Monitoring anticoagulation using the activated clotting time (ACT) in patients treated with heparin and undergoing percutaneous coronary intervention (PCI) is one of the most frequently used tests in invasive cardiology. However, despite its widespread use and guideline endorsement, uncertainty remains regarding the association of ACT with outcomes in contemporary practice. We reviewed all PCI procedures performed at the Mayo Clinic (Rochester, Minnesota) from October 2001 to December 2012 and evaluated the association between the ACT before device activation and in-hospital and 1-year outcomes. ACT values were grouped into tertiles for descriptive purposes and analyzed as a continuous variable for assessment of outcomes. We used logistic and Cox proportional hazards regression models to estimate the association of ACT and outcomes. Of the 12,055 patients who underwent PCI with an ACT value before device activation, 3,977 (33.0%) had an ACT <227, 4,046 (33.6%) had an ACT 227 to 285, and 4,032 (33.4%) had an ACT >285. Baseline and procedural characteristics were similar across ACT tertiles. In unadjusted analysis, higher ACT values were associated with death (p <0.001), bleeding (p = 0.024), procedural complication (p <0.001), and higher 1-year events (cardiac death, p <0.001; cardiac death/myocardial infarction, p = 0.022). After multivariable adjustment for baseline and procedural characteristics, ACT was not independently associated with in-hospital or 1-year ischemic, thrombotic, or bleeding outcomes. In conclusion, ACT values before device activation are not independently associated with clinically important outcomes in contemporary PCI practice.
Use of periprocedural anticoagulation is recommended to reduce the risk of thrombus formation on interventional devices and thrombotic complications during percutaneous coronary intervention (PCI). Monitoring anticoagulation with the activated clotting time (ACT) in patients treated with unfractionated heparin (UFH) is generally considered standard practice. Advantages to UFH are related to availability, clinician familiarity, low cost, point-of-care (POC) testing using ACT, and reversibility, if necessary. POC testing allows for targeting ACT values to recommended ranges, and ACT values generally increase linearly with UFH supplementation. Despite guideline-supported use of ACT monitoring in patients undergoing PCI, there are limited and often conflicting data regarding the association of ACT values and outcomes in contemporary PCI practice. We used the Mayo Clinic (Rochester, Minnesota) PCI registry to explore the relation of ACT before device activation with in-hospital and 1-year ischemic and bleeding events. Our research hypothesis was to determine whether ACT values were independently related to outcome in a contemporary patient cohort who also received high rates of dual antiplatelet therapy and coronary stenting.
Methods
Since 1979, all patients who underwent PCI at Mayo Clinic have been followed in an ongoing prospective clinical registry. The registry includes demographic, clinical, angiographic, and procedural variables, including in-hospital events. Mayo Clinic research staff contact patients by telephone after hospital discharge to ascertain information regarding medication compliance, hospitalizations, and events through use of standardized questionnaires. For long-term follow-up, patients are contacted annually to ascertain clinical events occurring after PCI. Events are confirmed through review of the medical record. For hospitalizations occurring outside Mayo Health System, external medical records are obtained and reviewed. As part of quality assurance, 10% of all records are randomly audited for data integrity.
Approval for this study was obtained from the Institutional Review Board of Mayo Clinic. Patients who underwent elective, urgent, or emergent PCI from October 2001 to December 2012 were included in this analysis. We excluded 1,005 patients who did not provide authorization for use of their medical records for research purposes. We excluded 1,479 patients (11%) who underwent PCI during our study period and did not have an ACT documented. We identified 15,872 PCI procedures performed in 13,091 patients who had either an ACT before device activation or a peak ACT. We used the ACT before first device activation as the primary predictor variable for the study, the rationale being that this is how ACT values are used in interventional cardiology practice and reflect how ACT monitoring was implemented in previous trials of antiplatelet therapy during PCI. This resulted in a final analysis population of 12,055 patients.
The Hemachron whole-blood microcoagulation system, models 401 and 801 (International Technidyne Corporation, Edison, New Jersey), was the POC device used in the catheterization laboratory at Saint Mary’s Hospital (Mayo Clinic) from 1989 to 2010. The models differ only with respect to an additional testing well in model 801. After December 2010, the Hemachron system was replaced in the cardiac catheterization laboratories by the i-STAT1 analyzer for ACT (Abbott Laboratories, Princeton, New Jersey) using the ACT Celite cartridge ( celite ACT) in nonwarm mode. Both devices (Hemachron and celite ACT) were designed for clinical use during in vitro diagnostic testing of whole blood for POC monitoring of anticoagulation intensity in UFH-treated patients who underwent medical procedures. Although some patients receive non-UFH anticoagulants, ACT is most commonly used to monitor patients receiving UFH. The Mayo Clinic Department of Laboratory Medicine and Pathology oversees POC testing in the cardiac catheterization laboratories at Mayo Clinic. Quality controls are outlined in an institutional Standard Operating Procedure document.
Intraprocedural events included bleeding or thrombotic complication. Intraprocedural thrombus was defined as a visible filling defect noted by the operator on the guiding catheter or intracoronary wire or device. Clinically, overt bleeding was defined as any hematoma, femoral access site bleed, gastrointestinal bleed, central nervous system bleed, or retroperitoneal hemorrhage. Procedural complication included in-hospital death, Q-wave myocardial infarction (MI), urgent/emergent coronary artery bypass grafting (CABG), or stroke/transient ischemic attack. MI was diagnosed in the setting of 2 of the following 3 criteria: (1) chest pain for at least 20 minutes; (2) elevation of creatinine kinase (or the MB fraction) >2× normal; and (3) new Q-wave on electrocardiogram. In-hospital deaths included all deaths during the index PCI hospitalization. Target vessel revascularization (TVR) was defined as attempted coronary revascularization by either percutaneous or surgical means. Events through 1 year after index PCI included cardiac death, MI, and coronary revascularization. Significant coronary artery disease was defined as a stenosis ≥70% by visual assessment. Patients with ≥50% in the left main coronary were considered to have 2-vessel disease if right dominant or 3-vessel disease if left dominant. Preprocedural shock was defined as prolonged systolic blood pressure <95 mm Hg in patients not on inotropes or counterpulsation and <110 mm Hg in patients on inotropes or counterpulsation. Moderate/severe renal insufficiency was defined by a serum creatinine level >3.0 mg/dl or a history of hemodialysis or renal transplant.
Data are summarized as mean ± SD for continuous variables or frequency and percentage for categorical variables. Because there was often ≥1 ACT value per unique PCI procedure, we used the ACT before first device activation as the primary predictor variable for the analysis. If multiple devices were used in a unique case (i.e., thrombectomy, angioplasty, and stenting), the ACT before the first device was used. Groups were categorized by tertiles of ACT values for descriptive statistics and modeled in its continuous form to test its association with clinical end points. Kaplan-Meier methods were used to estimate time to clinical event 1-year post-index PCI for all patients for the end points of all-cause death, cardiac death, MI, TVR, and the composite of cardiac death/MI/TVR. Cox proportional hazards models were used to test the association between ACT and outcome. Separate models were constructed for in-hospital end points and 1-year events. To better understand both unadjusted and adjusted association of ACT with outcomes, ACT was analyzed in its continuous form using 3 degree of freedom–restricted cubic splines. Splines were displayed by plotting predicted values against ACT. For covariate-adjusted models, the predicted values were calculated using covariates held at their sample median/mode. Covariates selected for the regression model included age, gender, hypertension, diabetes, congestive heart failure, MI (recent and previous), smoking status, previous PCI and CABG, preprocedural shock, intra-aortic balloon pump status, number of diseased coronary vessels, glycoprotein IIb/IIIa (GpIIb/IIIa) use, stent type (bare metal vs drug eluting), body mass index, hyperlipidemia, and history of malignancy. In addition, we performed tests for interaction for all end points by gender and access site (radial vs femoral).
Results
From October 2001 to December 2012, there were 15,872 PCI procedures performed in 12,055 patients who had an ACT documented before device activation. The clinical characteristics of the cohort by ACT tertile are listed in Table 1 . Patients in the highest ACT tertile were slightly older, more likely to be women, and had higher rates of hypertension and previous MI and revascularization. Table 2 summarizes the angiographic and procedural characteristics. There were similar numbers of diseased vessels and drug-eluting stent use across ACT tertiles. Patients in the highest ACT tertile had lower rates of intralesional thrombus, GpIIb/IIIa use, and pre-procedure Thrombolysis In Myocardial Infarction 0 flow. UFH was the dominant anticoagulant used to support PCI procedures.
| Variable | Activated Clotting Time | P | ||
|---|---|---|---|---|
| <227 (N=3977) | 227-285 (N=4046) | >285 (N=4032) | ||
| Age, (years±SD) | 66.0 ±12.9 | 67.3 ±12.2 | 68.0 ±11.7 | <0.001 |
| Body mass index, (kg/m 2 ±SD) | 30.2 ±6.1 | 30.3 ±6.0 | 29.9 ±5.8 | 0.007 |
| Men | 72.5% | 72.2% | 69.6% | 0.006 |
| Smoker | <0.001 | |||
| Former | 40.5% | 47.6% | 47.8% | |
| Current smoker | 24.2% | 17.8% | 13.4% | |
| Hypertension | 76.4% | 80.9% | 81.5% | 0.016 |
| Hyperlipidemia | 81.1% | 84.1% | 88.6% | <0.001 |
| Diabetes mellitus | 27.1% | 29.9% | 29.2% | 0.016 |
| Mod/severe renal insufficiency | 4.2% | 4.1% | 4.0% | 0.87 |
| Heart failure | 0.09 | |||
| Previous | 4.4% | 4.7% | 5.0% | |
| Current | 14.5% | 12.9% | 12.2% | |
| Prior myocardial infarction | 27.2% | 30.4% | 33.2% | <0.001 |
| Prior percutaneous coronary intervention | 32.1% | 35.7% | 41.7% | <0.001 |
| Prior coronary bypass | 17.8% | 21.8% | 26.4% | <0.001 |
| Stroke/transient ischemic attack | 11.7% | 11.8% | 11.2% | 0.71 |
| Peripheral artery disease | 11.4% | 12.6% | 12.6% | 0.20 |
| Peptic ulcer disease | 6.1% | 6.8% | 5.4% | 0.049 |
| Metastatic malignancy | 1.1% | 0.9% | 1.1% | 0.45 |
| Chronic lung disease | 11.8% | 12.5% | 11.6% | 0.45 |
| Ejection fraction ≤40% | 9.7% | 10.3% | 10.8% | <0.001 |
| Predominant presenting symptoms | <0.001 | |||
| Chest pain | 88.3% | 81.4% | 76.3% | |
| Heart failure | 1.9% | 2.3% | 2.2% | |
| Arrhythmia | 3.2% | 2.1% | 2.0% | |
| Exercise induced | 3.1% | 8.8% | 12.6% | |
| Other | 3.0% | 4.3% | 5.6% | |
| Pre-procedural shock | 5.5% | 2.7% | 1.7% | <0.001 |
| Acute myocardial infarction | <0.001 | |||
| ST-elevation myocardial infarction | 32.9% | 17.7% | 8.7% | |
| Non ST-elevation myocardial infarction | 26.5% | 17.9% | 13.2% | |
| Myocardial infarction <24 hours | 37.8% | 20.7% | 10.6% | <0.001 |
| Variable | Activated Clotting Time | P | ||
|---|---|---|---|---|
| <227 (N=3977) | 227-285 (N=4046) | >285 (N=4032) | ||
| Number of narrowed coronary arteries | 0.06 | |||
| 0 | 3.1% | 3.6% | 3.6% | |
| 1 | 31.6% | 32.9% | 32.4% | |
| 2 | 37.9% | 37.7% | 39.2% | |
| 3 | 27.4% | 25.8% | 24.7% | |
| Type C lesion | 58.1% | 53.4% | 50.0% | <0.001 |
| Thrombus in any lesion | 42.7% | 25.9% | 16.1% | <0.001 |
| TIMI 0 pre-procedure flow in any lesion | 27.1% | 16.9% | 11.3% | <0.001 |
| Urgency of percutaneous coronary intervention | <0.001 | |||
| Emergency | 34.8% | 19.2% | 9.4% | |
| Urgent | 47.4% | 44.6% | 45.7% | |
| Elective | 17.9% | 36.2% | 44.9% | |
| Number of vessels treated | <0.001 | |||
| 1 | 85.9% | 85.2% | 82.7% | |
| 2 | 13.3% | 13.8% | 16.2% | |
| 3 | 0.8% | 1.0% | 1.0% | |
| Number of stents placed (±SD) | 1.4 (0.9) | 1.4 (0.9) | 1.5 (0.9) | 0.002 |
| Drug-eluting stents | 69.0% | 70.2% | 73.6% | <0.001 |
| Use of any stent | 93.5% | 93.4% | 94.4% | 0.13 |
| Vascular sheath size | <0.001 | |||
| <6F | 3.8% | 4.2% | 4.6% | |
| 6F to <7F sheath | 70.7% | 64.9% | 61.1% | |
| 7F to <8F sheath | 18.1% | 22.6% | 24.7% | |
| ≥8F | 7.4% | 8.4% | 9.6% | |
| Vascular closure device | 2.5% | 3.0% | 3.9% | 0.002 |
| Femoral artery access | 90.1% | 92.3% | 94.0% | <0.001 |
| Peri-procedural antithrombotic use | ||||
| Aspirin | 99.1% | 99.0% | 99.2% | 0.64 |
| Clopidogrel | 97.8% | 98.1% | 98.8% | 0.008 |
| Glycoprotein IIb/IIIa inhibitor | 67.1% | 54.3% | 45.3% | <0.001 |
| Heparin | 99.8% | 99.7% | 99.4% | 0.014 |
| Low molecular weight heparin | 5.0% | 4.1% | 3.3% | 0.001 |
| Bivalirudin | 0.4% | 0.2% | 0.8% | <0.001 |
| Thrombolytic | 7.5% | 2.6% | 2.2% | <0.001 |
| Intervention performed in left main | 3.7% | 4.3% | 4.6% | 0.09 |
| Intervention performed in left anterior descending | 41.3% | 43.8% | 44.4% | 0.012 |
| Intervention performed in left circumflex | 26.0% | 28.5% | 29.8% | <0.001 |
| Intervention performed in right coronary | 37.9% | 33.5% | 32.3% | <0.001 |
| Intervention performed in vein graft | 6.9% | 6.8% | 8.3% | 0.019 |
In unadjusted analysis, higher ACT values were associated with death (p <0.001), bleeding (p = 0.024), procedural complication (p <0.001), and higher 1-year events (cardiac death, p <0.001; cardiac death/MI, p = 0.022). After adjustment for important baseline and procedural characteristics, ACT (modeled as a continuous variable) was not independently associated with any of the tested in-hospital or 1-year ischemic, thrombotic, or bleeding outcomes ( Table 3 ). Figure 1 illustrates the predicted probability for (A) in-hospital overt bleeding and (B) procedural complication across the range of ACT values, after adjustment for other risk covariates. No statistically significant association between ACT and risk-adjusted bleeding or ischemic outcomes was observed. There was a significant interaction between ACT and 1-year rates of MI, including composite of cardiac death/MI, by gender, with higher ACT values before PCI associated with worse outcome in women compared with men ( Table 3 ; Figure 2 ). The interaction for increased MI risk in women appeared to emerge when ACT values exceeded 350 seconds ( Figure 2 ). Findings were similar when we restricted the analysis to the period in which only the Hemochron device was used (2001 to 2010). To further explore the relation between ACT and outcomes among patients not treated with GpIIb/IIIa inhibitors (n = 5,361), we tested the interaction between GpIIb/IIIa inhibitors and ACT on all clinical end points. In the unadjusted analysis, there were nonsignificantly lower rates of bleeding in the lowest ACT tertile in non-GpIIb/IIIa compared with GpIIb/IIIa-treated patients. After adjusting for confounders, the sensitivity analysis demonstrated no significant interaction between ACT and ischemic, thrombotic, or bleeding events in non-GpIIb/IIIa compared with GpIIb/IIIa-treated patients (all interaction p values >0.26). Splines were generated and visually inspected and confirmed these results. We observed similar rates of use of dual antiplatelet therapy and evidence-based medications, including β blockers, angiotensin-converting enzyme inhibitors, and statins across ACT tertiles at hospital discharge (data not shown).
