Ali Saad, MD

Sandsmark DK, Messé SR, Zhang X, Roy J, Nessel L, Hamm LL, et al. Proteinuria, but Not eGFR, Predicts Stroke Risk in Chronic Kidney Disease: Chronic Renal Insufficiency Cohort Study. Stroke. 2015

In the CRIC study (chronic renal insufficiency cohort) they took about 4000 people with CKD (chronic kidney disease) and followed them prospectively to see if they had strokes. 143 stroke events (25 of them were intracerebral hemorrhage) occurred over 6.4 years and proteinuria was found to be a risk factor for stroke independent of eGFR (estimated glomerular filtration rate) (HR 3.1, p<0.0001). Albuminuria more specifically was an even greater risk factor (HR 3.6, p<0.0001). Albuminuria remained statistically significant at all abnormal value ranges after multivariable analysis while proteinuria remained significant only at the mid and high abnormal value ranges. Another important result is that using an ACEI/ARB to decrease proteinuria did not decrease stroke risk. 

These patients were not dialysis dependant or on immunotherapy for renal disease or vasculitis. The ones who had strokes were also more likely to be older, black, have a history DM, MI, stroke, tobacco use, or alcohol use. Although there was a dose dependant response of lower eGFR correlating with a greater risk of stroke, it didn’t reach statistical significance once demographics were controlled for in multivariable analysis.
Limitations of this study:
– the results don’t necessarily mean that low eGFR is not a stroke risk factor since there were no healthy controls and everyone had renal dysfunction
– stroke subtypes were not specified, but one would think that most of these people would have small vessel disease strokes
– history and meds were self reported
– not enough ICH to make any conclusions about the association between ICH and proteinuria. also there is no distinction between primary ICH and hemorrhagic conversion of an ischemic stroke
– TIAs were not included
– patients were only followed for about 6 years
– population was mostly black males in their 50s-70s
– strokes were defined as definite/probable by patient reporting which then triggered medical records review, but there is no mention of whether this was based on neuroimaging

How does this study change my practice?
I already cite CKD as a stroke risk factor to patients, there’s just more data to back it up now. I wouldn’t collect a 24 hour urine sample in order to help risk stratify someone as it isn’t practical. It’s hard enough getting people to agree to heart monitors. Perhaps albuminuria can be incorporated into a stroke risk calculator like the CHA2DS2-VASc score or something similar?