Chronic kidney disease and poor outcomes in ischemic stroke: is impaired cerebral autoregulation the missing link?

Abstract

Background: Chronic kidney disease increases stroke incidence and severity but the mechanisms behind this cerebro-renal interaction are mostly unexplored. Since both vascular beds share similar features, microvascular dysfunction could be the possible missing link. Therefore, we examined the relationship between renal function and cerebral autoregulation in the early hours post ischemia and its impact on outcome.

Methods: We enrolled 46 ischemic strokes (middle cerebral artery). Dynamic cerebral autoregulation was assessed by transfer function (coherence, phase and gain) of spontaneous blood pressure oscillations to blood flow velocity within 6 h from symptom-onset. Estimated glomerular filtration rate (eGFR) was calculated. Hemorrhagic transformation (HT) and white matter lesions (WML) were collected from computed tomography performed at presentation and 24 h. Outcome was evaluated with modified Rankin Scale at 3 months.

Results: High gain (less effective autoregulation) was correlated with lower eGFR irrespective of infarct side (p < 0.05). Both lower eGFR and higher gain correlated with WML grade (p < 0.05). Lower eGFR and increased gain, alone and in combination, progressively reduced the odds of a good functional outcome [ipsilateral OR = 4.39 (CI95% 3.15-25.6), p = 0.019; contralateral OR = 8.15 (CI95% 4.15-15.6), p = 0.002] and increased risk of HT [ipsilateral OR = 3.48 (CI95% 0.60-24.0), p = 0.132; contralateral OR = 6.43 (CI95% 1.40-32.1), p = 0.034].

Conclusions: Lower renal function correlates with less effective dynamic cerebral autoregulation in acute ischemic stroke, both predicting a bad outcome. The evaluation of serum biomarkers of renal dysfunction could have interest in the future for assessing cerebral microvascular risk and relationship with stroke complications.

Keywords: Cardiovascular disease; Chronic kidney disease; Glomerular filtration rate; Stroke; transcranial Doppler.

Fig. 1

Frequency spectra from 0.03 to 0.5 Hz of Coherence and Transfer Function Gain and Phase measured in acute stroke patients (< 6 h) accordingly to their estimated glomerular filtration rate (eGFR). Continuous and spotted lines within dark grey and spotted white strips corresponds to means±SE of groups of low (< 60 mL/min/1.73 m²) and high (≥60 mL/min/1.73 m²) subgroups of eGFR, respectively. Groups were compared with Mann-Whitney and P values presented at right superior corner of each plot. Lower GFR shows higher gain values bilaterally which indicate less effective dynamic cerebral autoregulation

Fig. 2

Box plots comparing median and quartiles distribution of cerebral autoregulation. Gain, at low (LF: 0.03–0.15 Hz) and high (HF: 0.15–0.5 Hz) frequency bands (subpanels A and B), and estimated glomerular filtration rate (GFR, subpanel C) by groups with and without chronic use of calcium channel blockers (CCB). Statistical P value of Mann-whitey test (a, b) and T-test (c) is presented

Fig. 3

Relationship between renal function subgroups of low (< 60 mL/min/1.73 m²) and high (≥60 mL/min/1.73 m²) subgroups of estimated glomerular filtration rate (eGFR) and outcome accordingly to modified Rankin scale (a), risk of hemorrhage (b) and the severity of white matter lesions assessed at 24-h head Computed Tomography, accordingly to vanSwieten scale [33]. Subgroups were compared with ordinal regression or logistic regression as appropriate

Fig. 4

Relationship between cerebral autoregulation within 6 h of symptoms by subgroups of low and high LF gain subgroups for infarct (≤0.90, > 0.9%²/mmHg², respectively) and non-infarct hemisphere (≤1.10, > 1.10%²/mmHg², respectively) and outcome accordingly to modified Rankin scale (a), risk of hemorrhage (b) and the severity of white matter lesions assessed at 24-h head Computed Tomography, accordingly to vanSwieten scale [33]. Subgroups were compared with ordinal regression or logistic regression as appropriate

Fig. 5

Interaction between cerebral autoregulation and renal function on outcome accordingly to modified Rankin scale (a), risk of hemorrhage (b) and on severity of white matter lesions assessed at 24-h head Computed Tomography, accordingly to vanSwieten scale [33]. Subgroups were created by splitting renal function into low (< 60 mL/min/1.73 m²) and high (≥60 mL/min/1.73 m²) estimated glomerular filtration rate (eGFR); cerebral autoregulation into low (Lo gain) and high (Hi gain) subgroups of LF gain in infarct (≤ 0.90 and > 0.90%²/mmHg²) and non-infarct hemisphere (≤ 1.10 and > 1.10%²/mmHg²). Notice that higher gain values represents worse levels of cerebral autoregulation. The interaction effect in outcome and white matter lesions was tested in multinomial logistic or ordinal regression models as appropriate