Predictors of and outcomes related to perioperative myocardial injury post‐tracheotomy

Abstract

Introduction

Perioperative myocardial injury (PMI) is associated with increased mortality. We describe risk factors for and outcomes of PMI in patients undergoing tracheotomy.

Methods

Retrospective study of patients undergoing tracheotomy from 2007 to 2016. PMI was defined by a postoperative cardiac troponin I (cTnI) > 99th percentile. Demographics and comorbidities were extracted from the electronic medical record and compared between patients with and without PMI. Significant univariate predictors were included in a multivariable logistic regression model to determine independent predictors of PMI. Thirty‐day and 1‐year mortality of patients with and without PMI were compared.

Results

Of 861 patients undergoing tracheotomy, 41 (4.76%) had PMI and 820 (95.24%) did not. PMI was associated with higher mortality at both 30 days (40.5% vs. 11.2%, p < 0.001) and 1 year (73.2% vs. 44.1%, p < 0.001). Patients with PMI were older (median age 65 vs. 60, p = 0.002) and more likely to have prior myocardial infarction (MI) (36.6% vs. 10.7%, p < 0.001) and chronic kidney disease (31.7% vs. 16.7%, p = 0.024). Cancer diagnosis was associated with a lower risk of PMI (24.4% vs. 41.8%, p = 0.041). Older age (odds ratio [OR] = 1.033, p < 0.001) and prior MI (OR = 3.686, p < 0.001) were independently associated with PMI.

Conclusion

Patients with PMI following tracheotomy had increased short‐ and long‐term mortality. Increased age and history of prior MI were independent predictors of PMI, while cancer was associated with a lower risk of PMI following tracheotomy. ICU patients likely have more acute ailments contributing to a higher risk of PMI and poorer outcomes compared to cancer patients requiring tracheotomy. We propose routine screening for PMI with cTnI in the postoperative period in all tracheotomy patients.

INTRODUCTION

The advent of tracheotomy dates as far back as 1500 BC, being listed among the first operative procedures performed in the field of surgery. Many considered it futile, risky, and even irresponsible in its nascence. Acceptance of this airway‐preserving procedure grew during the 17th and 18th centuries for its role in life‐or‐death situations of diphtheria and poliomyelitis cases, despite operative mortality rates of higher than 75%. ¹ Presently, with emergency cricothyroidotomy as the surgical rescue technique of choice, tracheotomy is most often performed in the controlled environment of the operating room, resulting in less than 1% tracheotomy‐related deaths. ²

General indications for tracheotomy include prolonged intubation, airway protection, management of secretions, upper airway obstruction, and adjunct management of head and neck surgery and trauma. There is a lack of robust evidence regarding the optimal timing and efficacy of the procedure. Moreover, considering the cosmetic and functional morbidity associated with tracheotomy, the tremendous variability of its utility by clinicians comes by no surprise. Tracheotomy is performed in less than 20% of indicated cases at a widely varying range of 0%–59% among different institutions. ³ , ⁴

Further muddying the waters are complications associated with the tracheotomy itself. The benefit of decreased mortality rates when compared to nonsurgical management of compromised airways is dampened by the risk of increased morbidity related to anesthesia and surgery. The risks of hemorrhage, aspiration, infection, luminal stenosis, and fistula formation exist. Perioperative myocardial injury (PMI), which is strongly associated with both in‐hospital and 30‐day mortality, is a concern in all cardiac and noncardiac surgeries.

PMI is defined by elevation in cardiac troponin levels in addition to clinical symptoms or electrocardiographic, sonographic, or angiographic evidence of cardiac ischemia. It has been reported to occur in 0.8%–5% of noncardiac surgeries and accounts for up to 40% of in‐hospital mortality. ⁵ , ⁶ , ⁷ PMI can present silently, especially in patients under sedation and is often recognized late in its course. There has been growing interest in the utility of using the elevation of cardiac biomarkers solely. Referred to as myocardial injury following noncardiac surgery (MINS), this is an independent, prognostically significant clinical entity. ⁶ Detecting and preventing potential cardiac complications by identifying patients with MINS instead of PMI is advantageous because it increases sensitivity. A meta‐analysis of 195 studies totaling 530,867 surgeries revealed the incidence of MINS to be 17.9% and an eight‐ and fourfold increased risk of in‐hospital and 1‐year mortality, respectively. ⁸

Tracheotomy is an invasive surgical procedure that improves mortality in patients with respiratory compromise but whose benefits must be balanced with procedural risks. To better understand and appropriately weigh the risks and benefits of undergoing an operation with high morbidity to preserve the airway, it is important to consider the cardiac complications associated with the procedure. In the present study, we investigated the predictors of and resultant mortality rate following PMI in patients who underwent tracheotomy within a single integrated health system over 10 years.

METHODS

This study is a retrospective review of all patients ≥18 years of age who underwent elective tracheotomy within the Geisinger Health System between 2007 and 2016. Patients were identified by query of an electronic health record and were included in this study if they had PMI, defined as a cardiac troponin I (cTnI) of >99th percentile (0.1 ng/dL) within the first 72 h following surgery and before discharge. ⁹ Notably, we defined PMI as elevated cTnI with or without clinical symptoms or electrocardiographic, sonographic, or angiographic evidence of cardiac ischemia, as these data points were not obtained. Patients were excluded if they had an elevated cTnI during the index hospitalization but before surgery. Data on demographics, prior MI, cancer diagnosis, diabetes, hypertension, prior or current smoking status, chronic kidney disease, prior cerebrovascular disease, and chronic obstructive pulmonary disease were compared between patients with and without PMI using the Mann–Whitney Rank Sum test for numerical variables and chi‐square analysis for categorical variables. Significant univariate predictors were included in a multivariable logistic regression model to determine independent predictors of PMI. Thirty‐day and 1‐year mortality of patients with and without PMI were compared using chi‐square analysis. For all comparisons, a p < 0.05 was considered significant. Statistical analyses were performed with Sigmastat software (Systat). The study was approved by the Institutional Review Board of Geisinger Health System.

RESULTS

A total of 861 patients undergoing elective tracheotomy within the Geisinger Health System between 2007 and 2016 were included in this study. Of these, 41 patients (4.8%) had PMI and 820 (95.2%) did not have PMI. Table 1 compares demographics and comorbidities between patients with and without PMI. Patients with PMI were older and more likely to have had a prior MI and chronic kidney disease (Figure 1). Patients with a cancer diagnosis were less likely to have PMI.

Table 1.

Patients with and without PMI compared using Mann–Whitney rank sum test for numerical variables and chi‐square analysis for categorical variables.

Items	No PMI (n = 820)	PMI (n = 41)	p Value
Age, year (25%–75%)	60 (51–70)	65 (61–75)	0.002
Sex, % male	61.3	68.3	0.525
Body mass index, kg/m2 (25%–75%)	27.8 (23.2–35.3)	30.2 (24.5–37.6)	0.184
Cancer, %	41.8	24.4	0.041
Cerebrovascular accident, %	11.6	22.0	0.081
Chronic kidney disease, %	16.7	31.7	0.024
Chronic obstructive pulmonary disease, %	29.0	22.0	0.429
Diabetes, %	50.4	58.5	0.394
Hypertension, %	51.7	56.1	0.696
Prior myocardial infarction, %	10.7	36.6	<0.001
Smoking, %	69.0	74.3	0.633

Note: Odds ratios are shown with significantly associated variables highlighted.

Abbreviation: PMI, perioperative myocardial injury.

Figure 1.

Patients with PMI were more likely to have had a prior MI (36.6% vs. 10.7%, p < 0.001) and/or chronic kidney disease (31.7% vs. 16.7%, p = 0.024). Patients with a cancer diagnosis were less likely to have PMI (24.4% vs. 41.8%, p = 0.041). CKD, chronic kidney disease; MI, myocardial infarction; PMI, perioperative myocardial injury.

A multivariable logistic regression was created with significant univariate predictors of PMI and revealed that older age (odds ratio [OR] = 1.033, p < 0.001) and prior MI (OR = 3.686, p < 0.001) were independently associated with PMI (Table 2). Cancer also remained independently associated with reduced PMI risk (OR = 0.394, p = 0.014). Patients admitted to intensive care unit services (critical care medicine or surgical critical care) suffered from PMI at a rate of 10.1% compared to 3.2% in non‐ICU patients (p < 0.01, Figure 2). Patients who suffered from PMI had a 30‐day mortality rate of 40.5%, compared to 11.2% in those without PMI (p < 0.001). The mortality rate at 1 year was 73.2% for patients with PMI and 44.1% for those without PMI (p < 0.001). These findings are displayed in Figure 3.

Table 2.

Odds ratio values for multivariate logistic regression of the univariate predictors, with significantly associated variables highlighted.

Items	Odds ratio (5‐95% CI)	p Value
Age	1.033 (1.008–1.059)	<0.001
Cancer	0.394 (0.188–0.829)	0.014
Chronic kidney disease	1.551 (0.756–3.182)	0.231
Prior myocardial infarction	3.686 (1.831–7.420)	<0.001

Abbreviation: CI, confidence interval.

Figure 2.

Patients admitted to ICU services (critical care medicine or surgical critical care) suffered from PMI at a rate of 10.1% compared to 3.2% in non‐ICU patients (p < 0.01). ICU, intensive care unit; PMI, perioperative myocardial injury.

Figure 3.

Patients with PMI had a 30‐day mortality rate of 40.5%, compared to 11.2% in those without PMI (p < 0.001). The mortality rate at 1 year was 73.2% for patients with PMI and 44.1% for those without PMI (p < 0.001). PMI, perioperative myocardial injury.

DISCUSSION

Tracheotomy remains a frequently performed procedure. While the procedure is well established, the timing, indications, and contraindications continue to be elucidated. We investigated a series of 861 patients who underwent tracheotomy. PMI occurred in 41 patients (4.8%), defined by cTnI of >99th percentile within 72 h following surgery with or without symptoms or other objective evidence of cardiac injury. PMI remains a significant predictor of 30‐day and 1‐year mortality. Identifying those patients at greatest risk for PMI could allow for better surgical decision‐making when evaluating patients for tracheotomy, as well as heightening postoperative awareness in high‐risk patients. We analyzed potential risk factors for PMI post‐tracheotomy. Our analysis revealed increased age, prior MI, and chronic kidney disease to be associated with increased risk of PMI. Cancer diagnosis was found to be associated with a decreased risk of PMI.

Other authors have also looked at the risk of PMI associated with either specific procedures or types of procedures. Puelacher et al. designed a prospective study of 2546 noncardiac surgical patients with increased cardiovascular (CV) risk. ⁹ High risk was defined as ≥65 years of age or ≥45 years with a history of coronary artery disease, peripheral artery disease, or stroke. PMI occurred in 16% (397) of cases as defined by an absolute high‐sensitivity cardiac troponin T increase of ≥14 ng/L from preoperative value or between two postoperative values. Like our study, the authors did not require symptoms or EKG changes to define PMI. Identification of PMI triggered a structured response including assessment for symptoms, 12‐lead electrocardiogram (EKG), and cardiology consultation. The incidence of PMI was increased in those with more CV comorbidities (p < 0.001) and ICU admission, similar to our findings. Thirty‐day mortality in patients with and without PMI was 9.8% and 1.6%, respectively (p < 0.001). At 1 year the mortality rate was 22.5% with PMI and 9.3% without (p < 0.001). Notably, Puelacher's study did not include head and neck surgical patients.

Perioperative CV complications occur in 3% of all noncardiac surgery. ¹⁰ The Revised Cardiac Risk Index is routinely used to stratify patients into low and high (≥1%) risk for CV events within 30 days of surgery. Initiating interventions such as statin therapy decreases the risk of perioperative cardiac complications. While this holds great utility in planned noncardiac surgery, tracheotomy patients are rarely afforded the opportunity for preoperative CV evaluation unless paired with head and neck oncologic surgery.

Nagele et al. investigated a cohort of 378 patients undergoing major head and neck cancer surgery. ¹¹ All patients were assessed by a preoperative anesthesia clinic, and a minority were additionally seen by a cardiologist at the discretion of the anesthesiologist. All patients had postoperatively TnI drawn, and 57 patients (15%) had abnormal elevation. These patients had a significantly longer hospital (8.5 vs. 10.1 days; p = 0.014) and ICU (3 vs. 4.5 days; p = 0.001) admissions. Risk of death increased eightfold at 60 days postoperatively and doubled at 1 year. Their analysis revealed renal insufficiency, coronary artery disease, peripheral vascular disease, hypertension, and prior chemoradiation therapy to be associated with elevated TnI postoperatively. This study underscores the need for routine preoperative risk stratification and postoperative monitoring in head and neck cancer patients.

Cancer diagnosis in our study paradoxically appeared to be protective of PMI. While we did not characterize all acute medical conditions in the ventilator‐dependent ICU cohort, we suspect their overall health status was poorer than the cancer cohort. Poor overall health likely accounts for the relatively high rate of PMI. The discrepancy in PMI rates between patients with and without cancer, especially in the context of prior work demonstrating a high rate of PMI in head and neck cancer patients, highlights a suspected difference in PMI screening. ¹¹ ICU patients may have been more likely to have monitored for CV complications than head and neck cancer patients admitted for routine postoperative care.

PMI is an often‐silent disease process that can remain undetected in the postoperative period. Several prior studies and the present study demonstrate an increase in mortality associated with PMI. We report increased age, chronic kidney disease, and prior MI as being associated with PMI. Patients in the ICU were more likely to suffer from PMI than patients undergoing tracheotomy in setting of cancer. Given the challenges in preoperatively risk stratifying most tracheotomy patients for CV complications given their acute ailments with ventilator dependence, we propose routine screening for PMI with cTnI in the first 72‐h following tracheotomy. cTnI is more sensitive than EKG or screening for symptoms. Early detection of PMI may improve outcomes including mortality in tracheotomy patients.

AUTHOR CONTRIBUTIONS

Randy W. Lesh contributed to conceptualization, data analysis, writing, reviewing, and editing. Jino Park contributed to writing. Vincent M. Desiato contributed to reviewing and editing. Martin Matsumura contributed to conceptualization, data analysis, writing, reviewing, and editing. Thorsen W. Haugen contributed to conceptualization, writing, reviewing, and editing.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflict of interest.

ETHICS STATEMENT

This research was conducted ethically in accordance with the World Medical Association Declaration of Helsinki. The Institutional Review Board of Geisinger Health System granted ethics approval (IRB# 2017‐0336) and exemption from obtaining written informed consent.

Lesh RW, Park J, Desiato VM, Matsumura M, Haugen TW. Predictors of and outcomes related to perioperative myocardial injury post‐tracheotomy. World J Otorhinolaryngol Head Neck Surg. 2025;11:412‐416. 10.1002/wjo2.218

DATA AVAILABILITY STATEMENT

Patient data were pulled from the Geisinger Health System electronic health records system. All data generated or analyzed during this study are included in this article. Further enquiries can be directed to the corresponding author.