Identifying CKD in populations - SCORED

Nine-Question Survey Identifies Patients at Risk of Having Chronic Kidney Disease

Source - Medscape

The Screening for Occult Renal Disease (SCORED) questionnaire, which contains questions about demographic variables and health conditions, but does not require any laboratory test values, might be useful to identify individuals who are highly likely to have underlying chronic kidney disease (CKD).

Study Highlights

• This cross-sectional analysis included 8530 men and women 20 years or older in the NHANES, a nationally representative, US-population-based survey conducted from 1999-2000 and 2001-2002.
• Based on the literature, potential determinants of CKD evaluated for incorporation into SCORED were age, sex, race, marital status, anemia, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, triglycerides, hypertension, diabetes mellitus, peripheral vascular disease, history of cardiovascular disease, history of congestive heart failure, proteinuria, smoking status, physical activity, body mass index, educational and income levels, and health insurance status.
• The SCORED prediction model used univariate and multivariate associations between this comprehensive set of risk factors and CKD, defined as a glomerular filtration rate less than 60 mL/minute per 1.73 m². Optimal characteristics of the model were examined with internal measures, and a model-based numeric scoring system was developed. The Atherosclerosis Risk in Communities study was used for external validation.
• Combining NHANES 1999-2000 and 2001-2002 resulted in a data set of 10,291 individuals 20 years or older. Data from 8530 individuals were analyzed after exclusion of 1472 individuals with missing serum creatinine measurements and 289 with other missing covariates.
• 601 (5.4%) of 8530 participants had CKD. Their mean age was 46 years, and 52% were women; 72%, white; 10%, black; and 14%, Hispanic.
• Only 9 variables had statistically significant associations with CKD. These were age (P < .001), female sex (P = .02), hypertension (P = .03), diabetes (P = .03), peripheral vascular disease (P = .008), history of cardiovascular disease (P = .001), history of congestive heart failure (P = .04), proteinuria (P < .001), and anemia (P = .003). Age 50 to 59 years was assigned 2 points, 60 to 69 years was assigned 3 points, and 70 years and older was assigned 4 points. All other predictive features were assigned 1 point.
• The multivariate model was well validated in the internal and external data sets, with area under the receiver operating characteristic curve of 0.88 and 0.71, respectively.
• Based on both diagnostic and qualitative criteria and practical implementation considerations, a SCORED value of 4 or higher was chosen as a cutoff point for screening. This cutoff value yielded a sensitivity of 92%, specificity of 68%, positive predictive value of 18%, and negative predictive value of 99%.
• Study limitations include a model heavily weighted toward common risk factors for kidney disease, inability to determine family history of kidney disease, cross-sectional design, inability to rule out the possibility of developing CKD in the future, possible underestimation of CKD, and use of a statistical method unable to investigate complicated effect modifications among the risk and protective factors.