The 2014 American College of Cardiology and American Heart Association guidelines on perioperative evaluation recommend differentiating patients at low risk (<1%) versus elevated risk (≥1%) for cardiac complications to guide appropriate preoperative testing. Among the tools recommended for estimating perioperative risk is the National Surgical Quality Improvement Program (NSQIP) Myocardial Infarction and Cardiac Arrest (MICA) risk calculator. We investigated whether the NSQIP MICA risk calculator would accurately discriminate adverse cardiac events in a cohort of adult patients undergoing elective orthopedic surgery. We retrospectively reviewed 1,098 consecutive, elective orthopedic surgeries performed at Hershey Medical Center from January 1, 2013, to December 31, 2014. Adverse cardiac events were defined as myocardial infarction or cardiac arrest within 30 days of surgery. The mean estimated risk for adverse cardiac events using the NSQIP MICA risk calculator was 0.54%, which was not significantly different (p = 1) compared with the observed incidence of 0.64% (7 of 1,098 procedures). The c-statistic for discriminating adverse cardiac events was 0.85 (95% CI 0.67 to 1) for the NSQIP MICA risk calculator and 0.9 (95% CI 0.75 to 1) for the Revised Cardiac Risk Index. In conclusion, the NSQIP MICA risk calculator is a good discriminator of adverse cardiac events in patients undergoing elective hip and knee surgery, performing comparably to the Revised Cardiac Risk Index.
In 2014, the American College of Cardiology and the American Heart Association released updated guidelines on the perioperative evaluation of patients undergoing noncardiac surgery. Compared with previous guidelines which based risk stratification primarily on risk factors derived from the Revised Cardiac Risk Index (RCRI), the 2014 guidelines include novel risk calculators derived from the National Surgical Quality Improvement Program (NSQIP) database. Similar to the RCRI, these calculators incorporate patient-and procedure-specific data to estimate risk for perioperative events. When developed, the NSQIP Myocardial Infarction and Cardiac Arrest (MICA) risk calculator was shown to be a good discriminator of perioperative MICA in noncardiac surgery with a c-statistic of 0.88, comparing favorably with the RCRI with a c-statistic of 0.75. Whether superior discrimination of adverse outcomes by the NSQIP MICA risk calculator is applicable to all noncardiac surgeries or only to certain surgical populations is unknown. The aim of this study was to determine whether the NSQIP MICA risk calculator could accurately discriminate perioperative MICA in patients undergoing elective hip and knee replacement surgery and to compare its discriminatory performance with the RCRI in this cohort.
Methods
We reviewed electronic medical record data from consecutive, elective total hip replacement and total knee replacement surgeries performed on adult patients at Penn State Milton S. Hershey Medical Center from January 1, 2013, to December 31, 2014. Preoperative clinic visits with anesthesia were reviewed for patient medical history, current medications, functional capacity, surgical procedure planned, and a numerical assessment of American Society of Anesthesia (ASA) physical status class.
We documented an RCRI score for all patients in this cohort. The RCRI score is based on the presence of risk factors including high-risk procedure type (none in this cohort), a history of coronary artery disease, cerebrovascular disease, congestive heart failure, preoperative insulin use, and preoperative creatinine >2. The medical history components of the RCRI were determined based on a review of problem lists, preoperative clinic notes, and previous cardiac procedures. Preoperative insulin use was determined from patient medications documented in the anesthesia preoperative note. Preoperative creatinine level was determined from the most recent creatinine level recorded before the operation in the electronic medical record.
We also documented an estimation of perioperative risk using the NSQIP MICA risk calculator for all patients in this cohort, where risk estimation is determined from the following factors: patient age, preoperative creatinine ≥1.5, ASA physical status class, preoperative function, and procedure type. Patient age was recorded as age at the time of surgery. ASA physical status was recorded as the numerical assessment documented in the preoperative anesthesia clinic visit. Preoperative creatinine level was determined as described previously. Preoperative function in the NSQIP MICA risk calculator refers to functional status relative to activities of daily living and is stratified as total independence, partial dependence, or total dependence. We recorded functional status as documented in the medical record from a standardized preoperative nursing assessment form. Procedure type was assigned as “orthopedic and nonvascular extremity” for NSQIP MICA risk calculator estimates in this study.
Perioperative adverse cardiac events were defined as cardiac arrest or myocardial infarction within 30 days of surgery. Length of stay, 30-day perioperative mortality, and 30-day readmission to Hershey Medical Center were also documented. Myocardial infarction was defined according to the most recent consensus criteria published in 2012, which defines a myocardial infarction as an increase and/or decrease of cardiac troponin above the 99th percentile laboratory upper reference limit that is accompanied by ischemic symptoms and/or objective findings of ischemia by electrocardiography (new significant ST-T changes, new left bundle branch block, or pathological Q waves), echocardiography (new regional wall motion abnormality), or coronary angiography (identification of intracoronary thrombus). Laboratory results were screened to identify any troponin above the 99th percentile within 30 days of surgery, with manual review of hospitalization records to determine whether the ischemic criteria described previously were met. Cardiac arrest was defined as a pulseless rhythm requiring cardiopulmonary resuscitation.
Patients were classified as low or elevated risk for adverse cardiac events based on calculated risk scores of <1% or ≥1%, respectively, as recommended by the 2014 guidelines. Fisher’s exact testing was used to compare 30-day adverse outcomes, and the t test was used to compare mean length of stay. Receiver operating characteristic analysis was performed on NSQIP MICA risk calculator estimates and assigned RCRI scores of 0 to 4 modeled as a continuous variable using XLSTAT (version 2015.2.02.18579 copyright Addinsoft 1995 to 2016). The Penn State College of Medicine Institutional Review Board approved this study.
Results
There were 1,098 elective hip and knee replacement procedures performed during the study period reviewed. Patient characteristics, including risks determinants based on the RCRI and NSQIP MICA risk calculator, are included in Table 1 .
| Women | 616 (60%) |
| Surgery Type | |
| Total hip replacement | 412 (37.5%) |
| Total knee replacement | 686 (62.5%) |
| Revised Cardiac Risk Index Determinants | |
| Coronary artery disease | 135 (12.3%) |
| Congestive heart failure | 33 (3%) |
| Cerebrovascular disease | 58 (5.3%) |
| Insulin-dependent diabetes mellitus | 31 (2.8%) |
| Chronic kidney disease, creatinine > 2 | 11 (1%) |
| Exercise Capacity | |
| Good (≥ 4 METS) | 744 (67.8%) |
| Poor (< 4 METS) or Unknown | 354 (32.2%) |
| NSQIP MICA Risk Determinants | |
| Age (years) | 63 ± 11 |
| Creatinine (mg/dL) | |
| >1.5 | 25 (2.3%) |
| <1.5 | 1,072 (97.6%) |
| Unknown | 1 (0.1%) |
| ASA Physical Status Class | |
| 1 | 36 (3%) |
| 2 | 516 (47.3%) |
| 3 | 512 (47%) |
| 4 | 26 (2.4%) |
| Not documented | 8 (0.7%) |
| Preoperative Function | |
| Totally independent | 782 (71.2%) |
| Partially dependent | 314 (28.6%) |
| Totally dependent | 2 (0.2%) |
Excluding 2 patients who died within 30 days, all other patients (1,096 of 1,096) had documented follow-up at 30 days or more after surgery. Sufficient data were present to estimate risk using the NSQIP MICA risk calculator in 1,091 of 1,098 patients. Calculated risk estimates using the NSQIP MICA risk calculator ranged from a minimum of 0% to a maximum of 8.27%, with a mean of 0.54% and standard error of 0.018%. There was no significant difference (p = 1) when comparing the observed incidence of perioperative MICA (7 of 1,098 patients, 0.64%) versus the expected incidence (0.54%) based on mean risk estimate of the cohort using the NSQIP MICA risk calculator.
There were 922 of 1,098 patients (84%) estimated to be at low risk (<1%) for perioperative myocardial infarction or cardiac arrest by the NSQIP MICA risk calculator, compared with 169 (15.4%) estimated to be at elevated risk. When applying the RCRI criteria, 885 of 1,098 of patients (80.6%) were estimated to be at low risk (RCRI score of 0), compared with 213 (19.4%) estimated to be at elevated risk with an RCRI score of 1 or greater.
The c-statistic for discriminating 30-day MICA for the NSQIP MICA risk calculator was 0.85, with a 95% CI of 0.67 to 1. The optimal cut point for discriminating perioperative myocardial infarction or cardiac arrest for the NSQIP MICA risk calculator was 0.69%, with a sensitivity of 100% and a specificity of 75%. The c-statistic for the RCRI was 0.9, with a 95% CI of 0.75 to 1. The optimal cut point for the RCRI was a score of 1 or greater, with a sensitivity of 100% and specificity of 81%. A comparison of perioperative outcomes discrimination using these cut points from the receiver operating characteristic analysis is presented in Table 2 .
| Perioperative Outcome | Total N=1098 | NSQIP MICA < 0.69% N=809 | NSQIP MICA ≥0.69% N=282 | P value | RCRI = 0 N=885 | RCRI ≥ 1 N=213 | P value |
|---|---|---|---|---|---|---|---|
| Myocardial Infarction and/or Cardiac Arrest | 7 (0.6%) | 0 (0%) | 7 (2.5%) | <0.0001 | 0 (0%) | 7 (3.3%) | <0.0001 |
| Death from any Cause | 2 (0.2%) | 0 (0%) | 2 (0.7%) | 0.07 | 1 (0.1%) | 1 (0.5%) | 0.35 |
| Troponin >99th Percentile | 24 (2.2%) | 5 (0.6%) | 19 (7.2%) | <0.0001 | 10 (1.1%) | 14 (7%) | <0.0001 |
| Readmission | 39 (2.6%) | 21 (2.7%) | 18 (6.8%) | 0.005 | 29 (3.3%) | 10 (4.7%) | 0.31 |
| Length of Stay, mean days | 2.6 | 2.5 | 2.9 | <0.0001 | 2.5 | 2.9 | 0.0002 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
