In this meta-analysis of 21 articles derived from 33 prospective studies of generally good quality, among over 280
000 people experiencing almost 8000 stroke events, we found that patients with a baseline estimated glomerular filtration rate (eGFR) of <60 ml/min/1.73 m2
had a risk of future stroke that was 43% greater than those with a normal baseline eGFR. This relation was consistent across diverse population subgroups—that is, those with or without traditional cardiovascular risk factors. The size and inclusion of only prospectively collected data strengthened the robustness of our findings, as selection bias, recall bias, and reverse causality were unlikely. In addition, all studies included in our meta-analysis reported a multivariate adjusted relative risk, which probably mitigated the possibility of known confounding influencing our results.
We used subgroup analyses to assess the varying influence of several factors on the association between eGFR <60 ml/min/1.73 m2 and risk of stroke. The magnitude of risk was larger when studies used an eGFR >90 ml/min/1.73 m2 as reference compared with >60 ml/min/1.73 m2, which raised the possibility that an eGFR of 60-90 ml/min/1.73 m2 may increase the risk of stroke compared with an eGFR >90 ml/min/1.73 m2. Our formal meta-analysis did not, however, show significantly increased risk of incident stroke among patients with an eGFR of 60-90 ml/min/1.73 m2. The explanation could be that such a rate is not sensitive enough as a marker of kidney disease to discriminate risk of stroke. We did, however, find a possible dose-response relation between eGFR and stroke at levels <60 ml/min/1.73 m2—that is, the risk of stroke was significantly greater for eGFR <40 ml/min/1.73 m2 than for levels of 40-60 ml/min/1.73 m2.
A meta-analysis based on observational studies cannot prove causality. However, based on these results it may not be unreasonable to regard the presence of a low eGFR as a marker for increased risk of stroke, prompting optimal application of established vascular risk reduction strategies such as control of blood pressure, statin use, and antiplatelet therapy.7
Interestingly we found that Asian people with a low baseline eGFR seemed to be at higher risk of future stroke. Indeed, in Asian populations, hypertension is a major risk factor of both stroke and death from renal causes,39 40
chronic kidney disease further increases the association of blood pressure with stroke,25
and meta-analysis showed that the risk of stroke associated with hypertension is consistently and significantly greater in Chinese than in white people.41
Furthermore, it has been suggested that Asian people tend to develop hypertension at earlier ages than other races,42
and it is conceivable that a longer history of hypertension may cause more profound damage of end organs and vessels thereby leading to a higher likelihood of vascular events within a given study period. A systematic review that linked reduced eGFR with increased risk of coronary heart disease was only among participants in Western countries and so did not have the means of exploring this issue.5
Although most of the studies we analysed adjusted for hypertension or blood pressure, none adjusted for the duration of hypertension, thereby limiting the extent to which we could fully adjust for hypertension as a confounder. As such, this potential disparity between races will need to be more comprehensively explored in future studies.
We also observed that the effect of reduced eGFR was more profound on the risk of fatal stroke than on all strokes, which probably points to the association of compromised kidney function with risk factors for generally poor clinical outcomes such as oxidative stress, widespread inflammation, electrolyte derangements, procoagulation, and presence of uraemic toxins.3
In fact, kidney disease even of mild severity has been shown to be an independent predictor of poorer clinical outcomes among people with stroke, including higher risk of all cause mortality and cardiovascular mortality.43 44
Also of note, the presence of albuminuria did not substantially further increase the risk of stroke among patients with a baseline eGFR of <60 or >60 ml/min/1.73 m2
. Our result should be interpreted with caution, however, as it was based on just three studies and the rate of albuminuria is low in people without diabetes. A recent meta-analysis showed that compared with people without albuminuria or a low eGFR, those with either condition had a higher risk of cardiovascular death and those with both conditions had the highest risk of cardiovascular death.6
Furthermore, meta-analyses have shown that albuminuria was independently associated with a higher risk of stroke even when the included studies had adjusted for eGFR or serum creatinine level.45 46
Limitations of this meta-analysis
Limitations of this meta-analysis must be considered. Firstly, meta-analyses may be biased if the literature search fails to identify all relevant studies or the selection criteria for including a study are applied in a subjective manner. To minimise these risks we carried out thorough searches across different databases using explicit criteria for study selection, data abstraction, and data analysis. Secondly, compared with studies of good quality, those of fair quality showed a stronger association between reduced eGFR and stroke. When we restricted analysis to good quality studies, the estimate of association slightly decreased. Thirdly, a large amount of heterogeneity was observed in the results of the various studies. Although subanalyses were done to identify this, heterogeneity persisted in many subgroups, suggesting that other factors might explain this result. Meta-regression by average baseline eGFR and other variables could have been a better way of exploring potential sources of heterogeneity. However, most included articles did not provide average baseline eGFR in each eGFR category, which prevented us from exploring further. In those studies that provided both age and sex adjusted and multivariate unadjusted estimates, the association between reduced eGFR and stroke was slightly, but significantly, attenuated after further adjustment. Such an attenuation in effect size suggests that residual confounding may have remained and that the summary result presented here may be a slight overestimation of the true magnitude of the association between reduced eGFR and risk of stroke. Despite these limitations, the results of this systematic review represent the most precise and accurate estimate of the strength of the relation between reduced eGFR and incident stroke currently available.
Our formal meta-analysis found a significant association between eGFR <60 ml/min/1.73 m2 and increased incident stroke across various populations, after adjustment for established cardiovascular risk factors. None the less, these results possibly underestimated the magnitude of this relation because a reduced eGFR often simultaneously exists with several traditional and novel vascular risk factors. Of major public health interest were our findings that Asian patients with a low eGFR were at higher risk for stroke than their non-Asian counterparts, that below an eGFR level of 60 ml/min/1.73 m2 a dose-response relation with risk of stroke might exist, and that fatal strokes were especially associated with low baseline eGFR.
At this juncture, a low baseline eGFR should be seen simply as a risk marker. Established evidence based strategies already proved to mitigate vascular risk, such as reduction of blood pressure, when promptly and appropriately applied are likely to avert future strokes in people with renal insufficiency. Specific patient subgroups with a low eGFR, such as people of Asian race, may particularly benefit.
What is already known on this topic
- Most patients with an estimated glomerular filtration rate (eGFR) <60 ml/min/1.73 m2 die of cardiovascular causes and not progression to end stage renal disease
- A recent meta-analysis showed that an eGFR <60 ml/min/1.73 m2 was associated with all cause and cardiovascular mortality in the general population
What this study adds
- People with a baseline eGFR <60 ml/min/1.73 m2 had an independent risk of future stroke that was 43% greater than those with a normal baseline eGFR
- A dose-response relation between eGFR <60 ml/min/1.73 m2 and risk of stroke was observed, with risk of stroke being significantly greater for levels <40 ml/min/1.73 m2 compared with 40-60 ml/min/1.73 m2
- Asian patients with an eGFR <60 ml/min/1.73 m2 were at higher risk of stroke than people of non-Asian ethnicity