Background: Cardiac surgery, even when planned, has the potential for adverse outcomes, such that several factors are taken into consideration to help surgeons and their patients discuss the potential risks weighed against the expected recovery. Preoperative functional status and its influence on cardiac surgery, if any, have not been adequately evaluated to date. This study aimed to examine the relationship between preoperative functional status and postoperative mortality and morbidity in elective open heart patients. Methods: Preoperative baseline data (n = 43 subjects) were obtained to calculate Society of Thoracic Surgeons (STS) mortality and morbidity risk scores and preoperative functional status was measured using the Late-Life Function and Disability Instrument (LLFDI). Follow-up data were abstracted at one year postoperative to calculate actual mortality and morbidity events. Ordinary least squares and negative binomial regression analyses were conducted to assess the relationship between the LLFDI preoperative score to the STS mortality and morbidity risk scores. Results: Mortality risk was significant, F (1, 39) = 4.75, p = 0.035, with an adjusted R2 = 0.086, and Function Total (measured by LLFDI) yielded a significant negative association with mortality risk, β = −0.329. Morbidity was found to be significant, F (1, 40) = 4.89, p = 0.033, with an adjusted R2 = 0.087 and Function Total yielded a significant negative association with morbidity risk, β = −0.328, as well. Estimation of the counts for postoperative complications as estimated by Function Total failed to reach significance (Wald χ2 = 0.34, p = 0.56), which provided a pseudo R2 = 0.009. Probabilities for frequencies of adverse events (major complications), therefore, could not be reliably calculated. Conclusion: Preoperative diminished functional status, as measured by the LLFDI, is associated with an increased risk of mortality and morbidity in patients undergoing elective cardiac surgery. The risks and benefits of cardiac surgery should be weighed carefully and include a patient's preoperative functional status, especially in the case of an elective procedure.
Coronary artery disease is the leading cause of death worldwide [
In research, “impaired functional status” has become somewhat synonymous with the term “frailty”, and its association with mortality and morbidity has been studied primarily in the inpatient, non-surgical population. Narain et al. studied older adult inpatients with varied diagnoses and concluded that, more than the admitting diagnosis [
From a clinical perspective, cardiac surgeons have universally accepted risk score assessment systems [
STS made a recommendation in May, 2011, that preoperative functional measures such as gait speed be added to the STS database for adult patients pending cardiac surgery, in order to aid in stratifying risk [
The gold standard for measuring physical function and capacity has been the six-minute walk test but it only accounts for the “physical” component [
As of 2010, the risk-adjusted mortality rate for isolated CABG was 2.1% [
A non-experimental design using a prospective cohort of subjects undergoing elective cardiac surgery from Saint Vincent Health Center between June and December 2010 was assembled. Preoperative baseline data was obtained to calculate mortality and morbidity risk and follow-up postoperative data was abstracted at one year to calculate actual mortality and morbidity events. Regression analysis was conducted to assess the relationship between the LLFDI preoperative score (independent/predictor variable) to the STS mortality and morbidity risk scores (dependent/outcome variables).
Subjects included in this study were at least 18 years old, able to communicate fluently in English, and underwent one of the following elective cardiac surgeries: initial or redo coronary artery bypass graft (CABG), valve repair/replacement, or any valve/CABG combination procedure. If elective cardiac surgery became emergency surgery or a subject failed to submit or sufficiently complete their preoperative LLFDI form, they were excluded/terminated from the study. Consecutive subjects were screened after they were scheduled to undergo cardiac surgery which involved CABG, valve repair/replacement, or valve and CABG combination surgery. Eligible subjects were asked to complete a LLFDI questionnaire preoperatively and this served as the predictor variable for this study. This study was approved by the human subjects review committees at Saint Vincent Health Center and Western Michigan University.
In this study, functional status was measured using only the Functional component of the LLFDI tool. The Disability component focuses more on recreational tasks and community socialization than the daily movements and limitations asked in the Function component [
The specific demographic and clinical patient information found in
Preoperative | Postoperative |
---|---|
Age | Specific cardiac surgical procedure |
Gender | Hospital length of stay |
Ethnicity | Postoperative bleeding (return to O.R.; blood products) |
Ejection fraction % (EF) | Deep sternal infection |
NYHA Classification | Intubation > 24 hours (and re-intubation) |
Creatinine level | Creatinine within 72 hours |
Prior cardiovascular surgery | Adverse arrhythmia (pacemaker/defibrillator required) |
Number of vessel disease | Total cross-clamp time |
Myocardial Infarct history | Mortality (all-cause, at 30-days & 1-year) |
Prior neurologic event | Neurologic event |
Co-morbidities | Co-morbidities (progressed/new diagnoses) |
Valve disease/insufficiency | |
Body Mass Index (BMI) |
calculator version 2.81 also provides a combined “mortality or morbidity risk” estimate (in percentage) from the same calculation method used to estimate mortality risk. For this study, the morbidity risk was calculated by subtracting the estimated “mortality risk” score from the “mortality or morbidity risk” score, since STS does not directly estimate morbidity on its own.
Actual adverse events, which comprise the basis for morbidity risk according to the STS, is a composite based on any of the following 5 major complications (found in “Postoperative Data” in
Because mortality and morbidity data is often extended beyond the patients’ hospitalization, including postoperative data at 30 days and as far out as one year [
All data were analyzed using the statistical package Stata 14 (StataCorp, College Station, TX). Study demographics were collected (
The study cohort consisted of 43 subjects with completed preoperative LLFDIs for analysis (
Functional status, mortality and morbidity risk scores were assessed for normality, however; both mortality and morbidity risk scores evidenced significant departure from normality when Shapiro-Wilk test was conducted. Quantile-normal plots were generated to determine best fit for transformation and log transformation was determined for both outcome variables. After the necessary variables were transformed, all the assumptions of normality were met for analysis purposes.
The regression of mortality risk on functional status was found to be significant, F (1, 41) = 4.96, p = 0.032, providing an adjusted R2 = 0.086. Function Total yielded a significant negative association with mortality risk, β = −0.328.
Patient Characteristics (based on n = 43) | LLFDI demographics (based on n = 150) | ||
---|---|---|---|
Gender | Female Male | 12 (28%) 31 (72%) | Female 23% Male 77% |
Age | Range 45 - 83 years (SD 9.74) M = 66.35* | Range 60 - 90 years | |
Race/Ethnicity | Caucasian 41 (95%) | 84% | |
Hispanic or Latino 1 (2%) | 5% | ||
African American 1 (2%) | 7% | ||
Surgical Approach | CABG―on pump 30 (72.1%) | --- | |
CABG― off pump 5 (11.6%) | --- | ||
CABG - both types combined, 35 (83.7%) | --- | ||
Valve replacement/repair alone or with CABG 8 (16.3%) | --- | ||
Preoperative Functional Status | LLFDI Function Total (score range 0 - 100) M = 61.39 (SD 9.41) | LLFDI Function Total (score range 0 - 100) M = 62.9 (SD 13.0) | |
Mortality Risk (%) | M = 1.47 (SD 1.31) | n/t | |
Morbidity Risk (%) | M = 10.23 (SD 4.27) | n/t | |
Late-Life Function and Disability Instrumentn/t = not tested. CABG = Coronary Artery Bypass Graft *compared to national mean age of 64.9 years for CABG (STS, 2010).
Regression diagnostics indicated that normality of residuals and homoscedasticity assumptions were met. Analysis to detect influential observations using Cook’s distance revealed the presence of two potentially influential cases. Both cases were deleted and the analysis was rerun. The equation for mortality (
The equation for morbidity risk was not significant, F (1, 41) = 2.66, p = 0.11, providing an adjusted R2 = 0.038. Function Total yielded a significant negative association with morbidity risk, β = −0.247. Normality of residuals and homoscedasticity assumptions were met. Analysis for the presence of influential
Variable | B | SE B | β | Adjusted R2? |
---|---|---|---|---|
Functional Total | −0.044 | 0.013 | −0.469* | 0.086 |
*p < 0.05; **p < 0.01; ╪Adjusted R2 was based on reverse transformation of the dependent variable; B = Slope; SE B = Standard Error of Slope; β = Standardized Slope.
observations using Cook’s distance revealed one potentially influential case. The regression equation for morbidity following deletion of this observation (
Results for the negative binomial regression analysis appear in
Preoperative functional status, as measured by LLFDI Function Total, yielded significant findings in predicting both mortality and morbidity risk in elective cardiac surgery. Although the LLFDI outcome measure has been widely used with the cardiac population (e.g., cardiac rehab post-bypass surgery, coronary heart disease, congestive heart failure) [
Despite the overall significance found with the LLFDI, there were two cases deemed the exception as they were flagged as unduly influencing the mortality risk results due to their individual functional scores: one was unexpectedly high and the other was extremely low. Consequently, the LLFDI was a poor predictor of mortality risk in these two cases. As stated earlier, there are numerous potential variables that can influence cardiac surgery complications and mortality. It is essential to isolate as many of the key influencing variables as possible when using surgical risk predictor tools such as STS, otherwise calculating the estimate may be flawed, and on elective procedures especially, accuracy is paramount. In an attempt to examine the effect of the LLFDI alone, there certainly could have been another unexplained covariate(s) acting or anomalies occurring within the accounted STS variables.
Variable | B | SE B | β | Adjusted R2? |
---|---|---|---|---|
Functional Total | −0.014 | 0.006 | −0.328* | 0.087 |
*p < 0.05; **p < 0.01; ╪Adjusted R2 was based on reverse transformation of the dependent variable; B = Slope; SE B = Standard Error of Slope; β = Standardized Slope.
Variable | B | SE B | Pseudo R2 |
---|---|---|---|
Functional Total | −0.031 | 0.053 | 0.009 |
*p < 0.05; ** p < 0.01; B = Slope; SE B = Standard Error of Slope.
The LLFDI overestimated morbidity in one case and consequently was a poor predictor of morbidity risk in this one case. Interestingly, this case was one of the two cases that was also problematic with regard to mortality risk. Overall, results of the regression analysis were an accurate reflection of the association between the variables in the vast majority of the cases. Furthermore, finding a significant association between the LLFDI Function Total and STS morbidity risk, as well as LLFDI Function Total and STS mortality risk, suggests that the STS cardiac risk score may need to be refined.
This study had limitations with attrition rate from obtaining consents to receiving preoperative LLFDIs of 29% (n = 22). One explanation may have been that subjects were met briefly and typically after they just received news of needing cardiac surgery without much time to process all of the information. Unfortunately, there is only a small window of time (roughly two hours) to approach these potential subjects in person after all of their consults and tests have been completed and before they are discharged. This issue could potentially be improved in future studies by conducting a follow-up call at home 1 - 2 days after the initial contact and prior to surgery.
Conducting this study at a single site with a sample of convenience (i.e., elective, primarily on-pump cases) without a control group certainly challenges the ability to generalize findings to the cardiac surgery population. Closer examination of the data revealed that all but five of the participating subjects that underwent CABG surgery (n = 35), were performed on-pump, and of the eight subjects who underwent valve repair or replacement, three of the valve procedures was a combination valve/CABG surgery. Additionally, two of the valve cases were mitral valve surgeries, which naturally carry higher mortality risk, however; all of these sub-groups were too small to do any comparative studies on mortality and morbidity influence. Regardless of the cardiac surgery performed, however, preoperative diminished functional status, as measured by the Late-Life Function and Disability Instrument, is associated with an increased risk of mortality and morbidity in patients undergoing elective cardiac surgery. The risks and benefits of cardiac surgery should be weighed carefully and include a patient’s preoperative functional status, especially in the case of an elective procedure.
The researchers involved in this study would like to thank all the participants who took part in this study and their families who assisted them along the way.
There were no conflicts of interests to disclose for any of the authors. This study was not funded by any grant or supported by any external source.
MacPhedran, A.K., Barker, D.B., Marbey, M.L., Fogarty, K. and Vangsnes, E. (2018) Is Preoperative Functional Status Associated with Postoperative Mortality and Morbidity in Elective Open Heart Patients? Health, 10, 654-666. https://doi.org/10.4236/health.2018.105051