Materials and Methods
Study Organization
The CKD-REIN study is coordinated by the Inserm and Université Paris-Sud Centre for Epidemiology and Population Health, the database is managed by the Agence de la Biomédecine who also runs the REIN registry, and biological samples are centrally stored in the Biobanque de Picardie. A partnership council meets annually to assess study progress; a steering committee (SC) comprising the 12 study partners meets monthly and takes any decisions regarding the study; an international scientific committee advises the SC on study protocol and research works; an industrial committee comprising representatives of industrial partners and three SC members provides input from an industry perspective; an ethic committee advises the SC on any ethical aspects. The CKDopps SC chair is member of the CKD-REIN SC and three members of the CKD-REIN SC are members of the CKDopps SC.
Study Design and Participants
CKD-REIN is a clinical-based prospective cohort enrolling patients with CKD Stage 3–4 receiving nephrologist-led care. A key goal is to obtain a representative sample of 3600 patients, of whom 1800 have CKD stage 3 and 1800 stage 4. Study eligibility requires two measures of eGFR between 15 and <60 mL/min/1.73 m at least 1 month apart without prior chronic dialysis or transplantation. Patients <18 years old or unable to give informed consent are excluded, as well as those who plan to move or decline participation.
Selection of Clinical Centres and Recruitment Goal
In order to describe actual CKD care at the national level and maximize variation in clinical practices and outcomes, we proceeded in two steps to select clinical settings. We first identified all of the 241 nephrology outpatient clinic settings which were classified by department and legal status: public, private-for-profit and private non-for-profit (Table 1). A total of 22 departments out of 95 located all over continental France were selected. We then included all clinic settings in seven departments which had four or less of them, and a sample in the others in order to obtain a representative sample with respect to legal status (Table 1). We also considered recruiting patients from renal care networks. These networks—a total of seven of which four are dedicated to non-end-stage CKD—were developed in order to improve the detection and secondary prevention of CKD by providing standardized monitoring, educational programmes and information to CKD patients and their physicians on a voluntary basis. It is worth while to evaluate their input in the context of this study. The mean recruitment goal is 72 patients per centre, according to settings, and 150 for renal care networks.
Patient Census and Enrolment
For each participating clinic and renal care network, a census will be developed for all outpatients seen during the year-long enrolment phase with finalization at the end of the year. The goals of this census are to facilitate the identification of potentially eligible study patients, to characterize the underlying CKD Stage 3–5 population in participating clinics as well as to assess the representativeness of the study participants with respect to all potentially eligible outpatients. The census considers all outpatients seen over the period with nephrologist-confirmed CKD diagnosis and eGFR <60 mL/min/1.73 m. Patients with CKD Stage 5, with no prior dialysis or transplantation, are listed to enable the assessment of the prevalence of nephrology-led CKD care in the seven departments including all nephrology settings. Data elements collected for the clinic census include patient gender, age, places of birth and residence, diabetes status, type of CKD, eGFR and the number of visits at the clinic during the past year. Whenever possible, these data will be collected from available medical and/or administrative electronic database.
After obtaining informed consent, eligible patients with CKD Stage 3–4 are enrolled during a routine nephrology visit. Trained clinical research assistants (CRAs) interview the patients and collect data from their medical records. A nurse takes blood and urine for both routine biological measurements and biobank. Participants are also asked to fill out a self-administered questionnaire either at the nephrology clinic or at home.
Follow-up
All cohort participants will be assessed annually for at least 5 years, including after the start of renal replacement therapy. Each year, data will be drawn from both medical records and health insurance files and patients will be asked to report clinical events. Interviews, self-administered questionnaires and routine biological measurements will be repeated at year 1, 3 and 5. For those lost to follow-up, we will search the national REIN and death registries for ESRD, vital status and causes of death. Passive follow-up through national registries will also be applied to all eligible patients in order to validate cohort representativeness.
Data Collection and Routine Laboratory Measurements
Data collection instruments include patient-level and provider-level questionnaires addressing both CKD-REIN and CKDopps objectives. Main topics considered in patient-level questionnaires are presented in Table 2. It is worth noting that we are planning to collect a large number of sociodemographic, environmental and lifestyle data in the same way as in the Constances study to enable further comparisons with this large population-based cohort. Several scales will be used to assess mental health, activities of daily living and family relationships. Provider-level questionnaires will collect data about clinic characteristics, organization and staffing, patient education, clinical management of blood pressure, CKD complications and vascular access, as well as transition to ESRD and transplant evaluation. Whether or not patients and their provider participate in a renal care network will be recorded. We will also search the national health insurance database to assess resource utilization indicators at the patient level, e.g. hospitalizations, clinic visits, laboratory measurements and imaging examinations, and drug prescriptions. The Etablissement français du Sang will provide comprehensive data on blood transfusions. Biological markers recommended by the French Health Authorities for routine care of CKD will be systematically measured in all patients at baseline and follow-up visits (Table 3).
Study Outcomes
CKD-REIN will follow up a number of renal and non-renal events. Renal events include progression to ESRD, treatment modality choice and timing of dialysis initiation, timing of conservative therapy if this option is chosen by the patient, timing of transplantation and wait-listing and episodes of AKI. Non-renal events include mortality and causes of death, cardiovascular disease (coronary heart disease, acute coronary syndrome, stroke, peripheral arterial disease, heart failure, dysrhythmia, sudden death, venous thromboembolism and pulmonary embolism), infections, fractures, cancer, cognitive impairment or sleep apnoea. Women health events including pregnancy outcomes, dysmenorrhoea, hormonal treatments and menopause will also be investigated. Quality of life will be assessed using validated adaptations of the 12-Item Short Form Survey (SF-12) and the Kidney Disease Quality of Life Instrument (KDQOL). Finally, adverse drug effects will be recorded from medical records and patient interviews, and hospitalizations from both medical records and health insurance files.
Sampling Collection and Storage
Fasting blood and second morning urine samples will be collected in all participants at enrolment and at study end, as well as at the 1- and 3-year follow-up in a subsample of 1200 patients. Serum, plasma, DNA and RNA will be stored at ultra-low temperature at the Biobanque de Picardie, an ISO 9001 and NFS 96900 certified biological resources centre.
Data entry and protection
The Agence de la Biomedecine maintains a national Information System for the management of the national transplant waiting list, the evaluation of transplant activities and the REIN registry. A web-based data collection system was developed for CKD-REIN using the same secured web portal and patient identification module. The use of a common patient identifier between CKD-REIN and the REIN registry will substantially facilitate the identification of ESRD events for patients lost to follow-up. The confidentiality, security and the integrity of data are covered by the Agence de la Biomedecine information system as it is a very sensitive issue for data regarding organ donation (secured portal, encryption, mirror copies of databases on servers located in another site).
Relevant quality control checks and queries will be executed at the national level by the CKD-REIN data manager and at the international level by Arbor Research. Data will be run through quality control programmes (e.g. meet range checks, units of measure, evaluation of non-response, etc.).
Statistical Analyses and Sample Size Considerations
Standard analyses will be carried out to describe patients and clinical practices, overall and by subgroups defined by age, gender and diabetes status. Multivariate regression models will be used to study the associations between patient characteristics, biomarkers, or clinical practices and the various study outcomes. We will systematically search for interactions between determinants and their impact on CKD outcomes, based on a priori hypotheses. In standard observational analyses to estimate treatment effects, the true effect may be distorted by indication bias. To lessen this bias, we will use a provider-based analytic approach, which is conceptually similar to instrumental variable analysis, to address indication bias; this methodology is commonly used in clinical studies such as the Dialysis Outcomes and Practice Pattern Study (DOPPS).
Regarding sample size, with 3600 patients, the minimum detectable hazard ratios for a time-to-event (e.g. survival) analysis comparing event rates from two equal groups with two-sided α = 0.05, 80% power, an average follow-up of 5 years and 10% loss of follow-up time are 1.25, 1.18 and 1.13, for event rates of 0.05, 0.10 and 0.20/year, respectively. The event rates of 0.05 and 0.1/year are based on published estimates of mortality before ESRD in patients with CKD Stage 3 or 4, respectively. ESRD rates may be higher than mortality rates, and composite ESRD or mortality may exceed 0.2/year.