Data collection

The data for this study were obtained from the PHARMO Data Network. This population-based network combines data from different primary and secondary healthcare settings, including data from the general practitioner (GP), in-hospital and outpatient settings, and pharmacy data in The Netherlands [6]. These data sources are linked at the patient level through validated algorithms.

The GP data include medical records – reported diagnosis and symptoms (based on the International Classification of Primary Care [ICPC] coding) [7], laboratory results, referrals to hospital care, and medication prescriptions – collectively shown to be representative for the Dutch population [8]. The prescription records include information on the product type, prescription date, dose, use, quantity, and route of administration. The hospital data contain admission data from the Dutch Hospital Data foundation of more than 80% of all Dutch hospitals, with balanced representation of the different types of hospitals in The Netherlands (e.g. academic, teaching and peripheral hospitals). Information on admission and discharge date, discharge diagnoses (based on WHO ICD coding [9]), and procedures (based on NZa (Nederlandse Zorgautoriteit [Dutch Healthcare Authority]) declaration codes and Dutch Classification of Procedures [10]) were available. The outpatient pharmacy data comprises healthcare products the GP or medical specialist prescribes.

Study population

For the objectives of this study, the following cohorts were selected:

1.

Prevalence cohort: To study the most recently available prevalence of CKD, all people with GP data were selected on 31 December 2019. This date was chosen due to the change in care during the coronavirus disease-19 (COVID-19) pandemic (in 2020). No other analyses were based on this cohort.

2.

Outcomes cohort: To determine the renal and cardiovascular complications over two years, people with GP data AND the possible availability of hospital data (because they live in the catchment area of the affiliated hospitals), were selected. The index date for this analysis was 31 December 2017, allowing two years follow-up until 31 December 2019. Due to the linkage of these databases, the need for geographical overlap, and the required availability of longitudinal data, the number of patients in this cohort was lower than in the prevalence cohort.

Privacy statement

Because only anonymized data from the PHARMO Data Network was used for this study, the study is not subject to ethics review according to the Medical Research Involving Human Subjects Act (WMO). The institutional review board of STIZON, Utrecht, Netherlands approved the study. The institutional review board of STIZON consists of representatives of all different healthcare providers and a separate privacy expert, including the GDPR data protection officer of STIZON. All studies require permission of the Compliance Committee and its decisions are based on the applicable Dutch and European legislations, including the Medical Treatment Contracts Act (WGBO).

Study design

This is a descriptive cohort study with a cross-sectional design to describe the prevalence of CKD (cohort 1) and a longitudinal design to describe the event rates of cardiorenal complications (cohort 2). The population of people with CKD was defined as follows at index date: 18 years of age or older and availability of at least 12 months of database history. In addition, presence of CKD was based on either a recorded diagnosis of CKD (as ICPC code U99.01 and/or as a free-text description in the GP data) prior to index date and/or a CKD stage defined by estimated GFR (eGFR) in combination with urine albumin/creatinine ratio (UACR) based on Kidney Disease Improving Global Outcomes (KDIGO) criteria [11], a CKD stage defined by the GP, or CKD stage recorded as a primary discharge diagnosis in the hospital data.

The stage of CKD was determined where possible, using the two most recent eGFR measurements in the two years before the index date, with a minimum of 90 days between these measurements. The CKD stages were defined according to the KDIGO classification [11]. When the two measurements fell into different stages, the best stage of the most recent two measurements was used. When eGFR was missing, or only one eGFR was available, the stage established by the GP was used when available. If insufficient information was available in the GP data to classify the CKD stage, it was based on the primary discharge diagnosis in the hospital data. The stage was considered unspecified if the stage could not be determined based on the above criteria.

Definitions of comorbidities

The diagnosis of type 1 or 2 diabetes mellitus (T1D and T2D) was determined step-by-step in the GP data:

1.

People with a recorded diagnosis of T1D (ICPC code T90.01) before or on the day of the first prescription of short-acting insulin (ATC code A10AB) were classified as T1D;

2.

People with a recorded diagnosis of T2D (ICPC code T90.02) within 12 months of the index date were classified as T2D;

3.

People younger than 30 years of age with an insulin prescription (ATC code A10A) were classified as T1D;

4.

People with a prescription of at least one blood glucose lowering therapy other than insulin (ATC code A10B) in the 6 months before the index date were classified as T2D;

People who did not meet the above criteria were considered unclassifiable and removed from the study. This concerned four patients.

A recorded diagnosis determined heart failure (ICPC code K77) and other comorbidities in the GP data before or on the index date. Cardiovascular disease at index date was defined as (a history of) ischemic heart disease, heart failure, atrial fibrillation, stroke and/or peripheral artery disease. Clinical parameters (UACR and systolic blood pressure) were determined based on the most recent value in the two years before the index date. Medication use was determined based on prescriptions in the year before the index date.

Complications

The occurrence of all-cause mortality, cardiovascular mortality (defined as death during hospitalization with a cardiovascular diagnosis), renal mortality (defined as death during hospitalization with renal diagnosis), myocardial infarction, cerebrovascular events, hospitalization for heart failure, hospitalization for CKD, cardiorenal syndrome (defined as a concurrent diagnosis of both heart failure and CKD during hospitalization), decreased in eGFR values of at least 57% [12], kidney transplantation, and peripheral arterial vascular disease was recorded in the two years of follow-up. These complications were determined by primary discharge diagnosis as stated in the hospital data, except for a decrease in eGFR values, which was established using GP data. Per complication, the first event was used to determine the incidence rate. Complications were assessed in the entire study population with CKD in cohort 2 and in the subgroups of people with the additional diagnosis of T2D and/or heart failure.

Statistics

Renal and cardiovascular complications are described as the incidence (incidence rate, (IR)) per 1,000 person-years. Since this is a descriptive study, no statistical hypothesis testing between groups was performed. A statistical analysis plan was written in advance for this study. The analyses were performed in SAS version 9.4.

Data availability

The data that support the findings of this study are available from STIZON via the PHARMO Institute but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available from the authors upon reasonable request and with permission of STIZON.