ORIGINAL ARTICLE
312
Marcio Roberto Moraes de Carvalho1, Nelson Albuquerque de Souza e Silva1, Gláucia Maria Moraes de Oliveira1, Carlos Henrique Klein2
Complications and hospital length of stay in coronary artery bypass graft surgery in public hospitals in Rio de Janeiro Complicações e tempo de internação na revascularização miocárdica em hospitais públicos no Rio de Janeiro
1. Medical College, Universidade Federal do Rio de Janeiro - UFRJ - Rio de Janeiro (RJ), Brazil. 2. Escola Nacional de Saúde Pública/ Fundação Oswaldo Cruz - Fiocruz – Rio de Janeiro (RJ), Brazil.
This study is a part of the main project “Mortality in High-Complexity Cardiovascular Procedures in Rio de Janeiro,” with financial support from Fundação de Apoio a Pesquisa do Estado do Rio de Janeiro (FAPERJ) and using human resources from Universidade Federal do Rio de Janeiro (UFRJ) and Fundação Oswaldo Cruz (Fiocruz). Conflicts of interest: None. Submitted on January 19, 2011 Accepted on August 1, 2011 Corresponding author: Márcio Roberto Moraes de Carvalho Rua Domingues de Sá, 410 Zip Code: 24220-091 - Niterói - Rio de Janeiro (RJ), Brazil. Phone: + 55 21 8177-8173 / Fax: + 55 21 2710-9249 E-mail:
[email protected]
ABSTRACT Objective: To evaluate associations between post-operative complications in patients who survive surgery and inhospital deaths and lengths of hospital stays of patients who undergo coronary artery bypass graft surgery Methods: Patients who underwent coronary artery bypass graft surgery and survived the operating theater were randomly selected. Information on complications and hospital lengths of stay until hospital discharge or death were retrospectively collected based on medical records and declarations of death. These aspects were estimated according to the presence of complications, frequency of complications, mortality, relative risk and attributable population risk. Mean hospital lengths of stay were compared using Wald’s statistics. Results: Medical records indicating deaths in the operating theater were excluded, and 86.9% of the included records reported complications; the greatest loss of information (43.9%) was related to kidney
failure. Hyperglycemia was estimated as the most frequent complication (74.6%), with an attributable risk of 31.6%. The population’s attributable risks were greater than 60% for low cardiac output (77.0%), kidney failure (64.3%) and cardiorespiratory failure (60.4%). Twelve different situations were identified for paired combinations of significant differences between average post-operative hospitalization times and complications, according to the outcome of discharge or death. Conclusion: Several complications were identified during the postoperative period of coronary artery bypass graft surgery, with different frequencies and impacts on mortality. Control of the myocardium at the risk of ischemia, hemodynamic stabilization and volume replacement strategies may be effective for controlling mortality rates and shortening hospital lengths of stay. Keywords: Myocardial revascularization; Cardiac surgical procedures; Length of stay; Mortality; Postoperative complications
INTRODUCTION Between 1999 and 2003, the in-hospital mortality rate after coronary artery bypass graft surgery (CABG) at four public hospitals in Rio de Janeiro, Brazil, ranged between 7.4% and 14.3%.(1) Of 23 preoperative factors were analyzed, the 7 following factors had significant associations with death: age, arterial hypertension, smoking status, dyslipidemia, stroke and involvement of left coronary trunk associated with obstruction of main coronary systems.(1) Stoica et al.(2) emphasized that understanding intraoperative complications favors an understanding of adverse outcomes, especially in low-risk patients, thus improving the prediction of the risk of death. Michalopoulous et al.(3) observed that postoperative morbidity and lengths
Rev Bras Ter Intensiva. 2011; 23(3):312-320
Complications and length of stay in coronary artery bypass surgery
of stay was related to patients’ ages. They concluded that analysis of preoperative variables might contribute to a better understanding of surgery patients’ mortality. This study aimed to assess the associations between postoperative complications in patients surviving CABG surgery who later died in the hospital and hospital lengths of stay at four public hospitals in Rio de Janeiro, Brazil, between January 1999 and December 2003. We previously published our research on preoperative features’ associations with mortality,(1) the use of the EuroSCORE in this population(4) and the association of intraoperative features with time of death within one year of follow-up after surgery.(5) METHODS Coronary artery bypass graft (CABG) surgery patients were identified in the city of Rio de Janeiro based on hospital admissions and their corresponding codes. Surgeries performed between January 1999 and December 2003 were included; concomitant valve interventions were excluded. For patients who underwent more than one CABG surgery, the final selection included only the most recent procedure. Overall, 2,692 subjects from the four hospitals in the city of Rio de Janeiro, Brazil, were identified. Two of the hospitals were university hospitals, and two were cardiology reference centers; each pair was constituted by one federal and one state hospital. Thus, the groups were designated A, B, C and D. The sample’s selection criteria and the number of CABG patients selected from the four hospitals have been described elsewhere.(1) In summary, we randomly selected 600 medical charts, 150 from each hospital, with different sample mixes in terms of the outcome, i.e., discharge or death, as recorded in the admission documents. Patients who died in the operating room were excluded. Data from hospital charts were retrospectively searched between 2006 and 2007 by cardiologist investigators, who were trained for data collection using a standardized form that included social and demographic data, hospital admissions information, risk factors or comorbidities, complementary tests, medical prescriptions, surgery conditions, previous angioplasty or CABG, postoperative complications and hospital outcomes. The collected data were transferred into electronic forms using Epidata v. 3.1 software.(6) The variables’ definitions and their categorization criteria were provided on sheets for the investigators’ instruction and reference. The project was approved by the ethics committee of Hospital Clementino
313
Fraga Filho, Universidade Federal do Rio de Janeiro, under number 102/05. For this study, variables related to complications, mean length of stay and outcome (discharge or death) were selected. The complications were assessed using the procedure established by Oliveira et al.,(7) and they corresponded to 10 sets, according to the following characteristics: hemodynamic or hemorrhagic; cardiologic ischemic; cardiologic non-ischemic; mechanic; respiratory; metabolic; neurologic; vascular; and multiorgan failure. Data on the outcomes, in-hospital mortality or discharge, were retrieved from the charts. In-hospital death was considered to be death during the hospital stay, confirmed by a declaration of death.(8) The time between the surgery and discharge or death was estimated in days using the difference between the dates of surgery and discharge or death for the patients surviving the operation room. Statistical analysis Relative frequencies (percentages) and means were calculated inversely to the selection for sample probability, weighted according to each individual patient, and corrected by the medical charts factor. The weights, according to the patient’s hospital and outcome of discharge or death stratum, were: hospital A, survivals: 3.24, deaths 1; hospital B, survivals 15.84, deaths 2.51; hospital C, survivals 4.14, deaths 1.07; and hospital D, survivals 7.41, deaths 1.08. The relative frequencies of each complication in the patients’ sets were estimated, as well as the percentage mortality in the subgroups with/without complications. These percentage differences were tested as population attributable risk (PAR)(9) using weighted chi-squared distribution; values of p < 0.05 were considered significant. The relative risks (RR) of death were estimated according to the presence or absence of each complication, with corresponding confidence intervals and PAR(9) calculated for each complication in the patients’ sets. Mean times from surgery until discharge or death were calculated according to the presence/absence of complications and outcomes. The combination of the presence/absence of a given complication and a discharge/death outcome resulted in four means. We tested the differences between the observed means for the following pairs: survival and death without a given complication; survival and death with that complication; survival with and without that complication; nonsurviving patients with/without that complication. These mean differences were tested with Wald’s statistics
Rev Bras Ter Intensiva. 2011; 23(3):312-320
314
Carvalho MRM, Silva NAS, Oliveira GMM, Klein CH
and were weighted, and those differences with p < 0.05 were considered significant. Considering the number of complications and the diverse number of observations, we elected to analyze the pairs using p < 0.01 denoting statistical significance. RESULTS From the originally planned 600 medical charts, 546 were actually collected (91.0%); 22 were excluded due to death during surgery. Of the 524 remaining
charts, in 508 (86.9%), information on the analyzed complications was identified. Information loss was more important for metabolic complications. Data on acute renal failure were only found in 294 (56.1%) of the medical charts, on hypoglycemia in 435 (83.0%) of the charts and on hyperkalemia in 461 (88.0%) of the analyzed medical charts. The weightings corresponding to each chart, by hospital and outcome, were maintained for the analysis. Table 1 shows frequencies, mortality, relative and population attributable risks, according to
Table 1 – Frequency of complications and their associations with mortality Complication Low output Renal failure Cardiorespiratory arrest Arrhythmia Intraaortic balloon Other CA device Hyperglycemia Acute myocardial infarction Hyperkalemia Bleeding Blood transfusion Septic shock Heart failure MODS Coma Pneumonia Stroke ARDS Other sites’ infection Unstable angina Seizure Mediastinitis Respiratory prosthesis >24 h Free wall rupture Peripheral vascular insufficiency Psychiatric conditions Pulmonary thromboembolism Sepsis Tamponade Acute mitral reflux Pericarditis Soft tissue infection Amputation Deep vein thrombosis Limb revascularization Inter-ventricular defect
%
Frequency 95%CI
19.1 11.3 8.1 17.0 12.4 1.2 74.6 8.6 8.1 9.4 31.0 2.1 7.4 1.4 1.3 6.9 2.3 1.6 4.8 4.2 0.8 1.3 12.4 0.2 2.2 2.6 0.3 1.9 0.5 0.1 0.6 2.3 0.2 0.3 0.8 -
15.8-22.5 7.9-14.7 6.7-9.6 13.0-20.9 9.4-15.5 0.4-1.9 69.2-80.0 6.2-11.0 5.3-11.0 6.3-12.4 25.2-40.0 1.3-3.0 4.8-10.1 0.8-2.1 0.7-1.8 4.1-9.7 1.3-3.4 0.7-2.5 2.7-6.8 2.0-6.5 0.1-1.5 0.4-2.1 9.6-15.2 0.0-0.5 0.4-4.1 0.9-4.2 0-0.6 1.1-2.7 0.02-0.9 0-0.4 0.0-1.3 0.7-4.0 0.0-0.4 0.0-0.7 0.0-2.1 -
Mortality (%) With Without complications complications 43.5 2.6 36.5 2.2 80.1 4.1 26.3 6.9 33.0 7.0 29.8 10.0 6.8 5.0 32.0 8.2 22.6 6.4 25.9 8.6 14.0 8.7 74.6 8.8 27.1 8.9 91.2 9.0 100.0 9.1 25.4 9.1 54.0 9.2 57.2 9.5 20.4 9.7 19.5 9.8 62.0 9.8 35.3 9.9 38.2 6.2 100.0 10.0 19.3 10.0 16.4 10.1 58.5 10.1 77.6 8.9 36.9 10.1 100.0 10.1 23.3 10.1 12.6 10.2 23.6 10.2 10.2 13.4 9.5 10.2 -
Relative risk
PAR
p value
RR
95%CI
%
