Abstract
Introduction:
Hypo- and hyperkalemia are associated with higher risk of ischemic stroke. However, this association has not been examined in an advanced CKD population.
Methods:
From among 102,477 US Veterans transitioning to dialysis between 2007–2015, 21,357 patients with 2 pre-dialysis outpatient eGFR <30 ml/min/1.73m2 90–365 days apart and at least 1 potassium (K) each in the baseline and follow-up period were identified. We separately examined the association of both baseline time-averaged K (chronic exposure) and time-updated K (acute exposure) treated as categorized (hypokalemia [K <3.5 mEq/L] and hyperkalemia [K >5.5 mEq/L] vs. referent [3.5–5.5 mEq/L]) and continuous exposure with time to first ischemic stroke event prior to dialysis initiation using multivariable-adjusted Cox regression models.
Results:
A total of 2,638 (12.4%) ischemic stroke events (crude event rate 41.9 per 1,000 patient years; 95% confidence interval [CI] 40.4–43.6) over a median (Q1–Q3) follow-up time of 2.56 (1.59–3.89) years were observed. The baseline time-averaged K category of hypokalemia (adjusted hazard ratio [aHR], 95% CI: 1.35, 1.01–1.81) was marginally associated with significantly higher risk of ischemic stroke. However, time-updated hyperkalemia was associated with significantly lower risk of ischemic stroke (aHR, 95% CI: 0.82, 0.68–0.98). The exposure-outcome relationship remained consistent when using continuous K levels for both the exposures.
Discussion/Conclusion:
In patients with advanced CKD, hypokalemia (chronic exposure) was associated with a higher risk of ischemic stroke; whereas, hyperkalemia (acute exposure) was associated with a lower risk of ischemic stroke. Further studies in this population are needed to explore the mechanisms underlying these associations.
Keywords: potassium, chronic kidney disease, ischemic stroke, dialysis
Introduction
The kidneys play a critical role in serum potassium (K) homeostasis; thus, patients with chronic kidney disease (CKD) are prone to dyskalemias (hypo- and hyperkalemia, especially the latter) [1, 2]. Dyskalemias in CKD occur due to various factors including lower eGFR, prevalence of diabetes mellitus (DM) and/or cardiovascular disease, use of medications such as renin-angiotensin-aldosterone system inhibitors (RAASi) and diuretics, and reduced dietary intake of K, [1–4] and are associated with adverse outcomes, including higher mortality and cardiovascular events, and healthcare burden [4–11].
Previous studies suggested that dyskalemias (especially hypokalemia) are associated with a higher risk of ischemic stroke [12–15]. Among the general population, [14] diuretics users, [12] and treated hypertensive patients, [15] hypokalemia was associated with a higher risk of ischemic stroke. Conversely, a study in the general population [13] showed that higher levels of K were associated with a higher risk of ischemic stroke. Patients with advanced CKD are at a higher risk of dyskalemias [4, 16] as well as ischemic stroke [17, 18]. However, the association between dyskalemias and ischemic stroke has not been examined in an advanced CKD population. The aim of our study was to assess this association in a large population of patients with advanced CKD prior to their transition to dialysis.
Materials and Methods
Study population
Longitudinal data from a historical cohort of United States (US) veterans transitioning to dialysis (Transition of Care in Chronic Kidney Disease [TC-CKD] cohort [n=102,477]) from October 1, 2007 through March 31, 2015 identified from the United States Renal Data System (USRDS) were used for this study [16, 19, 20]. An initial sample of 60,520 US veterans with pre-dialysis outpatient estimated glomerular filtration rate (eGFR) data was identified. Amongst these, 36,644 with two outpatient eGFR <30 ml/min/1.73m2 measured 90–365 days apart were identified, with the second eGFR serving as the index. Further, the sample was restricted to 23,363 with at least one-year each of look back period (baseline) and follow-up period (prior to dialysis initiation) from the index. Amongst these, 21,669 had at least 1 outpatient or inpatient K value each in the baseline and follow-up period. Finally, we excluded 312 patients (age <18 years at index [n=1], without medication data [n=235], and with ischemic stroke event at the index [n=76]) to yield a final sample size of 21,357 patients (shown in online supplementary Fig. 1).
Exposure
The exposures of interest were: a) baseline time-averaged K levels (average of all K measurements over the one-year baseline); and b) time-updated K levels through end of follow-up, categorized as hypokalemia (K <3.5 mEq/L), hyperkalemia (K >5.5 mEq/L) and referent (3.5≤ K ≤5.5 mEq/L), [19] and also treated as a continuous exposure.
Covariates
Patient demographic characteristics were extracted from the USRDS Patient and Medical Evidence file. Data on marital and smoking status was obtained from VA records [21, 22]. Preexisting comorbidities as of the index were identified from the VA Inpatient and Outpatient Medical SAS, and the VA/Centers for Medicare and Medicaid Services (CMS) databases with a diagnosis defined as the presence of 2 outpatient or 1 inpatient claims for the condition according to the International Classification of Disease, Ninth Revision, Clinical Modification (ICD-9-CM) diagnostic and Current Procedural Terminology codes. The Charlson Comorbidity Index (CCI) score was calculated using the Deyo modification for administrative datasets with kidney disease was excluded from the algorithm [23]. Data on prescribed medications were collected both in the baseline period and as a time-varying covariate for the follow-up period from inpatient and outpatient VA pharmacy dispensation records and CMS Medicare Part D. For the baseline period, patients were considered to be users if they had at least one 30-day supply dispensation for medications used for chronic therapy (RAASi, loop diuretics, K sparing diuretics, thiazide diuretics, non-steroidal anti-inflammatory drugs [NSAID], antiplatelet agents, aspirin, anticoagulants, lipid lowering medications, antiarrhythmics, digoxin, beta blockers, calcium channel blockers, insulin, oral hypoglycemic drugs, beta-2 agonist, and calcineurin inhibitors) and at least 1 dispensation of any day supply for sodium polystyrene sulphonate (SPS), trimethoprim, azole antifungals, and laxatives. For the time-updated K levels exposure, medications were treated as a time-varying covariate and were deemed to be present if the prescription date was the same as the K measurement date or if the K measurement fell within the time period covered by the day supply of the medication. Laboratory measurements and vital signs data were captured over the baseline period and were obtained from VA research databases as previously described [24, 25]. The estimated glomerular filtration rate (eGFR) was calculated by the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation, using outpatient serum creatinine values [26]. Further, sodium (Na) and eGFR levels were treated as time-varying covariates and captured on the same date as each individual time-updated K level, when K levels were treated as a time-updated exposure.
Outcomes
The outcome of interest was time to first ischemic stroke over the follow-up period. Ischemic stroke was ascertained based on ICD-9-CM diagnosis codes (433.01, 433.11, 433.21, 433.31, 433.81, 433.91, 434.00–434.91, and 436) [27] captured as the primary diagnosis at an outpatient or inpatient visit. Follow-up started from index eGFR and patients were followed through first ischemic stroke or censored at dialysis initiation for those who did not experience the stroke event, whichever occurred first.
Statistical analysis
Patient characteristics were summarized for the entire sample and by hypokalemia (K <3.5 mEq/L), hyperkalemia (K >5.5 mEq/L) and referent (3.5≤ K ≤5.5 mEq/L) categories based on baseline time-averaged K levels. Data were presented as counts (percentages), mean (SD) or median (25–75th percentile), and the differences between the K categories were assessed using chi-square tests, one-way analysis of variance and Kruskal-Wallis test, as appropriate. The association of baseline time-averaged K categories with time to ischemic stroke was assessed using multivariable-adjusted Cox regression models (models 1–4) which incrementally accounted for confounders based on theoretical considerations and availability in the database (shown in online supplementary Table 1). We also conducted an additional exploratory model (model 5) which adjusted for systolic blood pressure (SBP) and diastolic blood pressure (DBP) due to their potential mediatory effects. Similarly, the association of categorized time-updated K levels and ischemic stroke was assessed in incrementally multivariable-adjusted models as described in online supplementary Table 1, with additionally accounting for time-varying medications, eGFR, and Na and baseline average K levels in model 4 and model 5. Cox regressions using cubic splines and fractional polynomials by treating the K levels as a continuous exposure were used to assess nonlinearity in exposure-outcome relationship.
Missingness was observed for marital status (0.05%), smoking status and region of residence (0.07% each), Na (0.4%), SBP and DBP (0.7% each), body mass index (BMI) (10.8%), total cholesterol (12.3%), triglycerides (13.5%), high density lipoprotein (14.5%), low density lipoprotein (17.2%), and bicarbonate (22.3%) and time-varying Na (2.3%) and eGFR (6.5%).The main analysis (models 1–4 and model 5) for the exposure-outcome relationship was conducted using singly imputed data for missing baseline covariates derived from regression imputation. In the time-updated K level exposure models, time-varying Na and eGFR at each K were imputed using last observation carried forward method. We conducted subgroup analyses after categorizing patients by age, race, region of residence, prevalent DM, congestive heart failure, ischemic heart disease, and ischemic stroke/transient ischemic stroke [TIA], baseline use of SPS, RAASi, loop diuretics and antiplatelet agents, BMI and index eGFR using singly imputed data. Potential interactions between dyskalemia categories and selected subgroups were tested by including interactions terms. We conducted several sensitivity analyses. The exposure-outcome association was assessed by categorizing the K levels into granular categorizes as K <3.5 mEq/L, 3.5–<4.0 mEq/L, 4.0–<4.5 mEq/L [reference], 4.5–<5.0 mEq/L, 5.0–<5.5 mEq/L, 5.5–<6.0 mEq/L, and ≥6.0 mEq/L. Further, the exposure-outcome association was assessed amongst those without baseline history of ischemic stroke/TIA (n=18,096). Finally, analyses were repeated using multiple imputation (imputation n=25) and complete cases (n=12,388 for baseline K exposure [after excluding missing baseline covariates]; n=6,542 for time-updated K exposure [after excluding missing baseline and time-varying covariates]).
A two-sided p-value of <0.05 was used as a threshold of statistical significance. Analyses were conducted in SAS Enterprise guide v8.2 (SAS Institute; Cary, NC) and STATA/MP Version 15 (STATA Corporation, College Station, TX). The study was approved by the Institutional Review Boards of the Memphis and Long Beach VA Medical centers, with exemption from informed consent.
Results
Baseline characteristics
The mean (SD) age of the sample was 68.6 (10.4) years; 98.2% were males; 28.3% were Black; and 68.1% had DM (shown in Table 1). The most commonly used medications were lipid lowering agents (76.9%), RAASi (74.4%), and beta blockers (68.6%). Approximately 8% of the patients used SPS. The median (25–75th) index eGFR was 24.8 (20.9–27.6) ml/min/1.73m2. Patients had a median (25–75th) of 3 (2–6) K measurements, with a mean (SD) K of 4.5 (0.5) mEq/L in the baseline. Approximately 3% and 1.9% of the sample had average baseline K levels >5.5 mEq/L and <3.5 mEq/L, respectively. Those with average baseline K levels >5.5 mEq/L were more likely to be older, white, have prevalent hyperlipidemia and anemia, SPS users and have lower eGFR, bicarbonate, SBP, DBP, and BMI levels. Conversely, those with average baseline K levels <3.5 mEq/L were more likely to be younger, Black, users of loop, K sparing, and thiazide diuretics and insulin, and have higher eGFR, bicarbonate, SBP, DBP, and BMI levels.
Table 1.
Baseline characteristics
| Characteristic | All (N=21,357) | K <3.5 mEq/L (N=402) | K 3.5–5.5 mEq/L (N=20,287) | K >5.5 mEq/L (N=668) | p-value |
|---|---|---|---|---|---|
| Age at index | 68.6 (10.4) | 65.6 (10.2) | 68.6 (10.4) | 70.6 (10.2) | <0.0001* |
| Males | 20,963 (98.2) | 395 (98.3) | 19,912 (98.2) | 656 (98.2) | 0.98† |
| Race | <0.0001† | ||||
| White | 14,712 (68.9) | 190 (47.3) | 13,392 (68.9) | 530 (79.3) | |
| Blacks | 6,044 (28.3) | 203 (50.5) | 5,725 (28.2) | 116 (17.4) | |
| Other | 601 (2.8) | 9 (2.2) | 570 (2.8) | 22 (3.3) | |
| Married | 12,309 (57.7) | 216 (54.0) | 11,670 (57.6) | 423 (63.3) | 0.008† |
| Region | <0.0001† | ||||
| Northeast | 3,155 (14.8) | 36 (8.9) | 3,007 (14.8) | 112 (16.8) | |
| Mid-West | 4,776 (22.4) | 87 (21.6) | 4,562 (22.5) | 127 (19.0) | |
| South | 9,648 (45.2) | 229 (56.9) | 9,105 (44.9) | 314 (47.0) | |
| West | 3,459 (16.2) | 46 (11.4) | 3,332 (16.4) | 81 (12.1) | |
| Other | 305 (1.4) | 4 (1.0) | 267 (1.3) | 34 (5.1) | |
| Income ($) | 18,036 (6,492–34,662) | 17,734 (4,524–34,992) | 18,036 (6,684–34,675) | 18,042 (1,722–33,876) | 0.58‡ |
| Smoking status | 0.18† | ||||
| Current | 6,928 (32.5) | 122 (30.4) | 6,610 (32.6) | 196 (29.3) | |
| Past | 7,524 (35.3) | 134 (33.3) | 7,147 (35.3) | 243 (36.4) | |
| Never | 6,892 (32.3) | 146 (36.3) | 6,517 (32.1) | 229 (34.3) | |
| Comorbidities | |||||
| Diabetes mellitus | 14,752 (69.1) | 284 (70.7) | 14,004 (69.0) | 464 (69.5) | 0.77† |
| Congestive heart failure | 7,466 (34.9) | 153 (38.1) | 7,097 (34.9) | 216 (32.3) | 0.16† |
| Hypertension | 20,678 (96.8) | 392 (97.5) | 19,649 (96.9) | 637 (95.4) | 0.07† |
| Hyperlipidemia | 16,669 (78.1) | 289 (71.9) | 15,842 (78.1) | 538 (80.5) | 0.004† |
| Peripheral vascular disease | 6,579 (30.8) | 96 (23.9) | 6,257 (30.8) | 226 (33.8) | 0.003† |
| Cerebrovascular disease | 5,359 (25.1) | 87 (21.6) | 5,095 (25.1) | 177 (26.5) | 0.19† |
| Chronic lung disease | 6,308 (29.5) | 107 (26.6) | 6,013 (29.6) | 188 (28.1) | 0.31† |
| Peptic ulcer disease | 1,001 (4.7) | 14 (3.5) | 948 (4.7) | 39 (5.8) | 0.19† |
| Ischemic heart disease | 10,573 (49.5) | 195 (48.5) | 10,032 (49.5) | 346 (51.8) | 0.45† |
| Paraplegia/hemiplegia | 507 (2.4) | 8 (1.9) | 489 (2.4) | 10 (1.5) | 0.27† |
| Anemia | 10,145 (47.5) | 167 (41.5) | 9,624 (47.4) | 354 (52.9) | 0.001† |
| Atrial fibrillation | 2,456 (11.5) | 55 (13.7) | 2,321 (11.4) | 80 (11.9) | 0.35† |
| Liver disease | 1,563 (7.3) | 33 (8.2) | 1,491 (7.4) | 39 (5.8) | 0.26† |
| Malignancies | 4,238 (19.8) | 71 (17.7) | 4,044 (19.9) | 123 (18.4) | 0.34† |
| Ischemic stroke/Transient ischemic stroke | 3,261 (15.3) | 64 (15.9) | 3,109 (15.3) | 88 (13.2) | 0.29† |
| Charlson comorbidity index | 4 (2–6) | 3 (2–5) | 4 (2–6) | 4 (2–6) | 0.22‡ |
| Utilization measures | |||||
| Outpatient visits | 16 (9–28) | 17 (10–28) | 17 (9–28) | 12 (7–20) | <0.0001‡ |
| Hospital visits | 0 (0–1) | 0 (0–1) | 0 (0–1) | 0 (0–1) | <0.0001‡ |
| Emergency room visits | 0 (0–1) | 0 (0–1) | 0 (0–1) | 0 (0–0) | 0.009‡ |
| Nephrology visits | 0 (0–1) | 0 (0–1) | 0 (0–1) | 0 (0–0) | <0.0001‡ |
| Medications | |||||
| Renin-angiotensin-aldosterone system inhibitors | 15,892 (74.4) | 292 (72.6) | 15,114 (74.5) | 486 (72.8) | 0.42† |
| Loop diuretics | 11,986 (56.1) | 253 (62.9) | 11,428 (56.3) | 305 (45.7) | <0.0001† |
| K sparing diuretics | 1,955 (9.2) | 76 (18.9) | 1,840 (9.1) | 39 (5.8) | <0.0001† |
| Thiazide diuretics | 6,969 (32.6) | 228 (56.7) | 6,547 (32.3) | 194 (29.0) | <0.0001† |
| Sodium polystyrene sulphonate | 1,684 (7.9) | 2 (0.5) | 1,496 (7.4) | 186 (27.8) | <0.0001† |
| Nonsteroidal anti-inflammatory drugs | 7,188 (33.7) | 137 (34.1) | 6,892 (33.9) | 159 (23.8) | <0.0001† |
| Digoxin | 1,007 (4.7) | 22 (5.5) | 961 (4.7) | 24 (3.6) | 0.30† |
| Beta blockers | 14,650 (68.6) | 303 (75.4) | 13,931 (68.7) | 416 (62.3) | <0.0001† |
| Calcium channel blockers | 14,089 (65.9) | 307 (76.4) | 13,414 (66.1) | 368 (55.1) | <0.0001† |
| Anticoagulants | 1,839 (8.6) | 42 (10.5) | 1,757 (8.7) | 40 (5.9) | 0.02† |
| Antiplatelets | 3,300 (15.5) | 41 (10.2) | 3,157 (15.6) | 102 (15.3) | 0.01† |
| Antihyperlipidemic | 16,441 (76.9) | 297 (73.9) | 15,638 (77.1) | 506 (75.8) | 0.24† |
| Antiarrhythmics | 499 (2.3) | 16 (3.9) | 468 (2.3) | 15 (2.3) | 0.09† |
| Aspirin | 6,004 (28.1) | 106 (26.4) | 5,783 (28.5) | 115 (17.2) | <0.0001† |
| Insulin | 8,698 (40.7) | 177 (44.0) | 8,291 (40.9) | 230 (34.4) | 0.002† |
| Oral hypoglycemic | 7,506 (35.2) | 153 (38.1) | 7,103 (35.0) | 250 (37.4) | 0.20† |
| Calcineurin inhibitors | 227 (1.1) | 4 (1.0) | 215 (1.1) | 8 (1.2) | 0.93† |
| Trimethoprim | 392 (1.8) | 9 (2.2) | 368 (1.8) | 15 (2.3) | 0.59† |
| Azole antifungals | 2,304 (10.8) | 34 (8.5) | 2,225 (10.9) | 45 (6.7) | 0.0009† |
| Beta-2 agonists | 3,224 (15.1) | 57 (14.2) | 3,095 (15.3) | 72 (10.8) | 0.006† |
| Laxatives | 5,881 (27.5) | 101 (25.1) | 5,674 (27.9) | 106 (15.9) | <0.0001† |
| Vitals | |||||
| Body mass index, kg/m2 | 29.9 (6.1) | 31.4 (6.3) | 29.9 (6.1) | 29.0 (5.8) | <0.0001* |
| Systolic blood pressure, mm/Hg | 143.6 (16.3) | 147.7 (18.8) | 143.5 (16.2) | 143.8 (17.6) | <0.0001* |
| Diastolic blood pressure, mm/Hg | 74.9 (10.8) | 80.1 (13.2) | 74.9 (10.8) | 72.9 (11.0) | <0.0001* |
| Low density lipoprotein, mg/dl | 94.5 (36.6) | 96.8 (40.1) | 94.5 (36.6) | 93.5 (35.2) | 0.41* |
| High density lipoprotein, mg/dl | 40.0 (13.1) | 40.7 (13.2) | 39.9 (13.1) | 40.4 (11.8) | 0.48* |
| Total cholesterol, mg/dl | 172.1 (47.1) | 174.5 (50.1) | 172.1 (47.1) | 169.9 (43.7) | 0.35* |
| Triglycerides, mg/dl | 188.5 (145.3) | 184.6 (136.6) | 188.9 (146.0) | 177.7 (127.3) | 0.17* |
| Laboratory measures | |||||
| Index eGFR, ml/min/1.73m2 | 24.8 (20.9–27.6) | 24.1 (20.3–27.4) | 24.9 (20.9–27.6) | 23.4 (19.2–27.3) | <0.0001‡ |
| Average K, mEq/L | 4.5 (0.5) | 3.3 (0.2) | 4.5 (0.5) | 5.7 (0.2) | <0.0001* |
| Number of K measurements | 3 (2–6) | 3 (2–6) | 3 (2–7) | 2 (1–4) | <0.0001‡ |
| Bicarbonate, mEq/L | 24.9 (3.2) | 27.9 (3.9) | 24.9 (3.2) | 22.8 (3.1) | <0.0001* |
| Sodium, mEq/L | 139.5 (2.7) | 139.9 (2.8) | 139.5 (2.7) | 139.4 (2.7) | 0.0009* |
| At least 1 dyskalemia event | |||||
| K >5.5 mEq/L | 3,709 (17.4) | 2 (0.5) | 3,309 (14.9) | 668 (100) | <0.0001† |
| K <3.5 mEq/L | 2,431 (11.4) | 402 (100) | 2,029 (10.0) | 0 | <0.0001† |
| Both K < 3.5 and K > 5.5 mEq/L | 371 (1.7) | 2 (0.5) | 369 (1.8) | 0 | 0.0003† |
Data are presented as n (%), mean (standard deviation), median (Q1–Q3) unless otherwise needed eGFR: estimated glomerular filtration rate; K: potassium
one-way analysis of variance
chi-square test
Kurskal-Wallis test
Association of baseline time-averaged dyskalemia categories with ischemic stroke
A total of 2,638 (12.4%) ischemic stroke events (crude event rate 41.9 per 1,000 patient years; 95% confidence interval [CI] 40.4–43.6) occurred over a median (Q1–Q3) follow-up time of 2.56 (1.59–3.89) years. Crude event rates for the overall cohort and by baseline time-averaged K categories are shown in online supplementary Table 2. In the unadjusted analysis, only hypokalemia was associated with higher risk of ischemic stroke (hazard ratio [HR], 95% confidence interval [CI]: 1.37, 1.06–1.76) (shown in Fig. 1 [Model 1]). Similarly, in the multivariable-adjusted model, only hypokalemia (HR, 95% CI: 1.35, 1.01–1.81) was associated with higher risk of ischemic stroke (shown in Fig. 1 [Model 4]). The association for hypokalemia (HR, 95% CI: 1.34, 0.99–1.79) was attenuated after adjustment for SBP and DBP (shown in Fig.1 [Model 5]). Continuous K levels showed a non-linear association (p-value for quadratic term:0.002), with lower K levels associated higher risk of ischemic stroke (shown in Fig. 2).
Fig 1.
Association of baseline time-averaged potassium categories with time to ischemic stroke (n=21,357)
Models 1–5 account for confounders described in Supplementary Table 1.
CI: confidence interval
Fig 2.
Association of baseline continuous time-averaged potassium with time to ischemic stroke (n=21,357)
Dashed and solid lines represent hazard ratio and 95% confidence interval, respectively.
Model adjusted for confounders accounted in model 4 (fully-adjusted) as described in Supplementary Table 1.
Associations for the categorized K levels were robust to multiple imputation and complete case analyses (data not shown). In the multivariable-adjusted model, the association of granular K level categories showed a lower risk (vs K 4.0–<4.5 mEq/L) of ischemic stroke associated with K 4.5–<5.0 mEq/L and K 5.0–<5.5 mEq/L (shown in online supplementary Table 3 [Model 4]). No significant differences were observed across subgroups (shown in online supplementary Table 4). In the sensitivity analysis, amongst the 84.7% without baseline history of stroke, hypokalemia was associated with higher risk of ischemic stroke (shown in online supplementary Table 5 [Model 4]).
Association of time-updated dyskalemia categories with ischemic stroke
Over the follow-up time (median [Q1–Q3]: 2.56 [1.59–3.89] years), there were a total of 489,486 K measurements with median (Q1–Q3) of 15 (6–30) K measurements per patient, of which 21,382 (4.5%) and 27,318 (5.6%) were categorized as hypokalemia and hyperkalemia, respectively. Online supplementary Table 6 shows the distribution of the K categories using the last time-updated K level prior to ischemic stroke by the baseline time-averaged K categories amongst those who experienced a stroke event. Online supplementary Table 7 shows the distribution of time-updated K categories across all K measurements during follow-up by baseline time-averaged K categories. In the unadjusted analysis, hyperkalemia was associated with lower risk of ischemic stroke (HR, 95% CI: 0.75, 0.63–0.89; shown in Fig. 3 [Model 1]). The results were similar in the multivariable-adjusted model (HR, 95% CI: 0.82, 0.68–0.98; shown in Fig. 3 [Model 4]). Results were similar after adjusting for SBP and DBP (HR, 95% CI: 0.82, 0.68–0.98; shown in Fig. 3 [Model 5]). Continuous K levels showed a non-linear association (p-value for quadratic term: 0.009), with higher K levels associated lower risk of ischemic stroke (shown in Fig. 4).
Fig 3.
Association of time-updated potassium categories with time to ischemic stroke (n=21,357) Models conducted as described in Supplementary Table 1 with further accounting for baseline averaged potassium levels and time-varying medications, estimated glomerular filtration rate and sodium levels.
CI: confidence interval
Fig 4.
Association of continuous time-updated potassium with time to ischemic stroke (n=21,357)
Dashed and solid lines represent hazard ratio and 95% confidence interval, respectively.
Model adjusted for confounders accounted in model 4 (fully-adjusted) as described in Supplementary Table 1 with further accounting for baseline averaged potassium levels and time-varying medications, estimated glomerular filtration rate and sodium levels.
Association of time-updated K categories with time to ischemic stroke were similar to the main analysis when using multiple imputation and complete case analyses (data not shown). In the multivariable-adjusted model, the association of granular K level categories showed a lower risk (vs K 4.0–<4.5 mEq/L) of ischemic stroke associated with K 5.5–<6.0 mEq/L (shown in online supplementary Table 8 [Model 4]). In the subgroup analysis, significant differences were observed by region of residence (shown in online supplementary Table 9). In the sensitivity analysis, amongst the 84.7% without baseline history of stroke, lower risk of ischemic stroke associated with hyperkalemia could no longer be established (shown in online supplementary Table 10).
Discussion/Conclusion
In a nationally representative cohort of US veterans with advanced CKD who transitioned to dialysis, hypokalemia (chronic exposure) was associated with a higher risk of ischemic stroke irrespective of baseline history of ischemic stroke/TIA. Conversely, hyperkalemia (time-updated, acute exposure) was associated with a lower risk of ischemic stroke. Our results were robust to various other sensitivity analyses including granular K categories, complete case analysis, and multiple imputation.
Our results align with existing research that suggests that hypokalemia (chronic exposure) is associated with higher risk of ischemic stroke. Amongst diuretic users, [12] general population, [14] and a treated hypertensive population [15], hypokalemia was associated with a 2.5-, 2.1-, and 2-fold higher risk of ischemic stroke, respectively. On the other hand, in a general population [13] both K levels between 4.3–8.4 mmol/L and a per mmol/L increase in K levels were associated with 1.3-fold higher risk of ischemic stroke. However, all these studies are characterized by the use of baseline K levels (chronic exposure) to define dyskalemias and long follow-up times for outcome assessment (minimum follow-up 1 year and maximum median follow-up 26.9 years) [12–15].
To determine the short-term risk associated with dyskalemias, we also assessed the association of time-updated dyskalemias (acute exposure) with ischemic stroke where we observed that hyperkalemia (and higher levels of K) was associated with a lower risk of ischemic stroke. Previous studies have hypothesized that hypokalemia might be a marker of increased RAAS activity [28] and an increased activity in systemic and cerebral RAAS system could potentiate the effect of a stroke, by resulting in more extensive neurologic damage and neurologic deficits [14]. This may explain the association of hypokalemia (chronic exposure) with higher risk of ischemic stroke. Further, a number of studies (animal models and epidemiological research), [29–37] suggested that higher levels of K lead to vasodilation, whereas lower levels of K lead to vasoconstriction by exerting effects through the Na+,K+ATPase in the vascular smooth muscle cell. A recent study by Li et al. [38] in a rat model observed that elevated serum K levels alleviated cerebral ischemia-reperfusion injury. The lower risk of ischemic stroke associated with time-updated hyperkalemia levels could be explained by these biological changes exerted by higher levels of K, especially in the acute setting. Patients with advanced CKD are at a higher risk of hyperkalemia [4, 16] and the hemodynamic effects exerted by repeated hyperkalemia events may acutely potentiate lower blood pressure levels and hence lower the risk of stroke. It is also noteworthy that in our cohort, we observed that those in the hyperkalemia group (vs hypokalemia group) based on the baseline time-averaged K levels had lower SBP (143.8 vs 147.7 mmHg) and DBP (72.9 vs 80.1 mmHg) levels, which supports the hypothesis that vasodilatory effects of hyperkalemia could lower the risk of ischemic stroke. Further, higher dietary intake of K is associated with better blood pressure control and lower risk of ischemic stroke [39–41]. However, current guidelines recommend restriction of dietary intake of K due to the risk of hyperkalemia, [42] despite the lack of association between dietary K intake and serum K levels or hyperkalemia [43] and the known health benefits of high dietary K intake [2]. However, lack of dietary K intake data did not allow us to assess the association of dietary intake of K (in concurrence with serum K) with ischemic stroke.
Our study results need to be interpreted in light of several limitations. First, our cohort consisted of predominantly male US veterans, thus limiting generalizability to women or to a broader general population. Second, we cannot infer causality due to the observational nature of the study. Thirdly, we cannot eliminate the possibility of unmeasured confounding due to lack of access to dietary intake of K. Several studies have noted the beneficial effects of dietary K intake and lowering of blood pressure and stroke risk [39–41]. Finally, our source cohort is designed such that outcomes such as mortality are only observed following dialysis initiation, but for this study we assessed dyskalemia-ischemic stroke association in the pre-dialysis period. Thus, caution should be exercised when interpreting the association of time-updated hyperkalemia and lower risk of ischemic stroke, as dyskalemias are associated with short-term risk of death, [10, 44] which could not be ascertained due to the nature of our cohort.
In conclusion, in patients with advanced CKD transitioning to dialysis, hypokalemia as a chronic exposure was associated with a higher risk of ischemic stroke, while hyperkalemia as an acute exposure was associated with lower risk of ischemic stroke. Further studies are needed to explore this association and shed light on the possible mechanisms driving this relationship.
Supplementary Material
Supplementary Table 1. List of confounders for multivariable-adjusted regression models for baseline time-averaged K exposure
Supplementary Table 2. Crude event rates, follow-up time and number of events overall and by baseline time-averaged potassium categories (n=21,357)
Supplementary Table 3. Association of baseline time-averaged granular potassium categories with time to ischemic stroke (n=21,357)
Supplementary Table 4. Association of baseline time-averaged potassium categories with time to ischemic stroke by subgroups (n=21,357)
Supplementary Table 5. Association of baseline time-averaged potassium categories with time to ischemic stroke amongst those without baseline history of stroke (n=18,096)
Supplementary Table 6. Distribution of potassium categories using the last time-updated potassium level prior to ischemic stroke by baseline time-averaged potassium categories amongst those experiencing ischemic stroke (n=2,368)
Supplementary Table 7. Distribution of time-updated potassium categories by baseline time-averaged potassium categories (n=21,357)
Supplementary Table 8. Association of time-updated granular potassium categories with time to ischemic stroke (n=21,357)
Supplementary Table 9. Association of time-updated potassium categories with time to ischemic stroke by subgroups (n=21,357).
Supplementary Table 10. Association of time-updated potassium categories with time to ischemic stroke amongst those without baseline history of stroke (n=21,357)
Supplementary Figure 1. Sample selection criteria
Acknowledgement
CPK, KKZ, and ES are employees of the Department of Veterans Affairs. Opinions expressed in this article are those of the authors and do not necessarily represent the opinion of the Department of Veterans Affairs.
Conflict of Interest Statement
CPK received honoraria from Akebia, Ardelyx, Astra Zeneca, Bayer, Boehringer-Ingelheim, Cara Therapeutics, Reata, and Tricida. KK-Z has received honoraria and/or support from Abbott, Abbvie, ACI Clinical (Cara Therapeutics), Akebia, Alexion, Amgen, American Society of Nephrology, Astra-Zeneca, Aveo, BBraun, Chugai, Cytokinetics, Daiichi, DaVita, Fresenius, Genentech, Haymarket Media, Hofstra Medical School, International Federation of Kidney Foundations, International Society of Hemodialysis, International Society of Renal Nutrition & Metabolism, Japanese Society of Dialysis Therapy, Hospira, Kabi, Keryx, Kissei, Novartis, OPKO, National Institutes of Health, National Kidney Foundations, Pfizer, Regulus, Relypsa, Resverlogix, Dr Schaer, Sandoz, Sanofi, Shire, Veterans’ Affairs, Vifor, UpToDate, ZS-Pharma. JG has received research support from AstraZeneca, Merck & Co., and GlaxoSmithKline. YO has received research support from Relypsa/Vifor Pharma Inc. The remaining authors declare that they have no relevant financial interests.
Funding Sources
This study is supported by grant 5U01DK102163 from the National Institute of Health (NIH) to KKZ and CPK., and by resources from the US Department of Veterans Affairs. The data reported here have been supplied in part by the United States Renal Data System (USRDS). Support for VA/CMS data is provided by the Department of Veterans Affairs, Veterans Health Administration, Office of Research and Development, Health Services Research and Development, VA Information Resource Center (project numbers SDR 02-237 and 98-004).
Footnotes
Statement of Ethics
The study was approved by the Institutional Review Boards of the Memphis and Long Beach VA Medical Centers, with exemption from informed consent.
References
- 1.Kovesdy CP, Appel LJ, Grams ME, Gutekunst L, McCullough PA, Palmer BF, et al. Potassium Homeostasis in Health and Disease: A Scientific Workshop Cosponsored by the National Kidney Foundation and the American Society of Hypertension. Am J Kidney Dis. 2017;70(6):844–58. [DOI] [PubMed] [Google Scholar]
- 2.Palmer BF, Clegg DJ. Physiology and Pathophysiology of Potassium Homeostasis: Core Curriculum 2019. Am J Kidney Dis. 2019. [DOI] [PubMed] [Google Scholar]
- 3.Kidney Disease Improving GLobal Outcomes (KDIGO). Controverseries conference on potassium management. Miami, USA. October, 2018[Available from: https://kdigo.org/conferences/controversies-conference-potassium-management/. [Google Scholar]
- 4.Gilligan S, Raphael KL. Hyperkalemia and Hypokalemia in CKD: Prevalence, Risk Factors, and Clinical Outcomes. Adv Chronic Kidney Dis. 2017;24(5):315–8. [DOI] [PubMed] [Google Scholar]
- 5.Collins AJ, Pitt B, Reaven N, Funk S, McGaughey K, Wilson D, et al. Association of Serum Potassium with All-Cause Mortality in Patients with and without Heart Failure, Chronic Kidney Disease, and/or Diabetes. Am J Nephrol. 2017;46(3):213–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Fitch K, Woolley JM, Engel T, Blumen H. The Clinical and Economic Burden of Hyperkalemia on Medicare and Commercial Payers. Am Health Drug Benefits. 2017;10(4):202–10. [PMC free article] [PubMed] [Google Scholar]
- 7.Kim K, Thomsen RW, Nicolaisen SK, Hasvold LP, Palaka E, Sorensen HT. Healthcare resource utilisation and cost associated with elevated potassium levels: a Danish population-based cohort study. BMJ Open. 2019;9(4):e026465. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Korgaonkar S, Tilea A, Gillespie BW, Kiser M, Eisele G, Finkelstein F, et al. Serum potassium and outcomes in CKD: insights from the RRI-CKD cohort study. Clin J Am Soc Nephrol. 2010;5(5):762–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Kovesdy CP, Matsushita K, Sang Y, Brunskill NJ, Carrero JJ, Chodick G, et al. Serum potassium and adverse outcomes across the range of kidney function: a CKD Prognosis Consortium meta-analysis. Eur Heart J. 2018;39(17):1535–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Luo J, Brunelli SM, Jensen DE, Yang A. Association between Serum Potassium and Outcomes in Patients with Reduced Kidney Function. Clin J Am Soc Nephrol. 2016;11(1):90–100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Nakhoul GN, Huang H, Arrigain S, Jolly SE, Schold JD, Nally JV Jr, et al. Serum Potassium, End-Stage Renal Disease and Mortality in Chronic Kidney Disease. Am J Nephrol. 2015;41(6):456–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Green DM, Ropper AH, Kronmal RA, Psaty BM, Burke GL, Cardiovascular Health S. Serum potassium level and dietary potassium intake as risk factors for stroke. Neurology. 2002;59(3):314–20. [DOI] [PubMed] [Google Scholar]
- 13.Johnson LS, Mattsson N, Sajadieh A, Wollmer P, Soderholm M. Serum Potassium Is Positively Associated With Stroke and Mortality in the Large, Population-Based Malmo Preventive Project Cohort. Stroke. 2017;48(11):2973–8. [DOI] [PubMed] [Google Scholar]
- 14.Mattsson N, Kumarathurai P, Larsen BS, Nielsen OW, Sajadieh A. Mild Hypokalemia and Supraventricular Ectopy Increases the Risk of Stroke in Community-Dwelling Subjects. Stroke. 2017;48(3):537–43. [DOI] [PubMed] [Google Scholar]
- 15.Smith NL, Lemaitre RN, Heckbert SR, Kaplan RC, Tirschwell DL, Longstreth WT, et al. Serum potassium and stroke risk among treated hypertensive adults. Am J Hypertens. 2003;16(10):806–13. [DOI] [PubMed] [Google Scholar]
- 16.Dashputre AA, Sumida K, Potukuchi PK, Kar S, Obi Y, Thomas F, et al. Potassium trajectories prior to dialysis and mortality following dialysis initiation in patients with advanced chronic kidney disease. Neprhon. In Press. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Lee M, Saver JL, Chang KH, Liao HW, Chang SC, Ovbiagele B. Low glomerular filtration rate and risk of stroke: meta-analysis. BMJ. 2010;341:c4249. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Masson P, Webster AC, Hong M, Turner R, Lindley RI, Craig JC. Chronic kidney disease and the risk of stroke: a systematic review and meta-analysis. Nephrol Dial Transplant. 2015;30(7):1162–9. [DOI] [PubMed] [Google Scholar]
- 19.Sumida K, Dashputre AA, Potukuchi PK, Thomas F, Obi Y, Molnar MZ, et al. Laxative Use and Risk of Dyskalemia in Patients with Advanced CKD Transitioning to Dialysis. J Am Soc Nephrol. 2021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Dashputre AA, Potukuchi PK, Sumida K, Kar S, Obi Y, Thomas F, et al. Predialysis Potassium Variability and Postdialysis Mortality in Patients With Advanced CKD. Kidney Int Rep. 2021;6(2):366–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.McGinnis KA, Brandt CA, Skanderson M, Justice AC, Shahrir S, Butt AA, et al. Validating smoking data from the Veteran’s Affairs Health Factors dataset, an electronic data source. Nicotine Tob Res. 2011;13(12):1233–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Soohoo M, Moradi H, Obi Y, Rhee CM, Gosmanova EO, Molnar MZ, et al. Statin Therapy Before Transition to End-Stage Renal Disease With Posttransition Outcomes. J Am Heart Assoc. 2019;8(6):e011869. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45(6):613–9. [DOI] [PubMed] [Google Scholar]
- 24.Kovesdy CP, Norris KC, Boulware LE, Lu JL, Ma JZ, Streja E, et al. Association of Race With Mortality and Cardiovascular Events in a Large Cohort of US Veterans. Circulation. 2015;132(16):1538–48. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Kovesdy CP, Alrifai A, Gosmanova EO, Lu JL, Canada RB, Wall BM, et al. Age and Outcomes Associated with BP in Patients with Incident CKD. Clin J Am Soc Nephrol. 2016;11(5):821–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF 3rd, Feldman HI, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150(9):604–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Russo CA (Thomson Reuters) and Andrews RM (AHRQ). Hospital Stays for Stroke and Other Cerebrovascular Diseases, 2005. HCUP Statistical Brief #51. May2008. Agency for Healthcare Research and Quality, Rockville, MD. May 2008 [Available from: https://www.hcup-us.ahrq.gov/reports/statbriefs/sb51.pdf. [PubMed] [Google Scholar]
- 28.Brown MJ, Brown J. Does angiotensin-II protect against strokes? Lancet. 1986;2(8504):427–9. [DOI] [PubMed] [Google Scholar]
- 29.Haddy FJ. Potassium and blood vessels. Life Sci. 1975;16(10):1489–97. [DOI] [PubMed] [Google Scholar]
- 30.Webb RC, Bohr DF. Potassium-induced relaxation as an indicator of Na+-K+ ATPase activity in vascular smooth muscle. Blood Vessels. 1978; 15(1–3):198–207. [DOI] [PubMed] [Google Scholar]
- 31.Paller MS, Linas SL. Hemodynamic effects of alterations in potassium. Hypertension. 1982;4(5 Pt 2):III20–6. [DOI] [PubMed] [Google Scholar]
- 32.Dolson GM, Ellis KJ, Bernardo MV, Prakash R, Adrogue HJ. Acute decreases in serum potassium augment blood pressure. Am J Kidney Dis. 1995;26(2):321–6. [DOI] [PubMed] [Google Scholar]
- 33.Ma G, Mamaril JL, Young DB. Increased potassium concentration inhibits stimulation of vascular smooth muscle proliferation by PDGF-BB and bFGF. Am J Hypertens. 2000;13(10):1055–60. [DOI] [PubMed] [Google Scholar]
- 34.Pikilidou MI, Lasaridis AN, Sarafidis PA, Tziolas IM, Zebekakis PE, Dombros NV, et al. Blood pressure and serum potassium levels in hypertensive patients receiving or not receiving antihypertensive treatment. Clin Exp Hypertens. 2007;29(8):563–73. [DOI] [PubMed] [Google Scholar]
- 35.Haddy FJ, Vanhoutte PM, Feletou M. Role of potassium in regulating blood flow and blood pressure. Am J Physiol Regul Integr Comp Physiol. 2006;290(3):R546–52. [DOI] [PubMed] [Google Scholar]
- 36.Brace RA. Time course and mechanisms of the acute effects of hypokalemia and hyperkalemia on vascular resistance. Proc Soc Exp Biol Med. 1974;145(4):1389–94. [DOI] [PubMed] [Google Scholar]
- 37.Webb RC. Potassium relaxation of vascular smooth muscle from DOCA hypertensive pigs. Hypertension. 1982;4(5):609–19. [DOI] [PubMed] [Google Scholar]
- 38.Li N, Qin S, Xie L, Qin T, Yang Y, Fang W, et al. Elevated Serum Potassium Concentration Alleviates Cerebral Ischemia-Reperfusion Injury via Mitochondrial Preservation. Cell Physiol Biochem. 2018;48(4):1664–74. [DOI] [PubMed] [Google Scholar]
- 39.Filippini T, Naska A, Kasdagli MI, Torres D, Lopes C, Carvalho C, et al. Potassium Intake and Blood Pressure: A Dose-Response Meta-Analysis of Randomized Controlled Trials. J Am Heart Assoc. 2020;9(12):e015719. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Hunt BD, Cappuccio FP. Potassium intake and stroke risk: a review of the evidence and practical considerations for achieving a minimum target. Stroke. 2014;45(5):1519–22. [DOI] [PubMed] [Google Scholar]
- 41.Vinceti M, Filippini T, Crippa A, de Sesmaisons A, Wise LA, Orsini N. Meta-Analysis of Potassium Intake and the Risk of Stroke. J Am Heart Assoc. 2016;5(10). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Clase CM, Carrero JJ, Ellison DH, Grams ME, Hemmelgarn BR, Jardine MJ, et al. Potassium homeostasis and management of dyskalemia in kidney diseases: conclusions from a Kidney Disease: Improving Global Outcomes (KDIGO) Controversies Conference. Kidney Int. 2020;97(1):42–61. [DOI] [PubMed] [Google Scholar]
- 43.Ramos CI, Gonzalez-Ortiz A, Espinosa-Cuevas A, Avesani CM, Carrero JJ, Cuppari L. Does dietary potassium intake associate with hyperkalemia in patients with chronic kidney disease? Nephrol Dial Transplant. 2020. [DOI] [PubMed] [Google Scholar]
- 44.Brunelli SM, Du Mond C, Oestreicher N, Rakov V, Spiegel DM. Serum Potassium and Short-term Clinical Outcomes Among Hemodialysis Patients: Impact of the Long Interdialytic Interval. Am J Kidney Dis. 2017;70(1):21–9. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Table 1. List of confounders for multivariable-adjusted regression models for baseline time-averaged K exposure
Supplementary Table 2. Crude event rates, follow-up time and number of events overall and by baseline time-averaged potassium categories (n=21,357)
Supplementary Table 3. Association of baseline time-averaged granular potassium categories with time to ischemic stroke (n=21,357)
Supplementary Table 4. Association of baseline time-averaged potassium categories with time to ischemic stroke by subgroups (n=21,357)
Supplementary Table 5. Association of baseline time-averaged potassium categories with time to ischemic stroke amongst those without baseline history of stroke (n=18,096)
Supplementary Table 6. Distribution of potassium categories using the last time-updated potassium level prior to ischemic stroke by baseline time-averaged potassium categories amongst those experiencing ischemic stroke (n=2,368)
Supplementary Table 7. Distribution of time-updated potassium categories by baseline time-averaged potassium categories (n=21,357)
Supplementary Table 8. Association of time-updated granular potassium categories with time to ischemic stroke (n=21,357)
Supplementary Table 9. Association of time-updated potassium categories with time to ischemic stroke by subgroups (n=21,357).
Supplementary Table 10. Association of time-updated potassium categories with time to ischemic stroke amongst those without baseline history of stroke (n=21,357)
Supplementary Figure 1. Sample selection criteria