 ml/min/1.73 m2, the multivariable adjusted hazard ratios for any first CVD were 0.95 (95% CI 0.83 to 1.08), 0.84 (CI 0.70 to 1.00), and 1.00 (CI 0.79 to 1.27) among participants with GFR values of 75 to 89.9, 60 to 74.9, and <60 ml/min/1.73 m2, respectively. The analysis revealed a significantly increased risk of CVD death only in women with GFR values ≤60 ml/min/1.73 m2, with a hazard ratio of 1.68 compared with those with GFR values ≥90 ml/min/1.73 m2.
"We didn’t really see an overall association between kidney function and future cardiovascular disease, with the exception of cardiovascular deaths, which were increased in the group with impaired kidney function compared to those with normal kidney function," observed lead author Tobias Kurth, MD, ScD, lecturer in medicine at Harvard Medical School and associate epidemiologist at Brigham and Women’s Hospital in Boston.
The findings confirmed results from other studies and underscored the importance of knowing the populations involved in any analyses involving renal function and CVD, according to Kurth. "One really has to understand what is the target population. Is it elderly individuals with high cardiovascular risk factors or with already impaired kidney function? Because in that circumstance, kidney function likely has something to do with an increased risk for cardiovascular disease," he stressed. "But in primary prevention among apparently healthy individuals I think we have better predictors of CVD than kidney function."
While the study added to previous findings that kidney function is not a strong marker of future CVD in apparently healthy populations, it underscores the shortcomings of data based on serum creatinine levels, at least from older assays, and the inherent difficulties of the various methods for estimating GFR, according to Daniel Weiner, MD, MS, assistant professor of medicine at Tufts University School of Medicine in Boston. Until recent standardization efforts, serum creatinine concentrations could vary by as much as 17.7 μmol/L, which could have introduced misclassification bias into the study, particularly for samples with creatinine values <88.4 μmol/L, he indicated. "We’ve come a long way since the data for this study was collected. Creatinine assays and measurements across the country are so much better than even five years ago," he said. "A similar study to this one done now might actually show far more consistent information because there would be less bias and noise in terms of the error in measurement." Any shortcomings of the serum creatinine measurements would have been compounded in this study because serum creatinine measurements were only taken at baseline. "You have to recognize the results show an estimate of GFR and are based on a single measurement," he said.
Weiner also proposed that the data revealed a J-shaped association between kidney function and outcomes, so that women with the lowest serum creatinine concentrations, and therefore highest estimated GFR, trended to a higher risk of adverse outcomes than those with GFR values of 60 to 89.9 ml/min/1.73 m2, potentially indicating a heterogenous control group comprised of some healthy women with normal kidney function and some in which lower creatinine levels indicated those subjects had less muscle mass and some increased risk of adverse events. Kurth disagreed with this conclusion. "I would not go so far as making this statement out of our data," he said.
Kurth observed that there is "an ongoing discussion," in the medical and research communities about how to best estimate kidney function. A key question would be whether the method used resulted in a good distribution of estimated GFR values. "The extremes are easily identifiable. In between, it’s a bit difficult to make a cutoff, and here, ours was the commonly utilized cutoff of <60. We can discuss if we had used a different equation whether some people around the 60/61 cutoff would have been classified differently. That will always be a topic of discussion. However, we also used other estimation equations in our study and essentially found similar results." Kurth hopes to build on this analysis with new studies that will track any association between kidney function and CVD using newer methods to estimate kidney function.
Dr. Kurth has received within the last 2 years investigator-initiated research funding from the National Institutes of Health and Merck. Further, he is a consultant to i3 Drug Safety and World Health Information Science Consultants, LLC, and he has received honoraria from Genzyme, Merck, and Pfizer for educational lectures.