Race, genetic ancestry and estimate of renal function in CRF
Participants were enrolled in the Chronic Renal Insufficiency Cohort (CRIC) study, a prospective multicenter observational study of patients of various racial and ethnic backgrounds with CRF in the United States. In this study, which began in 2003, the initial cohort of 3,939 patients included an intentional sample of adults with a wide range of types and severities of CRF.17-19 In order to be representative of key elements of the U.S. population with CRF, the protocol specified that the cohort should be made up of approximately 40% of participants who identified as black, approximately 50% female, d ‘about 50% of people with diabetes and about 15 people with diabetes. % who identified as Hispanic. We randomly selected a subgroup of 1423 participants from the CRIC study to undergo direct urinary GFR measurement. 125I-iothalamate clearance. This approach involved stratified sampling to ensure representation across strata of diabetes status, stages of CRF, age, sex, race, and participating clinical centers.17-19 Our analytical sample included 1248 of these participants for whom the following data was also collected: race as reported by the participant, genetic ancestry markers and serum creatinine, serum cystatin C, and 24 hour urine creatinine level (Fig. S1) in the Supplementary Annex, available with the full text of this article on NEJM.org). The CRIC study was approved by the institutional review committees of the participating sites. All participants provided written informed consent, which included consent to participate in secondary analyzes such as the current study.
Study design and supervision
The general objectives of our study were to determine whether we could accurately estimate GFR without including race. We considered three alternatives (Fig. S2). First, we examined whether we could replace race with a quantitative measure of genetic ancestry in estimating GFR. Second, we assessed whether we could substitute for race in estimating GFR by taking into account determinants of serum creatinine that are unrelated to GFR and vary by race. Third, we assessed whether we could eliminate the need to take into account race – and genetic ancestry, given the results of our first round of analyzes – in estimating GFR by replacing the use of creatinine. serum by serum cystatin C as a marker of glomerular filtration.
The manuscript was written by the CRIC study researchers without external assistance. The study was designed by the CRIC Steering Committee. Data was collected by the seven CRIC clinical centers and data analysis was performed at Kaiser Permanente Northern California – University of California, CRIC Clinical Center in San Francisco, and CRIC Scientific and Data Coordinating Center at the University. from Pennsylvania. The decision to submit the manuscript for publication was made by all authors and approved by the Executive Publishing Committee and the Steering Committee of CRIC. The authors had full access to the data.
Variables used in this analysis included participant self-reported race, demographic characteristics, percentage of genetic African descent, body composition measures (body mass index [BMI], height, weight, body surface area, phase angle of bioelectric impedance analysis and bioelectric impedance analysis – quantified lean mass), daily reported and calculated dietary protein intake, 24 hour urinary creatinine excretion, estimates of tubular creatinine secretion calculated from measured creatinine clearance and GFR, serum creatinine level, serum cystatin C level, and measured GFR level. Specific analyzes and collection methods for all measurements, including a detailed description of the attribution of genetic ancestry, are provided in Table S1 and the Additional Methods section in the Supplementary Annex.
All analyzes were performed using SAS software, version 9.4 (SAS Institute) at Kaiser Permanente Northern California (Oakland) and replicated independently at the University of Pennsylvania (Philadelphia). We used standardized differences (Cohen’s D statistic for continuous and binary variables and Cramér’s V statistic for categorical variables) to compare distributions of African or European ancestry and other characteristics among participants who identified as black or not black. We then performed a series of analyzes to examine the three alternatives for estimating GFR without including race. The measured GFR underwent a logarithmic transformation when it came to the model result; therefore, the coefficients for race or ancestry were exponential and reported as percent changes for interpretability on the original score scale. Additionally, serum creatinine level and cystatin C level underwent log transformation in all models.
First, to examine the potential utility of genetic ancestry in estimating GFR, we used random sampling to classify data into developmental (67%) and validation (33%) data sets within each category of race or ethnic group, as reported by participants (Black, White or other race or ethnicity). We then combined the datasets to form a final development dataset and a validation dataset with equal distributions of race or ethnicity as reported by participants. For our analyzes, we combined the white and “other race or ethnic group” categories into a “non-black” category due to the relatively small number of participants of another race or ethnic group. As a sensitivity analysis, we also performed the analyzes using 10-fold cross-validation of the full sample. We first used the developmental dataset to derive linear regression GFR estimation equations for GFR measured with serum creatinine level, age, and sex (i.e. variables in the widely used chronic kidney disease epidemiology collaboration [CKD-EPI]ten and diet modification in renal failure [MDRD] study the equations11). Race and African ancestry were then added individually in separate models. We then applied all the derived equations to the validation dataset to estimate the GFR, and compared the performance using the mean square error (RMSE) calculated in the original GFR scale, the difference between the measured and estimated GFR (i.e. the statistical bias) and its interquartile range (i.e. the precision) and the proportion of GFR estimated to be less than 10% of the measured GFR ( called Pten) and less than 30% of the measured GFR (called P30) (i.e. precision). We constructed 95% confidence intervals using the 2.5th and 97.5th percentiles from 1000 bootstrap samples from the validation dataset. We also explored the possible added value of interaction terms between race or African descent and serum creatinine level.
Second, to assess whether the non-GFR determinants of serum creatinine were independently associated with race or genetic ancestry, we used the full sample to perform multivariate linear regression to separately examine the association of race. or African descent with measurements of body composition, dietary protein intake. , 24-hour urinary creatinine excretion and tubular creatinine secretion, adjusted for age, sex, and measured GFR level. The tubular creatinine secretion models were not adjusted for the measured GFR because the secretion estimates were based on the measured GFR in their calculation. We then assessed whether any of these non-GFR determinants of serum creatinine that differed by race or ancestry could substitute for black race or African descent in estimating GFR. We did this by evaluating how well these factors attenuated the strength of association of the black race or African descent coefficient in linear models for measured GFR that included serum creatinine level, black race, or black race. African descent, age and gender. We individually assessed each non-GFR determinant of serum creatinine that differed by race, as reported by participants, and we also evaluated all possible combinations to obtain a final model that maximizes the attenuation of the black race coefficient. or of African descent.
Third, we repeated the analyzes using serum Cystatin C as an alternative marker of glomerular filtration. We first used the full sample to examine independent associations of black and African descent with serum C cystatin, after adjusting for age, sex, and measured GFR. Then, we used the same developmental and validation datasets described above to derive (in the developmental dataset) GFR estimation equations with cystatin C, age, sex and either a term for black race or African descent, then to evaluate the performance of the model. (in the validation dataset).