Methods
Study eligibility
To be included in the CKD-PC meta-analysis, the study in principle has to have at least 1,000 adult (18+ years) participants and information at baseline on estimated GFR (eGFR) and urine albumin levels. The study has to have outcomes of interest with a minimum of 50 events. As recommended in clinical guidelines, we preferred urine albumin-to-creatinine ratio (ACR) as the measure of albuminuria. However, we also accepted urine albumin excretion as well as a qualitative measurement using dipstick, since a positive result is primarily due to increased albumin excretion and in some settings dipstick screening is widely used.
Analytic method for the papers
The participating studies were asked to prepare a dataset with relevant variables. To minimize heterogeneity, we circulated guidelines for definitions of variables (e.g. hypertension, diabetes, smoking) and dataset preparation. We instructed studies to use a complete data analysis for the two key risk factors, eGFR and albuminuria. For other variables we allowed for imputation with the mean value of the covariate or more sophisticated methods. Those studies unable to send the individual participant level data (de-identified) were sent a standard program (written in STATA) that was designed to automatically save all output needed for the meta-analysis including categorical/continuous analyses and tests of interaction. The participating studies sent their results to the coordinating center, where the data were pooled and analyzed across centers using STATA.
Statistical code
We have provided statistical code used in recent consortium analyses in the zip file below. We have included an overview file to provide more information on the codes. Please feel free to contact us with any questions at ckdpc@jhmi.edu.
To be included in the CKD-PC meta-analysis, the study in principle has to have at least 1,000 adult (18+ years) participants and information at baseline on estimated GFR (eGFR) and urine albumin levels. The study has to have outcomes of interest with a minimum of 50 events. As recommended in clinical guidelines, we preferred urine albumin-to-creatinine ratio (ACR) as the measure of albuminuria. However, we also accepted urine albumin excretion as well as a qualitative measurement using dipstick, since a positive result is primarily due to increased albumin excretion and in some settings dipstick screening is widely used.
Analytic method for the papers
The participating studies were asked to prepare a dataset with relevant variables. To minimize heterogeneity, we circulated guidelines for definitions of variables (e.g. hypertension, diabetes, smoking) and dataset preparation. We instructed studies to use a complete data analysis for the two key risk factors, eGFR and albuminuria. For other variables we allowed for imputation with the mean value of the covariate or more sophisticated methods. Those studies unable to send the individual participant level data (de-identified) were sent a standard program (written in STATA) that was designed to automatically save all output needed for the meta-analysis including categorical/continuous analyses and tests of interaction. The participating studies sent their results to the coordinating center, where the data were pooled and analyzed across centers using STATA.
Statistical code
We have provided statistical code used in recent consortium analyses in the zip file below. We have included an overview file to provide more information on the codes. Please feel free to contact us with any questions at ckdpc@jhmi.edu.
| ckd-pc_stata_code.zip |