Is Visible Infarction on Computed Tomography Associated With an Adverse Prognosis in Acute Ischemic Stroke?
To the Editor:
In the article by Wardlaw et al in the July issue of Stroke,1 the outcome of stroke patients with an evident lesion on CT scan was compared with the outcome of stroke patients with no evident lesion. The result was that the prognosis for a patient with a “clinical stroke” evident on CT scan also was worse than the prognosis for a patient showing no evidence on CT scan of the lesion causing the clinical stroke.
However, obvious, it may due to a bias.
How severe were the patients without the CT lesion? How were the various times of their CT scans distributed. Was it in a manner similar to that shown in the article’s Figure 2, for patients with CT lesions? In fact, patients with total anterior circulation infarcts (TACI) are more frequent (80%–95%) in the group with a positive CT, and vice-versa, lacunar infarcts (LACI) are more frequent in the group without positive CT scan (45%–60%): ie, the results may be due just to this unbalance. In fact, the sentence, “Patients who had a CT scan within a week of the stroke were more likely to be dead or dependent at 6 months than those scanned later …” (p 1317), does not indicate a lethal effect of the CT scan but rather that in the Lothian Area physicians tend to ask for CT scan earlier in the more severe cases.
Regarding the number of patients (n=124; 10% of the registry) who did not have a CT scan examination “for humane reasons” or for other reasons, I think that the AA made a little, but significant, error in calculating the percent of the patients who died within 14 days of stroke onset (p 1317): in fact, if the total number is 24 of 124, the result is 19%, instead of 15%.
Finally, do the AA agree that an outcome as severe as 89% of death or dependence for TACI and 40% of death or dependence for partial anterior circulation infarcts (PACI) may well justify incisive therapeutical attempts?
- Copyright © 1998 by American Heart Association
Wardlaw JM, Lewis SC, Dennis MS, Counsell C, McDowall M. Is visible infarction on computed tomography associated with an adverse prognosis in acute ischemic stroke? Stroke. 1998;29:1315–1319.
Our study set out to answer the question “Is visible infarction on CT associated with an adverse prognosis in acute ischemic stroke?” after all other possible confounding factors had been accounted for including time from symptom onset to CT, clinical stroke syndrome (TACI, LACI, PACI, POCI), severity of the stroke, and so on. We wished to avoid confounding due to problems such as the fact that patients with more severe strokes may reach the hospital more quickly and be scanned more quickly than those with milder strokes, as eluded to by Dr Lenzi in his letter. To that end, we developed a logistic regression model, detailed in the methods and in the tables and appendix, to take account of all these factors. This, therefore, as far as one possibly can, avoids the problem of bias and confounding so that when all these factors are taken into consideration, the bottom line is that a patient with a particular syndrome scanned at a particular time after their stroke with a particular severity of stroke, and so on, has a slightly greater risk of poor functional outcome if the scan shows a recent infarct corresponding with their symptoms than does an identical patient whose CT scan does not show a recent relevant infarct. It is well known that patients with a TACI are much more likely to have a visible infarct on their CT scan than those with a LACI, and so one has to take the stroke syndrome, along with everything else, into consideration to avoid bias.
We obviously had to exclude from the analysis any patients who had not had a CT scan, whatever the reason (some of them had MR instead, some of them died before the scan could be done, some of them were moribund on admission), but we were concerned to see whether exclusion of those patients (and there was nothing else we could do with them) would introduce significant bias to the patient group. The group not CT scanned on average were more likely to have died early on, because they were either moribund on admission or died before the scan was done, which two groups made up the majority. However, apart from acknowledging the slight difference in this patient population, there is not much else we can do about them. Dr Lenzi quite rightly points out a typographical error on page 1317 that should be 19%, not 15%, of those who did not have a CT scan died within 14 days; however this difference of 4% is not statistically or clinically significant.
Finally, the functional outcome for TACI and PACI patients in our cohort was very similar to that described in previous cohorts and is, as we all know, a clear indication of what a devastating condition stroke is and clear justification for urgently finding an effective treatment. We don’t see how anybody could possibly disagree with that, but it was really not anything to do with the content of our paper.