The purpose was to evaluate the short-term (30-day) and long-term (1,000-day) prognostic values of perioperative troponin T (TnT) and electrocardiographic (ECG) findings in hip fracture patients. A consecutive cohort of 200 patients (68 men) was enrolled. Blinded TnT levels and ECG were assessed on admission, before operation, and on first and second postoperative days. Median (interquartile range) follow-up time was 3.1 (0.3) years. TnT elevation was observed in 71 patients (35.5%): already before the operation in 36 patients (51%) and only after surgery in 35 patients (49%). New ischemic ECG changes were detected in 101 of 194 patients (52%), including 7 ST elevations. Patients with TnT elevation had higher short-term (17% vs 4.7%, respectively, p = 0.008) and long-term (61% vs 40%, respectively, p = 0.005) mortality, the short-term mortality being higher in patients with major (TnT ≥0.15 μg/L) compared with minor TnT elevation (24% vs 13%, respectively, p = 0.005). TnT elevation was the only independent predictor of short-term mortality (hazard ratio [HR] 3.87, 95% confidence interval [CI] 1.45 to 10.3, p = 0.007), whereas TnT elevation (HR 1.73, 95% CI 1.14 to 2.64, p = 0.01), increasing age (HR 1.03, 95% CI 1.01 to 1.06, p = 0.01), dementia (HR 1.84, 95% CI 1.22 to 2.78, p = 0.004), and atrial fibrillation (HR 1.86, 95% CI 1.18 to 2.93, p = 0.007) remained independent predictors of long-term mortality. ECG findings other than ST elevation did not affect mortality. In conclusion, elevated perioperative TnT level is a strong predictor of short-term and long-term mortality. Routine TnT measurements and earlier diagnosis together with appropriate treatment may improve survival of this fragile patient group.
Hip fractures are common in the aging population and are associated with a high mortality. Hip fracture patients often suffer from cardiovascular diseases, and cardiopulmonary diseases are the primary cause of death in this patients group. Patients underwent elective noncardiac surgery may develop a perioperative myocardial infarction, which is usually clinically silent and can be detected by measurement of cardiac troponins with a high specificity for cardiac injury. Recent studies including ours have shown that up to 1/3 of hip fracture patients develop a perioperative troponin elevation suggesting a myocardial infarction. On the contrary to studies that suggest that the operation is the main cause of the myocardial injury, our earlier report suggests that in hip fracture patients, more than 1/2 of the myocardial injuries occur already before the surgery. The purpose of the present analysis was to perform a comprehensive evaluation of how the magnitude and timing of troponin T (TnT) elevation and electrocardiographic (ECG) changes affect both short-term and long-term prognoses in these patients.
Methods
This study ( www.ClinicalTrials.gov , identifier NCT01015105 ) is part of a wider protocol in progress to assess thrombotic and bleeding complications of invasive procedures in Western Finland. A cohort of 200 consecutive low-trauma hip fracture patients referred to the Turku University Hospital during a period from October 19, 2009, to May 19, 2010, was formed. The only exclusion criterion was the patient’s refusal, and only 1 patient was excluded. Clinical and radiological examination was performed in the emergency room, and an on-call anesthesiologist evaluated the patients’ clinical condition and placed a lumbar epidural catheter for pain control. The patients received a mixture of a local anesthetic and opiate from the admission to the second postoperative morning. The patients were operated under spinal anesthesia with isobaric bupivacaine. Significant postoperative blood loss was substituted with red blood cell transfusions. Hypotension (blood pressure <100/60) was treated with rapid fluid challenge, vasopressors, and atropine as appropriate. Patient’s cardiac medications (excluding diuretics) were continued throughout the hospital period. Basic blood chemistry tests and chest x-ray studies were performed on admission and later according to the clinical need. Blinded TnT measurements and ECG recordings were performed for study purposes only, on admission, before operation, and on first and second postoperative days, and physicians were unaware of these results. Additional tests were performed for clinical purposes when clinically indicated. TnT levels were determined by a fourth-generation TnT assay (ECLIA; Roche Diagnostics GmbH, Mannheim, Germany), and the recommended diagnostic threshold of 0.03 μg/L was used to evaluate TnT elevation. TnT elevations from 0.03 to 0.15 μg/L were considered mild, and those ≥0.15 μg/L were considered major. ST elevations, ST depressions, and T-wave inversions were classified according to the guidelines of the European Society of Cardiology. Data on medical history, medication, and cardiac risks were collected from the electronic medical records. These data were also used to evaluate the revised cardiac risk index value (the Lee’s score) for each patient. Furthermore, each patient was assigned an American Society of Anesthesiologists’ physical status class. The patients were followed up until April 2013. The study was conducted in accordance with the Helsinki declaration. The study protocol was reviewed and approved by the Ethics Committee of the Hospital District of Southwest Finland, and all the included patients gave their informed consent.
Normality was tested using Kolmogorov-Smirnov and Shapiro-Wilk tests. Skewed variables presented as median and interquartile range, and categorical variables as percentage. Mann-Whitney U and chi-square tests were used for comparison of variables as appropriate. Survival analysis was performed using Kaplan-Meier’s method and Cox proportional hazards method. A Cox regression analysis with backward selection was performed to analyze the independent predictors of short-term and long-term mortality. A p value of <0.05 was considered statistically significant. All computations were carried out with SPSS software (V16.0; SPSS Inc., Chicago, Illinois).
Results
Baseline characteristics are listed in Table 1 . TnT levels during the hospitalization, and the complete follow-up data up to 1,000 days were available in all the 200 patients. Diagnostic TnT elevation as a sign of myocardial injury was detected in 71 of 200 patients (36%). During hospitalization, ECG was obtained in 194 patients (97%), and 101 patients (52%) had new ischemic ECG changes consisting of ST depression (n = 44, 23%), ST elevation (n = 7, 3.6%), or T-wave inversion (n = 50, 26%). TnT elevation was detected in 50 patients (50%) with ECG changes. Seven patients had ST elevation that developed in 4 patients already before the operation, and TnT elevation was detected in all these 7 patients. None of these patients experienced chest pain, and diagnosis of ST elevation myocardial infarction was reached only in 3 patients. Median (interquartile range) duration of hospitalization was 6.0 (4.0) days with no difference in the duration between the patients with and without TnT elevation.
| Variable | All Patients | Troponin T Elevation | p Value | |
|---|---|---|---|---|
| (n = 200) | Yes (n = 71) | No (n = 129) | ||
| Men | 68 (34%) | 27 (38%) | 41 (32%) | 0.44 |
| Age (years) | 80.8 [11] | 84.5 [7] | 78.8 [13] | <0.001 |
| Coronary artery disease ∗ | 65 (32%) | 31 (44%) | 34 (26%) | 0.02 |
| Prior myocardial infarction | 24 (12%) | 13 (18%) | 11 (9%) | 0.07 |
| Prior coronary revascularization | 16 (8%) | 10 (14%) | 6 (5%) | 0.03 |
| Hypertension | 103 (52%) | 38 (54%) | 65 (50%) | 0.77 |
| Atrial fibrillation | 43 (22%) | 18 (25%) | 25 (19%) | 0.37 |
| Diabetes mellitus | 36 (18%) | 15 (21%) | 21 (16%) | 0.44 |
| Heart failure | 28 (14%) | 16 (23%) | 12 (9%) | 0.02 |
| Renal failure | 12 (6%) | 8 (11%) | 4 (3%) | 0.03 |
| Preoperative ASA score | 2.32 ± 0.81 | 2.5 ± 0.71 | 2.2 ± 0.85 | 0.009 |
| Revised cardiac risk index (Lee’s score) | 0.76 ± 0.92 | 1.0 ± 0.97 | 0.62 ± 0.87 | 0.004 |
| Prior TIA or stroke | 37 (19%) | 14 (20%) | 23 (18%) | 0.85 |
| Dementia | 81 (41%) | 30 (42%) | 51 (40%) | 0.76 |
∗ History of coronary artery disease diagnosed by cardiac stress test or coronary angiography or prior myocardial infarction.
At 30-day follow-up, the overall mortality was 9%. The patients with perioperative TnT elevation had higher mortality than those without TnT elevation (17% vs 4.7%, p = 0.008; Figure 1 ). Mild and major TnT elevations were detected in 46 (23%) and 25 (13%) patients, respectively. Patients with major TnT elevation had significantly higher 30-day mortality compared with those with mild TnT elevation and those without TnT elevation (24% vs 13% vs 4.7%, p = 0.005, respectively; Figure 2 ). The 30-day mortality of patients with ST-elevation myocardial infarction was 29%. No significant difference was detected in the prognosis between the patients with no ischemic ECG change and those with T-wave inversion or ST depression.
Median (interquartile range) follow-up time was 3.1 (0.3) years, and the mortality in patients with perioperative TnT elevation remained constantly higher during the follow-up (61% vs 40%, respectively, p = 0.005) as shown in Figure 3 . A median time of death was 38 (374) versus 246 (639) versus 252 (637) days (p = 0.13) for patients with major versus minor versus no TnT elevation during index hospitalization. Ischemic ECG changes other than ST elevation did not affect the long-term mortality of patients with TnT elevation (data not shown).
Out of the 71 patients with a TnT elevation, the elevation was initially observed in 23 patients (32%) already on hospital admission, in 13 patients (18%) later in the preoperative phase, and in 35 patients (49%) postoperatively. Of the 36 patients with a preoperative TnT elevation, the TnT level remained elevated in 27 patients after the operation. The incidence of major TnT elevations was similar in the patients whose TnT elevation was observed initially before or after surgery. The patients with initial TnT elevation before surgery had also similar short-term and long-term mortality as those with only postoperative TnT elevation, but the 23 patients with an elevated TnT already on hospital admission had a higher 30-day mortality than those with a normal TnT on admission (26% vs 6.2%, respectively, p = 0.007).
Pre-existing coronary heart disease had no significant effect on 30-day or 1,000-day mortality (19% vs 15%, p = 0.75 and 71% vs 53%, p = 0.15, respectively).
TnT elevation during index hospitalization appeared as the only independent predictor of 30-day mortality in a Cox regression model including age, renal impairment, the presence of dementia or atrial fibrillation, red blood cell transfusions, new ECG changes, and revised cardiac risk index values (the Lee’s score) as covariates (hazard ratio [HR] 3.87, 95% confidence interval [CI] 1.45 to 10.32, p = 0.007). In contrast, increasing age (HR 1.03, 95% CI 1.01 to 1.06, p = 0.01), the presence of dementia (HR 1.84, 95% CI 1.22 to 2.78, p = 0.004), atrial fibrillation (HR 1.86, 95% CI 1.18 to 2.93, p = 0.007), and TnT elevation during index hospitalization (HR 1.73, 95% CI 1.14 to 2.64, p = 0.01) remained independent predictors of long-term mortality in a similar Cox regression model. Both preoperative and postoperative TnT elevations remained independent predictors of 30-day mortality when analyzed separately in a similar model. For the long-term mortality, preoperative TnT elevation appeared to be a stronger predictor compared with postoperative TnT elevation (HR 1.95, 95% CI 1.20 to 3.15, p = 0.007 vs HR 1.54, 95% CI 0.99 to 2.38, p = 0.055, respectively).
At discharge, the use of aspirin, statins, and β blockers was equally infrequent in patients with and without TnT elevation as listed in Table 2 . The use of these drugs was not a significant predictor of 30-day and 1,000-day mortalities in multivariate regression model (data not shown).
