Two-hundred eighty-two patients including 155 men and 127 women, aged 65.20 ± 12.70 years (17-90) were enrolled.
The mean NIHSS score was 13.16 ± 6.39 at time of admission, and it took averagely 55.99 ± 30.12 hours for tPA injection (5-184h). The NIHSS score improved significantly after the injection in 36 hours (p=0.027) with the mean thirty-six-hour score of 10.10 ± 8.93.
RDW values ranged from 10.4 to 20.5, with the mean value of 13.67 ± 1.17.
Mean values for RDW did not significantly correlate with the severity of stroke (p = 0.11). In mild form of stroke (NIHSS = 0-6), the mean value for RDW was 13.60 ± 0.22 and in stroke of moderate severity (NIHSS = 7-15), it was 13.58 ± 0.11. For patients with severe stroke (NIHSS > 16), the mean RDW was higher than the mild to moderate cases with mean value of 13.854 ± 0.12.
Table 1 demonstrates patients' characteristics according to the severity of stroke.
The patients were evaluated 7 days after the time of admission; however, the data was available only for 268 patients. The mean NIHSS was 2.7 ± 1.7 seven days after anti-thrombolytic therapy. The MRS was 3.14 ± 2.22 thirty-six hours after the treatment, which improved by 2.7 ± 1.73 in 7 days.
Classification of results based on the RDW values
The patients' characteristics according to the baseline RDW level is demonstrated in Table 2.
The categorization of RDW to <<12.9 and >13 values revealed insignificant difference in stroke severity score, accounting for the mean baseline NIHSS of 11.74 ± 6.39 in normal RDW values and 13.38 ± 0.49 in higher RDW group (p = 0.60). Similarly, the mean NIHSS of subjects with RDW<12.9 was lower than the patients with RDW>13 in each thirty-six-hour and seven-day evaluation, while the difference between two groups was statistically insignificant. The mean NIHSS was 8.19 ± 8.2 and 9.94 ± 8.28 in patients with normal and higher RDW values, respectively (p = 0.64). After seven days, NIHSS was 6.46 ± 7.28 in normal RDW group and was 8.52 ± 8.35 in increased RDW group (p = 0.058).
The categorization of final outcome according to RDW level demonstrated mean MRS of 2.74 ± 1.56 within thirty-six hours of tPA injection in the group of patients with normal RDW value, which was 3.25 ± 2.55 in increased RDW group. The final outcome results had a trend toward improvement in both RDW categories after seven days. The mean MRS was 2.33 ± 1.59 and 2.72 ± 1.75 in normal and increased RDW group, respectively.
Neither the thirty-six-hour, nor the seventh day MRS demonstrated significant difference between those with normal and higher RDW values.
- Length of stay at hospital
The length of stay at hospital in patients with RDW<<12.9 was 14.34 ± 18.5 and in those with RDW > 13 was 15.08 ± 15.9. The results didn't differ significantly between two groups (p = 0.96).
SICH occurred in 14 patients, in 6 patients severe symptoms led to diagnosis and in 8 of them hemorrhage was asymptomatic. In patients with normal RDW level, 2.04% had symptomatic hemorrhage and 2.04% had asymptomatic hemorrhage. Among patients with elevated RDW, asymptomatic and symptomatic hemorrhage occurred in 4.69% and 3.35% of the patients. The analysis with Pearson's test did not reveal a correlation between SICH and RDW.
Three-month Follow up
Out of 282 enrolled subjects, only 208 referred for the three-month follow up session, 98 of them had a good final outcome with mean baseline RDW of 13.57 ± 1.35 and 110 had poor outcome on MRS evaluation with score of 13.76 ± 1.06. The linear regression analysis didn’t address any significant regression between final outcome results and baseline RDW values (r = 0.04, p = 0.52).
The sensitivity, specificity and AUC of baseline RDW for predicting final outcome within 36 hours were 77.6%, 69.8% and 0.51, respectively (Figure 1.A). In 7-day follow up, the sensitivity, specificity and AUC of baseline RDW were 75.0%, 74.3%, and 0.48, respectively (Figure 1.B). And during the period of 3 months the sensitivity, specificity and AUC of baseline RDW for MRS prediction were 74.4%, 71.4% and 0.57, respectively (Figure 1.C).
Corrections for confounding factors
Multivariant linear regression analysis revealed only a significant correlation between age and stroke severity (p= 0.01) and outcome (p= 0.03). However, after corrections for the age there was still an insignificant relation between RDW and stroke severity and outcome (p= 0.20).