Background The frequency of onset of acute myocardial infarction follows a circadian pattern, with a peak incidence between 6:00 am and noon. Circadian variations have been defined for platelet aggregation, plasminogen-activator inhibitor, and a number of hemostatic and physiological factors, all of which might predispose toward clotting in the late morning and thrombolysis in the evening. Thus, the hypothesis for this retrospective analysis was that tissue-type plasminogen activator (TPA) has greater efficacy when administered between noon and midnight, as measured by coronary patency 90 minutes after initiation of treatment.
Methods and Results Seven hundred twenty-eight patients were enrolled in either of two studies in which TPA was administered under a uniform protocol for the treatment of acute myocardial infarction. Of these, 692 patients had qualifying arteriograms that allowed standardized assessment by a core angiographic laboratory of the primary end point of 90-minute patency. TPA has a circadian pattern of efficacy, with greater TIMI grade 3 patency when administered between noon and midnight (P<.001). When TPA was given within 2 hours of symptoms (n=127), the total patency was highest and there was a trend (P=.055) toward the greatest magnitude difference occurring between am and pm patency. The onset of myocardial infarction was confirmed to have a marked circadian variation with a peak incidence about 10:00 am. The peak efficacy of TPA was about 8:00 pm, representing a phase difference of about 10 hours after peak infarction incidence.
Conclusions There is a circadian variation in the ability of TPA to rapidly open coronary arteries, with highest efficacy between noon and midnight. This complements clinical and in vitro knowledge of increased morning thrombosis and is concordant with knowledge of a fibrinolytic profile that is more favorable for evening lysis. This finding has implications for understanding the circadian pathophysiology of myocardial infarction and for its chronotherapy.
The frequency of onset of acute myocardial infarction follows a circadian pattern, with a peak incidence between 6:00 am and noon.1 2 3 4 Circadian variations that may predispose toward thrombotic occlusion of coronary arteries in the late morning hours have been described for platelet aggregation,5 6 coronary flow,7 viscosity,8 cortisol levels,9 epinephrine levels,6 10 and activated partial thromboplastin time and thrombin time.11 Complementing the observations regarding increased clotting tendencies in the late morning are descriptions of circadian variations in levels of tissue-type plasminogen activator (TPA) and plasminogen-activator inhibitor–1 (PAI-1)12 13 14 and of plasma euglobulin fibrinolytic activity,14 15 all of which suggest a potential for enhanced fibrinolysis in the evening hours.
Knowledge of these circadian patterns of increased coronary thrombosis in the morning and of a hemostatic-fibrinolytic state that might be favorable for enhanced evening fibrinolysis led to the hypothesis regarding a potential circadian pattern in clinical efficacy of TPA. Review of the literature revealed one study that addressed this hypothesis. Becker et al16 found in a group of 28 patients that coronary artery patency 90 minutes after the beginning of treatment with TPA was 27% for treatment between midnight and noon compared with 82% for treatment between noon and midnight (P=.006).
Thus, the hypothesis for this retrospective analysis was that TPA has greater efficacy when administered after noon, as measured by coronary patency 90 minutes after initiation of treatment. A database with which to evaluate this hypothesis was available from a pair of large studies with a uniform protocol that administered TPA at a single weight-adjusted dose, with a primary end point of 90-minute patency evaluation.
The patients for this study were enrolled in either of two protocols17 18 using TPA (Duteplase; Burroughs Wellcome). Both studies had primary end points of perfusion status of the infarct-related artery 90 minutes after initiation of TPA infusion. Patient selection, drug dosing, and protocol requirements were identical between the two studies, except that one of the protocols17 called for a repeat catheterization at 24 to 48 hours to assess continued patency.
Inclusion criteria were coronary pain of more than 30 minutes’ duration, ability to begin TPA infusion within 6 hours of index pain, age of less than 76 years, and diagnostic ST-segment elevation in two contiguous leads (0.1 mV inferior or lateral, 0.2 mV precordial). Exclusion criteria were systolic pressure <80 mm Hg despite treatment with vasopressors, uncontrolled hypertension (diastolic pressure >120 mm Hg), prior cerebrovascular accident, prior coronary artery bypass graft surgery, left bundle-branch block, women of childbearing potential or pregnant, and a variety of exclusions related to increased risks of bleeding.
After screening for inclusion and exclusion criteria was performed, written informed consent was obtained in a manner approved by the institutional review board of each participating institution. Administration of TPA was begun as quickly as possible, usually in the emergency department. A total of 0.6 million International Units (MIU)/kg was given over 4 hours. Of this, 0.4 MIU/kg was given in the first hour, including 0.04 MIU/kg in the first minute. At 60 minutes, the dose was reduced to 0.07 MIU/kg per hour, continued for 3 additional hours. Lidocaine was given with the initial TPA bolus, as a loading dose of 1.0 to 1.5 mg/kg, followed by a 2-mg/min infusion. The patient was taken to the catheterization laboratory and 5000 U heparin IV was given when the sheaths were placed. The primary end point was perfusion status of the infarct-related artery 90 minutes after initiation of TPA infusion. Responders had heparin infusion at 1000 U/h begun at 90 minutes and subsequently adjusted to maintain activated partial thromboplastin time at 1.5 to 2 times control. Aspirin (325 mg/d) plus dipyridamole (75 mg TID) were begun when heparin was discontinued.
Data Collection and Definition
Demographic and clinical data were recorded during the protocol by the investigator and/or research nurse onto data collection forms that were later reviewed and edited in accordance with an established quality assurance program. Time of onset of infarction was defined based on history obtained from the patient and family. Perfusion status of the infarct-related artery at 90 minutes was graded as patent (TIMI 2 or 3 perfusion) or not patent (TIMI 0 or 1 perfusion) by analysis of the cine film by expert blinded reviewers at the core angiographic laboratory at the University of Washington, Seattle. Patients were excluded from this study if cine arteriography obtained at 90 minutes was not available for review by the core laboratory. No other patients were excluded.
The Cochran-Mantel-Haenszel test was used to assess the circadian pattern of patency as a function of time of treatment, controlling for elapsed time to treatment. The Cochran-Mantel-Haenszel rank correlation test was used to assess correlation between more rapid therapy and increased patency. χ2 testing was used to assess the circadian pattern of onset of myocardial infarction and to assess circadian differences in patency within individual elapsed time blocks. Second-order harmonic regression models19 were created for the frequency distribution of onset of myocardial infarction and for patency 90 minutes after TPA administration. These models are appropriate for analysis of data with periodicity, in this case a 24-hour cycle. This type of model was used by the MILIS investigators in their presentation of the circadian pattern of onset of infarction.1 Statistical analysis of the patency data fit to the harmonic regression was done with the F test, controlling for time to treatment. All P values are two-tailed.
The patency data in Figs 1⇓ and 2⇓, with the associated statistical analyses, are based on 12-hour time of treatment blocks demarcated at noon and at midnight. The reason for choosing those demarcation points was that Becker et al16 examined the same phenomenon in a small group of patients and used those time blocks. Inspection of our data as in Fig 3⇓ supports that selection of time blocks.
A total of 728 patients were enrolled in the two protocols analyzed for this study. Patients were entered at 26 sites between August 8, 1987, and August 24, 1989. Of these, 692 patients had qualifying arteriograms that showed standardized assessment of patency at 90 minutes and constituted the basis for this study. The other 36 patients were excluded because no arteriograms were available to the core cine laboratory (n=14) or the arteriograms available to the core laboratory were not obtained at approximately 90 minutes (n=22). The mean age was 56 years (age range, 32 to 76 years), and 592 (86%) were male. The distribution of the infarct-related arteries included the left anterior descending coronary artery (41.1%), right coronary artery (47.4%), left circumflex artery (11.4%), and left main artery (0.1%).
Greater TPA Efficacy With Shorter Time to Treatment
The efficacy of TPA in establishing patency 90 minutes after initiation of therapy is presented in Fig 1⇑. Overall 90-minute patency was 69% (95% confidence interval, 65% to 72%).
Shorter time to initiation of TPA was associated with increased 90-minute patency (P<.009, Cochran-Mantel-Haenszel rank correlation test) with 74% patency for those receiving TPA within 2 hours of onset of symptoms, 71% for TPA begun 2 to 4 hours after onset of symptoms, and 60% for TPA begun more than 4 hours after onset of symptoms.
Circadian Efficacy of TPA
The primary end point of coronary patency was defined at the time of data collection as TIMI 2 or 3 flow, consistent with scientific practice at that time. Current knowledge20 indicates that TIMI 3 patency represents a different physiology than TIMI 2 patency, with reported benefits that include decreased mortality, smaller infarct size, less congestive failure, and less recurrent ischemia. Therefore, TIMI 3 patency appears to be a better index of successful thrombolysis than combined TIMI 2 or 3 patency and is the basis for the primary analysis for this investigation. Patency is shown as a function of time of initiation of TPA, as well as a function of elapsed time between onset of symptoms and initiation of TPA, in Fig 1⇑. There was no evidence of a time of treatment by time to treatment interaction (Breslow-Day χ2=1.71, df=2, P=.43).
TPA has a clear circadian pattern of efficacy, with greater coronary patency when administered between noon and midnight (P<.001, Cochran-Mantel-Haenszel test controlling for time to treatment). TIMI 3 patency was 42% (95% confidence interval, 37% to 47%) for treatment between noon and midnight and 29% (95% confidence interval, 24% to 35%) for treatment between midnight and noon.
As measured by TIMI 2 or 3 patency, there was a trend toward greater efficacy when TPA was administered between noon and midnight (P=.07). Patency was 72% (95% confidence interval, 67% to 76%) for treatment between noon and midnight and 65% (95% confidence interval, 59% to 70%) for treatment between midnight and noon. TIMI 2 or 3 patency demonstrates a clear circadian pattern but without defining specific demarcation points of noon and midnight, as described below by use of a harmonic model (P=.036).
Fig 2⇑ displays the distribution of the 4 TIMI patency grades for midnight to noon and noon to midnight. Analysis of patency as a 4 grade ordinal variable shows a circadian pattern (P=.004). TIMI 3 patency as a fraction of combined TIMI 2 plus TIMI 3 patency is higher after noon (58%) than before noon (45%). The circadian variation is due to the prominent pattern of enhanced TIMI 3 patency from noon to midnight.
The subset of 127 patients who received TPA within 2 hours of symptoms represent an important group. Patency (TIMI 2 or 3) was 81% for treatment initiated between noon and midnight compared with 65% for treatment between midnight and noon (P=.055). Expressed conversely, lack of patency was 19% after noon compared with 35% before noon. This group that received TPA rapidly not only had the greatest patency but also had a strong trend toward the greatest magnitude difference between am and pm efficacy.
Circadian Pattern of Incidence of Infarction Compared With Circadian Pattern of Efficacy of TPA
The times of onset of symptoms of myocardial infarction showed a marked circadian variation as expected (Fig 3A⇑). The peak incidence at about 10:00 am was threefold to fourfold higher than the trough at about 2:00 am. Summarizing by 6-hour intervals, 15.5% of infarctions had onset between midnight and 6:00 am, 34.2% between 6:00 am and noon, 29.3% between noon and 6:00 pm, and 21.0% between 6:00 pm and midnight. The incidence of onset of myocardial infarction between 6:00 am and noon was 1.56 times greater than the average incidence during the other 18 hours. This increase is highly significant (P<<.0001). An excellent fit to the circadian pattern is obtained using a second-order harmonic regression, with R=.93: Infarctions per hour (percent of total)=4.17+0.03 sin(2πt/24)−1.69 cos(2πt/24)−0.64 sin(4πt/24)+0.48 cos(4πt/24), where t is time of day in hours.
The circadian pattern of efficacy of TPA is shown in the bottom panel of Fig 3⇑. Thus, Fig 3A⇑ shows the frequency distribution of an event throughout the day, and Fig 3B⇑ shows the variability of efficacy of an intervention throughout the day. The circadian patterns are out of phase, with peak incidence of infarction at about 10:00 am and the peak efficacy of TPA at about 8:00 pm, giving an estimated phase difference of approximately 10 hours.
TIMI 3 patency as a function of the time treatment is begun, for all 692 patients, was fitted with a second-order harmonic regression as shown in Fig 3B⇑, with R=.67: Patency=36.41−8.75 sin(2πt/24)+4.00 cos(2πt/24)−3.05 sin(4πt/24)+2.25 cos(4πt/24).
Patency defined as TIMI 2 or 3 flow was fitted with a second-order harmonic regression (figure not shown), with R=.65: Patency=68.67−4.45 sin(2πt/24)+0.75 cos(2πt/24)−3.75 sin(4πt/24)−1.03 cos(4πt/24). A harmonic model to look at circadian pattern while controlling for time to treatment has P=.036.
TPA has a clear circadian pattern of efficacy, as measured by the ability to rapidly produce a patent coronary artery. Efficacy is higher from noon to midnight whether measured by the presence of TIMI 3 patency (P<.001) or by analyzing patency as a 4 grade ordinal variable (P=.004). This circadian pattern is not only highly significant statistically but also of a magnitude to be clinically significant.
The data presented are derived from a pair of large studies in which TPA was administered by a uniform protocol that included acute coronary arteriography. Based on standard clinical practice at the time this study was designed, aspirin and β-blockers were not administered, allowing a study of the effect of TPA not confounded by those agents. Thus, this database is well suited to address the hypothesis of circadian efficacy of TPA therapy. The only prior information on this subject was based on 28 patients compared with 692 patients in the present study, but a similar conclusion was reached.16 The absolute patencies in this study are lower than for more recent protocols using accelerated dosing of TPA and additional adjuvant therapy but are very similar to patencies observed in comparable studies of the same era.
Shorter time to initiation of TPA results in a substantially increased 90-minute patency. This observation suggests that clots of longer duration have either altered the pattern of existing constituents (eg, greater cross-linking of fibrin polymer by factor XIII) or altered the composition of the thrombus (eg, greater proportion of platelets). Therefore, time to treatment was controlled in data analysis. Fresher thrombi provide not only the best opportunity for TPA success but also the best opportunity for the circadian pattern of TPA efficacy to be observed.
The a priori hypothesis for this investigation defined TPA efficacy as the variable being analyzed for a circadian pattern, and thus time of initiation of TPA treatment is the independent variable. An additional hypothesis could have defined sensitivity of a clot to lytic therapy as the variable to be analyzed for a circadian pattern, in which case time of infarction would have been the independent variable. Although the two hypotheses have much in common, the former was selected because of its practical implications and because time of treatment was precisely defined.
The mechanism for the circadian pattern of TPA efficacy cannot be determined from this study. Our understanding of thrombus formation and lysis and our knowledge regarding circadian patterns of a number of hemostatic and fibrinolytic factors suggest at least four possible mechanisms, as discussed in the following paragraphs. Two of these are the known circadian patterns of PAI-1 and of platelet activity. Other mechanisms might include diurnal variation in thrombus composition and diurnal variation in spontaneously patent coronary arteries; neither of the latter two phenomena has been demonstrated. Each of these mechanisms could manifest as increased thrombosis in the morning and/or increased fibrinolysis in the evening, with the clinical efficacy of TPA representing a balance between these processes. Included within these mechanisms may be a propensity to procoagulation in response to plasminogen activation,21 particularly since this protocol did not include early administration of aspirin or heparin.
PAI-1 levels are twofold12 14 to fourfold13 higher in the morning than in the evening. The level of circulating TPA in the active form is inversely proportional to PAI-1 levels, thus peaking in the evening. These observations suggest a greater tendency for fibrinolysis in the evening, as has been measured by twofold or more increases in euglobulin fibrinolytic activity.14 15 Patients with acute infarction were found to have preservation of this circadian pattern of PAI-1 activity but at a level more than twice that of controls.22 These circadian patterns of fibrinolysis have been invoked in the past to explain the known circadian pattern of increased frequency of onset of myocardial infarction in the morning and can now also be invoked to perhaps explain the circadian pattern of increased TPA efficacy in the evening. Although pharmacological doses of TPA might be expected to overwhelm physiological levels of PAI-1, it has been suggested that the actual PAI-1 levels faced by pharmacological doses of TPA may be much higher than the measured peripheral plasma levels of patients without infarction, due to markedly elevated local concentration of PAI-1 in the area of the thrombotic occlusion, to rapid induction of PAI-1 synthesis and release following the infusion of exogenous TPA, or to higher levels in patients with acute myocardial infarction.16 21 23 24 In patients receiving TPA for acute infarction, pretreatment levels of PAI-1 were higher in those patients whose infarct arteries were occluded at arteriography performed 3 days later.25
Platelet function has a circadian pattern, with greater aggregability in the morning.5 6 This pattern has been invoked to potentially explain increased onset of myocardial infarction in the morning but might also bear a causal relation to increased TPA efficacy in the evening. Fujii et al21 found that platelet activation released products that stimulated endothelial synthesis of PAI-1, thus potentially raising local levels of PAI-1 and providing one rationale for a linkage of circadian variations of these two systems.
Coronary thrombus composition might be different in the evening versus the morning, perhaps in conjunction with circadian variability of the clotting-lytic-platelet systems, but this has not been investigated. In particular, if evening clots were less platelet rich, they might be more subject to TPA lysis.
Pretreatment angiography was considered inappropriate in the era when this study was performed, so it is not possible to exclude that there is a circadian pattern in spontaneous opening of coronary arteries or in occurrence of infarction without a totally occluded artery. It was also considered unethical to use a placebo control to observe for such circadian phenomena. Becker et al16 not only noted the circadian pattern of patency in 28 patients treated with TPA but also noted no circadian pattern in 20 patients treated with streptokinase, providing limited evidence that the circadian efficacy of TPA is not a marker of a circadian pattern of spontaneously open coronary arteries.
Clinical demonstration of improved lysis in the evening supports previous knowledge of in vitro findings with the same pattern14 15 and complements clinical and in vitro knowledge of increased morning thrombosis. Of importance is that the definition of lysis was TIMI 3 patency, or “complete” lysis. Further investigation of the mechanisms for these circadian patterns is needed. The observations of this retrospective investigation need replication in other studies. Elucidation of the presence or absence of similar circadian patterns with streptokinase or other lytic agents would complement this knowledge. Prourokinase is more effective in the presence of plasma that is platelet rich26 and would thus potentially have a circadian efficacy pattern different than that of TPA, although the activity of prourokinase is promoted by endogenous TPA.26 Data needed to address these and related hypotheses may be available in existing databases.
The present study is well suited to establish the circadian pattern of early coronary patency because the number of patients is very large for an acute catheterization investigation. Only a limited amount of clinical information was gathered, intentionally. Mortality (4.5%), recurrent infarction, and peak creatine kinase did not differ between those patients treated between midnight and noon and those treated between noon and midnight. Clinical outcome end points cannot be definitively investigated in a study of this size, whereas the noninvasive megatrials might be well suited to establish a circadian pattern of clinical outcomes.
A number of potential therapeutic implications based on the circadian efficacy of TPA can be presented as hypotheses, each of which would require full investigation. Perhaps the dose of TPA should be adjusted based on the time of day, as has also been suggested for heparin.11 If other lytic agents do not possess a circadian pattern of efficacy, perhaps the time of day should play a role in selection of the most appropriate lytic agent. Likewise, the time of day might be considered as a factor in selecting between pharmacological thrombolysis and direct angioplasty. Variants of TPA such as those resistant to PAI-127 might avoid the nadir of circadian efficacy. Conjunctive pharmacological agents, including potent anticoagulants or antiplatelet agents more potent than aspirin,28 might also remove the nadir of the circadian pattern of efficacy of TPA.
This work was supported by Burroughs Wellcome Co. I am indebted to Judith K. LittleJohn, MD, for generous support and advice; to Suzanne Edwards, DrPH, for untiring statistical expertise; to Zoltan G. Turi, MD, for incisive critique of the manuscript; to Florence H. Sheehan, MD, and Greg Brown, MD, for core angiographic laboratory analysis; to Harvey L. Waxman, MD, for guidance throughout this project; to Marsha Jones, MD, for assistance; and to Mary Zimmerman and Theresa Cangelosi for secretarial expertise. I am indebted to the principal investigators and their colleagues: Edwin Alderman, MD, Stanford University, Stanford, Calif; Maurice Buchbinder, MD, University of California at San Diego; Samuel M. Butman, MD, University of Arizona; Marcus A. DeWood, MD, Deaconess Medical Center, Spokane, Wash; Sheldon Goldberg, MD, Thomas Jefferson University; Joel Gorfinkel, MD, Mt Carmel Medical Center, Columbus; Cindy L. Grines, MD, University of Kentucky; Abnash C. Jain, MD, West Virginia University; John M. Kalbfleisch, MD, St Francis Hospital, Tulsa; Ronald P. Karlsberg, MD, Brotman Medical Center, Culver City, Calif; Fareed Khaja, MD, Henry Ford Hospital, Detroit, Mich; Peter B. Kurnik, MD, Robert Wood Johnson Medical School, Camden, NJ; Michael A. Kutcher, MD, Bowman Gray School of Medicine; Warren K. Laskey, MD, University of Pennsylvania; William T. Maddox, MD, Memorial Mission Hospital, Asheville, NC; Raymond D. Magorien, MD, Ohio State University; Patrick A. McKee, MD, University of Oklahoma; D. Lynn Morris, MD, Lehigh Valley Hospital Center; James M. Perry, Jr, MD, Vanderbilt University; Neal Shadoff, MD, Presbyterian Hospital, Albuquerque, NM; Leo J. Spaccavento, MD, Wilford Hall USAF Medical Center; George J. Taylor, MD, St John’s Hospital, Springfield, Ill; Condon R. Vander Ark, MD, University of Wisconsin; George W. Vetrovec, MD, Medical College of Virginia; Chris J. White, MD, and Thomas Martyak, MD, Walter Reed Army Medical Center; John F. Williams, MD, University of Indiana; N. Kent Wise, MD, St Francis Hospital, Peoria, Ill; Joshua Wynne, MD, and Zoltan G. Turi, MD, Harper Hospital, Detroit, Mich.
Reprint requests to Peter B. Kurnik, MD, Cardiology Division, Cooper Hospital/University Medical Center, Camden, NJ 08103.
Presented in part at the American Heart Association Scientific Sessions, Anaheim, Calif, November 12, 1991.
- Received July 6, 1994.
- Revision received September 26, 1994.
- Accepted October 9, 1994.
- Copyright © 1995 by American Heart Association
Muller JE, Stone PH, Turi ZG, Rutherford JD, Czeisler CA, Parker C, Poole WK, Passamani E, Roberts R, Robertson T, Sobel BE, Willerson JT, Braunwald E, and the MILIS Study Group. Circadian variation in the frequency of onset of acute myocardial infarction. N Engl J Med. 1985;313:1315-1322.
Willich SN, Linderer T, Wegscheider K, Leizorovicz A, Alamercery I, Schroder R, ISAM Study Group. Increased morning incidence of myocardial infarction in the ISAM study: absence with prior β-adrenergic blockade. Circulation. 1989;80:853-858.
Hjalmarson A, Gilpin E, Nicod P, Dittrich H, Henning H, Engler R, Blacky AR, Smith SC, Ricou F, Ross J Jr. Differing circadian patterns of symptom onset in subgroups of patients with acute myocardial infarction. Circulation. 1989;80:267-275.
Ridker PM, Manson JAE, Buring JE, Muller JE, Hennekens CH. Circadian variation of acute myocardial infarction and the effect of low-dose aspirin in a randomized trial of physicians. Circulation. 1990;82:897-902.
Fujita M, Franklin D. Diurnal changes in coronary blood flow in conscious dogs. Circulation. 1987;76:488-491.
Decousus H, Croze M, Levi FA, Jaubert JG, Perpoint BM, DeBonadona JF, Reinberg A, Queneau PM. Circadian changes in anticoagulant effect of heparin infused at a constant rate. Br Med J. 1985;290:341-344.
Angleton P, Chandler WL, Schmer G. Diurnal variation of tissue-type plasminogen activator and its rapid inhibitor (PAI-1). Circulation. 1989;79:101-106.
Rosing DR, Redwood DR, Brakman P, Astrup T, Epstein SE. Impairment of the diurnal fibrinolytic response in man. Circ Res. 1973;32:752-758.
Becker RC, Corrao JM, Baker SP, Gore JM, Alpert JS. Circadian variation in thrombolytic response to recombinant tissue-type plasminogen activator in acute myocardial infarction. J Appl Cardiol. 1988;3:213-221.
Kalbfleisch J, Kurnik PB, Thadani U, DeWood MA, Kent R, Magorien RD, Jain AC, Spaccavento LJ, Morris DL, Taylor GJ, Perry JM, Kutcher MA, Gorfinkel HJ, LittleJohn JK for the Burroughs Wellcome Study Group. Myocardial infarct-artery patency and reocclusion rates after treatment with duteplase at the dose used in ISIS-3. Am J Cardiol. 1993;71:386-392.
Grines CL, Karlsberg R, Stadius M. Infarct vessel patency and bleeding complications after weight-adjusted dosing of a double chain tissue plasminogen activator preparation: final report. J Am Coll Cardiol. 1990;5:2A. Abstract.
Kleinbaum DG, Kupper L. Applied Regression Analysis and Other Multivariable Methods. North Scituate, Mass: Duxbury Press; 1978.
Lincoff AM, Topol EJ. Illusion of reperfusion: does anyone achieve optimal reperfusion during acute myocardial infarction? Circulation. 1993;87:1792-1805.
Fujii S, Hopkins WE, Sobel BE. Mechanisms contributing to increased synthesis of plasminogen activator inhibitor type 1 in endothelial cells by constituents of platelets and their implications for thrombolysis. Circulation. 1991;83:645-651.
Huber K, Rosc D, Resch I, Schuster E, Glogar DH, Kaindl F, Binder BR. Circadian fluctuations of plasminogen activator inhibitor and tissue plasminogen activator levels in plasma of patients with unstable coronary artery disease and acute myocardial infarction. Thromb Haemost. 1988;60:372-376.
Lucore CL, Sobel BE. Interactions of tissue-type plasminogen activator with plasma inhibitors and their pharmacologic implications. Circulation. 1988;77:660-669.
Lucore CL, Fujii S, Sobel BE. Dependence of fibrinolytic activity on the concentration of free rather than total tissue-type plasminogen activator in plasma after pharmacologic administration. Circulation. 1989;79:1204-1213.
Barbash GI, Hod H, Roth A, Miller HI, Rath S, Zahav YH, Modan M, Zivelin A, Laniado S, Seligsohn U. Correlation of baseline plasminogen activator inhibitor activity with patency of the infarct artery after thrombolytic therapy in acute myocardial infarction. Am J Cardiol. 1989;64:1231-1235.
Gurewich V, Emmons F, Pannell R. Spontaneous clot lysis in whole human plasma by endogenous tissue type and urokinase type plasminogen activators: demonstration of a promoting effect by t-PA and by platelets on urokinase. Fibrinolysis. 1988;2:143-149.
Runge MS, Quertermous T, Haber E. Plasminogen activators. Circulation. 1989;79:217-224.