This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Sloan, M. A. Articles by Califf, R. M. Search for Related Content PubMed PubMed Citation Articles by Sloan, M. A. Articles by Califf, R. M. Pubmed/NCBI databases Substance via MeSH Hazardous Substances DB HEPARIN

(Circulation. 1998;98:1376-1382.)
© 1998 American Heart Association, Inc.

Clinical Investigation and Reports

Prediction of 30-Day Mortality Among Patients With Thrombolysis-Related Intracranial Hemorrhage

Michael A. Sloan, MD; Cathy A. Sila, MD; Kenneth W. Mahaffey, MD; Christopher B. Granger, MD; W. T. Longstreth, Jr, MD, MPH; Peter Koudstaal, MD; Harvey D. White, MB, DSc; Joel M. Gore, MD; Maarten L. Simoons, MD; W. Douglas Weaver, MD; Cindy L. Green, MS; Eric J. Topol, MD; Robert M. Califf, MD; ; for the GUSTO-I Investigators

From the University of Maryland Medical System, Baltimore (M.A.S.); the Cleveland Clinic Foundation, Cleveland, Ohio (C.A.S., E.J.T.); Duke Clinical Research Institute, Durham, North Carolina (K.W.M., C.B.G., C.L.G., R.M.C.); University of Washington, Seattle (W.T.L., W.D.W.); Erasmus Universiteit, Rotterdam, the Netherlands (P.K., M.L.S.); Green Lane Hospital, Auckland, New Zealand (H.D.W.); and the University of Massachusetts Medical Center, Worcester (J.M.G.).

Correspondence to Michael A. Sloan, MD, Department of Neurosciences, Harbin Clinic, 1825 Martha Berry Blvd, Rome, GA 30165-1698.

	Abstract

Top Abstract Introduction Methods Results Discussion References

 
Background—Limited information exists on risk factors for mortality after thrombolysis-related intracranial hemorrhage. We wished to determine the characteristics associated with 30-day mortality after thrombolysis-related intracranial hemorrhage.

Methods and Results—We performed an observational analysis within a randomized trial of 4 thrombolytic therapies, conducted in 1081 hospitals in 15 countries. Patients presented with ST-segment elevation within 6 hours of symptom onset. Our population was composed of the 268 patients who had primary intracranial hemorrhage after thrombolysis. With univariable and multivariable analyses, we identified clinical and brain imaging characteristics that would predict 30-day mortality among these patients. CT or MRI were available for 240 patients (90%). The 30-day mortality rate was 59.7%. Glasgow Coma Scale score, age, time from thrombolysis to symptoms of intracranial hemorrhage, hydrocephalus, herniation, mass effect, intraventricular extension, and volume and location of intracranial hemorrhage were significant univariable predictors. Multivariable analysis of 170 patients with complete data, 98 of whom died, identified the following independent, significant predictors: Glasgow Coma Scale score (², 19.3; P<0.001), time from thrombolysis to intracranial hemorrhage (², 15.8; P<0.001), volume of intracranial hemorrhage (², 11.6; P<0.001), and baseline clinical predictors of mortality in the overall GUSTO-I trial (², 10.3; P=0.001). The final model had a C-index of 0.931.

Conclusions—This model provides excellent discrimination between patients who are likely to live and those who are likely to die after thrombolytic-related intracranial hemorrhage; this may aid in making decisions about the appropriate level of care for such patients.

Key Words: thrombolysis • hemorrhage • mortality • prognosis • myocardial infarction

	Introduction

Top Abstract Introduction Methods Results Discussion References

 
Intracranial hemorrhage is the most feared complication of thrombolytic therapy for acute myocardial infarction (AMI). Its frequency in large trials varies from 0.22% to 0.70%, depending in part on the thrombolytic agent used.^{1 2 3 4 5 6 7 8 9 10 11 12 13 14 15} Several of these trials have shown mortality rates from this complication of 48% to 74%.^{1 2 3 4 5 7 8 9 10 11 12 13 14 15} Higher doses of the thrombolytic agent,^{12 13} the type of thrombolytic agent used,⁴ a larger volume of intracerebral hemorrhage, significant mass effect with midline shift, and multiple bleeding sites¹⁰ have been associated with increased mortality after intracranial hemorrhage in past thrombolytic trials, but such analyses have included relatively few patients with this complication.

In the Global Utilization of Streptokinase and t-PA (alteplase) for Occluded Coronary Arteries (GUSTO-I) trial,¹⁴ 268 patients had intracranial hemorrhage.¹⁵ The baseline clinical predictors of 30-day mortality in the entire population have been reported.¹⁶ We wished to develop a model that also incorporated neurological and brain imaging variables to better predict 30-day mortality among patients who experience intracranial hemorrhage after thrombolysis for AMI.

	Methods

Top Abstract Introduction Methods Results Discussion References

 
The GUSTO-I study included 41 021 patients AMI who presented with ST-segment elevation within 6 hours of symptom onset from 1081 hospitals in 15 countries. Descriptions of the entire population, thrombolytic and adjunctive treatments, end points, data management and quality assurance, reporting, and classification of cerebrovascular events have been reported.^{14 15}

Classification of Intracranial Hemorrhage
Criteria for parenchymal intracerebral hemorrhage, subdural hematoma, intraventricular hemorrhage, and subarachnoid hemorrhage have been published.^{6 12 13 15 17} If multiple lesions or lesion types were present, the most important lesion, clinically speaking, was the largest or the one that best explained the patient's neurological state. The features of each hemorrhagic lesion on brain images were classified centrally by 2 investigators (C.A.S., M.A.S.), as described.¹⁷

Volume and Extent of Lesions
Each lesion was identified by brain imaging and measured by investigators blinded to patient treatment and outcome.¹⁷ A pencil line was drawn around the margin of the lesion on all available brain image slices. For parenchymal intracerebral hemorrhages, 2 methods of volume measurement were used: the modified (greatest vertical height x greatest anterior-posterior extent x greatest medial-lateral extent)/2 method^{13 18 19 20 21 22} or the 3-dimensional stereotactic/planimetric method (proprietary software run on a Sun Microsystems SparcStation), performed by the Center for Computer Assisted Neurosurgery, Department of Neurosurgery, Cleveland Clinic Foundation. For subdural hematomas, volumes were calculated on the basis of a further modification of the first method shown above.^{18 19 20 21 22} The greatest vertical height was calculated as for the intracerebral hemorrhages. The maximum anterior-posterior lesion length was estimated by drawing a straight line between the anterior and posterior borders of the lesion. The medial-lateral extent of the lesion was estimated from the slice portion that showed the greatest distance between the cortical surface and the inner table of the skull.²³ We used the volumes calculated by the 3-dimensional method for all analyses reported here.

Grading of intraventricular hemorrhage and hydrocephalus has been described.¹⁷ The presence of subarachnoid hemorrhage was graded semiquantitatively according to the method of Hijdra et al.²³ Brain edema was defined as the low-density region contiguous to the hematoma. The volume of edema was measured by the 3-dimensional volumetric method—drawing a pencil line around the border of the edema, measuring the volumes of the hematoma and edema, and subtracting hematoma volume from the total. Mass effect was graded according to the degree of ventricular compression and cisternal effacement (ambient cistern, quadrigeminal cistern). The presence and direction of any herniation were recorded: subfalcial, downward, or upward.

Clinical Features
Two independent investigators (M.A.S., K.W.M.) retrospectively determined the Glasgow Coma Scale score at the initial evaluation of 202 of the 268 patients with intracranial hemorrhage (75.4%), using all available clinical information. Patients described as being "in coma" or "unresponsive" and who had no additional discriminating information were assigned a score of 7. If seizures were associated with the clinical presentation, the best estimated Glasgow Coma Scale score before the seizure or after complete recovery from the postictal state was used. Agreement between the 2 evaluators was excellent; for the 174 patients classified by both, the median (25th, 75th percentile) absolute difference in scores was 0 (0,1), and the mean± SD absolute difference was 1.0±1.8.

Statistical Analysis
Baseline characteristics were summarized as frequencies and percentages for categorical data and as medians and interquartile ranges for continuous variables. ² and Fisher's exact tests were used to examine relations between categorical variables, whereas the Wilcoxon rank-sum test was used for continuous variables.

A multivariable logistic regression model²⁴ examined individual and joint relations between baseline clinical predictors of mortality in the GUSTO-I population,¹⁶ neurological and neuroimaging features, and death within 30 days of randomization for the 268 patients with intracranial hemorrhage during hospitalization. These baseline variables were examined graphically and statistically to assess the assumption that they related linearly to the log-odds of the outcome event (30-day mortality). Adjustments were made for nonlinear relations. We also tested for interactions among the significant variables. Complete data existed for 170 patients (63.4%); these data were used to construct the multivariable outcome prediction model.

Odds ratios and 95% CI for each variable in the full model were computed. For this analysis, total volume was truncated at a lower limit of 50 mL (the 25th percentile) to meet linearity assumptions. The predictive performance of the model was validated with bootstrapping techniques. The model was refitted for each of 50 bootstrap samples (which included samples of the same size as the original population, drawn randomly with replacement from the original sample), then tested on the original sample. This estimates the predictive accuracy of the model if applied to an independent sample of patients.^{25 26 27 28} The measure of the model's predictive discrimination was the area under the receiver-operator characteristic curve (the C-index). This curve shows all possible pairwise sensitivity and specificity values for the regression model. The C-index measures the concordance of predictions with actual outcomes^{29 30} ; an area of 1.00 represents a model with perfect discrimination. Calibration of the model was assessed by comparing the average model prediction with the observed mortality rate across deciles of risk.¹⁶ A nomogram was constructed using the coefficients of a simplified version of the regression model to predict the individual risk of mortality at 30 days in patients with intracranial hemorrhage after thrombolysis.³¹

	Results

Top Abstract Introduction Methods Results Discussion References

 
In 240 patients with intracranial hemorrhage, the most prominent lesion was a parenchymal intracerebral hemorrhage in 197 patients (82.1%), a subdural hematoma in 7 patients (2.9%), and both in 36 patients (15%). The parenchymal intracerebral hemorrhage group included subarachnoid and intraventricular hemorrhages (2 had intraventricular or subarachnoid hemorrhages only). The overall mortality rate from intracranial hemorrhage was 59.7%. Patients who died were older and had intracranial hemorrhage sooner after thrombolysis than those who survived (Tables 1 and 2). The median (25th, 75th percentiles) time from treatment to death for the 160 patients with intracranial hemorrhage who died was 54.4 hours (24, 134). The median time from stroke symptom onset to death for these patients was 40.5 hours (12, 116).

View this table: [in this window] [in a new window]   Table 1. Baseline Clinical Characteristics

View this table: [in this window] [in a new window]   Table 2. Neurological Event-Related Characteristics

Patients with intracranial hemorrhage were more likely to die within 30 days if they had a lower Glasgow Coma Scale score, a combination parenchymal-subdural hematoma, multiple parenchymal intracerebral hemorrhages, or a deep parenchymal intracerebral hemorrhage (especially within the putamen/internal capsule and brain stem; Table 2). For other neuroimaging characteristics, the presence of mass effect, herniation, hydrocephalus, and intraventricular hemorrhage were all significantly more common in patients who died. Intraventricular hemorrhage was an independent predictor of 30-day mortality, but there was no relation between intraventricular hemorrhage volume and 30-day mortality (median, 14 mL; interquartile range 4 to 36 mL for patients who died versus 11 mL [5 to 23 mL] for survivors, P=0.56).

Univariable and Multivariable Predictors of Mortality
In univariable analysis, a lower Glasgow Coma Scale score was most strongly associated with 30-day mortality, followed by shorter time from thrombolysis to hemorrhage; larger total hemorrhagic volume; and the presence of hydrocephalus, herniation, or mass effect (Table 3). The baseline clinical predictors of 30-day mortality in the overall GUSTO-I population and location of hemorrhage also were significantly associated with 30-day mortality.

View this table: [in this window] [in a new window]   Table 3. Univariable Predictors of Mortality After Thrombolysis-Related Intracranial Hemorrhage

The most significant independent predictors of 30-day mortality (in decreasing order of importance) were time from thrombolysis to hemorrhage, Glasgow Coma Scale score, total hemorrhagic volume, and the GUSTO-I clinical predictors (Table 4). Because these 4 variables accounted for 93% of the total prognostic information, we developed a reduced version of this model (Table 5). Because age was by far the most predictive variable in the overall GUSTO-I model,¹⁶ a simpler model that substituted age alone for the combined "GUSTO-I clinical predictors" yielded similar results (Table 5).

View this table: [in this window] [in a new window]   Table 4. Multivariable Predictors of Mortality After Thrombolysis-Related Intracranial Hemorrhage

View this table: [in this window] [in a new window]   Table 5. Multivariable Predictors of Mortality After Thrombolysis-Related Intracranial Hemorrhage

The coefficients from the simplified model were used to create a nomogram to predict 30-day mortality in individual patients (Figure 1). For an 80-year-old with a Glasgow Coma Scale score of 5 due to an 80-mL intracranial hemorrhage occurring 20 hours after thrombolysis, the estimated probability of 30-day mortality would be 14+16+93+5=128, or 90%. For a 60-year-old with a Glasgow Coma Scale score of 12 due to a 30-mL intracranial hemorrhage occurring 40 hours after thrombolysis, the estimated probability of 30-day mortality is 7+5+87+0=99, or 7.5%.

View larger version (39K): [in this window] [in a new window]   Figure 1. Nomogram to predict individual risk of mortality at 30 days in patients with intracranial hemorrhage after thrombolysis, using the simplified model given in Table 5. The C-index for the model is 0.923.

Figure 2 shows the calibration (reliability) of the full model (Table 4). This figure shows the observed 30-day mortality rates across deciles of predicted risk versus the predicted mortality rates of the full multivariable model. The validation of the full model with bootstrapping techniques showed that the C-index was overoptimistic by 0.01, yielding a bias-corrected C-index of 0.93.

View larger version (9K): [in this window] [in a new window]   Figure 2. Graph showing observed 30-day mortality after thrombolysis- related intracranial hemorrhage versus mortality predicted by the full multivariable logistic regression model (Table 4). Patients were grouped into deciles according to their predicted probability of death. The actual mortality rate for the patients within each decile (dots) is plotted against the average model prediction. The solid line reflects perfect calibration of the model predictors.

	Discussion

Top Abstract Introduction Methods Results Discussion References

 
We anticipated that several factors would be highly predictive of 30-day mortality among patients with intracranial hemorrhage after thrombolysis, such as Glasgow Coma Scale score, total intracranial hemorrhage volume, and age. We also discovered a powerful new factor: time from thrombolytic treatment to onset of symptoms of intracranial hemorrhage. With a model that incorporates these and other variables, mortality among these patients can be predicted with a high degree of accuracy.

Mortality associated with spontaneous parenchymal intracerebral hemorrhage has been found to vary from 83% for pontine to 16% for subcortical lobar sites.^{32 33 34} Independent predictors of mortality from spontaneous supratentorial intracerebral hemorrhage have included hemorrhage volume,^{19 20 35 36 37 38} low Glasgow Coma Scale score,^{19 20 35 36 37 38} pulse pressure,^{35 37 38} intraventricular hemorrhage,³⁵ and an interaction term between intraventricular hemorrhage and Glasgow Coma Scale score.³⁵ The reported sensitivity, specificity, and accuracy of these models is 62% to 97%, 87% to 97%, and 92%, respectively.^{19 20 35 37 38} These models allow development of highly accurate algorithms to predict 30-day mortality in patients with supratentorial parenchymal hemorrhage, which can aid in the selection of appropriate patients for aggressive medical or surgical treatment.^{19 20 35 37 38}

Predictors of increased mortality 30 days after intracranial hemorrhage in GUSTO-I are similar to those of spontaneous supratentorial parenchymal hemorrhage and oral anticoagulant therapy–related intracerebral hemorrhage. Univariable analyses show significant relations between 30-day mortality and GUSTO-I clinical predictors,¹⁶ Glasgow Coma Scale score, hematoma location, hematoma volume, time from thrombolysis to intracranial hemorrhage onset, hydrocephalus, herniation, mass effect, and intraventricular hemorrhage. In fact, the combined parenchymal-subdural lesion was the most ominous, being fatal in 28 of 36 patients (78%). Our results thus confirm and extend the results from the first 3 Thrombolysis and Angioplasty in Myocardial Infarction (TAMI) trials and the Duke Databank for Cardiovascular Disease.¹⁰

In multivariable analysis, Glasgow Coma Scale score and total hemorrhage volume were independent predictors of 30-day mortality, similar to results reported for spontaneous supratentorial parenchymal hemorrhage.^{19 20 35 37 38} In addition, a shorter time from thrombolysis to intracranial hemorrhage onset was a significant, independent predictor of 30-day mortality. Our data suggest some correlation between shorter time from thrombolysis to intracranial hemorrhage onset and intracranial hemorrhage volume. This implies a temporal link between the biological effects of thrombolysis, clinical concomitants of intracranial hemorrhage, and mortality. Alternatively, mild intracranial hemorrhages may not have been recognized as quickly as were severe ones. Although intracranial hemorrhage volumes were significantly larger in patients treated with combined thrombolytic therapy,¹⁷ this factor was not an independent predictor of mortality. We also found that the baseline clinical predictors of 30-day mortality in the overall GUSTO-I population, especially age, were independently associated with 30-day mortality from intracranial hemorrhage. These 4 factors accounted for 93.1% of the full model's discriminative power. In the simplified model, the larger overall sample size of GUSTO-I and the effects of Glasgow Coma Scale score and hemorrhage volume may have diminished the strong general effect of age on mortality after infarction.^{39 40}

Our findings have implications for the management of intracranial hemorrhage after thrombolysis. It is not known whether patients with severe intracranial hemorrhage soon after thrombolytic therapy should receive only supportive medical care⁴¹ or should be aggressively managed (treatment of increased intracranial pressure, ventriculostomy, neurosurgical evacuation). For now, the use of Figure 1 may aid in the clinician's empirical decision-making for such patients.

This study has several limitations. First, the models were derived from a retrospective analysis. Second, the use of best-estimated Glasgow Coma Scale scores may have introduced some misclassification bias. The arbitrary assignment of a Glasgow Coma Scale score of 7 to patients with "coma" or "unresponsiveness" may be considered too conservative, however, thus underestimating its contribution to the model. In addition, the excellent agreement between a cardiologist and a neurologist in independently assigning Glasgow Coma Scale scores suggests that misclassification bias may have been minimized. However, prospective determination of Glasgow Coma Scale scores would have been better. Third, missing data, particularly for Glasgow Coma Scale scores, may have limited the power of the analysis and introduced bias, although the sample size of our multivariable model (n=170) compares favorably with those of recent studies of mortality prediction for spontaneous supratentorial parenchymal hemorrhage.^{19 35 37 38} Finally, the treating clinician's perception of poor outcome from intracranial hemorrhage may have influenced the decision to provide aggressive support for these patients. As a result, the model reported here may partly reflect the clinician's behavior in treating these patients. This model should be validated in a larger, unselected population.

	Acknowledgments

 
This study was funded by Genentech (South San Francisco, Calif), Bayer (New York, NY), CIBA-Corning (Medfield, Mass), ICI Pharmaceuticals (Wilmington, Del), and Sanofi Pharmaceuticals (Paris, France).

Received August 26, 1997; revision received June 10, 1998; accepted June 16, 1998.

	References

Top Abstract Introduction Methods Results Discussion References

 

1. Gruppo Italiano per lo Studio della Streptochinasi nell'Infarto Miocardico (GISSI). Effectiveness of intravenous thrombolytic treatment in acute myocardial infarction. Lancet. 1986;1:397–402.[Medline] [Order article via Infotrieve]
   
2. Gruppo Italiano per lo Studio della Streptochinasi nell'Infarto Miocardico (GISSI). GISSI-2: a factorial randomised trial of alteplase versus streptokinase and heparin versus no heparin among 12 490 patients with acute myocardial infarction. Lancet. 1990;336:65–71.[Medline] [Order article via Infotrieve]
   
3. The International Study Group. In hospital mortality and clinical course of 20 891 patients with suspected acute myocardial infarction randomised between alteplase and streptokinase with or without heparin. Lancet. 1990;336:71–75.[Medline] [Order article via Infotrieve]
   
4. ISIS-3 (Third International Study of Infarct Survival) Collaborative Group. ISIS-3: a randomised comparison of streptokinase vs tissue plasminogen activator vs anistreplase and of aspirin plus heparin vs aspirin alone among 41 299 cases of suspected acute myocardial infarction. Lancet. 1992;339:753–770.[Medline] [Order article via Infotrieve]
   
5. Kase CS, Pessin MS, Zivin JA, del Zoppo GJ, Furlan AJ, Buckley JW, Snipes RG, LittleJohn JK. Intracranial hemorrhage after coronary thrombolysis with tissue plasminogen activator. Am J Med. 1992;92:384–390.[Medline] [Order article via Infotrieve]
   
6. Sloan MA. Thrombolysis and intracranial hemorrhage. In: Feldmann E, ed. Intracerebral Hemorrhage. Armonk, NY: Futura; 1994:99–149.
   
7. Maggioni AP, Franzosi MG, Farina ML, Santoro E, Celani MG, Ricci S, Tognoni G. Cerebrovascular events after myocardial infarction: analysis of the GISSI trial: Gruppo Italiano per lo Studio della Streptochinasi nell'Infarto Miocardico (GISSI). BMJ. 1991;302:1428–1431.
   
8. AIMS Trial Study Group. Effect of intravenous APSAC on mortality after acute myocardial infarction: preliminary report of a placebo-controlled clinical trial. Lancet. 1988;1:545–549.[Medline] [Order article via Infotrieve]
   
9. AIMS Trial Study Group. Long term effects of intravenous anistreplase in acute myocardial infarction: final report of the AIMS study. Lancet. 1990;335:427–431.[Medline] [Order article via Infotrieve]
   
10. Uglietta JP, O'Connor CM, Boyko CB, Aldrich H, Massey EW, Heinz ER. CT patterns of intracranial hemorrhage complicating thrombolytic therapy for acute myocardial infarction. Radiology. 1991;181:555–559.[Abstract/Free Full Text]
    
11. de Jaegere PP, Arnold AA, Balk AH, Simoons ML. Intracranial hemorrhage in association with thrombolytic therapy: incidence and clinical predictive factors. J Am Coll Cardiol. 1992;20:289–294.
    
12. Gore JM, Sloan M, Price TR, Randall AM, Bovill E, Collen D, Forman S, Knatterud GL, Sopko G, Terrin ML, and the TIMI investigators. Intracerebral hemorrhage, cerebral infarction, and subdural hematoma after acute myocardial infarction and thrombolytic therapy in the Thrombolysis In Myocardial Infarction Study: TIMI Phase II, Pilot and Clinical Trial. Circulation. 1991;83:448–459.[Abstract/Free Full Text]
    
13. Sloan MA, Price TR, Petito CK, Randall AMY, Solomon RE, Terrin ML, Gore J, Collen D, Kleiman N, Feit F, Babb J, Herman M, Roberts WC, Sopko G, Bovill E, Forman S, Knatterud GL, for the TIMI Investigators. Clinical features and pathogenesis of intracerebral hemorrhage following rt-PA and heparin therapy for acute myocardial infarction: the Thrombolysis in Myocardial Infarction (TIMI) II Pilot and Randomized Clinical Trial combined experience. Neurology. 1995;45:649–658.[Abstract/Free Full Text]
    
14. The GUSTO Investigators. An international randomized trial comparing four thrombolytic strategies for acute myocardial infarction. N Engl J Med. 1993;329:673–682.[Abstract/Free Full Text]
    
15. Gore JM, Granger CB, Simoons ML, Sloan MA, Weaver WD, White HD, Barbash GI, Van de Werf F, Aylward PE, Topol EJ, Califf RM, for the GUSTO-I Investigators. Stroke after thrombolysis: mortality and functional outcomes in the GUSTO-I trial. Circulation. 1995;92:2811–2818.[Abstract/Free Full Text]
    
16. Lee KL, Woodlief LH, Topol EJ, Weaver WD, Betriu A, Col J, Simoons M, Aylward P, Van de Werf F, Califf RM, for the GUSTO-I Investigators. Predictors of 30-day mortality in the era of reperfusion for acute myocardial infarction. Results from an international trial of 41 021 patients. Circulation. 1995;91:1659–1668.[Abstract/Free Full Text]
    
17. Gebel JM, Sila CA, Sloan MA, Granger CB, Mahaffey KW, Weisenberger J, Green CL, White HD, Gore JM, Weaver WD, Califf RM, Topol EJ, for the GUSTO-I Investigators. Thrombolysis-related intracranial hemorrhage: a radiographic analysis of 244 cases from the GUSTO-1 trial with clinical correlation. Stroke. 1998;29:563–569.[Abstract/Free Full Text]
    
18. Kwak R, Kadoya S, Suzuki T. Factors affecting the prognosis in thalamic hemorrhage. Stroke. 1983;14:493–500.[Abstract/Free Full Text]
    
19. Broderick JP, Brott TG, Duldner JE, Tomsick T, Huster G. Volume of intracerebral hemorrhage: a powerful and easy-to-use predictor of 30 day mortality. Stroke. 1993;24:987–993.[Abstract/Free Full Text]
    
20. Lisk DR, Pasteur W, Rhoades H, Putnam RD, Grotta JC. Early presentation of hemispheric intracerebral hemorrhage: prediction of outcome and guidelines for treatment allocation. Neurology. 1994;44:133–139.[Abstract/Free Full Text]
    
21. Hier DB, Davis KR, Richardson EP, Mohr JP. Hypertensive putaminal hemorrhage. Ann Neurol. 1977;1:152–159.[Medline] [Order article via Infotrieve]
    
22. Kase CS, Williams JP, Wyatt DA, Mohr JP. Lobar intracerebral hematomas: clinical and CT analysis of 22 cases. Neurology. 1982;32:1146–1150.[Abstract/Free Full Text]
    
23. Hijdra A, Brouwers PHAM, Vermeulen M, van Gijn J. Grading the amount of blood on computed tomograms after subarachnoid hemorrhage. Stroke. 1990;21:1156–1161.[Abstract/Free Full Text]
    
24. Hosmer DW, Lemeshow S. Applied Logistic Regression. New York, NY: John Wiley & Sons, Inc; 1989.
    
25. Harrell FE Jr, Pollock BG, Lee KL. Graphical methods for the analysis of survival data. In: Proceedings of the 12th Annual SAS Users Group International Conference. Cary, NC: SAS Institute Inc; 1987:1107–1115.
    
26. Harrell FE Jr, Lee KL, Pollock BG. Regression models in clinical studies: determining relationships between predictors and response. J Natl Cancer Inst. 1988;80:1198–1202.[Abstract/Free Full Text]
    
27. Chambers JM, Hastie JT. Statistical Models in S. Pacific Grove, Calif: Wadsworth & Brooks/Cole; 1992.
    
28. Harrell FR Jr. Transcan. S function for transformation and imputation using canonical variates. Available from http://www.statlib@lib.stat.cmu.edu; 1993. Request "send transcan from S."
    
29. Lusted LB. Decision-making studies in patient management. N Engl J Med. 1971;284:416–424.
    
30. Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology. 1982;143:29–36.[Abstract/Free Full Text]
    
31. Harrell FE Jr. Design: S plus functions for biostatistical/epidemiologic modelling, testing, estimation, validation, graphics, prediction, and typesetting by storing enhanced model design attributes in the fit. UNIX version available from http://www.statlib@lib.stat.cmu.edu; 1996.
    
32. Kase CS, Mohr JP. General features of intracerebral hemorrhage. In: Barnett HJM, Mohr JP, Stein BM, Yatsu FM, eds. Stroke: Pathophysiology, Diagnosis and Management. New York, NY: Churchill-Livingstone; 1986:497–523.
    
33. Caplan LR. Clinical features of spontaneous intracerebral hemorrhage. In: Kaufman HH, ed. Intracerebral Hematomas. New York, NY: Raven Press, Ltd.; 1992:31–47.
    
34. Feldmann E. Intracerebral hemorrhage. Stroke. 1991;22:684–691.
    
35. Tuhrim S, Dambrosia JM, Price TR, Mohr JP, Wolf PA, Hier DB, Kase CS. Intracerebral hemorrhage: external validation and extension of a model for prediction of 30 day survival. Ann Neurol. 1991;29:658–663.[Medline] [Order article via Infotrieve]
    
36. Portenoy RK, Lipton RB, Berger AR, Lesser ML, Lantos G. Intracerebral hemorrhage: a model for the prediction of outcome. J Neurol Neurosurg Psychiatry. 1987;50:976–979.[Abstract]
    
37. Tuhrim S, Dambrosia JM, Price TR, Mohr JP, Wolf PA, Heyman A, Kase CS. Prediction of intracerebral hemorrhage survival. Ann Neurol. 1988;24:258–263.[Medline] [Order article via Infotrieve]
    
38. Tuhrim S, Horowitz DR, Sachen M, Godbold JH. Validation and comparison of models predicting survival following intracerebral hemorrhage. Crit Care Med. 1995;23:950–954.[Medline] [Order article via Infotrieve]
    
39. White HD, Barbash GI, Califf RM, Simes RJ, Granger CB, Weaver WD, Kleiman NS, Aylward PE, Gore JM, Vahanian A, Lee KL, Ross AM, Topol EJ. Age and outcome with contemporary thrombolytic therapy: results from the GUSTO-I trial. Circulation. 1996;94:1826–1833.[Abstract/Free Full Text]
    
40. Maggioni AP, Maseri A, Fresco C, Franzosi MG, Mauri F, Santoro E, Tognoni G, on behalf of the Investigators of the Gruppo Italiano per lo Studio della Sopravvivenza nell'Infarto Miocardico (GISSI-2). Age-related increase in mortality among patients with first myocardial infarctions treated with thrombolysis. N Engl J Med. 1993;329:1442–1448.[Abstract/Free Full Text]
    
41. Sane DC, Califf RM, Topol EJ, Stump DC, Mark DB, Greenberg CS. Bleeding during thrombolytic therapy for acute myocardial infarction: mechanisms and management. Ann Intern Med. 1989;111:1010–1022.
    

This article has been cited by other articles:

				S. Schulman, R. J. Beyth, C. Kearon, and M. N. Levine Hemorrhagic Complications of Anticoagulant and Thrombolytic Treatment: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines (8th Edition) Chest, June 1, 2008; 133(6_suppl): 257S - 298S. [Abstract] [Full Text] [PDF]

				S. G. Goodman, V. Menon, C. P. Cannon, G. Steg, E. M. Ohman, and R. A. Harrington Acute ST-Segment Elevation Myocardial Infarction: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines (8th Edition) Chest, June 1, 2008; 133(6_suppl): 708S - 775S. [Abstract] [Full Text] [PDF]

				V. Menon, R. A. Harrington, J. S. Hochman, C. P. Cannon, S. D. Goodman, R. G. Wilcox, H. J. Schunemann, and E. M. Ohman Thrombolysis and Adjunctive Therapy in Acute Myocardial Infarction: The Seventh ACCP Conference on Antithrombotic and Thrombolytic Therapy Chest, September 1, 2004; 126(3_suppl): 549S - 575S. [Abstract] [Full Text] [PDF]

				E. M. Ohman, R. A. Harrington, C. P. Cannon, G. Agnelli, J. A. Cairns, and J.W. Kennedy Intravenous Thrombolysis in Acute Myocardial Infarction Chest, January 1, 2001; 119(1_suppl): 253S - 277S. [Full Text] [PDF]

				T. G. Phan, M. Koh, R. A. Vierkant, and E. F.M. Wijdicks Hydrocephalus Is a Determinant of Early Mortality in Putaminal Hemorrhage Stroke, September 1, 2000; 31(9): 2157 - 2162. [Abstract] [Full Text] [PDF]

				Predicting Mortality in Patients with ICH After Thrombolysis for AMI Journal Watch Emergency Medicine, December 1, 1998; 1998(1201): 9 - 9. [Full Text]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Sloan, M. A. Articles by Califf, R. M. Search for Related Content PubMed PubMed Citation Articles by Sloan, M. A. Articles by Califf, R. M. Pubmed/NCBI databases Substance via MeSH Hazardous Substances DB HEPARIN