Prescription of renin‐angiotensin‐aldosterone system inhibitors (RAASi) and its determinants in patients with advanced CKD under nephrologist care

Abstract Renin‐angiotensin‐aldosterone system inhibitors (RAASi) are recommended for chronic kidney disease (CKD) patients. In this study, we describe RAASi prescription patterns in the Chronic Kidney Disease Outcomes and Practice Patterns Study (CKDopps) in Brazil, Germany, France, and the United States (US). 5870 patients (mean age 66‐72 years; congestive heart failure [CHF] in 11%‐19%; diabetes in 43%‐54%; serum potassium ≥5 in 20%‐35%) were included. RAASi prescription was more common in Germany (80%) and France (77%) than Brazil (66%) and the United States (52%), where the prevalence of prescription decreases particularly in patients with CKD stage 5. In the multivariable regression model, RAASi prescription was least common in the United States and more common in patients who were younger, had diabetes, hypertension, or less advanced CKD. In conclusion, RAASi prescription patterns vary by country, and by demographic and clinical characteristics. RAASi appear to be underused, even among patients with strong class‐specific recommendations. Although the reasons for this variation could not be fully identified in this cross‐sectional observation, our data indicate that the risk of hyperkalemia may contribute to the underuse of this class of agents in moderate to advanced CKD.

All support was provided without restrictions on publications. All grants were made to Arbor Research Collaborative for Health and not to coauthors directly. None of the funders, apart from Vifor Fresenius Medical Renal Pharma Ltd, had any role in interpretation of data. None of the funders had any role in study design; collection and analysis of data; writing the report; or the decision to submit this report for publication. In France, CKDopps is part of the CKD-REIN cohort, which is funded by the Agence Nationale de la Recherche through the 2010 «Cohortes-Investissements d'Avenir » program and by the 2010 national Programme Hospitalier de Recherche Clinique. CKD-REIN is also supported through a public-private partnership with Amgen, Fresenius Medical Care, and GlaxoSmithKline (GSK) since 2012, Lilly France since 2013, Otsuka Pharmaceutical since 2015, Baxter and Merck Sharp & Dohme-Chibret (MSD France) from 2012 to 2017, and Sanofi-Genzyme from 2012 to 2015. In Germany, funding support for participation of German CKD clinics in CKDopps is provided by WiNe (Wissenschaftliches Institut für Nephrologie) of the Verband Deutsche Nierenzentren. In the United States and Brazil, support for the CKDopps Coordinating Center has been provided by Keryx. In Japan, CKDopps is part of the REACH-J-CKD Cohort, which is supported in part by a Grant-in-Aid for Research on Advanced CKD, Practical Research Project for Renal Diseases from Japan Agency for Medical Research and Development (AMED). diabetes, hypertension, or less advanced CKD. In conclusion, RAASi prescription patterns vary by country, and by demographic and clinical characteristics. RAASi appear to be underused, even among patients with strong class-specific recommendations.
Although the reasons for this variation could not be fully identified in this cross-sectional observation, our data indicate that the risk of hyperkalemia may contribute to the underuse of this class of agents in moderate to advanced CKD.

| INTRODUC TI ON
Given the high prevalence of diabetes, proteinuria, and congestive heart failure (CHF) in the chronic kidney disease (CKD) population, clinical practice guidelines support the prescription of renin-angiotensin-aldosterone system inhibitors (RAASi) for most patients with CKD as a strategy to limit CKD progression and manage cardiac comorbidities. [1][2][3] The prescription pattern of RAASi in patients with CKD is still unknown, particularly in general nephrology practice in a "real-world setting." Decreasing renal function and the associated interference with potassium (K) excretion is a major cause for potassium elevation. In clinical practice, however, the development of hyperkalemia is usually the result of a combination of factors superimposed on renal dysfunction, such as diabetes mellitus with high glucose levels or hyporeninemic hypoaldosteronism and advanced stages of heart failure with accompanying reductions in renal perfusion. 4 Additionally, serum potassium is influenced by the use of medications with hyperkalemic effects, such as angiotensin-converting enzyme inhibitors (ACEIs), angiotensin receptor blockers (ARBs), mineralocorticoid-receptor-antagonists (MRA), and medications that can lead a reduction in serum potassium level, such as diuretics, bicarbonate, and potassium binders. 5 As such, motivating hypotheses for this analysis are that the prescription of RAASi is relatively low in some patients with CKD, in contrast to practice guidelines and even in real-life studies based in academic practices. The main objectives of this study were as follows: (a) to describe RAASi use in the Chronic Kidney Disease Outcomes and Practice Patterns Study (CKDopps) overall and by country; by patient characteristics, including CKD stage, level of proteinuria, serum potassium level, diabetic/heart failure status; by use of other medications such as diuretics; and by clinic-level variation; and (b) to explore patient characteristics associated with RAASi use.

| Study design and analytic cohort selection
We conducted a cross-sectional analysis of data from the CKDopps.
CKDopps is an ongoing international prospective cohort study of non-dialysis patients with estimated glomerular filtration rate (eGFR) <60 mL/min/1.73 m 2 . Participants were sequentially or randomly selected from national samples of nephrologist-run CKD clinics in Brazil, France, Germany, Japan, and the United States (US). The CKDopps study design and protocol details have been published previously. 6 The ratio of enrolled patients with eGFR < 30 mL/ min/1.73 m 2 compared with patients with eGFR of 30-< 60 mL/ min/1.73 m 2 was 2:1 in Brazil and US clinics; 3:1 in German clinics; and 1:1 in French clinics. At study enrollment, patient demographics and comorbidities are captured, as well as laboratory data up to 6 months prior to enrollment, if available. Routine laboratory data are collected during longitudinal follow-up based upon how often these data are measured for clinical care, up to a monthly frequency.
Medication-related data are also collected longitudinally throughout study follow-up. In Brazil, France, and the United States, clinical data are captured via medical record abstraction, while in Germany, data are captured from electronic health records (EHRs) after data review performed by a study coordinator to assess data quality. The study was conducted with adherence to the Declaration of Helsinki and received research ethics board approval (study number: 14004-05).
All patients provided written informed consent before enrollment.

| Data and statistical analyses
Baseline cross-sectional data were used for the analyses.
d Among patients with valid answer to the question: During the last 3 mo, did your health care team recommend changes to potassium in your diet? (Brazil/US); Is there a recommendation to pay attention to your consumption of potassium? (France).
e Includes ACEi, ARB, direct renin inhibitors, and aldosterone receptor antagonists.

| RE SULTS
After the application of inclusion and exclusion criteria, 5870 participants were included in the analysis (Figure 1 The mean serum potassium level was higher in Brazil (4.75 mEq/L) than in France (4.54 mEq/L), Germany (4.60 mEq/L), and the United States (4.50 mEq/L) ( Table 1). Serum potassium levels, by country and by CKD stage, are presented in Figure 2. Serum potassium levels were higher in more advanced stages of CKD. RAASi prescription was more common in Germany (80%) and France (77%) than Brazil (66%) and the United States (52%) ( Table 1). The prevalence of RAASi prescriptions, by country and by CKD stage, is shown in Patients prescribed RAASi were younger in all countries except Germany, had less advanced CKD, had greater prevalence of diabetes and hypertension, had higher serum potassium level, and more frequently were prescribed diuretics (Table 1). According to the patient self-administered questionnaire, more than 50% of patients reported that the health care team did not recommend any change in potassium intake for the last 3 months prior to enrollment, despite the prescription of RAASi. The prevalence of RAASi prescription by indication for use is shown in Figure 4. RAASi prescription was more frequent in patients with albuminuria in Brazil and France, but similar despite the presence of albuminuria in other countries ( Figure 4A).
The prevalence of RAASi prescription was higher among those with diabetes than those without diabetes, despite the presence of albuminuria in France and the United States, but there was no noticeable difference in other countries ( Figure 4B). Patients with CHF and albuminuria had higher prevalence of RAASi prescriptions in Brazil, but this was not observed in other countries ( Figure 4C).
TA B L E 2 Prevalence ratios for RAASi a prescription from modified Poisson models with different levels of adjustment

| D ISCUSS I ON
Current guidelines support the prescription of RAASi for CKD patients, particularly those with diabetes, proteinuria, or heart failure.
The results of this multinational study indicate that RAASi are generally underused in advanced CKD, most strikingly in the United States. RAASi prescriptions were more frequent with the presence of diabetes and hypertension comorbidities and less frequent with renal dysfunction, particularly at advanced CKD stages.
There were significant differences in RAASi prescription patterns across the different countries captured in this study, with a consistently higher RAASi prescription in France and Germany compared with the United States and Brazil; these differences were particularly visible in more advanced stages of CKD. These findings contrast with the analysis of the Ramipril Efficacy in Nephropathy (REIN) trial, which shows that the benefit of RAASi in reducing the progression of CKD is not dependent of the degree of kidney dys- While hyperkalemia is only present in 2% of patients treated with RAASi due to hypertension with apparently normal kidney function, the incidence of hyperkalemia in patients with CKD is substantially higher. 10 In the Irbesartan Diabetic Nephropathy Trial (IDNT) of patients with type 2 diabetes and diabetic nephropathy, the incidence of hyperkalemia, as defined by a serum potassium >6.0 mmol/L, was 18.6%. 11 Previous studies identified CKD as an important risk factor for hyperkalemia. Interestingly, in patients with eGFR < 30 mL/min, the use of RAAS inhibitors was associated with an increased risk for hyperkalemia, and diuretics with an increased risk for hypokalemia. 12 In our study, hyperkalemia varied across different countries and particularly CKD stages.
As expected, mean serum potassium increases as GFR declines, as does the prevalence of hyperkalemia, as defined by serum potassium higher than 5.5 mEq/L. Hyperkalemia was more common in Brazil than the United States and Germany; several measures used to control hyperkalemia, such as prescription of diuretics, potassium-binding resins, and bicarbonate supplementation, as well as dietary advice varied across countries, were uncommon.
This finding contrasts with the importance of RAASi use and the high risk of hyperkalemia in this study's patient population.
Among medications studied, only the concomitant prescription of diuretics was associated with higher prevalence of RAASi prescription. Diuretics are commonly prescribed in the CKD population, particularly in moderate to advanced CKD for the treatment of hypertension and heart failure, but also as a strategy to control hyperkalemia. 13 The kaliuretic effects of these drugs may justify the higher prevalence of RAASi prescription in patients prescribed diuretics, since the increase of potassium excretion may reduce the potential of hyperkalemic events.
Given the high-level evidence in support of RAASi use in CKD patients, emphasis on ways to increase RAASi use is urgently war- Systems of care likely also influence RAASi use. Even assuming US nephrologists' intentions to prescribe RAASi to as many CKD patients as possible, disjointed models of care may limit achieving this goal. In the primary care clinic, emergency room, or at hospital discharge, decisions may be made without nephrologist consultation to discontinue RAASi, or to lower dose, in the face of moderate hyperkalemia or small increases in serum creatinine levels. In this context, better coordination of care seems essential to improving RAASi use, as well as other care and outcomes, for these medically challenging patients. Solutions will be undoubtedly complex and may require systems of accountable care, EHRs that span providers, 14,15 and focused implementation science. 16 The cross-sectional design of the analysis has some limitations in the interpretation of these analyses, since it does not allow for analysis of clinical decision-making, particularly related to events that motivated changes in prescription prior to the inclusion of patients in this study, for example, discontinuation, changes from dual to single therapy, and/or dose reduction that may occur in response to hyperkalemia. Further longitudinal investigation about the change in dose or change in prescription regimens in response to serum potassium levels is needed. Another limitation of the study is that the list of medications captured in the data analysis was limited due to the low prevalence of prescription, such as calcineurin inhibitors, sulfamethoxazole-trimethoprim, and non-steroidal anti-inflammatory agents.
RAASi prescription patterns in CKD vary by country, and by demographic and clinical characteristics. RAASi appear to be underused, especially in the United States, where only half were prescribed a RAASi, even among patients with strong class-specific recommendations, including albuminuria, diabetes, or heart failure. Higher potassium levels are associated with significantly higher RAASi prescription, before and after adjusting for covariates, an association which is likely to reflect the hyperkalemic effect of RAASi.
In conclusion, it is plausible that RAASi use may be less common in patients at higher risk of hyperkalemia; however, in cross-sectional analysis, this association may be confounded by: (a) RAASi discontinuation previous to the period of observation; and (b) introduction of anti-hyperkalemia measures (such as dose reduction, dietary advice, and potassium binder agents) not captured in the observation period. Therefore, prospective studies will need to address whether anti-hyperkalemia measures may allow greater RAASi use and potentially improve renal and cardiovascular outcomes in patients with moderate to advanced CKD.

ACK N OWLED G M ENTS
Heather Van Doren, senior medical editor at Arbor Research Collaborative for Health, provided editorial assistance on this manu-