Introduction: The widely used Cincinnati Prehospital Stroke Scale (CPSS) has a sensitivity for stroke identification of only about 50%. A previous study suggested that stroke was missed more frequently by EMS using the CPSS in patients without motor signs. The purpose of this study was to compare the elements of the CPSS to the corresponding items of the gold standard National Institutes of Health Stroke Scale (NIHSS).
Hypothesis: Speech and language components of the NIHSS are more sensitive than the corresponding speech component of the CPSS.
Methods: We retrospectively studied patients transported to University of Alabama at Birmingham by EMS with suspected stroke from January to May of 2016. We excluded patients who did not have the CPSS performed by EMS in the field or admission NIHSS performed by neurology residents and those without a discharge diagnosis of stroke. We compared the sensitivity of abnormal facial droop, arm drift, and speech of the CPSS to the corresponding NIHSS items (facial palsy ≥ 1, left or right motor arm ≥ 1, best language or dysarthria ≥ 1) for a discharge diagnosis of stroke by vascular neurologists.
Results: There were 232 subjects (White, 52%; Black, 44%; men, 51%: age=66, sd=16) There was no difference in the sensitivity of arm drift for discharge diagnosis of stroke between the CPSS (77.6%, 95% CI 70-85) and NIHSS (79.2%, 71-87), p=0.13. Compared to the NIHSS, the CPSS had significantly lower sensitivities on tests of facial weakness (51.2%, 42-60 vs. 65.3%, 55-75, p=0.006) and speech and language abnormalities (63.2%, 54-72 vs.71.6%, 61-80, p<0.0001).
Conclusions: CPSS items testing facial weakness and speech/language are significantly less sensitive compared to corresponding NIHSS items. This may relate to more rigorous testing of these items on the NIHSS, but may also relate to different examiners or settings. Prehospital scales derived from NIHSS items may not accurately reflect performance in the prehospital setting.