Author + information
- Received May 29, 2007
- Revision received August 14, 2007
- Accepted September 23, 2007
- Published online March 4, 2008.
- Pooja Khatri, MD⁎,⁎ (, )
- Robert A. Taylor, MD†,
- Vanessa Palumbo, MD‡,
- Venkatakrishna Rajajee, MD§,
- Jeffrey M. Katz, MD∥,
- Julio A. Chalela, MD¶,
- Ann Geers, RN⁎,
- Joseph Haymore, MS, ACNP#,
- Daniel M. Kolansky, MD⁎⁎,
- Scott E. Kasner, MD††,
- Treatment of Acute Stroke after Cardiac Catheterization (TASCC) Study Group
- ↵⁎Reprint requests and correspondence:
Dr. Pooja Khatri, University of Cincinnati Academic Health Center, Department of Neurology, 260 Stetson Street, Suite 2300, P.O. Box 670525, Cincinnati, Ohio 45267-0525.
Objectives The purpose of this study was to systematically compare clinical outcomes of patients treated with thrombolysis with those without treatment in a multi-year, multicenter cohort of strokes after cardiac catheterization.
Background Ischemic strokes after cardiac catheterization procedures, although uncommon, lead to the morbidity and mortality of thousands of patients each year. Despite the availability of Food and Drug Administration–approved thrombolytic therapy for acute ischemic stroke since 1996, thrombolysis remains unestablished in the setting of cardiac catheterization, owing to unique concerns regarding safety and efficacy.
Methods Consecutive cases of ischemic stroke after cardiac catheterization were abstracted retrospectively and reviewed by clinicians at 7 major North American academic centers with acute stroke teams. Safety and efficacy outcome measures were pre-defined.
Results A total of 66 cases of ischemic strokes after cardiac catheterization were identified over 3 to 4 years; 12 (18%) were treated with thrombolysis, consisting of 7 intravenous and 5 intra-arterial recombinant tissue plasminogen activator cases. Improvement in stroke symptoms, as measured by the primary efficacy measure of median change in National Institutes of Health Stroke Scale score from baseline to 24 h, was greater in treated versus nontreated cases (p < 0.001). Additional secondary measures of efficacy also showed better outcomes in the treated group. There were no significant differences in bleeding events, defined as symptomatic intracerebral hemorrhage, hemopericardium, or other systemic bleeding resulting in hemodynamic instability or blood tranfusions. Mortality rates were also similar.
Conclusions Thrombolysis might improve early outcomes after post-catheterization strokes and seems safe in this context. Emergent cerebral revascularization should be a routine consideration.
More than 2 million cardiac catheterization procedures are performed in the U.S. annually. Although strokes after cardiac catheterization (SCCs) are relatively rare, this high volume of cardiac catheterizations in the U.S. leads to thousands of SCCs each year. Rates of SCCs, including both ischemic and hemorrhagic subtypes, range widely from 0.07% to 7.0%. Large contemporary registries of exclusively diagnostic and invasive coronary procedures report rates from 0.07% to 0.38%, and smaller studies of other invasive studies report higher rates (1).
The majority of SCCs are likely ischemic infarcts, for which there is Food and Drug Administration–approved therapy. Intravenous thrombolytic therapy with recombinant tissue plasminogen activator (rt-PA) within 3 h of ischemic stroke symptom onset has been shown to be efficacious and cost-effective in the general stroke population (2,3). Unfortunately, this proven therapy is provided to only a small fraction of all ischemic strokes, primarily owing to delayed presentation to emergency departments (4). In addition, intra-arterial thrombolytic drugs and devices might be effective at later time windows (5,6). The SCCs typically occur in hospitalized patients under close observation and therefore offer the potential for early and rapid detection and treatment. However, there is debate about the safety and efficacy of thrombolysis in the peri- and post-cardiac catheterization setting (7,8).
No study has yet attempted to systematically collect data on all SCCs and characterize their treatment in either a single or multicenter cohort. Therefore, we designed a retrospective cohort study of consecutive SCCs managed at major academic stroke centers with organized stroke teams, and we compared outcomes of patients treated with thrombolysis to those without treatment. We hypothesized that (1) intravenous and intra-arterial thrombolytic therapy for strokes in the cardiac catheterization setting would be relatively safe, with complication rates comparable to SCCs not receiving thrombolytic therapy, and (2) patients treated with thrombolytic therapy would have better outcomes than those not treated with thrombolysis.
We invited 15 academic centers with stroke teams to contribute, and 7 centers (National Institutes of Health Stroke Center, New York Presbyterian Hospital-Weill, University of Calgary, University of California Los Angeles, University of Cincinnati, University of Iowa, University of Pennsylvania) elected to participate. A designated principal investigator at each participating center identified potential patients by reviewing all medical records or their existing stroke databases with both a stroke diagnosis (International Classification of Diseases of the World Health Organization-9th edition [ICD-9] codes: 997.02, 436, 433.0–433.3, 433.8–434.1, and 434.9) and a cardiac procedure (ICD-9 codes: 36.01–337.21–37.23, 37.26, 37.27, 37.34, 88.52–88.57) during the same admission. Cardiac procedures included left heart catheterization, coronary angiography, coronary angioplasty or stenting, left ventriculography, valvuloplasty, patent foramen ovale/atrial septal defect (PFO/ASD) closure, and electrophysiologic diagnostic and ablative studies. Seven centers collected cases from September 2000 to September 2003, and 1 center collected data from September 2000 to September 2004. At each center, all consecutive ischemic strokes within 36 h of cardiac catheterization were included, regardless of the indication for cardiac catheterization, post-stroke treatment decision, or outcome. A clinician collected the following de-identified information on standardized case report forms: demographic information, medical history, cardiac catheterization technical details, post-stroke course, stroke localization and etiology, complications, and clinical outcome. National Institutes of Health Stroke Scale (NIHSS) scores were determined with data in the medical records, which is considered a highly reliable approach (9). This protocol received institutional review board approval at each institution under expedited or exempted mechanisms. Data were entered into spreadsheets with double-entry.
The pre-defined primary efficacy outcome measure was the change in NIHSS score from baseline to 24 h in the rt-PA group, compared with non–rt-PA group. Categorical analyses were also performed, comparing the proportions of patients with complete resolution (NIHSS = 0) or ≥5-point improvement from baseline to 24 h in both groups. We also planned secondary analyses excluding the mildest stroke cases (NIHSS <5), in anticipation of these patients being excluded from thrombolytics (10). Additional pre-specified secondary outcome measures were 7- and 30-day NIHSS scores, discharge modified Rankin Scores (mRS), and death during hospital stay.
The pre-defined primary safety measures were the rate of significant bleeding events and the rate of mortality during the hospital stay, comparing the rt-PA to the non–rt-PA group. Data on all bleeding events were collected, including puncture site hemorrhage, retroperitoneal hemorrhage, hemopericardium, and symptomatic and asymptomatic intracerebral hemorrhage (ICH) within 48 h of stroke. Significant bleeding events were defined as those requiring blood transfusion or causing hemodynamic instability and those specifically consisting of hemopericardium or symptomatic ICH. The ICHs were defined as “symptomatic” if the patient had any symptoms attributed to a new ICH by the local principal investigator. All transfusions given during the hospital stay were recorded on case report forms to ensure that bleeding events were not missed due to lack of documentation.
Patient characteristics were compared with unpaired t tests (age) and Wilcoxon rank sum test for variables that were not normally distributed (NIHSS and time factors). Chi-square or Fisher exact tests were used for categorical variables, depending on individual cell sizes.
Power calculations were difficult a priori. Assuming alpha = 0.05, power = 80%, and planned convenience sample of approximately 100 patients, we estimated an ability to detect a 5-point difference in NIHSS (primary efficacy outcome) between rt-PA cases and control subjects if only 9 cases were treated with rt-PA. A smaller total would be needed if more patients received thrombolytics.
A total of 66 SCCs were identified; 12 (18%) were treated with thrombolysis: 7 with intravenous rt-PA, and 5 with intra-arterial rt-PA. Age, medical comorbidities, and cardiac procedure characteristics were similar between treatment and nontreatment groups (Table 1).
Comparing the treatment and nontreatment groups, rates of heparin use (92% vs. 69%; p = 0.31) and glycoprotein (GP) IIb/IIIa inhibitor use (25% vs. 20%; p = 0.49) were similar. One case in the no rt-PA group received bivalirudin during the cardiac catheterization procedure. In 55 patients whose activated partial thromboplastin times (aPTTs) were recorded before the diagnosis or treatment of SCC, 19 (35%) had a prolonged aPTT (>40 s), including 2 of 7 (29%) treated IV, 1 of 5 (20%) treated IA, and 16 of 43 (37%) not treated with thrombolytics (p = 0.89). In 58 patients whose international normalized ratios (INRs) were recorded before diagnosis or treatment of SCC, 3 (5%) had an abnormal INR (>1.7), including none treated with rt-PA and 3 of 46 (7%) not treated with thrombolytics (p = 1.00). Activated clotting time data were not recorded.
As expected, patients who were treated with thrombolysis had more severe strokes (p = 0.01). Median baseline NIHSS scores were 10.5 in the rt-PA group and 6 in the non–rt-PA group.
The median time from symptom onset to treatment was 142 (range 54 to 305) min in the treated group, including 90 (range 75 to 168) min in the intravenous group and 240 (range 54 to 305) min in the intra-arterial group. Three rt-PA cases received additional stroke therapies, consisting of intra-arterial clot retrieval with a microsnare device, intra-arterial abciximab, and a possible neuroprotective agent (in the context of a trial).
Efficacy results are summarized in Table 2. The median change in NIHSS score from baseline to 24 h, the primary outcome measure of the study, was 6 in the rt-PA group and 0 in the non–rt-PA group (p < 0.001). Either complete resolution (NIHSS = 0) or ≥5-point improvement in NIHSS score over the first 24 h was seen in 58% of treated patients and 15% of untreated patients (risk ratio [RR] = 3.9, 95% confidence interval [CI] 1.8 to 8.8; p = 0.003). The median change in NIHSS from baseline to 7 days was 6.5 in the rt-PA group and 1.5 in the untreated group (p = 0.004).
In secondary analyses excluding the mildest strokes (baseline NIHSS <5), a relative contraindication to rt-PA and a key determinant of clinical outcome (11), the initial difference in baseline NIHSS scores was no longer significant (p = 0.46) and the change in NIHSS score from baseline to 24 h was still significantly greater with thrombolysis (median 6 vs. 1; p = 0.008). In this subset, there was either complete resolution (NIHSS = 0) or ≥5-point improvement over the first 24 h in 58% of treated patients and 21% of untreated patients (RR = 2.8, 95% CI 1.2 to 6.2; p = 0.017). The change in NIHSS score from baseline to 7 days was marginally better with treatment (median 6.5 vs. 3; p = 0.07).
Rates of discharge mRS 0 to 1 (30% rt-PA vs. 28% non–rt-PA; p = 0.72) were not significantly different between the 2 groups. Analysis of change in NIHSS over 30 days was not performed, owing to few 30-day data points.
The primary safety end points did not differ between groups. Specifically, mortality rates were 8% in the rt-PA and 7% in the non–rt-PA groups (p = 0.91), including 1 neurologic death due to cerebral herniation in each group (8% vs. 2%; p = 0.33). There were no symptomatic ICHs. There were 6 asymptomatic hemorrhages, including 3 in the rt-PA group (Table 3) and 3 in the untreated group (25% vs. 6%; p = 0.037). Minor puncture site bleeding rates were not significantly different (8% rt-PA vs. 7% untreated; p = 0.92). There were no retroperitoneal hemorrhages, hemopericardium, other sites of bleeding, or transfusions. Three patients in the rt-PA group had an elevated aPTT after the cardiac procedure, but none had a hemorrhagic complication.
Thrombolysis has been established as an effective treatment for acute ischemic stroke, but its benefit and risk for ischemic strokes specifically in the cardiac catheterization setting have not previously been investigated (7). Bleeding risks of thrombolysis might be increased after cardiac catheterization, owing to the recent invasive procedure and the concurrent administration of antithrombotic medications. Moreover, thrombi leading to SCCs arguably might be less likely than other strokes to respond to thrombolysis on the basis of presumptions about their likely composition. For example, dislodged aortic arch plaque might be calcified and fibrin-dense and therefore not amenable to lysis. Alternatively, calcific thrombi might partially lyse, leading to downstream emboli (12). In addition, air emboli and metallic emboli might not respond to thrombolysis at all (1,7,13,14).
In contrast, fresh thrombus on the tip of a guidewire might be easier to destabilize than other clot etiologies. Limited evidence suggests that the broad category of cardioembolic etiologies might be particularly amenable to thrombolysis (15). In addition, regardless of etiology, superimposed fresh fibrin deposition on atheroma or other components might make these emboli at least partially responsive to thrombolysis. Autopsy studies of coronary artery bypass graft patients with strokes suggest that larger clots causing territorial infarcts tend to be particularly rich in platelets and fibrin. However, specific data on the thrombus composition of ischemic SCCs are lacking. Thrombolysis has been shown to be equally effective across the major stroke subtypes (small-vessel, large-vessel, and cardioembolic) (2), and therefore the pathogenesis of SCCs might be irrelevant with respect to this treatment decision.
Case reports and small series consisted of 39 SCCs reported in the published data as of April 2007, including 1 patient treated with intravenous thrombolysis alone, 2 patients receiving combined intravenous/intra-arterial rt-PA, and the rest treated via the intra-arterial approach (1,8,16–19). Independent function or full recovery was achieved in 23 (61%) of the 38 cases commenting on functional outcome. These reports must be interpreted with great caution, owing to likely publication bias toward positive clinical outcomes. We report the first systematically collected data on every consecutive SCC and its treatment in either a single or multicenter cohort, thereby possibly minimizing these biases.
In our cohort, 18% were treated with thrombolytics, a far greater proportion than in the general stroke population (20). Moreover, patients were treated at a median door-to-needle time of 90 min with intravenous rt-PA and 240 min with intra-arterial therapy, suggesting comparable or better speed than major thrombolysis trials (5,21). Our primary analysis showed a statistically and clinically significant early improvement in stroke severity in NIHSS score by 24 h. Change in NIHSS score at 24 h was 1 of the most powerful discriminators of the rt-PA effect in the NINDS (National Institute of Neurological Disorders and Stroke) trial (22,23) and is likely to be a good surrogate marker for better long-term clinical outcome (24). Similar improvements were observed from baseline to 7 days.
There were no symptomatic ICHs in our treated cases, as compared with 6.4% in the only randomized trial of intravenous rt-PA (NINDS Part 1 and 2) (2) and 10% in the only randomized trial of intra-arterial thrombolysis (PROACT II) (5). A series of intra-arterial rt-PA SCC cases at a single center (n = 22) reported 3 (14%) symptomatic ICHs (19), comparable to the PROACT II rate. Our cohort had a 25% asymptomatic hemorrhage rate, also similar to the rates (26% to 42%) seen in major intra-arterial trials (25).
There were also no significant systemic bleeding events and no increase in even minor puncture site bleeding rates, in the rt-PA group. This is consistent with the substantial published experience with intra-arterial stroke therapies requiring groin site arterial puncture (5,26) and case series studies of post-surgical thrombolysis (27,28). Our sample of patients treated with thrombolysis is too small to estimate bleeding risks with precision but, in combination with prior reports, suggests that this risk is low. We also cannot address specific subgroups known to have higher symptomatic ICH risk, such as elderly patients, although 2 large observational studies of ischemic stroke patients suggest that elderly patients benefit from thrombolysis despite increased ICH risk (1).
Antithrombotic medications administered during the cardiac catheterization procedure might influence the decision to treat with thrombolytics. Heparin, if associated with a prolonged aPTT, is a contraindication for intravenous rt-PA in the standard protocol (2). However, heparin is routinely administered in intra-arterial thrombolysis cases (5,26). Therefore, intra-arterial rt-PA or mechanical revascularization might be a consideration in this situation, depending on the dose and timing of heparin received. Furthermore, the risks of rt-PA for stroke in the presence of other antithrombotic agents commonly used in cardiac interventional procedures, such as GP IIb/IIIa inhibitors, is not known. The low intracranial and systemic hemorrhage rates in our cohort suggest that periprocedural use of these antithrombotic agents does not absolutely militate against thrombolysis. Ongoing stroke trials using rt-PA combined with GP IIb/IIIa inhibitors might further define these risks.
Our findings suggest that thrombolysis might improve early outcome of SCC and seems relatively safe in this context. This study has the limitations inherent in its retrospective design and modest sample size. For example, we could not control for all potential baseline differences between treated and untreated groups in this observational cohort. However, this systematically collected, multicenter cohort of consecutive SCC patients provides compelling data to suggest that post-catheterization stroke patients might be opportune candidates for acute intervention, and emergent cerebral revascularization should be a routine consideration. Future large, prospective registries to further characterize this population are recommended.
The authors thank: at the University of Cincinnati, Edward C. Jauch, MD (administrative, technical, or material support), Kathleen Alwell, RN (administrative, technical, or material support), and Julie Brock (administrative, technical, or material support); at the University of Calgary, Andrew Demchuk and Michael D. Hill (administrative, technical, or material support); and at New York Presbyterian Hospital-Weill Cornell Medical Center, Alan Segal (administrative, technical, or material support) and Kevin Carey (administrative, technical, or material support).
This work was supported by the University of Pennsylvania Research Foundation. Dr. Kasner is supported by National Institutes of Health/National Institute of Neurological Disorders and Stroke (NIH/NINDS) K23 NS02147 and the Inverso-Baglivo Foundation, and Dr. Khatra is supported by NIH/NINDS P50 NS44283.
- Abbreviations and Acronyms
- computed tomography
- intracerebral hemorrhage
- National Institutes of Health Stroke Scale
- recombinant tissue plasminogen activator
- stroke after cardiac catheterization
- Received May 29, 2007.
- Revision received August 14, 2007.
- Accepted September 23, 2007.
- American College of Cardiology Foundation
- Demaerschalk B.M.,
- Yip T.R.
- Kleindorfer D.,
- Kissela B.,
- Schneider A.,
- et al.,
- Neuroscience Institute
- Smith W.S.,
- Sung G.,
- Starkman S.,
- et al.,
- MERCI Trial Investigators
- Kasner S.E.,
- Cucchiara B.L.,
- McGarvey M.L.,
- Luciano J.M.,
- Liebeskind D.S.,
- Chalela J.A.
- Kleindorfer D.O.,
- Khatri P.,
- Katzan I.
- Johnston K.C.,
- Connors A.F. Jr..,
- Wagner D.P.,
- Haley E.C. Jr.
- Molina C.A.,
- Montaner J.,
- Arenillas J.F.,
- Ribo M.,
- Rubiera M.,
- Alvarez-Sabin J.
- Horowitz M.,
- Jovin T.,
- Levy E.,
- Anderson W.
- Presbitero P.,
- Gasparini G.L.,
- Pagnotta P.
- ↵Kleindorfer D, Lindsell C, Brass L, Koroshetz W, Broderick JP. National US estimates of rt-PA use: ICD-9 codes substantially underestimate. Stroke 2008. In press.
- Hacke W.,
- Donnan G.,
- Fieschi C.,
- et al.,
- ATLANTIS Trial Investigators,
- ECASS Trial Investigators,
- NINDS rt-PA Study Group Investigators
- Broderick J.P.,
- Lu M.,
- Kothari R.,
- et al.
- Bruno A.,
- Saha C.,
- Williams L.S.
- Khatri P.,
- Wechsler L.R.,
- Broderick J.P.
- IMS Study Investigators
- Chalela J.A.,
- Katzan I.,
- Liebeskind D.S.,
- et al.