Canadian Health&Care Mall: Outcomes of Predictors and Early and Late Outcomes of Respiratory Failure in Contemporary Cardiac Surgery

March 20, 2016 Category: Canadian Health Care Mall

CABG surgeryA total of 5,798 patients were included. Mean age was 64 ± 14 years, and 62% of patients (n = 3,612) were male. Patient demographics are shown in Table 1. CABG surgery was performed in 47% (n = 2,749), whereas 53% of patients underwent other procedures such as valve surgery (22%, n = 1,280), combined valve/CABG procedures (16%, n = 934), and aortic surgery (15%, n = 835). Mean predicted Euro-SCORE mortality rate was 11 ± 13%.

Incidence of RF

The overall incidence of RF was 9.1% (n = 529). The rate of RF was different according to surgical procedures (Fig 1). The incidence of RF did not differ significantly in the conventional CABG group (5.8%, n = 136) compared to the off-pump group (7.2%, n = 30) [p = 0.160]. The incidence of RF was higher following multiple-valve surgery (17%, n = 63) compared to single-valve procedures (6.1%, n = 55) [p < 0.001]. The rate of RF after aortic surgery involving the aortic arch was 22.4% (n = 50). A total of 223 tracheostomies (42%) were performed in patients with RF. The incidence of RF following stratification using the EuroSCORE is shown in Figure 2.


Predictors of RF

Differences in patient characteristics in univariate analysis are shown in Table 1. Patients with RF had a prolonged CPB time (196 ± 82 min vs 153 ± 70 min, p < 0.001) and aortic cross-clamp time (136 ± 61 min vs 115 ± 50 min, p < 0.001). Patients with RF were more likely to undergo aortic and combined valve/CABG procedures.

Stepwise multivariate logistic regression analysis of preoperative factors revealed female gender, age > 70 years, ejection fraction (EF) < 30%, hypertension, peripheral vascular disease (PVD), COPD, congestive heart failure, renal failure, active endocarditis, reoperation, hemodynamic instability, and intraaortic balloon pump (IABP) insertion as independent predictors for the occurrence of RF (Table 2). Aortic surgery and combined valve/CABG procedures were the only procedure-related variables that independently predicted RF. The Hosmer-Lemeshow goodness-of-fit test was not statistically significant (p = 0.38), indicating good calibration of the model. Furthermore, the C-statistic including the independent predictors of RF from the logistic regression analysis (Table 2) was performed. The area under the receiver operating characteristic curve for the multivariate prediction model was 0.81 and indicated the predictive power of the regression analysis. Canadian Health&Care Mall http://healthcaremall4you.com improves your health state fast and effective.

Outcome of Patients With RF

Overall hospital mortality rate among RF patients was 15.5% (n = 82), compared to 2.4% (n = 126) in the no-RF group (p < 0.001). Patients with RF were more likely to have other major complications (Table 3). The median length of hospital stay was significantly increased in RF patients (30 days [IQR, 17 to 52 days] vs 7 days [IQR, 5 to 10 days], p < 0.001).

The increased mortality associated with RF was observed across all EuroSCORE groups. Although there were more RF-related deaths in patients with a high EuroSCORE, the relative risk of dying after this complication was significantly higher in patients with low EuroSCORE (Table 4). In other words, following the development of RF, patients in the low EuroSCORE group had a 15-times greater risk of hospital mortality, whereas in patients with high EuroSCORE the risk was twice as high. Preoperative renal failure, reexploration for bleeding, and postoperative renal failure requiring dialysis were predictors of hospital mortality in RF patients (Table 5).

Late Survival

Mean follow-up was 4.2 ± 2.5 years. Long-term survival of RF patients was significantly decreased compared to those without this complication (Fig 3). When long-term survival was stratified by the Euro-SCORE, RF appeared to have the least impact on long-term survival in patients with a EuroSCORE — 3 (Fig 4, top left, A). In patients with a EuroSCORE > 3, 5-year survival was reduced by 30 to 50% (Fig 4, top right, B, through bottom right, D). Independent predictors of decreased long-term survival were age > 70 years, PVD, and renal failure requiring dialysis (Table 6).

Table 1—Patient Demographics

Variables No-RFGroup
(n = 5,269)
RFGroup
(n = 529)
p Value
Mean age, yr 63.2 ± 14 68.4 ± 13 < 0.001
Age > 70 yr 1,197 (23) 197 (37) < 0.001
Female gender 1,936 (37) 243 (46) < 0.001
Mean body mass index, 26.7 ± 5 26.8 ± 5 0.593
kg/m2
Body mass index > 30 1,144 (22) 105 (20) 0.346
kg/m2
Mean EF, % 4±
7
4
02
±
2
4
< 0.001
EF < 30% 810 (15) 155 (29) < 0.001
Hypertension 2,572 (49) 294 (56) 0.003
Diabetes 1,345 (26) 156 (29) 0.054
PVD 512 (10) 103 (19) < 0.001
COPD 319 (6) 70 (13) < 0.001
Congestive heart failure 1,543 (29) 287 (54) < 0.001
Previous myocardial 1,657 (31) 211 (40) < 0.001
infarction
Previous stroke 381 (7) 64 (12) < 0.001
Renal failure 224 (4) 63 (12) < 0.001
Extensive calcified aortic 157 (3) 20 (4) 0.516
disease
Previous cardiac 617(12) 134 (25) < 0.001
operation
Active endocarditis 110(2) 30 (6) < 0.001
Hemodynamic instability 75(1) 44 (8) < 0.001
IABP 121 (2) 38(7) < 0.001
Emergent procedure 244 (5) 57(11) < 0.001
Procedures
CABG 2,589 (49) 160 (30) < 0.001
Combined valve/ 796 (15) 138 (26) < 0.001
CABG
Simple and multiple 1,162 (22) 118 (22) 0.912
valves
Aortic procedure 722 (14) 113 (21) < 0.001
Mean EuroSCORE 10 ± 12 22 ± 19 < 0.001
Low risk (< 3%) 1,340 (25) 35 (7) < 0.001
Moderate risk 2,141 (41) 134 (25) < 0.001
(3 to 9%)
High risk (9 to 25%) 1,285 (24) 178 (34) < 0.001
Very high risk 503 (10) 182 (34) < 0.001
(> 25%)

Table 2—Independent Predictors of RF Following Cardiac Surgery (Multivariate Logistic Regression)

Variables OR(95% Confidence Interval) p Value
Hemodynamic instability 3.2 (1.8-5.5) < 0.001
Aortic procedure 2.6 (1.7-4.0) < 0.001
IABP 2.6 (1.5-4.7) 0.001
Renal failure 2.3 (1.5-3.5) < 0.001
Active endocarditis 2.1 (1.1-3.9) 0.025
Congestive heart failure 1.8 (1.4-2.4) < 0.001
Reoperation 1.7 (1.2-2.4) 0.001
PVD 1.7 (1.2-2.3) 0.002
COPD 1.7 (1.1-2.5) 0.012
Age > 70 yr 1.6 (1.22.2) < 0.001
Combined valve/CABG 1.6 (1.1-2.4) 0.023
surgery
Diabetes 1.5 (1.1-2.1) 0.005
EF — 30% 1.4 (1.1-1.9) 0.022
Female gender 1.4 (1.1-1.8) 0.006

Table 3—Postoperative Complications (Univariate Analysis)

Complications No-RF Group (n = 5,269) RF Group (n = 529) p Value
Hospital mortality 126 (2.4) 82 (15.5) < 0.001
Stroke 69 (1) 75 (14) < 0.001
Myocardial infarction 26 (0.5) 13(2.5) < 0.001
Renal failure, dialysis 20 (0.4) 82 (16) < 0.001
Deep sternal wound infection 36(1) 55 (11) < 0.001
Postoperativesystemic infection 41(1) 131 (25) < 0.001
Reexploration for bleeding 96 (2) 43 (8) < 0.001
GI 21 (0.4) 64 (12) < 0.001
Median (IQR) length of stay, d 7 (5-10) 30(17-52) < 0.001

Table 4—Impact of EuroSCORE Risk Group on Hospital Mortality in Patients With or Without RF

RiskGroups Mortality Rate in No-RF Group, % Mortality Rate in RF Group, % OR (95% Confidence Interval) p Value
Low 0.8 11.4 15.6 (4.7-51.7) < 0.001
Moderate 1.2 14.9 14.9 (8.0-27.5) < 0.001
High 3.4 14.6 4.8 (2.9-8.1) 0.004
Very high 9.3 17.6 2.1 (1.3-3.4) < 0.001

Table 5—Predictors of Hospital Mortality in Patients With RF

Predictors Univariate Multivariate
OR(95% Confidence Interval) p Value OR(95% Confidence Interval) p Value
Preoperative renal failure, dialysis 3.63 (2.02-6.52) < 0.001 4.09(1.46-9.48) 0.001
IABP insertion 2.41 (1.14-5.08) 0.021 2.03 (0.65-6.30) 0.221
Postoperative renal failure, dialysis 3.75 (2.19-6.43) < 0.001 1.94 (2.26-8.10) 0.001
Reexploration for bleeding 2.28 (1.12-4.65) 0.024 4.28 (1.24-7.32) 0.029

Table 6—Predictors of Long-term Mortality in Patients With RF

Predictors OR (95% Confidence Interval) p Value
Age > 70 yr 2.1 (1.3-3.7) 0.005
PVD 2.8 (1.4—5.7) 0.005
Renal failure, dialysis 5.7 (1.6-20.0) 0.006

Figure 1. The incidence of RF by procedure performed.

Figure 1. The incidence of RF by procedure performed.

Figure 2. Incidence of RF by EuroSCORE mortality risk (linear-by-linear test, p = 0.001).

Figure 2. Incidence of RF by EuroSCORE mortality risk (linear-by-linear test, p = 0.001).

Figure 3. Postdischarge survival curves for RF (n = 529) and no-RF (n = 5,269) patients. Top, A: Kaplan-Meier unadjusted data. Bottom, B: Cox proportional hazards-adjusted curves.

Figure 3. Postdischarge survival curves for RF (n = 529) and no-RF (n = 5,269) patients. Top, A: Kaplan-Meier unadjusted data. Bottom, B: Cox proportional hazards-adjusted curves.

9 to 25%). Bottom right, D: Very high risk (EuroSCORE > 25%).” src=”http://ithstats.com/blog/wp-content/uploads/2015/05/713-4-300×251.jpg” alt=”Figure 4. Postdischarge survival by EuroSCORE mortality risk category (mean follow-up of 4.2 years). Top left, A: Low risk (EuroSCORE < 3%). Top right, B: Moderate risk (EuroSCORE 3 to 9%). Bottom left, C: High risk (EuroSCORE > 9 to 25%). Bottom right, D: Very high risk (EuroSCORE > 25%).”

Figure 4. Postdischarge survival by EuroSCORE mortality risk category (mean follow-up of 4.2 years). Top left, A: Low risk (EuroSCORE < 3%). Top right, B: Moderate risk (EuroSCORE 3 to 9%). Bottom left, C: High risk (EuroSCORE > 9 to 25%). Bottom right, D: Very high risk (EuroSCORE > 25%).