Author + information
- Received August 12, 1998
- Revision received November 6, 1998
- Accepted December 23, 1998
- Published online April 1, 1999.
- William H Matthai Jr., MD, FACC∗,* (, )
- Peter B Kurnik, MD, FACC†,
- William C Groh, MD, FACC∗,
- William J Untereker, MD, FACC∗ and
- Jamie E Siegel, MD‡
- ↵*Reprint requests and correspondence: Dr. William H. Matthai, Jr., Cardiology Division, Presbyterian Medical Center, 39th and Market Streets, Philadelphia, Pennsylvania 19104
This study evaluated changes in antithrombin (AT) activity around the time of percutaneous transluminal coronary revascularization (PTCR) with unfractionated heparin anticoagulation and the effects these changes had on major thrombotic complications of PTCR.
Heparin is used during PTCR to prevent thrombosis. However, heparin, a cofactor for AT, causes AT activity to fall. AT activity <70% is associated with thrombosis. There is a prothrombotic state after heparin discontinuation that has not been well explained.
Antithrombin activity was sampled at the start and end of PTCR and the next two mornings in 250 consecutive patients. We recorded occurrence of major thrombotic events, defined as 1) major thrombotic complications of PTCR; 2) major in-lab thrombus formation; or 3) subacute occlusion. Discriminant analysis was employed to evaluate the relationship of AT activity to these events. Change in AT activity and its relationship to heparin was evaluated. Evidence of restenosis at six months was obtained.
There were 14 major thrombotic events. Antithrombin activity <70% was strongly (p = 0.006) associated with these events. The AT activity fell significantly through the morning after PTCR when 21% of patients had AT activity <70%; AT activity did not normalize until >20 h after heparin discontinuation. Pre-PTCR use of heparin led to lower AT activity in proportion to duration of heparin use. There was no relationship between AT activity and restenosis.
Low AT activity may contribute to major thrombotic complications of PTCR. The way heparin is used before and after PTCR is important to development of low AT activity.
Intense anticoagulation during percutaneous transluminal coronary revascularization (PTCR) is required to minimize the risk of thrombotic complications during and shortly after the procedure. This anticoagulation is almost universally accomplished with heparin. In addition, heparin pretreatment for several days has been suggested to reduce the chance of an adverse outcome, especially in the subset of patients with unstable angina (1). When heparin is discontinued, a transient hypercoagulable state has been described that has not been well explained (2–4).
Heparin has indirect antithrombotic activity; it functions as a cofactor for antithrombin (AT), increasing the activity of that molecule 1000-fold. However, treatment with heparin causes AT levels to fall (5–7), and AT deficiency has been associated with thrombosis (8). The effects of heparin on AT activity around the time of PTCR and the clinical implications of these effects on the results of PTCR, particularly with respect to thrombotic complications, have not been evaluated.
Two hundred fifty consecutive patients undergoing PTCR by the faculty at Cooper Medical Center were enrolled between May 18 and December 22, 1994. Elective, urgent, and emergent procedures were included. Patients were enrolled regardless of the type of procedure planned or actually undertaken. However, if a procedure was unsuccessful owing to inability to cross the lesion with a wire and balloon, that patient was removed from the protocol.
The protocol was approved by the Institutional Review Committee of Cooper Medical Center. Because the protocol did not affect the angioplasty procedure in any way, no investigational drug or device was used as part of the protocol, and only minimal extra phlebotomy was required by the protocol, individual informed consent was not required.
The angioplasty procedure was not affected by the protocol in any way. All patients were on antiplatelet therapy at the time of PTCR, consisting of aspirin or, in the few patients with aspirin allergies, ticlopidine. An activated clotting time (ACT; Hemochron, Technidyne, Edison, New Jersey) of greater than 250 s was confirmed prior to the first dilation. Additional heparin was administered to achieve a goal ACT of 300 to 350 s. The ACT was recorded at the conclusion of the procedure at which time mean ACT was 339 ± 64 s, and 75% of the population had an ACT greater than 300 s. The type of revascularization procedure performed (balloon, stent, laser, directional atherectomy) was determined by the attending interventional cardiologist. At the time of this protocol, stents were generally placed only for threatened closure or inadequate PTCR result and not as primary therapy. Either a noncalcium chelating form of meglumine diatrizoate (Hypaque 76™), a high-osmolality contrast agent, or iohexol (Omnipaque™), a nonionic low osmolality agent, was used as contrast media. Heparin therapy prior to PTCR was at the discretion of the attending cardiologist, but duration of heparin therapy prior to the procedure was recorded. Arterial and venous sheaths were generally maintained in place the night after PTCR with a goal aPTT of 65 to 100 s. Heparin was discontinued the following morning for sheath removal and was restarted following sheath removal at the discretion of the attending cardiologist.
Blood samples were drawn from the unheparinized venous sheath at the beginning of PTCR, at the end of PTCR and the following morning just before heparin was discontinued for sheath removal. A fourth sample was drawn the second morning after PTCR from a fresh venupuncture if the patient was still hospitalized. The AT activity was measured with a chromogenic assay using anti-Xa activity (Chromogenix, Molndel, Sweden) on an MLA 1000 (Medical Laboratory Automation, Pleasantville, New York) with an established reference interval of 76% to 138%. All specimens were run on the same lot of reagent. An AT activity <70% was defined as clinically important (8).
Clinical outcome events
Primary clinical outcome events are defined in Table 1. These events were chosen because they are thrombotic complications with clinical significance. In addition, they were chosen because they are all angiographic events and are less subjective than either evidence of thrombus at the PTCR site not severe enough to require additional intervention or postprocedure chest pain without angiographic correlation. Thrombus was defined as a filling defect surrounded by contrast. Events that were clearly due to major coronary dissections, such as long spiral dissections, were not included as end point events. All events were reviewed by two investigators with no knowledge of AT activity, and if it was believed that thrombus could be playing a role in the event, that event was counted as a thrombotic event. Patients were followed until discharge after PTCR, and only events during that hospitalization were included in the analysis.
A secondary end point was restenosis at six months, which was defined either angiographically or clinically. Restenosis was defined angiographically as cross-sectional diameter stenosis >50% at the site of the PTCR. Clinical restenosis was defined as cardiac death, myocardial infarction in the distribution of the dilated artery or evidence of ischemia on stress testing in the distribution of the dilated artery. Angiographic follow-up or stress testing was not required by the protocol.
Continuous variables were analyzed with the Student ttest for normally distributed continuous variables and with the Mann-Whitney Utest for variables not normally distributed. Repeated measures of continuous variables were analyzed by repeated measures analysis of variance (ANOVA), and the Newman-Keuls procedure was used to evaluate differences between groups. Chi-square analyses were used for discrete variables. Repeated measurements of discrete variables were evaluated with Cochran’s Q; the McNemar test with a Bonferroni correction was used to evaluate differences between groups. Independent variables that were significantly related to the occurrence of an outcome event were included in a discriminant analysis, which was done using the discriminant procedure of the SPSS-X statistical package with cutoffs of p = 0.3 for entering and removing variables from the model. The significance of the F to remove at the last step provided p values for the variables in the model. Results were analyzed by patient and not by lesion because the major point of interest, AT activity, is not independent between two lesions in the same patient. If a patient had more than one lesion dilated, either the first lesion dilated or the lesion with an outcome event was used for analysis. In no case was there more than one outcome event in a single PTCR procedure.
Patient sample and primary outcome events
Two hundred fifty consecutive patients undergoing PTCR were successfully enrolled in the protocol, and 310 lesions were dilated. Patient demographics and procedures performed are outlined in Table 2. Antithrombin activity was measured in 245 patients before PTCR, 246 at the end of the procedure, and in 242 and 200 on the first and second mornings after the procedure, respectively.
There were 14 primary outcome events. Nine patients had acute, in-lab thrombotic occlusion or reduced coronary flow due to angiographically present thrombus that required further treatment. There were three cases of out-of-lab subacute closure. One patient had failed PTCR due to the requirement of emergency bypass surgery, and one patient died suddenly from cardiac causes the day after PTCR.
As seen in Figure 1, for the sample as a whole, mean AT activity declined over the course of the procedure, declined further as heparin was continued the evening of PTCR, and rose toward baseline between the first and second morning after the procedure when heparin had been discontinued in many of the patients. A significant difference occurred across the four time periods (p < 0.001, repeated measures ANOVA) and between each and every other point (p < 0.05, Newman-Keuls procedure). The percent of patients with AT activity <70%, a range believed to confer clinical importance, is shown in Figure 2. There was a significant difference (p < 0.05, Cochran’s Q) across the four time periods, and more patients had AT activity <70% after PTCR and the following morning than before PTCR (p < 0.001, McNemar’s test). An AT activity <70% was measured in 29.2% of the population in at least one sample. There was no relationship between the type of procedure performed or choice of contrast agent and AT activity.
Antithrombin activity was lower in those who received preprocedural heparin, and the degree of decline was directly related to the duration of time that a patient was on heparin prior to the start of the procedure (Table 3). Not only was this finding true for the mean AT activity at the start of PTCR, but mean AT activity at the end of the procedure and the morning after PTCR was also lower the longer a person had been on preprocedural heparin. When the incidence of AT activity <70% is analyzed, results are similar; the longer a patient was on heparin prior to PTCR, the more likely that AT activity would be <70%. There were 22 patients who started the PTCR procedure with an AT activity <70%; 21 of those 22 patients had received heparin for greater than 48 h prior to the procedure.
There is a more rapid drop in AT activity over the course of the PTCR procedure itself than in other periods of similar duration. The mean decrease in AT activity was 7.5% over the 1 to 2 h of the procedure (90.1 ± 18.0% to 82.6 ± 15.8%) and only 4.1% in the period from the end of the procedure to the next morning (82.6 ± 15.8% to 78.5 ± 12.9%). Forty-one patients had AT activity <70% by the end of PTCR. The change in AT activity during the procedure was independent of the time the patient had been exposed to heparin prior to the procedure.
Heparin was discontinued in all patients before sheath removal, generally the morning after PTCR. In 194 patients, heparin was restarted shortly thereafter based on the angiographic appearance of the PTCR result or for other clinical indications and was continued for variable periods of time. In many patients, heparin was discontinued before the AT level was drawn on the second morning after PTCR, allowing us to examine recovery of AT activity after discontinuation of heparin (Fig. 3). The AT activity gradually returned toward normal, but mean AT activity continued to be significantly less than normal unless heparin had been discontinued for more than 20 h.
Association of outcome events and antithrombin activity
Univariate association of AT activity with primary outcome events is shown in Table 4. There was a strong association of AT activity <70% with a thrombotic event, and mean AT activity was lower in those patients with events than in those without events. The association of other variables with outcome events is shown in Table 5. The presence of thrombus at the start of PTCR and use of nonionic contrast media were particularly strongly associated with these thrombotic events. Although AT activity <70% was defined as the critical level in the protocol and all analyses used this level, secondary analysis reveals that AT activity below the reference range of the laboratory (<76%) was measured in 11 of 14 patients in the sample drawn closest to the event. Antithrombin activity <70% was measured in 10 of 14 patients with an event (Table 4)and of these 10, it was drawn in the sample taken closest to the event in 6 patients. In the other four patients, it was 75% or less in the sample drawn closest to the event.
Independent variables related to outcome events by univariate analysis were included in a multivariate analysis to clarify further the role these variables played in the occurrence of thrombotic events. Many of the variables with a univariate relationship to a primary outcome event were not independent of each other, and therefore many of these variables were not included in the multivariate analysis. In particular, the presence of thrombus at the start of PTCR, a diagnosis of unstable angina or recent infarction, and use of preprocedural heparin were all strongly related to each other. A discriminant analysis with AT activity (as a dichotomous variable, <70% or ≥70%), presence or absence of thrombus at the start of the procedure, contrast agent used (nonionic or ionic), initial percent stenosis, and lesion length (≤10 mm or >10 mm) were included in the analysis. The occurrence of AT activity <70% at any time continued to be strongly associated with outcome events (p = 0.006). Use of nonionic contrast (p = 0.0002), thrombus before PTCR (p = 0.02), and lesion length (p = 0.02) also continued to have a statistically strong relationship to these major thrombotic events.
Not surprisingly, the population with AT activity <70% was more likely to have unstable angina or recent infarction than were those with AT activity ≥70%. Clinically, this is the population in whom heparin use would be expected. Other than a greater likelihood of high cholesterol in those with AT activity <70% (p = 0.045), there were no differences in clinical variables between the two groups.
Six-month follow-up was available in 93% of the study sample. Angiographic follow-up was obtained in 59 patients (23.6%). There was clinical or angiographic evidence of restenosis in 24%. There was no relationship between AT activity and restenosis (Table 6).
This study is the first to examine and demonstrate a strong correlation between AT activity during the period of angioplasty and major thrombotic complications of that procedure. We have demonstrated that AT activity declines by an unexpectedly large amount during the PTCR procedure itself and continues to decline over the evening after PTCR when heparin is continued. In a significant percentage of this population, AT activity drops to a level that is considered clinically important. The degree that AT activity is lower than normal when measured before, just after or the morning after PTCR is directly correlated with the duration of preprocedure heparin. Following discontinuation of heparin, AT activity recovers slowly and remains lower than normal during the same period in which a relative hypercoagulable state has been demonstrated in other studies after the discontinuation of heparin. This study also confirmed the deleterious impact preprocedural thrombus (9)and lesion length (10)have on outcome, and we have added evidence to support the contention that use of low osmolality nonionic contrast media in angioplasty may be associated with thrombotic complications (11).
Low AT activity and reduced local antithrombotic effect
Low AT activity may lead to reduced local anticoagulant effect at the time of PTCR. Although it is widely recognized that the effect of the heparin/AT complex on clot-bound thrombin (12,13)or on thrombin bound to soluble fibrin (14)is less than that on fluid-phase thrombin, it is important to understand that this is a relative effect and that inhibitory activity continues to increase as heparin concentration increases. Therapeutic heparin concentrations of 0.2 to 0.4 U/ml completely inhibit fluid-phase thrombin but only produce 20% to 40% inhibition of clot-bound thrombin, but a heparin concentration of 2 U/ml will inhibit 70% of clot-bound thrombin (12). In another study, growth of thrombus on damaged porcine carotid artery in the presence of fresh thrombus was inhibited in direct proportion to the dose of administered heparin up to the point of 100% inhibition at very high doses (13).
Because both heparin and AT are required for an effective antithrombotic effect, we theorize that if either AT or heparin concentration is inadequate, expected anticoagulation may not result. Perhaps the full benefit of AT-dependent therapy cannot be realized unless AT is adequate at the PTCR site. Whereas heparin concentration in the peri-PTCR period may be sufficient to easily inhibit fluid-phase thrombin as reflected in an elevated ACT, low AT activity may result in inadequate inhibition of clot bound thrombin with resultant local ongoing thrombus generation and platelet activation at the site of PTCR. Heparin concentration at the PTCR site may be adequate, but low AT activity may lead to local thrombosis even in the setting of continued systemic heparin anticoagulation. Heparin-resistant thrombin activity, manifested as elevated coronary sinus fibrinopeptide A levels, has been associated with angiographic evidence of coronary thrombus and ischemic complications of PTCR (15). Patients with threatened or abrupt coronary occlusion during PTCR have been shown to have elevated local thrombin activity when compared with those without an event despite similar heparin concentrations and ACT (16). Experimental models evaluating heparin effect have been performed only in the presence of normal AT activity, but increased antithrombotic effects have been shown to be directly related to dose of administered AT in vitro (17). Neither low AT activity in the presence of high heparin concentrations nor the effects of high AT levels in the presence of stable heparin concentrations have been studied.
Effects of heparin on AT activity
The effects of standard doses of heparin on AT levels have been evaluated in small studies with conflicting results (5–7). Although it has been suggested that AT levels fall rapidly after heparin administration but do not continue to decline and that they recover rapidly after heparin discontinuation (7), the current study confirms the findings of Marciniak (5)and Rao et al. (6)that AT activity declines progressively over at least the first several days of heparin therapy and recovers relatively slowly after heparin discontinuation.
The progressive decline in AT activity over 48 h and slow recovery may have clinical relevance to management of patients about the time of PTCR. Clinical studies evaluating heparin use prior to PTCR have produced conflicting results. Laskey et al. (1)suggested that 48 h of preprocedural heparin reduces the incidence of periprocedural complications in patients with unstable angina or visible thrombus (18), whereas Klein et al. (19)found that the group that had more than 48 h of preprocedural heparin had more complicating myocardial infarctions and deaths. Our study shows that patients who receive heparin for greater than 48 h have significantly lower AT activities than do those with a shorter duration of pre-PTCR heparin; this suggests that extended infusions of heparin prior to PTCR may be detrimental in some patients owing to the effects of heparin on AT. In addition, prolonged heparin therapy after PTCR may lead to further decline in AT activity, supporting the current practice of early heparin discontinuation after PTCR, although there are no good data to suggest this practice may reduce thrombotic events.
There is a period of relative hypercoagulability following discontinuation of heparin therapy in acute coronary syndromes (2–4). Our data have shown that AT activity remains lower than normal over this prothrombotic period after heparin discontinuation, raising the possibility that low AT activity may be playing a role in this prothrombotic state and in those patients with thrombotic events after PTCR. Each of the four patients in our study with events after PTCR had AT activity <70% before the event. D-dimer levels, which are markers of thrombin activity, have been shown to be lower in patients in whom AT activity was normalized with AT concentrates prior to discontinuation of heparin than in those in whom AT activity was not normalized (20).
Our findings unexpectedly demonstrate a surprisingly large fall in AT activity during the PTCR procedure itself. The AT levels do not drop during PTCR with direct thrombin inhibitors (21), suggesting that the procedure itself does not cause AT levels to fall. We cannot exclude the possibility that the marked decrease in AT activity seen may be explained by the very large doses of heparin used during PTCR, but Rao et al. (6)showed that a 100 U/kg heparin bolus did not cause AT levels to decline. Our data raise the possibility that the large decrease in AT activity from the start to the end of PTCR may be due to other mechanisms unique to PTCR with heparin anticoagulation, which may increase AT clearance or consumption and may increase the risk of thrombosis. The large decline in AT activity over the course of the procedure itself explains why prolonged heparin infusion alone was not a predictor of a thrombotic event.
Antithrombin and restenosis
The lack of long term benefit on the incidence of restenosis is consistent with other trials of antithrombotic therapy. Reduction of local thrombus formation with any number of antithrombotic agents, including hirudin (22,23)and AT (24), has reduced restenosis in animal models, but these benefits have not been confirmed in clinical studies (25,26). The mechanism of restenosis is multifactorial, and any role for antithrombotic therapy has yet to be defined.
The major limitation of this study is the lack of confirmatory evidence that the thrombotic events that occurred were a result of low AT activity. It is our hypothesis that low AT activity reduces the effect of heparin on clot-bound thrombin. These data cannot answer those questions. It is well documented that low AT activity is associated with a systemic hypercoagulable state in other populations (8,27), but low AT activity in those populations may not have the same effect as low AT activity in this population. Nevertheless, we hypothesize that a systemic hypercoagulable state is not a prerequisite to the thrombotic events we have reported; inadequate AT activity at the PTCR site may result in a local hypercoagulable state. If low AT activity is not the primary cause of the thrombotic events, it appears to be, at the least, a marker of increased risk. Once validated, low AT activity may guide a physician toward more aggressive antithrombotic therapy in the period around PTCR.
Demonstration of a strong relationship between AT activity and these thrombotic events could be achieved if administration of AT concentrates or other antithrombotics was shown to reduce the incidence of thrombotic complications in those with low AT activity. Antithrombin concentrates have been used in two studies of PTCR. Schachinger and colleagues (28)showed that a mean dose of 730 U of AT did not reduce the incidence of thrombotic events in patients undergoing PTCR. The AT activity was not measured during this trial, and the infusion rate of AT used would not be expected to normalize a low AT activity. In a trial of 50 PTCR patients with unstable angina, Grip et al. (20)administered either placebo or AT with a goal AT activity of 120%. They found no difference in clinical events or in acute markers of coagulation activation but did find that thrombin activation after discontinuation of anticoagulation as measured by the D-dimer was less pronounced in the population that received AT. Inadequate AT may have been given in the first trial, and the second trial was not powered to evaluate the clinical significance of AT supplementation.
This study has shown a strong relationship between low AT activity and major thrombotic complications of PTCR. The effects of heparin therapy and PTCR on AT activity in a large population have been carefully evaluated, and the decline in AT activity was found to be significant in a large portion. We believe these data may explain, in part, thrombotic complications of PTCR. The study has also raised a number of intriguing questions regarding the role for measurement of AT activity as a marker for patients at risk for thrombosis, for the appropriate use of heparin or other antithrombotic agents before, during, and after PTCR, and suggests that there may be a role for AT administration in selected patients. It is likely that unfractionated heparin will continue to be used for the immediate future during PTCR and in the management of acute coronary syndromes, and this study provides more complete understanding of the effects of the antithrombotic therapy in current usage.
We wish to thank Ed J. Gracely, PhD, for his statistical advice and assistance, and Sharon Hanlon, RN, our study coordinator.
☆ This study was supported by a grant from Baxter Healthcare Corporation, Hyland Division.
- activated clotting time
- percutaneous transluminal coronary revascularization
- Received August 12, 1998.
- Revision received November 6, 1998.
- Accepted December 23, 1998.
- American College of Cardiology
- Granger C.B,
- Miller J.M,
- Bovill E.G,
- et al.
- ↵Blajchman MA, Austin RC, Fernandez Rachubinski F, Sheffield WP. Molecular basis of inherited human antithrombin deficiency. Blood 1992;80;9:2159–71.
- Lincoff A.M,
- Popma J.J,
- Ellis S.G,
- Hacker J.A,
- Topol E.J
- Ryan T.J,
- Faxon D.P,
- Gunnar R.M,
- et al.
- Grines C.L,
- Schreiber T.L,
- Savas V,
- et al.
- Meyer B.J,
- Badimon J.J,
- Mailhac A,
- et al.
- Weitz J.I,
- Leslie B,
- Hudoba M
- Oltrona L,
- Eisenberg P.R,
- Lasala J.M,
- Sewall D.J,
- Shelton M.E,
- Winters K.J
- Winters K.J,
- Smith S.C,
- Eisenberg P.R
- Grip L,
- Blomback M,
- Egberg N,
- Olsson A,
- Svane B,
- Velander M
- Kobayashi H,
- Suzuki S,
- Sakamoto S,
- Matsuo T
- Unterberg C,
- Sandrock D,
- Nebendahl K,
- Buchwald A.B
- Gallo R, Padurean A, Toschi V, et al. Prolonged thrombin inhibition reduces restenosis after balloon angioplasty in porcine coronary arteries. Circulation 1198;97:581–8.
- Hedin U,
- Frebelius S,
- Sanchez J,
- Dryjski M,
- Swedenborg J
- DeStefano V,
- Finazzi G,
- Mannucci P
- Schachinger V,
- Allert M,
- Kasper W,
- Just H,
- Vach W,
- Zeiher A.M