Controversies: Stroke Prevention in Chronic Kidney Disease

Abstract

Risk of both ischemic and hemorrhagic stroke is increased in the chronic kidney disease (CKD) population, particularly in end-stage kidney disease patients. Uremic factors that contribute to stroke risk include blood pressure variability, vascular calcification, build-up of vascular toxins, chronic inflammation, platelet dysfunction and increased brain microbleeds. This paper discusses the controversial evidence for stroke prevention strategies including blood pressure control, statins, antiplatelet agents, and anticoagulation in the CKD population. Only a few randomized clinical trials included patients with advanced CKD, thus evidence is derived mostly from observational cohorts and real-world data. Overall, targeting a lower systolic blood pressure below 120 mmHg and statin prescription do not appear to decrease stroke risk in CKD. Antiplatelet agents have not shown a clear benefit for secondary stroke prevention, but aspirin may reduce incident stroke in hypertensive CKD stage 3B-5 patients. Observational data suggests that the factor Xa inhibitor apixaban has a favorable profile over warfarin in dialysis patients with atrial fibrillation; apixaban being associated with lower stroke risk and fewer major bleeding events.

Introduction

Chronic kidney disease (CKD) affects 9.1% of the global population¹ and accounts for over $60 billion in healthcare costs in the USA.² CKD is an independent risk factor for cardiovascular and cerebrovascular events, and increased mortality.³ Both ischemic and hemorrhagic stroke risk increase as kidney function declines. A meta-analysis of 83 studies reported an inverse linear relationship whereby risk of stroke increased by 7% (RR: 1.07, CI: 1.04–1.09) for every 10 mL/min/1.73 m² decrease in estimated glomerular filtration rate (eGFR).⁴ In end-stage kidney disease (ESKD) patients on hemodialysis, the incidence of ischemic stroke is 2.5-fold higher than in age- and sex-matched controls, while the incidence of hemorrhagic stroke is increased 5-fold⁵. This paper discusses the controversial evidence for prevention of ischemic stroke in the CKD population including blood pressure control, statins, antiplatelet agents, and anticoagulation. Primary vs secondary stroke events are described if these were specified in the studies discussed.

Factors unique to the CKD milieu that increase stroke risk

The uremic milieu increases the risk of both ischemic and hemorrhagic stroke. Factors in CKD patients that contribute to stroke risk include:¹ blood pressure (BP) fluctuations due to activation of the renin-angiotensin-aldosterone axis, anemia, and fluid shifts with ultrafiltration on dialysis;² arterial medial calcification driven in part by deficiency of normal calcification inhibitors such as fetuin-A and α-klotho;³ build-up of vascular toxins such as urea and phosphorus;⁴ breakdown of arterial elastin;⁵ vascular inflammation driven by toxins derived from metabolites of the altered gut microbiome (indoxyl sulfate, trimethylamine N-oxide, etc.);⁶ platelet dysfunction; and⁷ higher prevalence of brain microbleeds. Some of these factors (BP variability, brain microbleeds) disproportionately increase the risk for hemorrhagic rather than ischemic stroke, thus complicating the use of therapies that can increase bleeding events such as antiplatelets and anticoagulation. For a more in-depth discussion of the factors listed above, the reader is directed to prior review papers.^6–8

Treatment of hypertension

Data is limited, as CKD patients were historically excluded from BP lowering trials. The largest CKD cohort to date is from SPRINT (Systolic Blood Pressure Intervention Trial) where CKD stage 3–4 patients (eGFR of 20 to <60) comprised 28% of the study population (approximately 2600 patients). The main finding from SPRINT was that a BP target <120 as compared to <140 mmHg significantly lowered the composite cardiovascular events primary outcome (composite of myocardial infarction, acute coronary syndrome, stroke, heart failure, or death from cardiovascular causes) but there was no clear effect on stroke on secondary outcomes analysis (HR: 0.89, 95% CI: 0.63–1.25).⁹ The lack of impact on stroke outcomes remained true in the subset of CKD patients, where stroke HR was 0.99 (95% CI: 0.57–1.70).¹⁰ The earlier PROGRESS (Perindopril Protection against Recurrent Stroke Study) trial included 1757 CKD patients with prior stroke history, and reported a significant decrease in secondary stroke events with perindopril-based therapy.^{11, 12} However, this study has been criticized for reporting outcomes as combined from two treatment arms and had no pre-specified BP targets.¹³

Further, data from large observational cohorts suggest potential harm when BP is lowered below systolic 120 mmHg in CKD patients. In secondary analysis of 1,549 CKD patients from the Atherosclerosis Risk in Communities Study (ARIC) and Cardiovascular Health Study (CHS) longitudinal cohorts, multivariate adjusted analysis showed a J-shaped relationship between BP and stroke outcomes. CKD patients on antihypertensives with systolic BP <120 mmHg were at significantly increased risk for incident stroke compared the reference group of CKD patients with systolic BP 120–129 mmHg (HR: 2.62, 95% CI: 1.22 to 5.66).¹⁴ Similar to the non-CKD patients, stroke risk was also increased with BP > 130 mmHg.

There is no randomized controlled trial (RCT) examining BP targets and stroke outcomes in dialysis patients. Large observational studies in dialysis have reported weak positive or no independent association between stroke risk and measures of BP.¹⁵ BP trials may be particularly difficult to design in the ESKD population where there is a prevalence of structural heart disease and vascular stiffness, and where BP can fluctuate widely due to volume expansion between dialysis treatments.¹⁵

Overall, the American College of Cardiology / American Heart Association (ACC/AHA) 2017 hypertension guidelines recommend a target BP <130/80 mmHg in adult CKD patients based on demonstrated benefit in reducing composite cardiovascular events;¹⁶ evidence specifically for stroke reduction is lacking. The Kidney Disease Improving Global Outcomes (KDIGO) Practice Guideline for the Management of Blood Pressure in CKD, last released in 2012, recommended lowering systolic BP to ≤ 140 mmHg (level B evidence).¹⁷

Treatment of dyslipidemia

Statins

Statins have not been shown in RCTs to reduce stroke risk in pre-dialysis CKD and ESKD patients. The Pravastatin Pooling Project (PPP) combined data from three placebo-controlled RCTs and included 4,491 patients with coronary heart disease and CKD stage 3 (eGFR 30–60 mL/min/1.73 m²). The primary end-point of time to myocardial infarction, coronary death, or coronary revascularization was significantly reduced with pravastatin 40 mg daily, but there was no benefit in regard to stroke.¹⁸

There have been three major statin RCTs in ESKD patients that examined the composite end-point of cardiovascular events or death. The consensus is that statins have little or no beneficial effects on stroke or other cardiovascular events in dialysis patients, despite clinically relevant reductions in serum cholesterol levels¹⁹ (Table 1). The German Diabetes and Dialysis Study (4D) compared atorvastatin 20 mg daily to placebo in 1,255 dialysis patients with type 2 diabetes mellitus and observed higher stroke mortality with statin therapy (RR: 2.03, 95% CI: 1.05–3.93).²⁰ The rate of nonfatal stroke was similar with/without statin therapy. The AURORA multi-national trial (A Study to Evaluate the Use of Rosuvastatin in Subjects on Regular Hemodialysis: An Assessment of Survival and Cardiovascular Events) enrolled 2,776 dialysis patients and compared rosuvastatin 10 mg daily to placebo; there was no difference in stroke incidence or mortality. Further, there was no relationship between the composite primary cardiovascular end point and baseline low-density lipoprotein (LDL) cholesterol levels.²¹

Table 1.

Randomized controlled trials have not shown benefit with statin therapy for stroke prevention in end-stage kidney disease (dialysis) patients.

Location	n	Median follow-up	Intervention vs placebo	Stroke outcomes
German Diabetes and Dialysis Study (4D)
Germany	1255 dialysis patients with type 2 diabetes mellitus	4 years	Atorvastatin 20 mg	Higher stroke mortality in statin group RR 2.03 [95% CI 1.05–3.93] p = 0.04
AURORA (An Assessment of Survival and Cardiovascular Events)
280 centers in 25 countries	2776 dialysis patients	3.8 years	Rosuvastatin 10 mg	Same stroke incidence 1.2 vs 1.1 per 100 patient-years p = 0.42
SHARP (Study of Heart and Renal Protection)
380 hospitals in 18 countries	9270 chronic kidney disease patients which included 3023 dialysis patients	4.9 years	Simvastatin 20 mg with ezetimibe 10 mg	Less ischemic stroke RR 0.72 [95% CI 0.57–0.92] p = 0.007 Lacked sufficient power to assess effects separately in dialysis patients

The 4D and AURORA trials did not detect any benefit with statins for stroke prevention in dialysis patients. The SHARP trial detected an overall benefit with combination simvastatin plus ezetimibe therapy in lowering stroke events in a large chronic kidney disease cohort; there was no statin-only arm, and the study was not adequately powered to assess outcomes in the dialysis subgroup. Subsequent meta-analyses that included the 4D, AURORA and SHARP cohorts concluded that there is no statistically significant difference between statin and placebo groups in terms of cardiovascular outcomes.

The multi-national SHARP trial (Study of Heart and Renal Protection) did observe a benefit with decreased cardiovascular events, but was not sufficiently powered to make conclusions in the dialysis subgroup. SHARP compared simvastatin 20 mg plus ezetimibe 10 mg daily to placebo in over 9,000 CKD patients and included 3,023 dialysis patients.²² There was a reduction in ischemic stroke in the overall population (114 [2·5%] vs 157 [3·4%]; RR: 0,72, 95% CI: 0.57–0.92). The lack of a statin-only arm makes it unclear if there is a benefit of adding ezetimibe independent of LDL-lowering effects. Subsequent meta-analyses that included the 4D, AURORA and SHARP cohorts concluded that there was no statistically significant difference between statin and placebo groups in terms of cardiovascular outcomes.¹⁹

PCSK9 Inhibition

The FOURIER (Further Cardiovascular Outcomes Research with PCSK9 Inhibition in Subjects with Elevated Risk) trial randomized individuals with clinically evident atherosclerosis and dyslipidemia (low-density lipoprotein cholesterol (LDL-C) ≥70 mg/dL or non–high-density lipoprotein cholesterol ≥100 mg/dL) to evolocumab or placebo.²³ The trial included 4,443 patients with CKD stage 3 or 4. Patients with an eGFR below 20 mL/min/1.73 m² or a history of kidney transplant were excluded. Absolute risk reduction at 30 months for the primary endpoint (composite of cardiovascular death, myocardial infarction, stroke, hospitalization for unstable angina, or coronary revascularization) was not significant in patients with CKD stage ≥3 (−1.5% with 95% CI: ™3.8% to 0.8%). However, absolute risk reduction for the key secondary endpoint (composite of cardiovascular death, myocardial infarction, or stroke) was more significant in the CKD stage 3/4 patients compared with non-CKD individuals (−2.5% with 95% CI: −4.7% to −0.4% vs −1.7% with 95% CI: −2.8% to 0.5% respectively). The impact of PCSK9 inhibition specifically on stroke was inconclusive, with secondary analysis demonstrating an absolute risk reduction of −0.4% (95% CI: −1.6% to 0.8%). Finally, PSCK9 inhibition did not impact eGFR decline at 30 months compared to placebo.²³

Antiplatelet agents

Antiplatelet drugs increase the risk for major and minor bleeding but do not significantly reduce the risk for stroke or mortality in CKD.²⁴ This is likely related to the complex combination of platelet defects in uremic patients that promote both bleeding and thrombosis.²⁵ Pro-bleeding defects include dysfunctional von Willebrand factor, anemia, decreased thromboxane A2 and cyclic adenosine monophosphate (cAMP), and conformational changes in glycoprotein IIb/IIIa. Pro-thrombotic defects include resistance to antiplatelet therapy (particularly relevant to clopidogrel), increased plasma fibrinogen, and preactivated glycoprotein IIb/IIIa.²⁵

Evidence for antiplatelet therapy for primary stroke prevention in CKD is limited. One large RCT did suggest a benefit for aspirin use in hypertensive CKD stage 3b-5 patients (eGFR <45 mL/min/1.73 m²). The Hypertension Optimal Study (HOT) randomized participants aged 50–80 years with a diastolic BP of 100–115 mmHg from 26 countries in Europe, North America, South America and Asia to aspirin 75 mg daily vs placebo; the study included >3,500 non-dialysis CKD patients.²⁶ For the total study population, no significant benefit was observed from aspirin therapy in terms of stroke or mortality. However, in the subset of patients with eGFR <45 mL/min/1.73 m² (536 patients) there was a significant decrease in cardiovascular events including decreased stroke risk (HR: 0.21, 95% CI: 0.06–0.75). The cardiovascular benefits have to be weight with increased bleeding events: “Among every 1,000 persons with eGFR <45 mL/min/1.73 m² treated for 3.8 years, 76 major cardiovascular events and 54 all-cause deaths will be prevented while 27 excess major bleeds will occur”.²⁶

There is no RCT in the ESKD population examining aspirin and stroke outcomes. Observational data from the multi-national Dialysis Outcomes and Practice Patterns Study (DOPPS) showed that aspirin did not decrease cardiovascular events.²⁷ A meta-analysis by the Antithrombotic Trialists’ Collaboration included a subgroup report of the effect of antiplatelet therapy on cardiovascular risk derived from RCTs primarily investigating antithrombotic therapy and hemodialysis vascular access patency.²⁸ The meta-analysis reported that aspirin reduced major cardiovascular events (nonfatal myocardial infarction, nonfatal stroke, or vascular death) in hemodialysis patients by 41% (RR: 0.59, 95% CI: 0.40–0.89) – however the risk of additional serious bleeding events would outweigh the stroke risk reduction. Herrington et al. applied summary risk ratios from this meta-analysis to event rates from the SHARP dialysis cohort, and estimated that treating 1,000 dialysis patients with vascular disease with aspirin for 5 years could prevent 16 ischemic strokes and 75 myocardial infarctions or revascularizations – but would incur an additional 19 intracranial bleeds and 53 serious extracranial bleeds.¹⁵

Adding clopidogrel on top of aspirin therapy has no clear benefits in CKD, likely related to a high prevalence of clopidogrel resistance (up to 50–80% in ESKD). Exact mechanisms of clopidogrel resistance are unclear but proposed factors include increased platelet turnover, diminished oral absorption, and altered drug metabolism.²⁵ RCTs and observational data from large CKD cohorts have reported either no effect or worse stroke outcomes with clopidogrel therapy (Table 2).^{25,27,29–32}

Table 2.

Clinical trials with the antiplatelet agent clopidogrel (added to aspirin therapy) have shown no clear benefit for stroke prevention in the chronic kidney disease population.

Study	n	Follow-up period	Clopidogrel intervention	Clopidogrel and stroke outcomes
CURE (Clopidogrel in Unstable Angina to Prevent Recurrent Events) Keltai et al. 2007	Post hoc analysis of 12,253 patients after acute coronary syndrome, analyzed in tertiles of kidney function	12 months	300 mg loading dose then 75 mg daily vs placebo; all patients were on aspirin	No benefit in lowest tertile of kidney function for composite primary outcome (CV death, MI, stroke) RR 0.89 [95% CI 0.76–1.05]
CREDO (Clopidogrel for the Reduction of Events During Observation) Best et al. 2008	Post hoc analysis of 2002 patients after elective PCI, including 331 CKD patients with estimated creatinine clearance <60 mL/min	12 months	75 mg daily for 28 days vs 300 mg loading dose then 75 mg daily for 1 year; all patients were on aspirin	Similar rate of composite primary outcome (death, MI, stroke) 13.1% vs 17.8% p = 0.24
CHARISMA (Clopi-dogrel for High Atherothrombotic Risk and Ischemic Stabilization, Management, and Avoidance) Dasgupta et al. 2009	Post hoc analysis of 15,603 patients including nondiabetic, diabetic without nephropathy, and diabetic nephropathy patients	Median follow-up 28 months	75 mg daily vs pla-cebo; all patients were on aspirin	Increased risk of all-cause and CV mortality only in the diabetic nephropathy subgroup
DOPPS (Dialysis Outcomes and Practice Patterns Study) Sood et al. 2013	48,144 dialysis patients in 12 countries; 3–25% antiplatelet agents including clopidogrel	1.5 years (average)	Observational cohort; clopidogrel grouped with ticlopidine, dipyridamole and pentoxifylline for outcomes analysis	increased risk of death from stroke, all-cause and CV mortality
Taiwan dialysis cohort Chen et al. 2014	1936 dialysis patients admitted for first ischemic stroke; 146 patients were started on clopidogrel	Minimum follow-up 3 years	Retrospective cohort study	Clopidogrel had no effect on risk of second ischemic stroke or all-cause mortality, but increased risk of MI with HR 1.564 [95% CI 1.232–2.985]

This is in part due to high prevalence of clopidogrel resistance at lower levels of kidney function; this prevalence is 50–80% in the dialysis population. Mechanisms of uremic clopidogrel resistance may be related to increased platelet turnover, diminished oral absorption, and altered metabolism. CKD = chronic kidney disease, CV = cardiovascular, MI =myocardial infarction, PCI = percutaneous coronary intervention.

Oral anticoagulation agents

The inherent risk for serious bleeding events with oral anticoagulation is of particular concern in ESKD patients where, as noted in the Introduction above, the incidence of hemorrhagic stroke is increased 5-fold as compared to the general population.⁵ Figure 1 summarizes the intravenous, subcutaneous and oral formulations of anticoagulation agents available in the United States.

Fig. 1.

Anticoagulation formulations available in the United States. In terms of oral formulations, warfarin was the only oral anticoagulant on the market for over 50 years. Since 2010 the FDA has approved four direct oral anticoagulants (DOACs). Dabigatran, a direct thrombin inhibitor, is primarily metabolized by the kidneys and its use is avoided in advanced chronic kidney disease. The factor Xa inhibitors are primarily metabolized in the liver, and includes apixaban which is the only DOAC approved for use in end-stage kidney disease patients. IV = intravenous, SQ = subcutaneous.

From 1954 to 2010, warfarin was the only oral anticoagulant available. In the general population, warfarin is more effective than aspirin for stroke prevention in atrial fibrillation and reduces stroke risk by 64% (95% CI: 49–74%).^33,34 ESKD patients were historically excluded from anticoagulation RCTs, thus data for the CKD population is derived mainly from observational real-world outcomes. Two large meta-analyses that included over 40,000 CKD patients with atrial fibrillation were published in 2016.^35,36 In non-dialysis CKD patients, warfarin decreased risk of ischemic stroke/thromboembolism (HR: 0.70, 95% CI: 0.54–0.89).³⁶ However, in the hemodialysis population, warfarin either had no effect on stroke risk³⁶ or was associated with increased stroke risk.³⁵ Further, major bleeding events were significantly increased (HR: 1.30, 95% CI: 1.08–1.56).³⁶

A more recent meta-analysis of warfarin outcomes in ESKD examined 15 studies published between 2008–2019 which encompassed 47,480 dialysis patients with atrial fibrillation; 10,445 (22%) were on warfarin.³⁷ Mean follow-up period was 2.6 years. Warfarin therapy had no impact on mortality or ischemic stroke outcomes; but was associated with significantly increased hemorrhagic stroke events (HR: 1.49, 95% CI: 1.03–1.94).³⁷ In our single center experience, warfarin use in ESKD was associated with higher mortality and bleeding events compared ESKD patients with indications for anticoagulation therapy but were not initiated on warfarin; there was no difference on stroke risk.³⁸

Aside from increasing the risk of major bleeding, warfarin may worsen vascular calcification in CKD. There is a well-known association between CKD and increased arterial medial calcification due to uremic factors that include phosphorus toxicity, elastin degradation, and deficiency of endogenous inhibitors of calcification.³⁹ Warfarin affects all vitamin K-dependent pathways by inhibiting vitamin K epoxide reductase, which is necessary for recycling of vitamin K. While producing the desired effect (inhibition of clotting factors II, VII, IX and X), warfarin indiscriminately impacts other beneficial proteins such as the calcification inhibitor matrix glutamate protein (MGP). MGP relies on vitamin K-dependent carboxylation to be activated; warfarin thus results in a functional MGP deficiency.⁴⁰ Calcific uremic arteriolopathy (formerly called calciphylaxis) is a rare but often fatal syndrome of calcification in cutaneous blood vessels leading to skin necrosis, that occurs in 5% of ESKD patients.⁴¹ Of note, ~50% of such patients with calcific uremic arteriolopathy are on warfarin.⁴²

Since 2010, four direct oral anticoagulants (DOACs) have emerged on the market, with apixaban gaining prominence in the CKD population. Dabigatran, a direct thrombin inhibitor, is primarily metabolized by the kidneys and was FDA-approved in 2010. The factor Xa inhibitors are primarily metabolized in the liver and include rivaroxaban, apixaban and edoxaban (FDA-approved in 2011, 2012 and 2015 respectively). There were four large Phase III trials comparing DOACs with warfarin in a total of >71,000 patients with atrial fibrillation: RE-LY (dabigatran), ROCKET-AF (rivaroxaban), ARISTOTLE (apixaban) and ENGAGE AF-TIMI 48 (edoxaban). CKD stage 4/5 patients were excluded from these trials. Overall, DOACs demonstrated a greater net clinical benefit in CKD stage 3 with stroke and bleeding outcomes.^43,44 In CKD patients, DOACs theoretically confer additional advantage over warfarin because they do not inhibit the vitamin K–dependent carboxylation activation of MGP.

Apixaban, with its lower degree of renal clearance, was approved for use in dialysis patients based upon a single-dose pharmacokinetic study involving eight hemodialysis patients.⁴⁵ A more rigorous pharmacokinetics study later raised concerns about drug accumulation in ESKD, as the standard-dose apixaban 5 mg twice a day led to average blood trough levels that were above the 90th percentile compared to reference levels in persons with normal kidney function.⁴⁶

Concerns for drug safety with standard-dose apixaban in ESKD appear to conflict with efficacy data. Analysis of the US Renal Data System (2010–2015 period) encompassing 25,523 ESKD patients with atrial fibrillation (~10% on apixaban, matched 1:3 to warfarin based on prognostic score) noted no difference in the risks of stroke/systemic embolism between apixaban and warfarin (HR: 0.88, 95% CI: 0.69–1.12.⁴⁷ However, on sensitivity analysis, standard-dose apixaban (n = 1034) was associated with significantly lower risk of stroke/systemic embolism and death as compared with lower-dose apixaban 2.5 mg twice a day or warfarin.⁴⁷ Apixaban was associated with a significantly lower risk of major bleeding overall as compared to warfarin.

There are three anticoagulation RCTs in dialysis patients, and these compare apixaban to warfarin. The AXADIA trial is ongoing in Germany.⁴⁸ The RENAL-AF trial, which randomized dialysis patients in the United States to apixaban vs warfarin, was stopped in 2019 due to lack of funding after 155 of a planned 760 patients were enrolled and produced inconclusive results.⁴⁹ The SAFE-D trial in Canada includes a placebo group and thus will be informative as to whether anticoagulation per se (be it DOAC or warfarin) definitively modifies ischemic stroke risk in ESKD (ClinicalTrials.gov Identifier: NCT03987711).

Conclusions

Lowering BP below systolic 120 mmHg and statins have not shown clear benefit with stroke prevention in the CKD population. The multi-national HOT trial suggests that aspirin reduces incident stroke risk in CKD patients with eGFR < 45 mL/min/1.73 m² however benefits have to be weighed against increased bleeding events. Antiplatelets have not shown clear benefit for secondary stroke prevention in CKD. In CKD patients with atrial fibrillation who may benefit from anticoagulation, observational data suggests that DOACs have a favorable efficacy and safety profile over warfarin.