Medical records-based chronic kidney disease phenotype for clinical care and “big data” observational and genetic studies

Abstract

Chronic Kidney Disease (CKD) represents a slowly progressive disorder that is typically silent until late stages, but early intervention can significantly delay its progression. We designed a portable and scalable electronic CKD phenotype to facilitate early disease recognition and empower large-scale observational and genetic studies of kidney traits. The algorithm uses a combination of rule-based and machine-learning methods to automatically place patients on the staging grid of albuminuria by glomerular filtration rate (“A-by-G” grid). We manually validated the algorithm by 451 chart reviews across three medical systems, demonstrating overall positive predictive value of 95% for CKD cases and 97% for healthy controls. Independent case-control validation using 2350 patient records demonstrated diagnostic specificity of 97% and sensitivity of 87%. Application of the phenotype to 1.3 million patients demonstrated that over 80% of CKD cases are undetected using ICD codes alone. We also demonstrated several large-scale applications of the phenotype, including identifying stage-specific kidney disease comorbidities, in silico estimation of kidney trait heritability in thousands of pedigrees reconstructed from medical records, and biobank-based multicenter genome-wide and phenome-wide association studies.

Genetics in chronic kidney disease: conclusions from a Kidney Disease: Improving Global Outcomes (KDIGO) Controversies Conference

“…heritability of CKD to be in the range of 25%-44%, with higher estimates for patients of African ancestry. 26 These estimates are generally consistent with traditional family-based heritability studies of CKD and glomerular filtration rate. [27][28][29] The relatively high heritability of CKD is likely attributable to both monogenic causes as well as complex or polygenic factors.…”

“…124 The development of standardized, scalable, and portable computable phenotypes is time consuming and represents many challenges, 125 but it can empower future genetic studies by automated identification of kidney-disease patients in large EHR databases. 26,126 Notably, just as important as accurately defining those with a disease is defining those without to serve as healthy controls in genetic studies, which is often harder. We envision that computable phenotyping can be used to find patients with or without CKD, hypertension, kidney stones, and glomerular disease, as well as patients who have received a kidney biopsy or kidney transplant.…”

“…We envision that computable phenotyping can be used to find patients with or without CKD, hypertension, kidney stones, and glomerular disease, as well as patients who have received a kidney biopsy or kidney transplant. In nephrology, computable phenotyping is underway, 26,[127][128][129] with CKD phenotyping perhaps best positioned for widespread implementation, given the availability of new algorithms based on International Classification of Diseases codes and laboratory values that are routinely measured in clinical practice. 26,126 In addition to genomic discovery, EHR-linked genetic research may allow for recontacting of patients with a specific genotype for detailed clinical and molecular studies.…”

GWAS defines pathogenic signaling pathways and prioritizes drug targets for IgA nephropathy

“…We performed meta-analysis of PheWAS results (meta-PheWAS) across two large biobank-based datasets: Electronic Medical Records and Genomics-III (eMERGE-III) and the UK Biobank (UKBB). The eMERGE-III network provides access to EHR information linked to GWAS data for 102,138 individuals; detailed quality control analyses of genetic data have been described previously 82–84 . Briefly, GWAS datasets were imputed using the latest multiethnic Haplotype Reference Consortium (HRC) panel using Michigan Imputation Server 85 .…”

“…The eMERGE-III network provides access to EHR information linked to GWAS data for 102,138 individuals; detailed quality control analyses of genetic data have been described previously [82][83][84] . Briefly, GWAS datasets were imputed using the latest multiethnic Haplotype Reference Consortium (HRC) panel using Michigan Imputation Server 85 .…”