Long-Term Mortality After Pneumonia in Cardiac Surgery Patients

Abstract

Background:

The role that intensive care unit (ICU)-acquired pneumonia plays in the long-term outcomes of cardiac surgery patients is not well known. This study examined the association of pneumonia with in-hospital mortality and long-term mortality after adult cardiac surgery.

Methods:

A total of 2750 patients admitted to our ICU after cardiac surgery from January 2003 to December 2009 are the basis for this observational study. Patients who developed ICU-acquired pneumonia were matched with patients without it in a 1:2 ratio. The matching criteria were age, urgent or scheduled surgery, surgical procedure, and the propensity score for pneumonia. Multiple regression analysis was used to find predictors of hospital mortality. The relationship between pneumonia and long-term survival was analyzed with Kaplan-Meier survival estimates and a risk-adjusted Cox proportional regression model for patients discharged alive from hospital.

Results:

Pneumonia was diagnosed in 32 (1.2%) patients and there were 19 cases per 1000 days of mechanical ventilation. Patients with pneumonia had a significantly higher hospital mortality rate (28% vs 6.2%, P = .003) and a higher mortality at the end of follow-up (53% vs 19%, P < .0001) than those without it. Regression analysis showed that pneumonia was a strong predictor of hospital mortality. Five-year survival was as follows: pneumonia, 62%; control, 81%; and cohort patients, 91%. The Cox model showed that, after adjusting for confounding factors, patients with pneumonia (hazard ratio = 3.96, 95% confidence interval [CI]: 1.41-11.14) had poorer long-term survival.

Conclusion:

Pneumonia remains a serious complication in patients operated for cardiac surgery and is associated with increased hospital mortality and reduced long-term survival.

Keywords

lung infection outcomes postoperative care statistic propensity matching surgery complications

Pneumonia develops in 1.2% to 9.7% of patients who undergo major cardiac surgery and is associated with severe outcomes during hospital stay^1

–5 and increased health care costs.⁶ However, knowledge about the effects of hospital-acquired pneumonia (HAP) on the outcome of cardiac surgery patients after hospital discharge is scant. Few available studies^7
–9 reporting the long-term outcome of intensive care unit (ICU) patients who develop ventilator-associated pneumonia (VAP) show a high mortality rate, prolonged period of mechanical ventilation and hospitalization, and increased medical costs. Interpretation of the results of studies is limited by the heterogeneous cohorts of patients included and because patients with VAP frequently have underlying disease that leads to poor long-term outcome of death even if VAP did not occur.

Several studies on the effects of community-acquired pneumonia (CAP) after hospital discharge have clearly demonstrated that it is also associated with long-term mortality.^10,11 The claim that comorbid conditions preceding pneumonia explain such increased long-term mortality is contentious.¹² However, it has been stated that severe cardiac complications that develop after discharge are responsible for such bad outcomes in patients with CAP.¹³

The purpose of this study was to determine the effects of ICU-acquired pneumonia on hospital and long-term mortality in adult patients undergoing cardiac surgery.

Materials and Methods

Patients

Between January 2003 and December 2009, all patients who underwent major cardiac surgery in our hospital and were admitted to the cardiac surgery ICU were prospectively enrolled in a clinical registry. Our hospital is a 900-bed general hospital serving as a cardiac surgery referral center for a population of 1 000 000 inhabitants. The local institutional review board approved the study and waived the requirement for informed consent.

We studied adult patients (age > 17 years) operated under extracorporeal circulation (ECC) or surgical revascularization without ECC and who stayed in the ICU at least for 48 hours. Cardiac transplantation is not performed in our hospital. Exclusion criteria were the presence of pneumonia or tracheobronchitis up to 15 days before cardiac surgery. The population and data collection methods have been previously described in detail.¹⁴ Briefly, clinical preoperative morbidities, operative, and postoperative data were prospectively recorded according to a preestablished protocol by a cardiac surgeon and an intensivist. The surgical risk was evaluated by the logistic European System for cardiac operative risk evaluation (EuroSCORE). Preoperative anemia was defined as hemoglobin <12 g/dL for women and <13 g/dL for men as measured by the laboratory on the day before surgery. Preoperative renal dysfunction was defined as an estimated glomerular filtration rate (GFR) <60 mL/min/1.73 m². Estimated GFR was calculated following the abbreviated Modification of Diet in Renal Disease formula:

\begin{aligned} G F R & = 186.3 \times (s e r u m c r e a t i n i n e)^{- 1.154} \\ \times {a g e}^{- 0.203} \times 0.742 (i f f e m a l e) \end{aligned}

Postoperative morbidity in the ICU was defined as cardiac, renal, neurologic, pulmonary, gastrointestinal, and vascular complications and postoperative bleeding as any bleeding that required surgical reoperation after initial departure from the operating theater. Postoperative acute myocardial infarction (AMI) was defined as a new Q on electrocardiogram (ECG) and/or typical acute ischemic ECG changes. Stroke was defined as a focal brain injury persisting for at least 24 hours confirmed by computed tomography scan. Acute renal failure was defined according to the Acute Kidney Injury Network criteria¹⁵: a 50% increase in serum creatinine level or a serum creatinine increase ≥0.3 mg/dL compared with baseline. We also recorded the requirements for hemofiltration during the ICU stay.

Pneumonia in the ICU was clinically diagnosed upon the appearance of new or progressive and persistent pulmonary infiltrates on 2 or more chest serial radiographs and 2 or more of the following criteria: purulent respiratory secretions, leukocytosis (>12 000 white blood cell [WBC]/µL) or leukopenia (<4000 WBC/µL), fever (>38°C), or worsening oxygen saturation (a reduction in the Pao ₂/Fio ₂ ≥ 15%). After pneumonia was clinically diagnosed, a respiratory sample was obtained for culture and empirical antimicrobial therapy began promptly.

Microbiology

Sampling of tracheal aspirates for respiratory tract infection assessment was done whenever pneumonia was clinically suspected. Specimens of endotracheal aspirates and blood cultures were transported to the laboratory immediately after collection. A positive tracheal culture aspirate was confirmed when yielded ≥10⁵ colony-forming units/mL of a pathogenic microorganism.

Surgical and Postoperative Management

All patients underwent cardiac surgery through a median sternotomy. Antimicrobial prophylaxis for surgery consisted of 1 g cefazolin administered before surgery and every 8 hours thereafter up to 48 hours (patients with allergy to cefazolin received 1 g of vancomycin before surgery and every 12 hours thereafter). All patients routinely received tranexamic acid or epsilon-aminocaproic acid after induction of general anesthesia. All red blood cell (RBC) units transfused were leukoreduced.

Postoperative respiratory care included tracheal aspiration through the orotracheal tube, endotracheal cuff pressure maintained between 20 and 30 cm of water registered during each shift, patient placement in a semirecumbent position (30°-40° elevation of the head of the bed), oral care rinsing with chlorhexidine, and early extubation in patients awake and stable. Early mobilization and chest physiotherapy were implemented. Antacids were routinely administered.

Outcomes

The primary outcomes were in-hospital and long-term mortality of patients discharged alive from hospital. Mortality following hospital discharge was calculated from data on patient life status, as of December 31, 2011, provided by the registry of Spanish National Institute of Deaths and the Regional Registry of Deaths (Servei Balear d’Estadistica). We used patients who died in hospital to validate the accuracy of both sources. The cause of death was not assessed.

Secondary outcome measures included ICU length of stay (the period between cardiac surgery and discharge of the ICU) and hospital length of stay.

Matching and Propensity Score

The propensity score is a measure of the probability of developing pneumonia after cardiac surgery. Propensity scores were estimated for each patient using an extensive, nonparsimonious logistic regression model based on baseline covariates of the whole cohort of cardiac surgery patients (n = 2750). Selection of the variables for inclusion in the logistic regression model to identify factors associated with pneumonia was based on clinical judgments and P values <.05 obtained in the univariate analysis (results not shown).

Each patient with pneumonia was matched to 2 patients without it on fulfilling these items: age (±5 years), urgent or scheduled surgery, surgical procedure, and the closest propensity score within 0.02. Surgical procedure was classified as isolated coronary artery bypass graft (CABG), valve, combined, and other surgery (162 of 217 patients were operated of proximal aorta and aortic arch surgery). Matched control patients were removed from the data file for the next matching. Patients with tracheobronchitis were excluded from the control group during matching. We assess balance within the matched pairs using the standardized differences in covariate means and proportions.

Statistical Analysis

Summary statistics are presented as percentages in the case of categorical variables, as means with standard deviations in the case of continuous variables, and as the median and interquartile range if their distribution was skewed. Percentages were compared between both the groups with the use of the chi-square test for categorical variables and means with t-test and Mann-Whitney U test for continuous variables.

We used a multiple logistic regression analysis to identify predictors of hospital mortality for the whole group of patients entering variables in which the difference between both the groups had a significant level of P < .1 in the univariate analysis.

Survival curves were estimated with the Kaplan-Meier method in patients discharged alive from the hospital. Censoring occurred at the time of death or at the end of study. We use a log-rank test to assess differences in survival curves between the groups.

Analysis of the factors affecting survival of patients discharged alive from hospital was performed using Cox regression method and the proportional hazards assumption to determine whether pneumonia was associated with mortality after adjusting for potential confounding variables. Given that we used a matching analysis, we defined confounding variables as those chosen by their clinical relevance as predictors of mortality in the literature and simultaneously associated both with pneumonia (P ≤ .2) and mortality in the control group (P ≤ .3). The inclusion of the confounding variables in the adjusted Cox regression models was done when the change of the effect of pneumonia on mortality was greater than 10%. Of the different models obtained, we choose the one that best estimated the β coefficient (ie, lowest standard error). All statistical tests were 2-tailed and the level of significance was set at P < .05. The statistical analysis was done with SPSS, v 18.0 (SPSS Inc, Chicago, Illinois, USA).

Results

A total of 2835 patients were included in our clinical database. We excluded 2 patients with pneumonia before surgery and 83 patients staying in the ICU less than 48 hours. We were able to evaluate 2750 patients. Table 1 shows selected baseline characteristics of these patients. The cumulative incidence of pneumonia was 1.2% (32 of 2750) and incidence density of 19 cases per 1000 days of mechanical ventilation. The incidence of pneumonia was different according to surgical procedures, being greater (4.6%) in the other group patients than in isolated CABG surgery (0.5%), isolated valve surgery (1%), and combined CABG and valve surgery (1.4%). Hospital mortality was greater in patients with pneumonia than in patients without pneumonia (28% vs 1.1%, P < .0001).

Table 1.

Summary of Demographics and Clinical Diagnoses of Cardiac Surgery Patients.^a

Age, years (mean ± SD)	65 ± 11.7
Sex, male (n, %)	1857 (67.5)
Height, cm (mean ± SD)	164 ± 9.4
Weight, kg (mean ± SD)	76 ± 13.7
Smoking (n, %)	493 (18)
Arterial hypertension (n, %)	1725 (63)
Diabetes (n, %)	853 (31)
COPD (n, %)	359 (13)
Vascular disease (n, %)	223 (8)
Chronic renal failure (n, %)	758 (28)
Cerebrovascular disease (n, %)	241 (8.8)
Previous myocardial infarction (n, %)	746 (27)
Pulmonary hypertension (n, %)	121 (4.4)
Severe LV dysfunction (EF < 30%; n, %)	100 (3.6)
Previous cardiac surgery (n, %)	134 (4.9)
Preoperative anemia (n, %)	1006 (37)
Surgery (n, %)
1. CABG	1162 (42)
2. VR	884 (32)
3. VR + CABG	487 (18)
4. Other	217 (8)
Type of surgery (n, %)
1. Scheduled	2523 (92)
2. Urgent	227 (8)
Surgery without ECC (n, %)	93 (3.4)
Euroscore, % (median, IQ range)	4.4 (2.4-7.8)
RBC transfusion (n, %)	1987 (72)
Length of MV, hours (median, IQ range)	5,5 (4.2-74)
Postoperative acute myocardial infarction (n, %)	79 (3)
Postoperative pneumonia (n, %)	32 (1.2)
Acute kidney injury (n, %)	536 (19.5)
Reoperation (n, %)	55 (2)
Hospital mortality (n, %)	39 (1.4)
Hospital length of stay, days (median, IQ range)	11 (8-18)
ICU length of stay, days (median, IQ range)	2 (2-4)

Abbreviations: CABG, coronary artery bypass graft; COPD, chronic obstructive pulmonary disease; ICU, intensive care unit; IQ, interquartile; LV, left ventricular; EF, ejection fraction; ECC, extracorporeal circulation; MV, mechanical ventilation; RBC, red blood cell; VR, valvular replacement; SD, standard deviation.

^an = 2750.

Table 2 shows baseline characteristics of the matched study patients. Most of all clinical covariates were well balanced between patients with ICU-acquired pneumonia and control patients except for previous cardiac surgery. There were differences (Table 3) in variables related to the operative characteristics such as greater logistic EuroSCORE, longer ECC time, and increased number of RBC units transfused in patients with pneumonia. In all, 1 (3.1%) patient did not receive transfusion of RBCs in the pneumonia group and 16 (31%) in the control group. Patients with pneumonia develop more severe postoperative complications (Table 4), either cardiac or noncardiac excluding AMI. The length of mechanical ventilation was much longer in patients with pneumonia than in control patients.

Table 2.

Baseline Characteristics of Patients With Pneumonia and Matched Control Patients.

	Pneumonia, n = 32	Control Patients, n = 64	P
Propensity score (mean ± SD)	0.0314 ± 0.03182	0.0301 ± 0.0266	.81
Age, years (mean ± SD)	67 ± 12.6	68 ± 11.6	.84
Sex, male (n, %)	25 (78)	48 (75)	.73
Height, cm (mean ± SD)	165 ± 10.4	165 ± 8.5	.82
Weight, kg (mean ± SD)	77 ± 12.7	76 ± 10.2	.60
BMI (mean ± SD)	29 ± 5.1	28 ± 3.7	.46
Smoking (n, %)	6 (19)	6 (9)	.19
Arterial hypertension (n, %)	23 (72)	47 (73)	.9
Diabetes (n, %)	9 (28)	21 (33)	.6
COPD (n, %)	9 (28)	22 (34)	.44
Vascular disease (n, %)	2 (6)	5 (8)	.78
Chronic renal failure (n, %)	7 (22)	11 (17)	.6
Cerebrovascular disease (n, %)	1 (3)	3 (5)	.72
Previous AMI (n, %)	8 (25)	14 (22)	.70
Severe PH (n, %)	4 (12)	11 (17)	.55
Severe LV dysfunction (EF < 30%; n, %)	4 (12)	7 (11)	.8
Normal LV function (EF > 50%; n, %)	19 (59)	42 (66)	.5
Previous cardiac surgery (n, %)	6 (19)	1 (2)	.002
Preoperative anemia (n, %)	12 (37)	30 (47)	.4
Basal Hb (g/dL; mean ± SD)	12.9 ± 2.0	12.9 ± 1.5	.96
Basal creatinine (mg/dL; mean ± SD)	1.53 ± 0.87	1.26 ± 0.82	.13

Abbreviations: AMI, acute myocardial infarction; BMI, body mass index; COPD, chronic obstructive pulmonary disease; EF, ejection fraction; Hb, hemoglobin; PH, pulmonary hypertension; LV, left ventricular; SD, standard deviation.

Table 3.

Operative Characteristics of Cardiac Surgery Between Propensity-Matched Groups.

	Pneumonia, n = 32	Control Patients, n = 64	P
Surgery, n			1
1. CABG	6	12
2. VR	9	18
3. VR + CABG	7	14
4. Other^a	10	20
Type of surgery (n, %)			.48
1. Scheduled	24 (75)	52 (81)
2. Urgent	8 (25)	12 (19)
Logistic Euroscore (mean ± SD)	18 ± 16.8	11 ± 8.5	.09
Median (IQ range)	10 (5 – 30)	9 (4 – 21)	.09
ECC time, minutes (mean ± SD)	147 ± 55.2	108 ± 40.9	<.001
Ischemia time, minutes (mean ± SD)	105 ± 47.6	81 ± 28.2	.001
Number of coronary grafts (mean ± SD)	2.4 ± 1.3	2.1 ± 1.0	.5
Red blood cells (units; mean ± SD)	8.8 ± 5.4	4.0 ± 3.5	<.0001

Abbreviations: CABG, coronary artery bypass graft; VR, valvular replacement; IQ, interquartile; SD, standard deviation.

^aproximal aorta and aortic arch surgery in 28 of 30, 93%.

Table 4.

Postoperative Complications in the ICU Between Propensity-Matched Groups.

	Pneumonia, n = 32	Control Patients, n = 64	P
MV, hours, median (IQ range)	172 (33-640)	6 (4-17)	<.0001
MV > 72 hours (n, %)	9 (28)	3 (5)	<.0001
Atrial fibrillation (n, %)	12 (35)	10 (16)	.002
Acute myocardial infarction (n, %)	2 (6)	4 (6)	1
Cardiogenic shock (n, %)	5 (16)	2 (3)	.02
Septic shock (n, %)	10 (31)	1 (2)	<.001
Acute kidney injury (n, %)	19 (59)	5 (8)	<.001
Hemofiltration (n, %)	8 (25)	0	<.0001
Reintubation (n, %)	21 (66)	3 (5)	<.0001
Tracheostomy (n, %)	8 (25)	1 (1.6)	<.001
Hemorrhagic shock (n, %)	6 (19)	1 (2)	.002
Reoperation (n, %)	7 (22)	1 (2)	.002

Abbreviations: ICU, intensive care unit; IQ, interquartile; MV, mechanical ventilation.

Microbiology

Pneumonia was diagnosed after 5.6 ± 4.61 days (median 4 days) of cardiac surgery. Considering the day of hospital admission, early pneumonia (within 4 days of hospital stay) was diagnosed in 22% (7 of 32) of the patients. Overall, 63% were caused by Enterobacteriaceae and 27% were caused by Pseudomonas aeruginosa. Pneumonia was polymicrobial in 26% of the cases.

Outcomes

Hospital mortality rate among patients with pneumonia (Table 5) was 28% and 6.2% in the matched control group (P < .0001). In early-onset pneumonia, hospital mortality was similar to late-onset pneumonia, 32% versus 25%, respectively. At the end of follow-up, 17 (53%) of 32 patients with pneumonia died against 12 (19%) of 64 in controls (P < .0001).

Table 5.

Outcomes Variables Between Propensity-Matched Groups.

	Pneumonia, n = 32	Control patients, n = 64	P
ICU stay (days; mean ± SD)	26 ± 21.0	4.6 ± 5.3	<.0001
Median (IQ range)	19 (8-41)	3 (2-5)	<.0001
Hospital stay (days; mean ± SD)	44 ± 26.8	16 ± 11.7	<.0001
Median (IQ range)	41 (24-58)	11 (9-18)	<.0001
Hospital mortality, n (%)	9 (28)	4 (6.2)	.003

Abbreviations: ICU, intensive care unit; IQ, interquartile; SD, standard deviation.

Risk Factors for Hospital Mortality

The multiple logistic regression analysis of the whole cohort (n = 2750) found the following variables to be independent predictors of hospital mortality: older age, pneumonia, AMI, RBC units transfusion and acute kidney injury. The following variables were predictors of pneumonia according to the multiple logistic regression analysis of the whole cohort: severe left ventricular dysfunction (HR 3.47, 95% confidence interval [CI] 1.08-11.1, P = .037), chronic pulmonary obstructive disease (HR 2.68, 95% CI 1.2-6.1, P = .19), and number of RBC transfusion (HR 1.26 95% CI 1.18-1.34, P < .0001).

Survival curves are shown in Figure 1. Patients with pneumonia had a 1-year and 5-year survival rate of 83% and 62%, respectively, compared to a 1-year and 5-year survival rate of 98% and 81%, respectively, in control patients. The rest of the cohort (n = 2628 patients) had a 1-year and 5-year survival rate of 98% and 91%, respectively. Patients with pneumonia discharged alive from hospital (n = 23) showed a median follow-up time of 3.3 years, those in the control group (n = 60) of 4.4 years, and those in the cohort patients of 4.8 years.

Figure 1.

Survival curves of patients discharged alive from hospital: cohort of patients without pneumonia (….), matched controls (___), and patients with pneumonia (- - - - - -). Number at risk at 0, 1 year, 3 years, and 5 years was 2628, 2577, 2090, and 1248 in cohort patients; 60, 59, 43, and 25 in controls; and 23, 19, 13, and 6 in patients with pneumonia. Kaplan-Meier analysis showed that the survival time of matched control was significantly higher than in patients with pneumonia (P < .0001).

According to the Cox proportional hazards model applied to matched control and patients with pneumonia discharged alive from hospital, after adjustment for confounding variables such as AMI and preoperative anemia, risk of long-term mortality associated with pneumonia was high, HR 3.96 (95% CI, 1.41-11.14).

Discussion

This study shows that patients undergoing cardiac surgery and acquiring pneumonia in the ICU have a higher hospital mortality rate and a lower 5-year survival than matched control patients without pneumonia.

The reported incidence of postoperative pneumonia in patients undergoing major cardiac surgery in the recent years is very variable, ranging from 9.7% to 1.2%.^3,5,16,17 Our incidence is lower than those reported by previous studies.^1,3,17 We believe that these differences can be attributed to the different populations studied, different definitions of pneumonia, and different study designs and postoperative management. It has been repeatedly pointed out¹⁸ that the practical difficulty of making the diagnosis of postoperative pneumonia in cardiac surgery patients is challenging because many conditions among critically ill patients produce similar clinical signs: congestive heart failure, atelectasis, excessive fluid administration, acute adult respiratory distress syndrome, fibrosis, and chronic obstructive pulmonary disease.

Microorganisms causing pneumonia in the ICU vary considerably according to the different case mix of patients admitted into different ICUs. Common pathogens in our study included Enterobacteriaceae and P aeruginosa which are similar to other series.^3,16 Contrary to other studies, we did not find methicillin-resistant isolates or Acinetobacter spp. Polymicrobial pneumonia was present in 27% of the episodes in our study, very similar to a recent European study.¹⁶

In this study, hospital mortality in patients with pneumonia was multifactorial and higher than in matched control patients (28% vs 6.2%). Hospital mortality rates in patients undergoing cardiac surgery who develop VAP range from 49% to 24%.^3,17,19 Gaynes et al²⁰ determined that pneumonia occurred in 3.8% of patients undergoing isolated CABG operation with a hospital mortality rate of 26.6% but these patients were younger than ours. Our mortality was greater than that predicted by logistic EuroSCORE, when it is known that this score tends to overestimate hospital mortality of patients undergoing cardiac surgery.^14,21

Pneumonia was also associated with increased duration of mechanical ventilation, length of ICU, and hospital stay, similar to previous studies.^1,4,16 We did not find a better prognosis for early-onset pneumonia as others.²² However, both the populations studied and the design of the study were different since we did not adjust for the length of mechanical ventilation at the time the condition was diagnosed. Recent studies have demonstrated that the attributable mortality of VAP is much lower than previously assumed.^23,24 Pneumonia is a risk factor that can be modified and efforts to prevent it are worthwhile since in our study affected patients were nearly 4 times more likely to die than those who were not affected. A large contemporary study²⁵ has revealed that RBC transfusion was strongly associated with pneumonia following a dose–response relationship; therefore, a measure that might prevent postoperative pneumonia would be to restrict transfusion, especially in patients who are at high risk of it.

Studies regarding the long-term survival of ICU patients with pneumonia are scarce and all of them suggest a high long-term mortality. In patients with pneumonia who survived their hospital stay, Ranes et al⁸ estimated mortality at 1 and 5 years as 26% and 45%, respectively, and in-hospital mortality as 42%. In a more recent study,⁷ patients with HAP treated in the ICU and patients with VAP surviving hospital discharge showed a mortality rate at 1 and 5 years of 35% and 41%, respectively. However, in these studies there were no cardiac surgery patients or only a small proportion of them. In our study, 5-year mortality rate in patients with pneumonia was 38% and doubled the risk of dying similar to that of controls. Long-term survival of cardiac surgery patients with respiratory failure (mechanical ventilation needed >72 hours postoperatively) was very low, with a 5-year survival of 48%.²⁶

Long-term mortality has been studied in large series of patients with CAP either hospitalized or not. In hospitalized patients, Johnstone et al¹⁰ found that mortality at 1 year and at the end of study was 28% and 53%, respectively. A similar study¹¹ found the same mortality rate. Explanations for this bad outcome include the presence of subclinical inflammation¹² and of higher d-dimer and thrombin–antithrombin complexes when patients leave the hospital alive.²⁷ Major cardiac complications after CAP may play an important role in short- and long-term outcomes.^28,29 In another study, patients with CAP hospitalized in medical wards showed a 7-year mortality of 53%, considerably higher than in an age-matched and sex-matched general population. One explanation is that pneumonia may represent a marker of poor general health status. Causes of death in the follow-up were mostly comorbidity related and were different from those in controls.³⁰

The present report is the first describing the association between pneumonia and long-term mortality after cardiac surgery. Studies that developed long-term survival prediction models for cardiac surgery failed to evaluate the specific adverse impact of pneumonia as a risk factor and often they also missed some other relevant predictor factors like preoperative anemia and chronic renal failure defined by an estimated GFR. We cannot state that pneumonia is one of the potential mechanisms exposing these patients to increase long-term mortality since we were not able to collect information of the cause of death during the follow-up.

Propensity score matching is often used to balance 2 treatment groups but the method can also be used to balance patients across other exposures.³¹ Although we consider propensity score matching a strength of our study, it may not balance unmeasured confounders. We were able to match all patients with pneumonia with control patients without pneumonia, with acceptable matching in most covariates before surgery. The stratification of patients into surgical procedure subgroups provided well defined and homogeneous study groups. The relatively low number of patients with pneumonia included and the observational nature of our study are 2 limitations that restrict its external validity. The selection of patients as well as the management of the perioperative period may vary very widely in cardiac surgical units, specially the duration of mechanical ventilation after cardiac surgery, an important factor leading to pneumonia. Finally, we were not able to provide information on quality of life and cause of death after discharge.

Conclusions

This study shows that patients undergoing cardiac surgery who acquired pneumonia in the ICU have a bad hospital mortality and a poor long-term prognosis. Pneumonia is an independent predictor of late mortality following cardiac surgery.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Merck, Sharp and Dohme- Spain signed an agreement with Health Service of Balearic Islands to partly fund our data bank.

References

Bouza

Perez

Munoz

. Ventilator-associated pneumonia after heart surgery: a prospective analysis and the value of surveillance. Crit Care Med. 2003;31(7):1964–1970.

De Santo

Bancone

Santarpino

. Microbiologically documented nosocomial infections after cardiac surgery: an 18-month prospective tertiary care centre report. Eur J Cardiothorac Surg. 2008;33(4):666–672.

Kollef

Sharpless

Vlasnik

Pasque

Murphy

Fraser

. The impact of nosocomial infections on patient outcomes following cardiac surgery. Chest. 1997;112(3):666–675.

Leal-Noval

Marquez-Vacaro

Garcia-Curiel

. Nosocomial pneumonia in patients undergoing heart surgery. Crit Care Med. 2000;28(4):935–940.

Riera

Ibanez

Herrero

. Respiratory tract infections after cardiac surgery: impact on hospital morbidity and mortality. J Cardiovasc Surg (Torino). 2010;51(6):907–914.

Safdar

Dezfulian

Collard

Saint

. Clinical and economic consequences of ventilator-associated pneumonia: a systematic review. Crit Care Med. 2005;33(10):2184–2193.

Karhu

Ala-Kokko

Ylipalosaari

Ohtonen

Laurila

Syrjala

. Hospital and long-term outcomes of ICU-treated severe community- and hospital-acquired, and ventilator-associated pneumonia patients. Acta Anaesthesiol Scand. 2011;55(10):1254–1260.

Ranes

Gordon

Chen

. Predictors of long-term mortality in patients with ventilator-associated pneumonia. Am J Med. 2006;119(10):897–899.

Sligl

Eurich

Marrie

Majumdar

. Age still matters: prognosticating short- and long-term mortality for critically ill patients with pneumonia. Crit Care Med. 2010;38(11):2126–2132.

10.

Johnstone

Eurich

Majumdar

Jin

Marrie

. Long-term morbidity and mortality after hospitalization with community-acquired pneumonia: a population-based cohort study. Medicine (Baltimore) 2008;87(6):329–334.

11.

Mortensen

Kapoor

Chang

Fine

. Assessment of mortality after long-term follow-up of patients with community-acquired pneumonia. Clin Infect Dis. 2003;37(12):1617–1624.

12.

Yende

D'Angelo

Kellum

. Inflammatory markers at hospital discharge predict subsequent mortality after pneumonia and sepsis. Am J Respir Crit Care Med. 2008;177(11):1242–1247.

13.

Corrales-Medina

Suh

Rose

. Cardiac complications in patients with community-acquired pneumonia: a systematic review and meta-analysis of observational studies. PLoS Med. 2011;8(6):e1001048.

14.

Riera

Herrero

Ibanez

. Mid-term survival of patients undergoing major cardiac surgery. Rev Esp Cardiol. 2011;64(6):463–469.

15.

Mehta

Kellum

Shah

. Acute kidney injury network: report of an initiative to improve outcomes in acute kidney injury. Crit Care. 2007;11(2):R31.

16.

Hortal

Munoz

Cuerpo

Litvan

Rosseel

Bouza

. Ventilator-associated pneumonia in patients undergoing major heart surgery: an incidence study in Europe. Crit Care. 2009;13(3):R80.

17.

Tamayo

Alvarez

Martinez-Rafael

. Ventilator-associated pneumonia is an important risk factor for mortality after major cardiac surgery. J Crit Care. 2012;27(1):18–25.

18.

Klompas

. Does this patient have ventilator-associated pneumonia? JAMA. 2007;297(14):1583–1593.

19.

Hortal

Giannella

Perez

. Incidence and risk factors for ventilator-associated pneumonia after major heart surgery. Intensive Care Med. 2009;35(9):1518–1525.

20.

Gaynes

Bizek

Mowry-Hanley

Kirsh

. Risk factors for nosocomial pneumonia after coronary artery bypass graft operations. Ann Thorac Surg. 1991;51(2):215–218.

21.

Bhatti

Grayson

Grotte

. The logistic EuroSCORE in cardiac surgery: how well does it predict operative risk? Heart. 2006;92(12):1817–1820.

22.

Valles

Pobo

Garcia-Esquirol

Mariscal

Real

Fernandez

. Excess ICU mortality attributable to ventilator-associated pneumonia: the role of early vs late onset. Intensive Care Med. 2007;33(8):1363–1368.

23.

Timsit

Zahar

Chevret

. Attributable mortality of ventilator-associated pneumonia. Curr Opin Crit Care. 2011;17(5):464–471.

24.

Bekaert

Timsit

Vansteelandt

. Attributable mortality of ventilator-associated pneumonia: a reappraisal using causal analysis. Am J Respir Crit Care Med. 2011;184(10):1133–1139.

25.

Horvath

Acker

Chang

. Blood transfusion and infection after cardiac surgery. Ann Thorac Surg. 2013;95(6):2194–2201.

26.

Filsoufi

Rahmanian

Castillo

Chikwe

Adams

. Predictors and early and late outcomes of respiratory failure in contemporary cardiac surgery. Chest. 2008;133(3):713–721.

27.

Yende

D'Angelo

Mayr

. Elevated hemostasis markers after pneumonia increases one-year risk of all-cause and cardiovascular deaths. PLoS One. 2011;6(8):e22847.

28.

Corrales-Medina

Madjid

Musher

. Role of acute infection in triggering acute coronary syndromes. Lancet Infect Dis. 2010;10(2):83–92.

29.

Corrales-Medina

Musher

Wells

Chirinos

Chen

Fine

. Cardiac complications in patients with community-acquired pneumonia: incidence, timing, risk factors, and association with short-term mortality. Circulation. 2012;125(6):773–781.

30.

Bruns

Oosterheert

Cucciolillo

. Cause-specific long-term mortality rates in patients recovered from community-acquired pneumonia as compared with the general Dutch population. Clin Microbiol Infect. 2011;17(5):763–768.

31.

Ahmed

Zannad

Love

. A propensity-matched study of the association of low serum potassium levels and mortality in chronic heart failure. Eur Heart J. 2007;28(11):1334–1343.