A Clinical Workflow for Cost-Saving High-Rate Diagnosis of Genetic Kidney Diseases

Abstract

Significance statement: To optimize the diagnosis of genetic kidney disorders in a cost-effective manner, we developed a workflow based on referral criteria for in-person evaluation at a tertiary center, whole-exome sequencing, reverse phenotyping, and multidisciplinary board analysis. This workflow reached a diagnostic rate of 67%, with 48% confirming and 19% modifying the suspected clinical diagnosis. We obtained a genetic diagnosis in 64% of children and 70% of adults. A modeled cost analysis demonstrated that early genetic testing saves 20% of costs per patient. Real cost analysis on a representative sample of 66 patients demonstrated an actual cost reduction of 41%. This workflow demonstrates feasibility, performance, and economic effect for the diagnosis of genetic kidney diseases in a real-world setting.

Background: Whole-exome sequencing (WES) increases the diagnostic rate of genetic kidney disorders, but accessibility, interpretation of results, and costs limit use in daily practice.

Methods: Univariable analysis of a historical cohort of 392 patients who underwent WES for kidney diseases showed that resistance to treatments, familial history of kidney disease, extrarenal involvement, congenital abnormalities of the kidney and urinary tract and CKD stage ≥G2, two or more cysts per kidney on ultrasound, persistent hyperechoic kidneys or nephrocalcinosis on ultrasound, and persistent metabolic abnormalities were most predictive for genetic diagnosis. We prospectively applied these criteria to select patients in a network of nephrology centers, followed by centralized genetic diagnosis by WES, reverse phenotyping, and multidisciplinary board discussion.

Results: We applied this multistep workflow to 476 patients with eight clinical categories (podocytopathies, collagenopathies, CKD of unknown origin, tubulopathies, ciliopathies, congenital anomalies of the kidney and urinary tract, syndromic CKD, metabolic kidney disorders), obtaining genetic diagnosis for 319 of 476 patients (67.0%) (95% in 21 patients with disease onset during the fetal period or at birth, 64% in 298 pediatric patients, and 70% in 156 adult patients). The suspected clinical diagnosis was confirmed in 48% of the 476 patients and modified in 19%. A modeled cost analysis showed that application of this workflow saved 20% of costs per patient when performed at the beginning of the diagnostic process. Real cost analysis of 66 patients randomly selected from all categories showed actual cost reduction of 41%.

Conclusions: A diagnostic workflow for genetic kidney diseases that includes WES is cost-saving, especially if implemented early, and is feasible in a real-world setting.

Graphical abstract

Figure 1

Diagnostic workflow for the diagnosis of kidney diseases. CAKUT, congenital anomalies of the kidney and urinary tract; US, ultrasound scanning.

Figure 2

Diagnostic rate and disease reclassification. (A) Percentage of diagnosis confirmed (gray), negative (white), and modified (reticulated) by WES and reverse phenotyping. (B) Inner pie chart: distribution of patients according to the eight clinical categories based on pre-WES clinical evaluation (podocytopathies, collagenopathies, CKDu, tubulopathies, ciliopathies, CAKUT, syndromic CKD, metabolic kidney disorders). Outer pie chart: percentage of diagnosis confirmed (solid), modified (reticulated), and negative (white) in patients belonging to the eight clinical categories. (C) Percentage of diagnosis obtained with WES alone (light green), WES coupled with reverse phenotyping in the patient (striped), and WES coupled with reverse phenotyping in the patient and in the family (double striped) of patients belonging to the eight clinical categories. (D) On the left side, the suspected and on the right side, the genetic diagnosis in those patients that underwent disease reclassification after WES. CAKUT, congenital anomalies of the kidney and urinary tract; CKDu, CKD of unknown origin; RPF, reverse phenotyping in the family; RPP, reverse phenotyping in the patient; WES, whole-exome sequencing.

Figure 3

Diagnostic rate and genetic findings in different age groups. (A) Distribution of genetic diagnosis according to age in the eight clinical categories (podocytopathies, collagenopathies, CKDu, tubulopathies, ciliopathies, CAKUT, syndromic CKD, metabolic kidney disorders). Each dot represents a patient. Dots are colored according to the genetic diagnosis. Age is reported in years. (B) Diagnostic rate according to age in the study population (476 patients). For each age group (congenital, pediatric, adults), we show confirmed (gray), modified (reticulated), and negative (white) results. Age is reported in years. (C) Genetic findings according to age in 319 patients with a genetic diagnosis. Age is reported in years. CAKUT, congenital anomalies of the kidney and urinary tract; CKDu, CKD of unknown origin.

Figure 4

Cost analysis. (A) Modeled diagnostic trajectories for patients with suspected genetic kidney diseases based on clinical categories. In the late WES model, the patient undergoes investigations included in tiers 1, 2, and 3 until reaching a final diagnosis. The diagnostic workup is based on current guidelines, available literature, and local clinical practice for each clinical category. Tier 1 includes baseline investigations that allow clinicians to suspect inherited nephropathies as belonging to broad clinical categories (podocytopathies, collagenopathies, tubulopathies, ciliopathies, syndromic CKD, metabolic kidney disorders). Tier 2 and 3 include increasingly complex and/or expensive investigations. Tier 3 includes first-choice genetic testing for each clinical category, excluding WES. Genomic investigations were modeled according to literature, current guidelines, and the genetic architecture of each clinical category of kidney diseases. In the late WES model, if the diagnostic workup (including genomic and nongenomic investigations of tiers 2 and 3) turns out negative, the patient undergoes WES. In the early WES model, the patient goes through only tier 1 and then directly to WES. In both models, patients are selected with the clinical criteria adopted in this study. (B) Comparison between mean costs per diagnosis of the late WES model versus the early WES model in the study population. We excluded patients belonging to the CKDu category because of the lack of guidelines allowing us to model the diagnostic trajectory for this clinical condition. (C) Strategy applied for real-life costs analysis and comparison with the early WES model. Real-life diagnostic pathway included all the real costs retrieved from the clinical history of a subgroup of patients (n=66) with all available clinical information and recorded data in the administrative database. The clinical categories (podocytopathies, collagenopathies, tubulopathies, ciliopathies, syndromic CKD, metabolic kidney disorders) were equally represented. Costs of the real-life diagnostic pathway were compared with those derived from the hypothetical application of the early WES model to the same patients. (D) Comparison between mean costs per diagnosis of the real-life diagnostic pathway versus the early WES model in 66 patients. Icons included in Panels A and C are from the website Noun Project (thenounproject.com) and the credits go to their creators. CKDu, CKD of unknown origin; CAKUT, congenital anomalies of the kidney and urinary tract; WES, whole-exome sequencing.