Diagnosis of Chronic Kidney Disease Using Retinal Imaging and Urine Dipstick Data: Multimodal Deep Learning Approach

The Modification of Diet in Renal Disease (MDRD) and Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equations are two commonly used formulae to estimate glomerular filtration rate (GFR) in adults. The CKD-EPI equation is recommended in current international and local guidelines for the diagnosis and management of chronic kidney disease (CKD), unless an alternative equation has been shown to have superior accuracy. Validation and comparison of the equations in local populations are therefore required. Previous studies have reported on the accuracy of these prediction equations in black South Africans and those of Indian ancestry. To evaluate the MDRD and CKD-EPI equations in South African (SA) adults of mixed ancestry. In all participants, GFR was measured (mGFR) from plasma clearance of 99mTc-diethylenetetraaminepenta-acetic acid (99mTc-DTPA), using a standardised technique. Serum creatinine assays were isotope dilution mass spectrometry traceable. GFR was estimated (eGFR) using the MDRD and CKD-EPI equations, with and without the black ethnicity factor. The agreement, bias, precision and accuracy of each equation was determined. Eighty adults were included (30 male, median age 39 years, median GFR 59 mL/min/1.73 m2). Sixty-eight had a diagnosis of CKD, 10 were potential kidney donors, and 2 were healthy volunteers. Both equations, without the black ethnicity factor, had good agreement with measured GFR. The equations tended to overestimate GFR, with bias of 1.6 and 7.9 mL/min/1.73 m2 for the MDRD and CKD-EPI equations, respectively. The interquartile ranges of the differences were 15.9 and 20.2 mL/min/1.73 m2, and as a measure of accuracy, the percentages of estimates that fell within 30% of the mGFR (P30) were 80% and 72.5% (p=0.18). For identification of individuals with a GFR <60 mL/min/1.73 m2, the sensitivity of MDRD eGFR was 97.3% and that of CKD-EPI eGFR was 97.1%. The MDRD and CKD-EPI equations have shown satisfactory and comparable performance in this SA mixed-ancestry adult population, with the MDRD equation marginally less biased than the CKD-EPI.

An Endorsement of the Removal of Race From GFR Estimation Equations: A Position Statement From the National Kidney Foundation Kidney Disease Outcomes Quality Initiative

Indexing Estimates of GFR to Body Surface Area in Low-Resource Settings With a High Burden of Malnutrition: Evidence From Guatemala

Using Deep Learning Models to Characterize Major Retinal Features on Color Fundus Photographs

In Reply to ‘Creatinine-Based GFR Estimating Equations in Kidney Transplant Recipients’ and ‘Assessing Kidney Function in Transplant Recipients: Time to Work Together and Address the Most Relevant Questions’

Effect of removing race from glomerular filtration rate-estimating equations on anticancer drug dosing and eligibility: a retrospective analysis of National Cancer Institute phase 1 clinical trial participants.

Accuracy of the MDRD (Modification of Diet in Renal Disease) Study and CKD-EPI (CKD Epidemiology Collaboration) Equations for Estimation of GFR in the Elderly

Accurate Assessment of Kidney Function in Indigenous Australians: The Estimated GFR Study

Background:Accurate estimation of the glomerular filtration rate (GFR) and staging of chronic kidney disease (CKD) are important. Currently, there is no research on the differences in several estimated GFR equations for staging CKD in a large sample of centenarians. Thus, this study aimed to investigate the differences in CKD staging with the most commonly used equations and to analyze sources of discrepancy.Methods:A total of 966 centenarians were enrolled in this study from June 2014 to December 2016 in Hainan province, China. The GFR with the Modification of Diet in Renal Disease (MDRD), Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) and Berlin Initiative Study 1 (BIS1) equations were estimated. Agreement between these equations was investigated with the κ statistic and Bland-Altman plots. Sources of discrepancy were investigated by partial correlation analysis.Results:The κ values of the MDRD and CKD-EPI equations, MDRD and BIS1 equations, and CKD-EPI and BIS1 equations were 0.610, 0.253, and 0.381, respectively. Serum creatinine (Scr) explained 10.96%, 41.60% and 17.06% of the variability in these three comparisons, respectively. Serum uric acid (SUA) explained 3.65% and 5.43% of the variability in the first 2 comparisons, respectively. Gender was associated with significant differences in these 3 comparisons (P < 0.001).Conclusions:The strengths of agreement between the MDRD and CKD-EPI equations were substantial, but those between the MDRD and BIS1 equations and the CKD-EPI and BIS1 equations were fair. The difference in CKD staging of the first 2 comparisons strongly depended on Scr, SUA and gender, and that of CKD-EPI and BIS1 equations strongly depended on Scr and gender. The incidence at various stages of CKD staging was quite different. Thus, a new equation that is more suitable for the elderly needs to be built in the future.

BackgroundAccurately determining renal function is essential for clinical management of HIV patients. Classically, it has been evaluated by estimating glomerular filtration rate (eGFR) with the MDRD-equation, but today there is evidence that the new Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation has greater diagnostic accuracy. To date, however, little information exists on patients with HIV-infection. This study aimed to evaluate eGFR by CKD-EPI vs. MDRD equations and to stratify renal function according to KDIGO guidelines.MethodsCross-sectional, single center study including adult patients with HIV-infection.ResultsFour thousand five hundred three patients with HIV-infection (864 women; 19%) were examined. Median age was 45 years (IQR 37–52), and median baseline creatinine was 0.93 mg/dL (IQR 0.82–1.05). A similar distribution of absolute measures of eGFR was found using both formulas (p = 0.548). Baseline median eGFR was 95.2 and 90.4 mL/min/1.73 m2 for CKD-EPI and MDRD equations (p < 0.001), respectively. Of the 4503 measurements, 4109 (91.2%) agreed, with a kappa index of 0.803. MDRD classified 7.3% of patients as “mild reduced GFR” who were classified as “normal function” with CKD-EPI. Using CKD-EPI, it was possible to identify “normal function” (>90 mL/min/1.73 m2) in 73% patients and “mild reduced GFR” (60–89 mL/min/1.73 m2) in 24.3% of the patients, formerly classified as >60 mL/min/1.73 m2 with MDRD.ConclusionsThere was good correlation between CKD-EPI and MDRD. Estimating renal function using CKD-EPI equation allowed better staging of renal function and should be considered the method of choice. CKD-EPI identified a significant proportion of patients (24%) with mild reduced GFR (60–89 mL/min/1.73 m2).

BackgroundOur hypothesis was that both the Chronic Kidney Disease-Epidemiology Collaboration (CKD-EPI) and Modification of Diet in Renal Disease (MDRD) equations would underestimate directly measured GFR (mGFR) to a similar extent in people with diabetes and preserved renal function.MethodsIn a cross-sectional study, bias (eGFR – mGFR) was compared for the CKD-EPI and MDRD equations, after stratification for mGFR levels. We also examined the ability of the CKD-EPI compared with the MDRD equation to correctly classify subjects to various CKD stages. In a longitudinal study of subjects with an early decline in GFR i.e., initial mGFR >60 ml/min/1.73 m2 and rate of decline in GFR (ΔmGFR) > 3.3 ml/min/1.73 m2 per year, ΔmGFR (based on initial and final values) was compared with ΔeGFR by the CKD-EPI and MDRD equations over a mean of 9 years.ResultsIn the cross-sectional study, mGFR for the whole group was 80 ± 2.2 ml/min/1.73 m2 (n = 199, 75 % type 2 diabetes). For subjects with mGFR >90 ml/min/1.73 m2 (mGFR: 112 ± 2.0, n = 76), both equations significantly underestimated mGFR to a similar extent: bias for CKD-EPI: -12 ± 1.4 ml/min/1.73 m2 (p < 0.001) and for MDRD: -11 ± 2.1 ml/min/1.73 m2 (p < 0.001). Using the CKD-EPI compared with the MDRD equation did not improve the number of subjects that were correctly classified to a CKD-stage. No biochemical or clinical patient characteristics were identified to account for the under estimation of mGFR values in the normal to high range by the CKD-EPI equation. In the longitudinal study (n = 30, 66 % type 1 diabetes), initial and final mGFR values were 102.8 ± 6 and 54.6 ± 6.0 ml/min/1.73 m2, respectively. Mean ΔGFR (ml/min/1.73 m2 per year) was 6.0 by mGFR compared with only 3.0 by MDRD and 3.2 by CKD-EPI (both p < 0.05 vs mGFR)ConclusionsBoth the CKD-EPI and MDRD equations underestimate reference GFR values >90 ml/min/1.73 m2 as well as an early decline in GFR to a similar extent in people with diabetes. There is scope to improve methods for estimating an early decline in GFR.

Comparison of CKD Awareness in a Screening Population Using the Modification of Diet in Renal Disease (MDRD) Study and CKD Epidemiology Collaboration (CKD-EPI) Equations

GFR Estimation Using Standardized Cystatin C in Kidney Transplant Recipients

ObjectiveAs a standard indicator of renal function, the glomerular filtration rate (GFR) is vital for the prognostic analysis of elderly patients with coronary artery disease (CAD). Thus, the search for the calculation equation of GFR with the best prognostic ability is an important task. The most commonly used Modification of Diet in Renal Disease (MDRD) equation and the Chinese version (CMDRD) of the MDRD equation has many shortcomings. The newly developed Mayo Clinic quadratic (Mayo) and Chronic Kidney Disease (CKD) Epidemiology Collaboration (CKD-EPI) equations may overcome these shortcomings. Because the populations involved in these equation-related studies are almost completely devoid of subjects > 70 years of age, there are more debates on the performance of these equations in the elderly. This study was designed to compare the prognostic abilities of different calculation formulas for the GFR in elderly Chinese patients with CAD.MethodsThis study included 1050 patients (≥60 years of age) with CAD. The endpoint was all-cause mortality over a mean follow-up period of 417 days.ResultsThe median age was 86 years (60–104 years). The median values for the MDRD-GFR, CMDRD-GFR, CKD-EPI-GFR, and Mayo-GFR were 66.0, 69.2, 65.6, and 75.8 mL/minute/1.73 m2, respectively. The prevalence of GFR < 60 mL/minute/1.73 m2 based on these measures was 39.3%, 35.4%, 43.0%, and 28.7%, respectively. Their area under the curve values for predicting death were 0.611, 0.610, 0.625, and 0.632, respectively. Their cut-off points for predicting death were 54.1, 53.5, 48.0, and 57.4 mL/minute/1.73 m2, respectively. Compared with the MDRD-GFR, the net reclassification improvement values of the CMDRD-GFR, CKD-EPI-GFR, and Mayo-GFR were 0.02, 0.10, and 0.14, respectively.ConclusionThe prognostic abilities of the CKD-EPI and Mayo equations were significantly superior to the MDRD and CMDRD equations; the Mayo equation had a mild, but not statistically significant superiority compared with the CKD-EPI equation in elderly Chinese patients with CAD.

Since the derivation of the Modification of Diet in Renal Disease (MDRD) equation for estimating glomerular filtration rate (GFR), investigators determined that it cannot be used for drug dosing. In 2009, the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) derived an equation that was more accurate than the MDRD estimation of GFR. Therefore, questions exist about which method should be preferred in making dosage adjustments for renally eliminated antimicrobials. To determine whether a difference exists when making antimicrobial dosage adjustments in patients with CKD based on estimation of GFR using the CKD-EPI and Cockcroft-Gault equations. A database of 409 patients with CKD admitted to a tertiary care facility was used. GFR was calculated using both the CKD-EPI equation(s) and the Cockcroft-Gault equation and compared using correlation and Bland-Altman methodology. Dosage discordance rates of antimicrobials were determined. Average GFRs for all patients using the Cockcroft-Gault and CKD-EPI equations were 34.8 +/- 12 mL/min and 39.9 +/- 13 mL/min, respectively (5.09 [95% CI 4.60 to 5.59]; p < 0.001). The correlation coefficient between the 2 estimations was high (r = 0.91). The Bland-Altman plot yielded limits of agreement of 15.3 and -5.1; thus, the CKD-EPI estimation may range from 5.1 mL/min below to 15.3 mL/min above the Cockcroft-Gault estimation for 95% of the cases. A discordance rate of 15-25% existed among the recommended dosing adjustments of the selected antimicrobials when comparing the Cockcroft-Gault and CKD-EPI estimations. Though this study did not determine which equation should be selected to dose adjust antimicrobials, it demonstrated statistically significant differences between the Cockcroft-Gault and CKD-EPI equations. The clinical significance of these differences is uncertain in the absence of data assessing clinical outcomes that result from the use of the discordant doses. Clinical judgment should be employed when making renal dosage adjustments of antimicrobials.