Risk Calculator Provides Early Warning of Chronic Kidney Disease

One of the most influential risk factors was diabetes

With Joseph Coresh MD, PhD, and Elena A. Christofides MD

Given that chronic kidney disease is associated with increased risk of all-cause mortality, cardiovascular disease, and end stage renal disease,1 it is imperative that clinicians have an efficient method of identifying patients who are at risk in order to initiate steps to help prevent these adverse health outcomes.

Data from large scale study used to create risk assessment tool for CDK.

Need to expand routine risk assessment of kidney function

This recognition was behind efforts initiated at Johns Hopkins School of Public Health to develop a convenient risk assessment tool to foster a more proactive effort among clinicians to identify patients with an increased likelihood of developing chronic kidney disease (CKD).

Researchers at the Chronic Kidney Disease Prognosis Consortium in Baltimore, Maryland, analyzed data from more than five million individuals across 28 countries, including some 780,000 with diabetes, to determine the degree of risk for CKD (measured by estimated glomerular filtration rate [eGFR]).2

“The idea is that CKD, in all of its stages, is very common and therefore it is important to know who is at risk for CKD to optimize prevention efforts,” said co-principal investigator Joseph Coresh, MD, PhD, Professor of Epidemiology, Medicine, and Biostatistics at the Johns Hopkins University School of Medicine, in Baltimore, Maryland.

The researchers developed a set of equations to calculate the 5-year risk of developing a reduced eGFR, which incorporated a range of sociodemographic and clinical factors (e.g., comorbidities, medications, laboratory test results) and reported a “demonstrated high discrimination” in populations both with and without diabetes.2

“We found that the most influential risk factors for developing chronic kidney disease were diabetes, hypertension, older age, female sex, current level of kidney function, and level of urine albumen,” said Dr. Coresh. The findings were published online in the Journal of the American Medical Association.

While kidney failure risk equations may be helpful in improving patient care in those with established CKD, there is a paucity of work in developing predictive tools to identify patients who are at high lifetime risk for CKD—an estimated 59.1% of the United States population.3

The research team sought to develop a “simple risk assessment tool that helps clinicians quickly identify patients at increased risk of reduced eGFR,” drawing on participants from the general public as well as those known to be at high risk for cardiovascular disease, Dr. Coresh said.   

Methodology used in developing the CDK risk assessment tool

Eligible participants had to be at least 18 years old, with an eGFR of >60 mL/min/1.73 m2 at baseline. Moreover, participants were excluded if they had end stage kidney disease and had to have more than one serum creatinine value recorded during the follow-up period.2

The model included 31 cohorts of individuals without diabetes and 15 cohorts of individuals with diabetes. The researchers developed one risk tool for the general population and another specifically geared to individuals with type 2 diabetes.

Variables included in the risk tool for those without diabetes are found below:

  • Demographic (age, sex, race/ethnicity)
  • eGFR
  • Medical history (i.e., CVD, ever-smoking, hypertension, body mass index [BMI], and lab values including serum creatinine and albuminuria)

For participants with diabetes, additional factors were incorporated:

  • Medications (i.e., insulin, oral medications only, no antidiabetes medication)
  • Hemoglobin A1c (HbA1c) values
  • Calculation of the interaction between antidiabetes medications and HbA1c

The Harrell C statistic, estimated within each cohort, was used to evaluate model discrimination.2

CKD risk calculators provide “very good to excellent” accuracy

Of the individuals without diabetes (n=4,441,084 with mean age 54 years +/-16, 38% female), 14.9% of the total population exhibited evidence of incident CKD during the 4.2 year follow-up. Of these, CKD was subsequently confirmed by eGFR measurements in 56.7% of these participants.

In patients with a diagnosis of diabetes at the outset (n=781,627 with a mean age 62 years +/- 11, 13% female), 40.1% showed evidence of incident CKD during the 3.9 year follow-up, with 67.7% of the cases confirmed by subsequent eGFR measurements.2

In both cohorts, regardless of the presence or absence of diabetes, older age, female sex, race, hypertension, history of CVD, lower eGFR values, and higher urine albumin:creatinine ratio resulted in a significantly greater incident low eGFR (<60mL/min/1.73m2).

It's worth noting that among your patients, women and African-Americans are more likely to develop CKD, Dr. Coresh said.

On the other hand, smoking proved to be a significant risk factor for incident eGFR only in individuals who did not have diabetes, while elevated HbA1c and use and type of diabetes medicines were significantly associated with low eGFR (<60 mL/min/1.73 m2) in patients who had diabetes.2

The authors reported that the median Harrell C statistic for the 5 year predicted probability of all low eGFR occurrences reflected “good discrimination” across all participants (0.845 [IQR, 0.789-0.890 in those without diabetes and 0.801 [IQR, 0.750-0.819] in those with diabetes).

Patients with subsequently confirmed eGFR events of <60 mL/min/1.73 m2 had similar median Harrell C statistics as those in the 5 year model (0.869 [IQR, 0.823-0.897 and 0.808 [IQR, 0.794-0.836], respectively).2

As might be expected, the absolute risk was generally higher among individuals with diabetes, compared to those without disease. On the other hand, older age and elevated albuminuria were both significantly associated with higher absolute risk, regardless of diabetes status.

CDK risk assessment tool validated

To verify the accuracy of the risk calculator, the researchers applied the data to an entirely new set of individuals (18 study populations in 9 external validation cohorts drawn from the OptumLabs Data Warehouse), which included over 2 million individuals (n=2,253,540).2

Dr. Coresh and colleagues were able to confirm that the 5 year predicted probability of all eGFR events of <60 mL/min/1.73 m2 was comparable to that of the non-diabetic cohort and diabetic cohort used in developing the risk calculator (0.84 and 0.81, respectively).

The calibration revealed that 89% of those validation cohorts had a slope of observed to predicted risk between 0.80 and 1.25,2 according to Dr. Coresh; as such, “the accuracy of the 5 year risk score calculator fell into the ‘very good’ range in those with diabetes and the ‘excellent’ range in those without.”

Key finding—measure urinary albumin at least annually to capture CDK

“One take-home message of our findings for clinicians is that urinary albumin should be measured annually in all patients with diabetes, as is recommended by the National Kidney Foundation for individuals with hypertension, older age, or a family history of kidney disease,”4 said Dr Coresh.

While lifestyle modification is helpful to lessen all diabetes-related complications, and to forestall chronic disease in all individuals, “it could in particular be a mainstay in prevention for people who do not have albuminuria and want to reduce their future risk for developing kidney disease,” he said.

After all, weight loss and physical activity are “particularly powerful when achieved to reduce the risk of diabetes, metabolic abnormalities, and progression to kidney disease.” In addition, “salt reduction remains a necessary intervention for blood pressure control and avoiding medications that are toxic to the kidneys should be heeded,” he said.

“In addition, albuminuria may serve as an indicator of when to consider prescribing SGLT2 [sodium glucose cotransporter-2] inhibitors in patients with diabetes because they are more kidney-specific, and may even work outside of diabetes,” he said.

The risk calculator is available at www.ckdpcrisk.org.

No predictive model has been a “blind spot” in chronic disease management

“The need to create a model of kidney decline addresses a critical blind spot in our future patient management,” said Elena A. Christofides, MD, a Columbus, Ohio-based endocrinologist. She is also a clinical instructor and chair of the department of internal medicine at Mount Carmel Health Systems of Columbus and clinical associate professor at Ohio University College of Medicine.

Dr. Christofides described the quantity of data used to perform the analyses as impressive, adding that it is “encouraging to see the model’s applicability to novel populations after the initial calculations.”

“It is a worthy endeavor to continue this line of inquiry,” she said. “However, there were several places where information was impugned, which limits the validity of the formula,” echoing concerns raised by an accompanying editorial6 by Sri Lekha Tummalapalli, MD, MBA, and Michelle M. Estrella, MD, MHS, both of the University of California San Francisco School of Medicine.

 “An important consideration in interpreting this work is that these risk prediction equations assume that key health factors have been measured, particularly eGFR, but in clinical settings, these laboratory measurements may have been performed based on the physician’s perception of the patient’s CKD risk and thus may affect the generalizability of the risk equation,” they said.

Despite this concern, Drs. Tummalalpalli and Estrella view the study as a critical step toward reducing the global burden of kidney disease by shifting the focus from secondary to primary prevention, since the well-validated Kidney Failure Risk Equation only stratifies risk of kidney failure in patients who already have moderate-to-severe CKD.5

Therefore, “identifying individuals before the onset of kidney disease would present a much larger opportunity to influence disease trajectory,” they wrote.

The research was supported with grants from the National Kidney Foundation and the National Institute of Diabetes, Digestive, and Kidney Diseases. Neither Dr. Coresh nor Dr. Christofides had any financial conflicts regarding their comments for this study.

Continue Reading:
Metformin May Lessen Diabetes Complications, Improve Cognitive Decline
close X