Article Prevalence and Factors Associated with ... - CiteSeerX

Article

Prevalence and Factors Associated with Hyperkalemia in Predialysis Patients Followed in a Low-Clearance Clinic Pantelis A. Sarafidis, Rochelle Blacklock, Eleri Wood, Adam Rumjon, Shanique Simmonds, Jessica Fletcher-Rogers, Rachel Ariyanayagam, Aziza Al-Yassin, Claire Sharpe, and Katie Vinen

Summary Background and objectives Recent studies evaluated the prevalence of hyperkalemia and related risk factors in patients with CKD of various stages, but there is limited relevant information in predialysis patients. This study aimed to examine the prevalence and factors associated with hyperkalemia in the structured environment of a low-clearance clinic.

Academic Department of Renal Medicine, King’s College Hospital, London, United Kingdom

Design, setting, participants, & measurements In a cross-sectional fashion over a prespecified period of 4 months, information on serum potassium and relevant laboratory variables, comorbidities, medications, and dietician input in patients with advanced CKD under follow-up in the low-clearance clinic of our department was recorded. Univariate and multiple logistic regression analyses were used to identify factors associated with serum potassium$5.5 meq/L.

Correspondence: Dr. Katie Vinen, Academic Department of Renal Medicine, King’s College Hospital, Denmark Hill, SE5 9RS London, United Kingdom. Email: katie. [email protected]

Results The study population consisted of 238 patients aged 66.264.2 years with estimated GFR of 14.564.8 ml/min per 1.73 m2. The prevalence of hyperkalemia. defined as potassium.5.0, $5.5, and $6.0 meq/L., was at 54.2%, 31.5%, and 8.4%, respectively. In univariate comparisons, patients with potassium$5.5 meq/L had significantly higher urea and lower estimated GFR and serum bicarbonate; also, they were more often using sodium bicarbonate and had received potassium education and attempts for dietary potassium lowering. Use of angiotensin-converting enzyme inhibitors or angiotensin receptor blockers was not associated with hyperkalemia. In multivariate analyses, estimated GFR,15 ml/min per 1.73 m2 and sodium bicarbonate use were independently associated with hyperkalemia. Conclusions The prevalence of hyperkalemia in predialysis patients with CKD is high. Even at this range of renal function, low estimated GFR seems to be the most important factor associated with hyperkalemia among the wide range of demographic, clinical, and laboratory characteristics studied. Clin J Am Soc Nephrol 7: 1234–1241, 2012. doi: 10.2215/CJN.01150112

Introduction Hyperkalemia is an established complication of reduced renal function in patients suffering from either CKD or acute renal failure (1,2), and it is long considered a potentially life-threatening condition because of the risk of ventricular arrhythmias and cardiac arrest when serum K+ is severely elevated (3,4). Although decreasing renal function and the associated interference with potassium excretion is a major cause for potassium elevation, in clinical practice, 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, advanced stages of heart failure with accompanying reductions in renal perfusion, concurrent high-potassium diet, use of potassium-based salt substitutes, and use of medications interfering with potassium homeostasis like angiotensin-converting enzyme inhibitors (ACEIs), angiotensin receptor blockers (ARBs), aldosterone receptor antagonists, b-blockers, and others (5–7). 1234

Copyright © 2012 by the American Society of Nephrology

For several years, one of the main concerns of physicians treating patients with CKD was to balance between the undisputed benefits of ACEIs and ARBs toward renal function preservation in proteinuric nephropathies (8) and the associated risk of hyperkalemia with these agents (9,10). This issue became even more important because of the accumulating evidence suggesting the potential benefits of the ACEI/ARB combination or adjunct aldosterone blockade toward renoprotection (11,12) as well as the data suggesting beneficial effects of medications that interfere with potassium homeostasis on other conditions commonly present in patients with CKD, such as the effect of aldosterone blockers on chronic heart failure or resistant hypertension (13,14) or the use of b-blockers for cardioprotection and effective hypertension control (15). Because of the increasing complexity of the above field, recent studies have attempted to delineate the relationship between CKD and hyperkalemia. These studies estimated either the potassium levels and www.cjasn.org Vol 7 August, 2012

Clin J Am Soc Nephrol 7: 1234–1241, August, 2012

prevalence of hyperkalemia (among other CKD complications) with decreasing levels of renal function (16–18) or the incidence of hyperkalemia associated with CKD stage, medication use, and other hyperkalemic factors (19,20). However, relevant studies are relatively few and suffer from some limitations, such as the retrospective nature, the variable definitions of hyperkalemia, and the different type of analyses used. In addition, none of them have simultaneously assessed the effect of all factors previously suggested to interfere with potassium levels in individuals with CKD. Most importantly, there is a paucity in current literature on hyperkalemia and associated factors in predialysis patients not in the general population but followed in a structured nephrology environment, an issue that is perhaps more relevant to everyday clinical practice. Thus, the aim of this study was to examine the prevalence and potential determinants of hyperkalemia in a population of predialysis CKD patients, such as patients followed in a low-clearance clinic (LCC) of a tertiary university hospital.

Materials and Methods Study Design and Patients This is a cross-sectional study in patients with predialysis CKD under regular follow-up in the LCC of our department. The LCC generally accepts referrals from other nephrology clinics of patients with estimated GFR (eGFR) below 20 ml/min per 1.73 m 2 or anticipated start of dialysis within 1 year. It is a structured, multidisciplinary clinic aiming to optimize the care of patients with advanced CKD to prepare them for their preferred dialysis method or pre-emptive transplantation or offer them conservative treatment accordingly. Patients are seen in the clinic by nephrology physicians (specialists or registrars) or specialist nurses at intervals ranging from 1 week to 3 months. In routine practice, all laboratory tests obtained in the clinic are checked by the LCC staff on the same or next day; in addition, the central laboratory immediately informs the renal team of any case with serum potassium$6.5 meq/L. Patients with potassium$6.5 meq/L are contacted by phone by the department staff and asked to attend the local acute and emergency department for a repeat potassium check, electrocardiogram, and proper treatment, if necessary; patients with potassium between 6 and 6.5 meq/L are usually invited for a repeat check the following day and receive individualized treatment. For the purpose of this study, we collected data at a single time point (i.e., at the first scheduled visit during a prespecified period of 4 months from May 15 to September 15, 2011) for all patients regularly attending the clinic, with the exception of those patients on the conservative care pathway who were being seen only in their own homes or in outreach clinics and not in the main clinic (to avoid the confounding effect of delayed potassium measurement at the laboratory on the study results). At each regular LCC clinic visit, blood samples were drawn by a dedicated phlebotomist under a standard process within the outpatient clinic environment. During the study 4-month data-recording period, bloods were also delivered to the laboratory at regular intervals (i.e., every 1 hour) during the clinic running time by LCC staff.

Hyperkalemia in Predialysis Patients, Sarafidis et al.

1235

Data Collection All study data for each subject were recorded during the relevant clinic on the electronic clinical database of our department and a purpose-built data-collecting sheet. We collected routine data on demographics, cause of CKD, comorbidities (hypertension, diabetes, coronary heart disease, stroke, heart failure, liver disease, and history of malignancy), and blood biochemical parameters (potassium, sodium, bicarbonate, urea, creatinine, and others) for each participant. Furthermore, we captured information on use of medications that could interfere with potassium regulation, including ACEIs, ARBs, renin inhibitors, aldosterone blockers, b-blockers, thiazides, loop diuretics, nonsteroidal anti-inflammatory drugs or cyclooxygenase inhibitors, insulin and oral hypoglycemic agents, heparin, trimethoprim, pentamidine, cyclosporine, tacrolimus, digoxin, resin K+ exchangers, sodium docusate, lactulose, senna, and sodium bicarbonate. In addition to the above data, to identify additional causes of hyperkalemia, the attending physician/specialist nurse recorded (at each prespecified visit) information on suspected causes that could possibly lead to acute on chronic renal failure within the past 4 weeks (dehydration/hypotension, infection/sepsis, postsurgery, postcontrast-related procedure, urinary obstruction, or other) as well as other possible causes of hyperkalemia within the last week (blood transfusion, crush injury, surgery, or other). Finally, at each study visit, our dietetics team recorded information relevant to the dietician input for each participant, including the date of the most recent assessment, the reason for referral (standard, hyperkalemia, or other), and the inclusion of potassium education and attempt for potassium dietary manipulation at that assessment. Dietary potassium manipulation was defined as possible when consumption of high-potassium foods was identified during the individualized diet assessment at the most recent dietician appointment, not possible when potassium-rich foods were not identified, and not applicable when dietary assessment for potassium was not performed in the specific appointment (i.e., performed previously and/or the appointment was for another reason). Definitions Hyperkalemia was defined as serum potassium.5.0 meq/L. Hypokalemia was defined as serum potassium,3.5 meq/L. In addition to the above conventional thresholds, we examined the percentage of patients with serum potassium levels$5.5, .5.5, $6.0, and .6.0 meq/L for reasons of relevance with nephrology clinical practice and comparison with previous studies. For the same reasons, we used the level of $5.5 meq/L for the univariate and multivariate analyses of factors associated with hyperkalemia. GFR was estimated from serum creatinine levels with the use of the Modification of Diet in Renal Disease equation, which is described in ref. 21. Statistical Analyses Statistical analyses were performed with Statistical Package for Social Sciences version 17.0 for Windows (SPSS Inc., Chicago, IL). Continuous variables are presented as mean 6 SD or median and range, and categorical variables are presented as absolute and relevant frequencies. In univariate analyses, clinical and laboratory parameters were compared for patients stratified by potassium levels using

1236

Clinical Journal of the American Society of Nephrology

chi-squared or Fisher exact tests (categorical variables) and independent t or relevant nonparametric tests if necessary (continuous variables). In addition, multiple logistic regression analysis was performed to assess the association of several demographic and clinical characteristics with hyperkalemia. Variables were tested for interaction and included in the model if P,0.20 in univariate analysis. We report adjusted odd ratios with 95% confidence intervals (CIs). Probability values of P,0.05 (two-tailed) were considered statistically significant.

Results Demographic and Clinical Characteristics From a total of 317 active patients under regular followup in the LCC of our Department of Renal Medicine as of May 1, 2011, 72 patients were excluded, because they were

on the conservative care pathway and regularly seen in outreach clinics or their own homes. Another three patients were not included, because they missed their clinic appointments during the prespecified follow-up period; additionally, three patients started dialysis, and one patient died before their clinic visit. Thus, the final study population consisted of 238 patients, and the clinical and demographic characteristics are shown in Table 1. Hyperkalemia Prevalence The mean level of potassium in the population studied was 5.160.6 meq/L. The prevalence of hyperkalemia defined as serum potassium.5.0 meq/L in the overall population was 54.2%. Another 45.4% of patients had normal potassium levels (3.5–5.0 meq/L), whereas only one patient (0.4%) had potassium levels,3.5 meq/L. In addition, a total of 31.5%,

Table 1. Demographic, clinical, and routine biochemical characteristics of the total study cohort and patients with and without elevated potassium levels

All Patients (n=238) Age (years) Sex (n) male female Ethnicity (n) white black African/black Caribbean Indian Chinese other/mixed/unknown Follow-up time in LCC (days) Primary cause of kidney disease (n) diabetic kidney disease hypertension/renovascular disease glomerulonephritis tubulointerstitial disease polycystic kidney disease unknown Comorbidities (n) diabetes hypertension dyslipidemia coronary heart disease previous stroke peripheral vascular disease heart failure history of malignancy Body weight (kg) Body mass index (kg/m2) Systolic BP (mmHg) Diastolic BP (mmHg) Urea (mmol/L) Creatinine (mmol/L) eGFR (ml/min per 1.73 m2) Sodium (meq/L) Potassium (meq/L) Serum bicarbonate (meq/L)

66.2614.2

Potassium,5.5 meq/L (n=163) 67.1614.2

Potassium$5.5 meq/L (n=75) 64.2614.2

P Valuea 0.15

139 (58.4%) 99 (41.6%)

89 (54.6%) 74 (45.4%)

50 (66.7%) 25 (33.3%)

110 (46.2%) 56 (23.5%) 14 (5.9%) 2 (0.8%) 56 (23.5%) 422 (8–5528)

70 (42.9%) 42 (25.8%) 9 (5.5%) 2 (1.2%) 40 (24.5%) 358 (8–3549)

40 (53.3%) 14 (18.7%) 5 (6.7%) 0 (0%) 16 (21.3%) 463 (15–5528)

104 (43.7%) 77 (32.4%)

69 (42.3%) 60 (36.8%)

35 (46.7%) 17 (22.7%)

14 (5.9%) 20 (8.4%) 9 (3.8%) 14 (5.9%)

6 (3.7%) 14 (8.6%) 5 (3.1%) 9 (5.5%)

8 (10.7%) 6 (8.0%) 4 (5.3%) 5 (6.7%)

0.13

36 (48.0) 73 (97.3%) 34 (45.3%) 12 (16.0%) 3 (4.0%) 4 (5.3%) 6 (8.0%) 7 (9.3%) 79.1 (54.5–166) 27.4 (20.1–64.0) 134.1614.9 71.7610.0 26.366.6 4266133 13.564.0 139.563.1 5.860.3 22.563.6

0.48 0.35 0.26 1.00 0.09 0.44 0.79 1.00 0.70 0.19 0.09 0.33 0.002 0.09 0.04 0.25 ,0.001 0.002

106 (44.5%) 226 (95.0%) 122 (51.3%) 39 (16.4%) 21 (8.8%) 19 (8.0%) 17 (7.1%) 22 (9.2%) 78.4 (42.6–166) 28.0 (18.2–64.0) 136.9617.1 72.7611.1 24.267.1 4036142 14.464.6 139.862.9 5.160.6 23.663.7

LCC, low-clearance clinic; eGFR, estimated GFR. a P value for comparisons between potassium groups.

70 (42.9) 153 (93.9%) 88 (54.0%) 27 (16.6%) 18 (11.0%) 15 (9.2%) 11 (6.7%) 15 (9.2%) 77.2 (42.6–137.5) 28.4 (18.2–54.2) 138.2617.9 73.2611.6 23.267.1 3926146 14.864.8 140.062.8 4.860.4 24.163.6

0.09

0.47 0.67


26.5%, 8.4%, and 5.9% of patients had potassium levels$5.5, .5.5, $6.0, and .6.0 meq/L, respectively. The distribution of serum potassium levels in the population studied is shown in Figure 1. Within the final study population, 37 patients were on the conservative care pathway but were routinely attending the main LCC clinic. The mean level of potassium in this subgroup was 4.960.5 meq/L (versus 5.160.7 meq/L in the nonconservative pathway patients; n=201, P=0.02). The prevalence of hyperkalemia defined as potassium.5.0, $5.5, and .5.5 in the conservative care patients was at 40.5%, 16.2%, and 8.1%, respectively; no patient in this subgroup had potassium levels$6.0 meq/L.

Factors Associated with Hyperkalemia In univariate comparisons (Table 1), individuals with serum potassium$5.5 meq/L were found to be slightly younger (64.2614.2 versus 67.1614.2 years) and more often males (66.7% versus 54.6%) compared with the individuals with potassium,5.5 meq/L, but the relevant differences did not reach statistical significance. The two groups also did not differ significantly with respect to time of follow-up in the LCC, ethnicity distribution, burden of comorbidities, or primary cause of kidney disease (although patients with high potassium were less likely to have hypertension as the primary cause). Body weight and BP levels were also not different between groups; however, systolic BP displayed a trend to lower levels in individuals with high potassium. As shown in Table 1, there were important differences in routine biochemistry between the groups, with individuals with high potassium having significantly higher levels of urea, slightly lower eGFR, and lower serum bicarbonate levels. Table 2 depicts information on the use of medications that could interfere with potassium regulation. Although the percentage of patients with high potassium receiving a renin–


1237

angiotensin–aldosterone system (RAAS) blocker was higher and the percentage of them receiving a thiazide or loop diuretic was lower than patients with normal potassium, the relevant differences did not reach statistical significance in univariate analysis. Similarly, the proportion of patients using b-blockers, insulin, oral hypoglycemic agents, and laxative compounds (including lactulose, senna, and sodium docusate) did not differ between groups. Thus, the only significant difference was the use of sodium bicarbonate, which was higher in patients with high potassium (65.3% versus 45.4%, P=0.008). Of note, none of the study subjects were receiving renin inhibitors, nonsteroidal anti-inflammatory drugs or cyclooxygenase inhibitors, heparin, trimethoprim, pentamidine, cyclosporine, tacrolimus, or resin K+ exchangers at the time of the study. There were no differences between the two groups with regards to clinical suspicion of reasons that could have caused acute decline in renal function during the preceding 4 weeks, which was rather rare at the time of the study (Table 3). With regards to RAAS blockers, no start or dose increase took place during that period. In addition, there was only one patient with a recent possible cause of hyperkalemia other than low renal function (resolving rectus sheath hematoma). With regard to the dietician input at the most recent appointment, hyperkalemia as a reason for referral was slightly more frequent in the high-potassium group. In addition, significantly higher percentages of subjects from the high-potassium group had received potassium education during the appointment and had an attempt for lowering of potassium intake in the diet, because relevant sources of potassium were identified. We also performed multiple logistic regression analysis, including serum potassium$5.5 meq/L as the dependent variable and several demographic, clinical, and laboratory factors that were previously identified from univariate analyses as independent variables. As shown in Table 4, advancing age and having diabetes or hypertension as the cause of CKD were associated with lower odds ratios of

Figure 1. | Distribution of serum potassium levels in the study population.

1238


Table 2. Use of medications that could interfere with potassium regulation in patients with and without elevated potassium levels

Potassium,5.5 meq/L (n=163)

Potassium$5.5 meq/L (n=75)

P Value

67 (41.1%) 49 (30.1%) 81 (49.7%) 17 (10.4%) 2 (1.2%) 81 (49.1%) 6 (3.6%) 89 (53.9%) 38 (23.0%) 19 (11.5%) 2 (1.2%) 17 (10.3%) 16 (9.7%) 5 (3.0%) 75 (45.4%)

32 (42.6%) 31 (41.3%) 45 (60.0%) 9 (12.0%) 1 (1.3%) 34 (45.3%) 1 (1.3%) 33 (44.0%) 25 (33.3%) 8 (10.7%) 2 (2.7%) 4 (5.3%) 7 (9.3%) 3 (4.0%) 49 (65.3%)

0.89 0.10 0.16 0.44 1.00 0.58 0.44 0.16 0.11 1.00 0.59 0.23 1.00 0.71 0.008

ACEIs (n) ARBs (n) ACEI or ARB alone (n)a ACEI and ARB combination (n) Aldosterone blockers (n) b-Blockers (n) Thiazide diuretics (n) Loop diuretics (n) Insulin (n) Oral hypoglycemic agents (n) Digoxin (n) Lactulose (n) Senna (n) Sodium docusate (n) Sodium bicarbonate (n)

ACEIs, angiotensin-converting enzyme inhibitors; ARBs, angiotensin receptor blockers. a One patient was receiving an ARB/spironolactone combination.

Table 3. Clinical factors predisposing to hyperkalemia and factors related to dietician input in patients with and without elevated potassium levels

Suspected cause of acute renal failure (n) dehydration/hypotension infection/sepsis postsurgery postcontrast-related procedure obstruction other Reason for dietician appointment (n)a routine appointment hyperkalemia other reason Education on potassium (n)a Dietary potassium manipulation (n)a,b possible not possible not applicable

Potassium,5.5 meq/L (n=163)

Potassium$5.5 meq/L (n=75)

P Value

6 (3.6%) 0 (0%) 1 (0.6%) 2 (1.2%) 0 (0%) 3 (1.8%) 0 (0%)

3 (4.0%) 1 (1.3%) 0 (0%) 0 (0%) 1 (1.3%) 0 (0%) 1 (1.3%)

0.90 0.14 0.50 0.33 0.14 0.28 0.14

148 (93.1%) 2 (1.3%) 9 (5.7%) 75 (47.2%)

66 (89.2%) 5 (6.8%) 3 (4.0%) 55 (73.3%)

61 (38.4%) 60 (37.7%) 38 (23.9%)

49 (66.2%) 14 (18.9%) 11 (14.7%)

0.07 ,0.001 ,0.001

a

At the time of the study, dietician consultation had taken place for 159 of 165 and 74 of 75 subjects in the two study groups. Dietary potassium manipulation was defined as possible when consumption of high-potassium foods was identified during the individualized diet assessment at the most recent dietician appointment, not possible when potassium-rich foods were not identified, and not applicable when dietary assessment for potassium was not performed in the specific appointment. b

hyperkalemia, but the differences were of borderline statistical significance. Similarly, male sex, use of RAAS blockers, dietary potassium education, and potassium manipulation displayed nonsignificant trends of increased odds ratios of hyperkalemia. Thus, the only variables that were independently associated with higher odds of hyperkalemia in multivariate analysis were eGFR,15 ml/min per 1.73 m2 (2.06; 95% CI=1.02–4.18), use of sodium bicarbonate (3.04; 95% CI=1.46–6.31), and unsurprisingly, referral to dieticians for hyperkalemia (10.9; 95% CI=1.04–115.2).

Discussion The prevalence of hyperkalemia in this population was high. More than one-half of the patients had potassium levels above the conventional limit of 5.0 meq/L, and around onequarter of the patients had potassium levels.5.5 meq/L. Among the wide range of demographic, clinical, and laboratory characteristics studied, individuals with hyperkalemia displayed only nonsignificant trends toward younger age and higher proportions of males and RAAS inhibitor use. In contrast, even within a population where all individuals had a priori low renal function, a lower level of eGFR was



Table 4. Odds ratios of serum potassium‡5.5 meq/L in the study population

Odds Ratio (95% CI) Age (years) ,55 55–75 .75 Sex female male Primary cause of kidney disease diabetic kidney disease hypertension/ renovascular disease glomerulonephritis tubulointerstitial disease polycystic kidney disease unknown eGFR (ml/min per 1.73 m2) $15 ,15 Serum bicarbonate (meq/L) $25 ,25 RAAS blocker use none ACEI or ARB ACEI and ARB combination spironolactone Loop diuretic use no yes Sodium bicarbonate use no yes Reason for dietician appointment routine appointment hyperkalemia other reason Education on potassium no yes Dietary potassium manipulation not possible possible not applicable

P Value

1.00 (reference) 0.46 (0.19–1.11) 0.78 (0.31–1.97)

0.08 0.59

1.00 (reference) 1.71 (0.85–3.43)

0.13

0.30 (0.07–1.34) 0.26 (0.06–1.09)

0.12 0.07

1.00 (reference) 0.34 (0.06–1.79)

0.20

0.43 (0.06–3.06)

0.40

0.41 (0.06–2.71)

0.35

1.00 (reference) 2.06 (1.02–4.18)

0.05

1.00 (reference) 1.30 (0.65–2.62)

0.46

1.00 (reference) 1.85 (0.86–3.99) 1.40 (0.40–4.90)

0.12 0.59

2.10 (0.13–32.78)

0.60

1.00 (reference) 0.69 (0.34–1.40)

0.30

1.00 (reference) 3.04 (1.46–6.31)

0.003

1.00 (reference) 10.93 (1.04–115.2) 0.05 0.57 (0.13–2.57) 0.46 1.00 (reference) 2.34 (0.76–7.21)

0.14

1.00 (reference) 2.02 (0.70–5.85) 1.42 (0.50–4.02)

0.20 0.51

CI, confidence interval; eGFR, estimated GFR; RAAS, renin– angiotensin–aldosterone system; ACEIs, angiotensin-converting enzyme inhibitors; ARBs, angiotensin receptor blockers.

independently associated with high serum potassium. Use of sodium bicarbonate and dietician referral for hyperkalemia were also independently related to high potassium levels, but this finding most likely reflects inverse causation. A few studies that examined the prevalence of various CKD complications with decreasing renal function also

1239

provided some data on hyperkalemia. In a cross-sectional analysis of retrospectively collected electronic medical records of 1216 individuals with various levels of renal function (17), the prevalence of potassium.5 meq/L rose from about 10% to 18% and 22%, and the actual levels of potassium rose from 5.2 to 5.3 and 5.5 meq/L for subjects with eGFR=30–60, 15–30, and ,15 ml/min per 1.73 m2, respectively. In this study, however, patients with renal function,15 ml/min per 1.73 m2 were a very small percentage of the study population (5.7%), and no information was provided on whether they were on replacement therapy. In a French cohort of 1038 outpatients with stages 2–5 CKD not on dialysis, the prevalence of hyperkalemia defined as serum potassium.5 mmol/L or treatment with ion exchange resin was 17% in the total population and around 42% in the subset of 184 individuals with eGFR,20 ml/min per 1.73 m2 (18), which is a finding close to our observations. A Japanese study analyzed retrospective data on the highest potassium levels of 9117 nephrology and diabetes outpatients and excluded patients on dialysis, with serum creatinine above 5 mg/dl, or using cation exchange resin or diuretics. Serum potassium was found to increase with serum creatinine levels; however, no direct comparisons can be made with our study, because subjects with creatinine above 3 mg/dl were only 0.6% of the population (16). Two other recent studies have examined the incidence of hyperkalemia in individuals with CKD. The first study was a posthoc analysis of the African-American Study on Kidney Disease (AASK), which included 1094 African-American nondiabetic adults with hypertensive CKD and eGFR=20– 65 ml/min per 1.73 m2 that were followed for 3.0–6.4 years. With hyperkalemia defined as serum potassium.5.5 meq/L during follow-up visits, a total of 80 events (1.2% of laboratory records) in 51 patients was identified (19). Finally, the largest study of the field included a retrospective analysis of records of 245,808 individuals (including 70,873 individuals with CKD stages 3–5) hospitalized over 1 year in the Veterans Health Administration Hospitals in the United States; a total of 66,259 hyperkalemic events was captured (3.2% of total laboratory values) (20). To the best of our knowledge, our work is the first study of the field that is prospectively designed to simultaneously capture data on most known factors that could interfere with potassium metabolism in CKD patients. Among a wide range of parameters studied, in univariate analyses, individuals with high potassium had significantly lower eGFR and serum bicarbonate levels, and they were more often treated with sodium bicarbonate and had received dietary potassium education and manipulation. They were also younger, more often males, and more often treated with RAAS blockers, but these differences did not reach statistical significance. Previous studies have only assessed the possible effects of a smaller number of factors. In studies with cross-sectional analyses, the work by Takaichi et al. (16) observed serum potassium levels to be higher in patients with diabetes and patients using an ACEI/ARB up to serum creatinine of 2 mg/dl and not for groups with lower renal function; this study, however, is limited by the use of only creatinine values and the absence of multivariate analysis. In the aforementioned French study, the odds ratio of hyperkalemia in multivariate analysis was significantly higher for males and subjects using an ACEI/ARB, it was lower with increasing

1240


eGFR, and it was marginally lower for black individuals. Age, cause of CKD, diabetes, body mass index, and BP control did not show independent associations (18). With regards to longitudinal studies, in multivariate analysis of the AASK study, the risk of incident hyperkalemia increased considerably with decreasing eGFR (patients with eGFR=20–30 ml/min per 1.73 m2 had a 6.8-fold higher risk than those patients with eGFR.50 ml/min per 1.73 m2) and use of ramipril (versus amlodipine or metoprolol). It was slightly higher with increasing age, and it was lower for females. BP, glucose, and body mass index levels and use of nonsteroidal anti-inflammatory agents at baseline did not affect hyperkalemia incidence (19). Finally, in the Veterans study, low renal function was again the most potent risk factor of incident hyperkalemia (with patients with CKD stage 5 having an 11-fold higher risk than those patients with eGFR.60 ml/min per 1.73 m2); use of ACEIs/ARBs and diabetes conferred a mildly elevated risk, whereas again, female sex was associated with lower risk (20). The results of our multivariate analysis showed similar patterns, with low eGFR being independently associated with hyperkalemia, and male sex and ACEI/ARB use suggesting an increased risk that did not reach statistical significance. The strong effect of low renal function on hyperkalemia development has been known for many years and is consistent through the literature (22,23). Similarly, hyperkalemia is a long-established side effect of ACEI/ARB as well as aldosterone blockers use that was noticed repeatedly in different settings (10,24–26). Interestingly, we noticed nonsignificant lower odds of hyperkalemia with increased age (as in the French study in ref. 18), whereas most other studies have shown older age to slightly increase hyperkalemia risk (19,20). Although the slight differences noted in our multivariate analysis could be simply related to the smaller sample, which leads to lower power, it cannot be excluded that a number of factors in our study play an important role for true attenuation of the effect of RAAS blockade, sex, and age on potassium levels. For example, the study population in our study included only predialysis patients and was clearly different from the previous studies (the French cohort had only 18% of patients with eGFR,20 ml/min per 1.73 m2, the AASK study had no patients below that level, and the Veterans study had only 6.7% of total patients with stage 5 CKD) (18–20). Furthermore, our cohort came from a different clinic environment with close follow-up of patients and use of standardized protocols of multidisciplinary care. Among these protocols, routine dietician support, including tailored potassium advice and attempts for dietary potassium manipulation after assessment of individual diets (including the diets of our African, Afro-Caribbean, and Indian populations, which sometimes contain high-potassium foods that can be missed if not specifically inquired about), may have played an important role in potassium homeostasis, as suggested by recent studies (27). Finally, our analysis included a much larger set of variables that could have contributed to attenuation of previously observed relationships. This study has strengths and limitations. It was carefully designed to capture a wide range of demographic, laboratory, and biochemical factors that could affect potassium levels. In addition, data collection included, for the first time, all medications that were known to interfere with potassium regulation, dietician input, and recent changes in clinical status that

could affect potassium. Also, particular care was taken to standardize blood sampling and fast shipping of samples to the laboratory to minimize the occurrence of pseudohyperkalemia, which could interfere with the results. However, the analysis used a cross-sectional design, and thus, cause and effect associations between the factors examined and hyperkalemia could not be established; this finding is particularly relevant for the associations of sodium bicarbonate use and referral to a dietician for hyperkalemia, which are more likely the result of and not a cause for high potassium. Similarly, because of the cross-sectional design, we could not capture information on the reason for sodium bicarbonate prescription (acidosis or hyperkalemia) or whether ACEIs or ARBs had been previously discontinued because of hyperkalemia, thus attenuating any possible association. Furthermore, although patients were distributed across different categories of important variables (i.e., medication use), the study sample was modest for the number of variables included in the logistic regression analysis, and it could not be excluded that, with a larger sample, different associations may appear. This study was also a single-center study, and therefore, it is not clear whether the observed prevalence of hyperkalemia would generalize to other LCCs or practices. Finally, this project was not designed to evaluate the longitudinal associations of serum potassium with incident cardiac complications and death in predialysis patients. This field remains of major interest, as recent data support previous hypotheses that patients with advanced stages of CKD are less susceptible to hyperkalemia-associated cardiac toxicity than those patients with normal renal function, because their hyperkalemia is chronic and therefore, better tolerated than an acute increase in potassium levels (1,20,28). In conclusion, this study shows that hyperkalemia is very common in stable predialysis patients with CKD. Although a number of factors may be contributing to this high prevalence, it seems that, even at this low level of renal function, the actual eGFR is the most important factor associated with high potassium levels. Additional studies are needed to evaluate the clinical significance of elevated serum potassium for morbidity and mortality in this patient population. Acknowledgment The authors thank the staff of the low-clearance clinic, Department of Renal Medicine, King’s College Hospital, for help in performing this study. Disclosures None. References 1. Allon M: Hyperkalemia in end-stage renal disease: Mechanisms and management. J Am Soc Nephrol 6: 1134–1142, 1995 2. Acker CG, Johnson JP, Palevsky PM, Greenberg A: Hyperkalemia in hospitalized patients: Causes, adequacy of treatment, and results of an attempt to improve physician compliance with published therapy guidelines. Arch Intern Med 158: 917–924, 1998 3. Weiner ID, Wingo CS: Hyperkalemia: A potential silent killer. J Am Soc Nephrol 9: 1535–1543, 1998 4. Perazella MA: Drug-induced hyperkalemia: Old culprits and new offenders. Am J Med 109: 307–314, 2000 5. Palmer BF: A physiologic-based approach to the evaluation of a patient with hyperkalemia. Am J Kidney Dis 56: 387–393, 2010


6. Siamopoulos KC, Elisaf M, Katopodis K: Iatrogenic hyperkalaemia— points to consider in diagnosis and management. Nephrol Dial Transplant 13: 2402–2406, 1998 7. Sarafidis PA, Georgianos PI, Lasaridis AN: Diuretics in clinical practice. Part II: Electrolyte and acid-base disorders complicating diuretic therapy. Expert Opin Drug Saf 9: 259–273, 2010 8. Sarafidis PA, Khosla N, Bakris GL: Antihypertensive therapy in the presence of proteinuria. Am J Kidney Dis 49: 12–26, 2007 9. Weir MR: Are drugs that block the renin-angiotensin system effective and safe in patients with renal insufficiency? Am J Hypertens 12: 195S–203S, 1999 10. Palmer BF: Managing hyperkalemia caused by inhibitors of the renin-angiotensin-aldosterone system. N Engl J Med 351: 585–592, 2004 11. Kunz R, Friedrich C, Wolbers M, Mann JF: Meta-analysis: Effect of monotherapy and combination therapy with inhibitors of the renin angiotensin system on proteinuria in renal disease. Ann Intern Med 148: 30–48, 2008 12. Toto RD: Aldosterone blockade in chronic kidney disease: Can it improve outcome? Curr Opin Nephrol Hypertens 19: 444–449, 2010 13. Sica DA, Hess M: Pharmacotherapy in congestive heart failure: Aldosterone receptor antagonism: Interface with hyperkalemia in heart failure. Congest Heart Fail 10: 259–264, 2004 14. Sarafidis PA, Bakris GL: Resistant hypertension: An overview of evaluation and treatment. J Am Coll Cardiol 52: 1749–1757, 2008 15. Bakris GL, Hart P, Ritz E: Beta blockers in the management of chronic kidney disease. Kidney Int 70: 1905–1913, 2006 16. Takaichi K, Takemoto F, Ubara Y, Mori Y: The clinically significant estimated glomerular filtration rate for hyperkalemia. Intern Med 47: 1315–1323, 2008 17. Drion I, Joosten H, Dikkeschei LD, Groenier KH, Bilo HJ: eGFR and creatinine clearance in relation to metabolic changes in an unselected patient population. Eur J Intern Med 20: 722–727, 2009 18. Moranne O, Froissart M, Rossert J, Gauci C, Boffa JJ, Haymann JP, M’rad MB, Jacquot C, Houillier P, Stengel B, Fouqueray B; NephroTest Study Group: Timing of onset of CKD-related metabolic complications. J Am Soc Nephrol 20: 164–171, 2009 19. Weinberg JM, Appel LJ, Bakris G, Gassman JJ, Greene T, Kendrick CA, Wang X, Lash J, Lewis JA, Pogue V, Thornley-Brown D, Phillips RA; African American Study of Hypertension and Kidney Disease Collaborative Research Group: Risk of hyperkalemia in


20.

21.

22.

23. 24. 25.

26. 27.

28.

1241

nondiabetic patients with chronic kidney disease receiving antihypertensive therapy. Arch Intern Med 169: 1587–1594, 2009 Einhorn LM, Zhan M, Hsu VD, Walker LD, Moen MF, Seliger SL, Weir MR, Fink JC: The frequency of hyperkalemia and its significance in chronic kidney disease. Arch Intern Med 169: 1156–1162, 2009 Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D; Modification of Diet in Renal Disease Study Group: A more accurate method to estimate glomerular filtration rate from serum creatinine: A new prediction equation. Ann Intern Med 130: 461–470, 1999 Textor SC, Bravo EL, Fouad FM, Tarazi RC: Hyperkalemia in azotemic patients during angiotensin-converting enzyme inhibition and aldosterone reduction with captopril. Am J Med 73: 719–725, 1982 Gennari FJ, Segal AS: Hyperkalemia: An adaptive response in chronic renal insufficiency. Kidney Int 62: 1–9, 2002 Reardon LC, Macpherson DS: Hyperkalemia in outpatients using angiotensin-converting enzyme inhibitors. How much should we worry? Arch Intern Med 158: 26–32, 1998 Juurlink DN, Mamdani MM, Lee DS, Kopp A, Austin PC, Laupacis A, Redelmeier DA: Rates of hyperkalemia after publication of the Randomized Aldactone Evaluation Study. N Engl J Med 351: 543–551, 2004 Khosla N, Kalaitzidis R, Bakris GL: Predictors of hyperkalemia risk following hypertension control with aldosterone blockade. Am J Nephrol 30: 418–424, 2009 Lu XH, Su CY, Sun LH, Chen W, Wang T: Implementing continuous quality improvement process in potassium management in peritoneal dialysis patients. J Ren Nutr 19: 469–474, 2009 Surawicz B, Chlebus H, Mazzoleni A: Hemodynamic and electrocardiographic effects of hyperpotassemia. Differences in response to slow and rapid increases in concentration of plasma K. Am Heart J 73: 647–664, 1967

Received: January 31, 2012 Accepted: April 24, 2012 Published online ahead of print. Publication date available at www. cjasn.org. See related editorial, “Hyperkalemia in Predialysis Patients,” on pages 1201–1202.