Influence of time elapsed from end of emergency surgery until admission to intensive care unit, on Acute Physiology and Chronic Health Evaluation II (APACHE II) prediction and patient mortality rate
The intensive care unit is the part of the hospital that is dedicated to providing a system of continuous surveillance for seriously ill patients who are potentially recoverable or at risk.1 The high-complexity features of these services and their high cost make it difficult to offer enough beds in intensive care units to cope with the growing demand. This demand is a reflection of increased life expectancy, longer survival of patients with diseases that used to be lethal, progress in diagnosing and treating various diseases, introduction of new therapeutic procedures and better approaches, and pre-hospital treatment. All these factors bring into hospitals larger number of traumatized patients who would previously have died at the scene of the accident or on the way to hospital.
The need to adapt the number of intensive care unit beds for this growing demand has made evaluation of the prognosis an important aspect of clinical investigation. In Brazil, the prognostic index most used is APACHE II (Acute Physiology And Chronic Health Evaluation II),2 which has been evaluated and validated for use in several countries.3-12 Many Brazilian studies have evaluated its application to intensive care units, and APACHE II has been approved for use in quantifying the severity of patients conditions. On the other hand, its calibration for predicting hospital mortality has not been good, and it is recommended that safeguards regarding differences between studied populations, correction factors and standard mortality rate should also be taken into consideration.13-20
Because APACHE II is the most used index, the need to evaluate the possible influences on its forecasting capacity continues to have significant importance. One of the factors that certainly interferes in the evaluation of Brazilian intensive care units is the great difference between the demand for high-complexity services and the limited capacity to absorb such patients.20-22 The lack of vacant beds, in conjunction with real need, especially in acute emergency cases, creates a situation in which, after completion of the surgical intervention, patients stay in the surgical unit awaiting transfer to the intensive care unit. This may have implications for the quality of treatment and, consequently, on such patients evolution and prognosis.
To verify whether the elapsed time before transfer from the operating theater to the intensive care unit interferes in the predictive accuracy of the APACHE II index, length of stay in the intensive care unit and hospital, and hospital mortality, among patients who have undergone emergency surgery.
All the patients admitted to the intensive care service of Hospital Santa Casa de Misericórdia de São Paulo between August 2002 and July 2003 were evaluated. This is a tertiary teaching hospital with approximately 820 active beds, of which 600 are for adults. At the time when this study was conducted, the intensive care unit had 15 beds for clinical or surgical adult patients. The study protocol had previously been submitted for consideration by the institutions research ethics committee and had been approved.
We selected 104 patients admitted to the intensive care unit directly from the surgical unit, following emergency surgery. Of these, five patients were excluded because they had not come from the hospitals emergency service, i.e. they had already been hospitalized in other units of the hospital. Four patients were excluded because they had already been in the intensive care unit during the same hospitalization. One other patient was also excluded because some of the information needed for the study was missing. The evaluation started after admission to the intensive care unit, and thus there was no interference from the hospitalization criteria. The patients were prospectively followed up until death or hospital discharge. Retrospective evaluation was undertaken to obtain data regarding the period prior to admission to the intensive care unit. The patients were classified in one of two groups, according to the elapsed time between the end of the operation and the admission to the intensive care unit. The “up-to-12 h” group waited for up to 12 hours before admission, and the “over-12 h” group waited for more than 12 hours.
The 94 patients finally included in the study were evaluated with regard to the following information: gender, age, diagnosis, APACHE II, risk of hospital death, length of stay in the intensive care unit, length of hospital stay, and evolution until hospital discharge or death. The APACHE II index was calculated within the first 24 hours following admission to the intensive care unit. The expected mortality was calculated by the APACHE II system and was compared with the observed mortality. The standardized mortality rate was thus obtained by means of dividing the observed by the expected mortality. The death risks were compared with the observed mortality. The sensitivity, specificity and percentage of correct classification were calculated for all of the risk levels. A single observer did the data checking and processing.
Regression analysis was performed to correlate between mortality and the time elapsed between the end of the operation and admission to the intensive care unit. Students t test was used for comparing the averages of continuous measurements. The chi-squared test, with the Yates correction, was used for comparing the proportions of categorized measurements, except in subgroups with five patients or less, in which case Fishers exact test was used. The chi-squared result was used as the trend. The regression analysis was performed to identify the risk factors for death. The predictive capability of the APACHE II index was assessed using the receiver operating characteristic curve, through a 2 x 2 decision matrix and linear regression analysis.
The significance level of p < 0.05 was considered statistically significant. The Statistical Package for the Social Sciences software, version 10.01, and Epi Info version 6.04 were used.
Among the 94 patients included in the study, 23 (24.5%) were admitted to the intensive care unit within 12 hours after the end of the surgery, and these 23 patients formed the up-to-12 h group. The majority of the patients (71; 75.5%) were in the over-12 h group. There was male prevalence in both groups, but no significant difference in gender between the groups (p = 0.8539). With regard to age, in the up-to-12 h group it ranged from 16 to 76 years, with an average of 48 years (standard deviation, SD = 17), while in the over-12 h group it ranged from 18 to 97 years, with an average was 56 years (SD = 19), without significant difference between the groups (p = 0.0920).
The patient distribution in the diagnosis categories according to the APACHE II classification. There was no difference between the groups regarding the most frequent diagnoses, or in the numbers of patients in each category. The APACHE II score ranged from 2 to 41, with an average for the up-to-12 h group of 20.1 (SD = 9.6), while in the over-12 h group it was 21.1 (SD = 7.8), without significant difference between them (p = 0.6129). The greatest concentration of patients (52%) was in the APACHE II range of 16 to 25.
The general mortality was 47.9%, and the mortality in the over-12 h group was significantly greater than for the up-to-12 h group (54.9% versus 26.1%; p = 0.018). The observed general mortality was similar to the expected mortality. For the patients in the up-to-12 h group, the observed mortality was similar to the predicted mortality, while for the over-12 h group, the observed mortality was greater than the expected mortality (54.9% versus 39.7%; p = 0.015). The overall standardized mortality rate was 1.21, while it was 0.69 for the up-to-12 h group and 1.36 for the over-12 h group.
The average length of stay in the intensive care unit was 14.7 days for the up-to-12 h group and 12.8 days for the over-12 h group, without statistical difference between them (p = 0.9077). The average length of hospital stay was 39.0 days for the up-to-12 h group and 28.1 days for the over-12 h group, without statistical difference between them (p = 0.4462).
The death risk factors identified through the regression logistic analysis were: age, APACHE II, elapsed time before admission to the intensive care unit and length of hospital stay (p < 0.02). The receiver operating characteristic curve, built up from the sensitivity and the complement of the specificity of the death risk, shows an area under the curve of 0.729.The best percentage of correct classification was obtained with a decision criterion of 0.6, and was 67%. The calibration curve, stratified in 10% risk bands, obtained an r2 value of 0.46.