Search

Objectives To define standardized endpoints to aid the design of trials that compare antibiotic therapies for bloodstream infections (BSI). Methods Prospective studies, randomized trials or registered protocols comparing antibiotic therapies for BSI, published from 2005 to 2016, were reviewed. Consensus endpoints for BSI studies were defined using a modified Delphi process. Results Different primary and secondary endpoints were defined for pilot (small-scale studies designed to evaluate protocol design, feasibility and implementation) and definitive trials (larger-scale studies designed to test hypotheses and influence clinical practice), as well as for Staphylococcus aureus and Gram-negative BSI. For pilot studies of S. aureus BSI, a primary outcome of success at day 7 was defined by: survival, resolution of fever, stable/improved Sequential Organ Failure Assessment (SOFA) score and clearance of blood cultures, with no microbiologically confirmed failure up to 90 days. For definitive S. aureus BSI studies, a primary outcome of success at 90 days was defined by survival and no microbiologically confirmed failure. For pilot studies of Gram-negative BSI, a primary outcome of success at day 7 was defined by: survival, resolution of fever and symptoms related to BSI source, stable or improved SOFA score and negative blood cultures. For definitive Gram-negative BSI studies, a primary outcome of survival at 90 days supported by a secondary outcome of success at day 7 (as previously defined) was agreed. Conclusions These endpoints provide a framework to aid future trial design. Further work will be required to validate these endpoints with respect to patient-centred clinical outcomes.Copyright © 2016 European Society of Clinical Microbiology and Infectious Diseases

Purpose: Patients who survive an episode of critical illness continue to experience significant mortality after hospital discharge. This study assessed the accuracy of physician prediction of 2-year mortality and compared it with 2 objective prognostic models. Methods: Sensitivity (probability of a prediction of death in patients who died within 2 years) and specificity (probability of a prediction of survival in patients who survived at least 2 years) of physicians' 2-year prediction were compared with those from 2 objective prognostic models, Acute Physiology and Chronic Health Evaluation (APACHE) II and Predicted Risk Existing Disease Intensive Care Therapy (PREDICT). Results: Physician prediction of 2-year mortality was available for 2497 (94.8%) intensive care unit admissions. Specificity was high (85.2%; 95% confidence interval [CI], 83.7-86.4), but sensitivity (65.0%; 95% CI, 61.1-68.8) and positive predictive value (57.4%; 95% CI, 53.6-61.2) were relatively low, suggesting overpessimistic prediction of 2-year mortality. Age, Charlson comorbidity index, and APACHE score were independent risk factors for an inaccurate physician prediction. The diagnostic odds ratio for the physician predictions was at least comparable with the APACHE and PREDICT models, which both had very good discrimination of mortality at 2-year follow-up. Conclusions: Physicians tended to overpredict the risk of 2-year mortality of critically ill patients, but accuracy was comparable with 2 objective prognostic models.

Healthcare utilisation can affect quality of life and is important in assessing the cost-effectiveness of medical interventions. A clinical database was linked to two Australian state administrative databases to assess the difference in incidence of healthcare utilisation of 19 921 patients who survived their first episode of critical illness. The number of hospital admissions and days of hospitalisation per patient-year was respectively 150% and 220% greater after than before an episode of critical illness (assessed over the same time period). This was the case regardless of age or type of surgery (i.e. cardiac vs non-cardiac). After adjusting for the ageing effect of the cohort as a whole, there was still an unexplained two to four-fold increase in hospital admissions per patient-year after an episode of critical illness. We conclude that an episode of critical illness is a robust predictor of subsequent healthcare utilisation.

Critical care service is expensive and the demand for such service is increasing in many developed countries. This study aimed to assess the changes in characteristics of critically ill patients and their effect on long-term outcome. This cohort study utilised linked data between the intensive care unit database and state-wide morbidity and mortality databases. Logistic and Cox regression was used to examine hospital survival and five-year survival of 22,298 intensive care unit patients, respectively. There was a significant increase in age, severity of illness and Charlson Comorbidity Index of the patients over a 16-year study period. Although hospital mortality and median length of intensive care unit and hospital stay remained unchanged, one- and five-year survival had significantly improved with time, after adjusting for age, gender, severity of illness, organ failure, comorbidity, ‘new’ cancer and diagnostic group. Stratified analyses showed that the improvement in five-year survival was particularly strong among patients admitted after cardiac surgery (P=0.001). In conclusion, although critical care service is increasingly being provided to patients with a higher severity of acute and chronic illnesses, long-term survival outcome has improved with time suggesting that critical care service may still be cost-effectiveness despite the changes in case-mix.

Purpose: The aim of this study is to assess the effect of comorbidities on risk of readmission to an intensive care unit (ICU) and the excess hospital mortality associated with ICU readmissions. Materials and Methods: A cohort study used clinical data from a 22-bed multidisciplinary ICU in a university hospital and comorbidity data from the Western Australian hospital morbidity database. Results: From 16 926 consecutive ICU admissions between 1987 and 2002, and 654 (3.9%) of these patients were readmitted to ICU readmissions within the same hospitalization. Patients with readmission were older, more likely to be originally admitted from the operating theatre or hospital ward, had a higher Acute Physiology and Chronic Health Evaluation (APACHE)-predicted mortality, and had more comorbidities when compared with patients without readmission. The number of Charlson comorbidities was significantly associated with late readmission (>72 hours) but not early readmission (≤72 hours) in the multivariate analysis. Both early and late ICU readmissions were associated with an increased risk of hospital mortality (odds ratio, 1.68; 95% confidence interval, 1.18-2.39; P = .004; odds ratio, 1.45; 95% confidence interval, 1.05-1.99; P = .022, respectively) after adjusting for age, admission source, type of admission, the APACHE-predicted mortality, and the number of Charlson comorbidities and APACHE chronic health conditions. Conclusions: Comorbidity was a risk factor for late ICU readmission. Comorbidities could not account for the excess mortality associated with ICU readmissions.

Introduction Interventional ICU trials have followed up patients for variable duration. However, the optimal duration of follow-up for the determination of mortality endpoint in such trials is uncertain. We aimed to determine the most logical and practical mortality end-point in clinical trials of critically ill patients. Methods We performed a retrospective analysis of prospectively collected data involving 369 patients with one of the three specific diagnoses (i) Sepsis (ii) Community acquired pneumonia (iii) Non operative trauma admitted to the Royal Perth Hospital ICU, a large teaching hospital in Western Australia (WA cohort). Their in-hospital and post discharge survival outcome was assessed by linkage to the WA Death Registry. A validation cohort involving 4609 patients admitted during same time period with identical diagnoses from 55 ICUs across Australia (CORE cohort) was used to compare the patient characteristics and in-hospital survival to look at the Australia-wide applicability of the long term survival data from the WA cohort. Results The long term outcome data of the WA cohort indicate that mortality reached a plateau at 90 days after ICU admission particularly for sepsis and pneumonia. Mortality after hospital discharge before 90 days was not uncommon in these two groups. Severity of acute illness as measured by the total number of organ failures or acute physiology score was the main predictor of 90-day mortality. The adjusted in-hospital survival for the WA cohort was not significantly different from that of the CORE cohort in all three diagnostic groups; sepsis (P = 0.19), community acquired pneumonia (P = 0.86), non-operative trauma (P = 0.47). Conclusions A minimum of 90 days follow-up is necessary to fully capture the mortality effect of sepsis and community acquired pneumonia. A shorter period of follow-up time may be sufficient for non-operative trauma.

Vitamin deficiency is believed to be common in critical illness. Water soluble and antioxidant vitamins are those most frequently used for supplementation in these patients. There are no data to confirm the prevalence of vitamin deficiencies in high-risk emergently admitted intensive care patients, nor their association with hospital mortality. One hundred and twenty-nine consecutive, critically ill patients who were emergently admitted to intensive care were enrolled in this prospective observational cohort study. Patient data including diagnosis, source of admission and severity of illness scores were prospectively collected. Within the first 48 hours of admission, concentrations of C-reactive protein, Vitamins A, E, B1, B12 and folate were measured on arterial blood. Multivariate stepwise logistic regression modelling was performed to examine the association of vitamin concentrations with hospital mortality. Fifty-five patients (43%) had a biochemical deficiency of one of the five vitamins on admission to the intensive care unit. A total of 18 patients died (14%) during their hospital stay (15 of those in the intensive care unit). Moderate correlations with C-reactive Protein concentrations were demonstrated for Vitamins B12, A and E (Spearman’s r=0.309, -0.541 and -0.299, P=0.001, 0.001 and 0.007 respectively). Hospital mortality was significantly associated with age, APACHE II score, admission and maximum Sequential Organ Failure Assessment scores and admission source in the univariate analyses. Multivariate analysis did not demonstrate an association between biochemical deficiency and mortality. Biochemical deficiencies of water-soluble and antioxidant vitamins are common on admission in unplanned or emergency admissions to the intensive care unit, but we could not demonstrate an independent association with hospital mortality.

It is unknown whether biochemical vitamin deficiencies in critical illness are associated with severity of illness, organ dysfunction, inflammation or mortality. This nested cohort study recruited 98 patients admitted as emergencies to the intensive care unit, who had a stay of greater than 48 hours. Patient data were prospectively collected. Within the first 48 hours of admission, concentrations of C-reactive protein, vitamins A, E, B1, B12 and folate were measured on arterial blood. These measures were then repeated at least once during the later (>48 hours) period of their stay. Seventy patients (71%) had completed vitamin studies eligible for inclusion in the analysis. Ten patients died (14.3%) during their hospital stay and mortality was associated with age, admission source and severity of illness scores. Vitamin B12 concentration was weakly associated with C-reactive protein concentrations on admission to the intensive care unit (r on days one and two=0.4 [P=0.002], 0.36 [P=0.04], respectively) and with the Sequential Organ Failure Assessment score between days two and four (Spearman’s r=0.361 [P=0.04], 0.42 [P=0.02] and 0.48 [P=0.02], respectively). Vitamin A concentration was weakly associated with the C-reactive protein concentrations on days one and five (Spearman’s r=-0.5 [P=0.001], -0.4 [P=0.03], respectively). Change in deficiency status of any of the vitamins over time in the first week of intensive care admission did not appear to influence mortality. We conclude that while weak correlations were identified between vitamins A and B12 and C-reactive protein and Sequential Organ Failure Assessment scores, the importance of these associations and their relationship to hospital mortality remain to be determined.

Background: Long-term survival outcome of critically ill patients is important in assessing effectiveness of new treatments and making treatment decisions. We developed a prognostic model for estimation of long-term survival of critically ill patients. Methodology and Principal Findings: This was a retrospective linked data cohort study involving 11,930 critically ill patients who survived more than 5 days in a university teaching hospital in Western Australia. Older age, male gender, co-morbidities, severe acute illness as measured by Acute Physiology and Chronic Health Evaluation II predicted mortality, and more days of vasopressor or inotropic support, mechanical ventilation, and hemofiltration within the first 5 days of intensive care unit admission were associated with a worse long-term survival up to 15 years after the onset of critical illness. Among these seven pre-selected predictors, age (explained 50% of the variability of the model, hazard ratio [HR] between 80 and 60 years old = 1.95) and co-morbidity (explained 27% of the variability, HR between Charlson co-morbidity index 5 and 0 = 2.15) were the most important determinants. A nomogram based on the pre-selected predictors is provided to allow estimation of the median survival time and also the 1-year, 3-year, 5-year, 10-year, and 15-year survival probabilities for a patient. The discrimination (adjusted c-index = 0.757, 95% confidence interval 0.745–0.769) and calibration of this prognostic model were acceptable. Significance: Age, gender, co-morbidities, severity of acute illness, and the intensity and duration of intensive care therapy can be used to estimate long-term survival of critically ill patients. Age and co-morbidity are the most important determinants of long-term prognosis of critically ill patients.

Objective: The objective was to assess the ability of potential clinical predictors and inflammatory markers within 24 h of intensive care unit (ICU) discharge to predict subsequent in-hospital mortality. Design and setting: A prospective cohort study of 603 consecutive patients who survived their first ICU admission, between 1 June and 31 December 2005, in a 22-bed multidisciplinary ICU of a university hospital. Measurements and results: A total of 26 in-hospital deaths after ICU discharge (4.3%) were identified. C-reactive protein (CRP) concentrations at ICU discharge were associated with subsequent in-hospital mortality in the univariate analysis (mean CRP concentrations of non-survivors = 174 vs. survivors = 85.6 mg/l, p = 0.001). CRP concentrations remained significantly associated with post-ICU mortality (a 10-mg/l increment in CRP concentrations increased the odds ratio [OR] of death: 1.09, 95% confidence interval [CI]: 1.03-1.16); after adjusting for age, the Acute Physiology and Chronic Health Evaluation (APACHE) II predicted mortality, and the Delta Sequential Organ Failure Assessment (Delta SOFA) score. The area under the receiver operating characteristic curve of this multivariate model to discriminate between survivors and non-survivors after ICU discharge was 0.85 (95% CI: 0.73-0.96). The destination and timing of ICU discharge, and the Discharge SOFA score, white cell counts and fibrinogen concentrations at ICU discharge were not significantly associated with in-hospital mortality after ICU discharge. Conclusions: A high CRP concentration at ICU discharge was an independent predictor of in-hospital mortality after ICU discharge in our ICU.

OBJECTIVE: To determine if circuit life is influenced by a higher pre-dilution volume used in CVVH when compared with a lower pre-dilution volume approach in CVVHDF. DESIGN: A comparative crossover study. Cases were randomized to receive either CVVH or CVVHDF followed by the alternative treatment. SUBJECTS: All patients >or= 18 yrs of age who required CRRT while in ICU were eligible to participate, but excluded if coagulopathic, thrombocytopenic or unable to receive heparin. Based on an intention-to-treat, 45 patients were randomized to receive either CVVH or CVVHDF followed by the alternative treatment. SETTING: A 24-bed, tertiary, medical and surgical adult intensive care unit (ICU). INTERVENTION: Blood flow rate, vascular access device and insertion site, hemofilter, anticoagulation and machine hardware were standardized. An ultrafiltrate dose of 35 ml/ kg/h delivered pre-filter was used for CVVH. A fixed pre-dilution volume of 600 mls/h with a dialysate dose of 1 L was used for CVVHDF. RESULTS: Thirty-one patients received CVVH or CVVHDF out of 45 participants followed by the alternative technique. There was a significant increase in circuit life in favor of CVVHDF (median=16 h 5 min, range=40 h 23 min) compared with CVVH (median=6 h 35 min, range=30 h 45 min). A Mann-Whitney U test was performed to compare circuit life between the two different CRRT modes (Z=-3.478, p0.09 (coefficient of the determination=0.117) or a linear relationship which could be associated with circuit life (p=0.228). CONCLUSION: Pre-diluted CVVHDF appeared to have a longer circuit life when compared to high volume pre-diluted CVVH. The choice of CRRT mode may be an important independent determinant of circuit life.

Many deaths among patients treated in Intensive Care Units (ICUs) occur following the withdrawal or withholding of life support. Following limitation of life support, most of these patients die in the ICU or ward after the decision to limit life support is made, although some may survive to hospital discharge. This study described the characteristics of patients who had life support limitations in ICU and their subsequent in-hospital and out-of-hospital survival using linked data from the state’s death registry. Among 26,019 ICU admissions between 1987 and 2002 there were 396 patients (1.5%) who had life support limitations. The hospital mortality of the patients who had life support limitations was 97.7% and this accounted for 16.2% of the hospital mortality of all ICU admissions. Of the 396 patients who had life support limitations, 315 patients (79.5%) died in the ICU, 72 patients (18.2%) died in the wards and nine patients (2.3%) were discharged from hospital. Of these nine patients who survived to hospital discharge, four died within 10 days of hospital discharge and a further two died within six months. There were two patients, both with significant neurological disabilities at hospital discharge, who survived for longer than three years after hospital discharge. Long-term survival in critically ill patients who had life support limitations was very rare in this ICU.

Noninvasive ventilation by facemask is commonly used for patients with respiratory failure. We evaluated the pressure exerted by two types of facemask on the faces of 12 healthy volunteers while they were being given different levels of continuous or bi-level positive airway pressure ventilation. The mean (SD) pressure recorded on the bridge of the nose was much higher than that on the cheek (nose: 65.8 (21.2) vs cheek 15.4 (7.2) mmHg, p < 0.0001). Progressive tightening of the harness and increasing of the volume of air in the facemask cushions increased the pressure on the bridge of the nose, and the effect of these two factors was additive. Some commercially available facemasks can produce substantial pressure on the bridge of the nose and this explains why pressure complications on the bridge of the nose are common during noninvasive ventilation.

This study compared the performance of the Acute Physiology and Chronic Health Evaluation (APACHE) II score with two organ failure scores in predicting hospital mortality of critically ill patients. A total of 1311 consecutive adult patients in a tertiary 22-bed multidisciplinary intensive care unit (ICU) in Western Australia were considered. The APACHE II score had a better calibration and discrimination than the Max Sequential Organ Failure Score (Max SOFA) (area under receiver operating characteristic (ROC) curve 0.858 vs 0.829), Admission SOFA (area under ROC 0.858 vs 0.791), and the first day or cumulative 5-day Royal Perth Hospital Intensive Care Unit (RPHICU) organ failure score (area under ROC 0.858 vs 0.822 and 0.819, respectively) in predicting hospital mortality. The APACHE II score predicted hospital mortality of critically ill patients better than the SOFA and RPHICU organ failure scores in our ICU.

We investigated whether replacing the Acute Physiology and Chronic Health Evaluation (APACHE) II weighted comorbidity score with other measures of prior comorbidity would improve the prediction of hospital mortality in critically ill patients. Clinical data of 24 303 critically ill patients were linked to the Western Australian hospital morbidity database to identify prior comorbidities. Minor comorbidities as described in the Charlson comorbidity index and Elixhauser comorbidities were prevalent in critically ill patients. Among 24 303 admissions, 3615 (14.9%), 10 223 (42.1%), and 11 597 (47.7%) patients had at least one comorbidity as defined in the APACHE II score, Charlson comorbidity index, and Elixhauser comorbidities, respectively. The ability of comorbidity alone to discriminate between hospital survivors and non-survivors was poor. Replacing the APACHE II weighted comorbidity score with other more comprehensive measures of comorbidity did not significantly improve the discrimination of the APACHE II score.

Introduction: The Acute Physiology and Chronic Health Evaluation (APACHE) II score is widely used in the intensive care unit (ICU) as a scoring system for research and clinical audit purposes. Physiological data for calculation of the APACHE II score are derived from the worst values in the first 24 hours after admission to the ICU. The collection of physiological data on admission only is probably logistically easier, and this approach is used by some ICUs. This study compares the performance of APACHE II scores calculated using admission data with those obtained from the worst values in the first 24 hours. Materials and Methods: This was a retrospective cohort study using prospectively collected data from a tertiary ICU. There were no missing physiological data and follow-up for mortality was available for all patients in the database. The admission and the worst 24-hour physiological variables were used to generate the admission APACHE II score and the worst 24-hour APACHE II score, and the corresponding predicted mortality, respectively. Results: There were 11,107 noncardiac surgery ICU admissions during 11 years from 1 January 1993 to 31 December 2003. The mean admission and the worst 24-hour APACHE II score were 12.7 and 15.4, and the derived predicted mortality estimates were 15.5% and 19.3%, respectively. The actual hospital mortality was 16.3%. The overall discrimination ability, as measured by the area under the receiver operating characteristic curve, of the admission APACHE II model (83.8%, 95% confidence interval = 82.9–84.7) and the worst 24-hour APACHE II model (84.6%, 95% confidence interval = 83.7–85.5) was not significantly different (P = 1.00). Conclusion: Substitution of the worst 24-hour physiological variables with the admission physiological variables to calculate the admission APACHE II score maintains the overall discrimination ability of the traditional APACHE II model. The admission APACHE II model represents a potential alternative model

Objective: To investigate the potential beneficial and adverse effects of early post-pyloric feeding compared with gastric feeding in critically ill adult patients with no evidence of impaired gastric emptying. Design: Randomised controlled studies comparing gastric and post-pyloric feeding in critically ill adult patients from Cochrane Controlled Trial Register (2005 issue 3), EMBASE and MEDLINE databases (1966 to 1 October 2005) without any language restriction were included. Two reviewers reviewed the quality of the studies and performed data extraction independently. Measurements and results: Eleven randomised controlled studies with a total of 637 critically ill adult patients were considered. The mortality (relative risk [RR] 1.01, 95% CI 0.76-1.36, p = 0.93; I2 = 0%) and risk of aspiration or pneumonia (RR 1.28, 95% CI 0.91-1.80, p = 0.15; I2 = 0%) were not significantly different between patients treated with gastric or post-pyloric feeding. The effect of post-pyloric feeding on the risk of pneumonia or aspiration was similar when studies were stratified into those with and those without the use of concurrent gastric decompression (RR ratio 0.95, 95% CI 0.48-1.91, p = 0.89). The risk of diarrhoea and the length of intensive care unit stay (weighted mean difference in days -1.46, 95% CI -3.74 to 0.82, p = 0.21; I2 = 24.6%) were not statistically different. The gastric feeding group had a much lower risk of experiencing feeding tube placement difficulties or blockage (0 vs 9.6%, RR 0.13, 95% CI 0.04-0.44, p = 0.001; I2 = 0%). Conclusions: Early use of post-pyloric feeding instead of gastric feeding in critically ill adult patients with no evidence of impaired gastric emptying was not associated with significant clinical benefits.

Purpose: The purpose of this study is to assess the ability of potential clinical predictors and inflammatory markers to predict intensive care unit (ICU) readmission during the same hospitalization. Materials and Methods: A nested case-control study utilized prospectively collected de-identified data of a 22-bed multidisciplinary ICU in a university hospital. Results: There were 1405 consecutive ICU admissions in 2004, and of these, 18 were regarded as ICU readmissions (1.3%). The destination and timing of ICU discharge, the Sequential Organ Failure Assessment scores, white cell counts, and fibrinogen concentrations at discharge were not associated with ICU readmission. C-reactive protein (CRP) concentration within 24 hours before ICU discharge was associated with ICU readmission (mean CRP concentrations of cases vs controls, 177.8 vs 56.5 mg/L, respectively; P < .0001). The results remained unchanged after adjustment with the propensity scores. The area under the receiver operating characteristic curve for the CRP concentrations to predict ICU readmission was 0.884 (95% confidence interval, 0.765-0.999; P < .0001). Patients readmitted to the ICU had a higher predicted mortality in their second ICU admission (34.9% vs 26.1%; P < .01) and a longer total hospital stay (33.3 vs 20.3 days; P < .003) than patients without ICU readmission. Conclusions: A high CRP concentration within 24 hours before ICU discharge is associated with a higher risk of readmission to the ICU.

Objective: To investigate the short-term outcome of critically ill Indigenous patients. Design and participants: Retrospective cohort study using de-identified audit data from a tertiary intensive care unit (ICU) in Western Australia for the 11-year period 1 January 1993 to 31 December 2003. Main outcome measures: Hospital mortality (crude, and adjusted for severity of illness). Results: Of 16 757 ICU patients, 1076 (6.4%) were identified as Indigenous. The Indigenous patients were younger and more commonly had chronic liver and renal diseases. Indigenous people represented 3.2% of the population of Western Australia in 2001, but represented 3.1% and 9.5% of all elective and emergency ICU admissions, respectively. Diagnoses of sepsis, pneumonia, trauma, and cardiopulmonary arrest were common among critically ill Indigenous patients. Following emergency admission, the crude hospital mortality for Indigenous patients was higher (22.7% v 19.2%; crude odds ratio, 1.24; 95% CI, 1.04-1.47) than for non-Indigenous patients. The crude hospital mortality of critically ill Indigenous patients was lower than that predicted by the APACHE II prognostic model and was similar to that of non-Indigenous patients after adjusting for severity of illness and chronic health status. Conclusions: The pattern of critical illness affecting Indigenous Australians in Western Australia was different from that affecting non-Indigenous patients. The crude hospital mortality was high, but similar to that of non-Indigenous Australians after adjusting for severity of illness and chronic health status.

Introduction: The Acute Physiology and Chronic Health Evaluation (APACHE) II score is widely used in the intensive care unit (ICU) as a scoring system for research and clinical audit purposes. Physiological data for calculation of the APACHE II score are derived from the worst values in the first 24 hours after admission to the ICU. The collection of physiological data on admission only is probably logistically easier, and this approach is used by some ICUs. This study compares the performance of APACHE II scores calculated using admission data with those obtained from the worst values in the first 24 hours. Materials and Methods: This was a retrospective cohort study using prospectively collected data from a tertiary ICU. There were no missing physiological data and follow-up for mortality was available for all patients in the database. The admission and the worst 24-hour physiological variables were used to generate the admission APACHE II score and the worst 24-hour APACHE II score, and the corresponding predicted mortality, respectively. Results: There were 11,107 noncardiac surgery ICU admissions during 11 years from 1 January 1993 to 31 December 2003. The mean admission and the worst 24-hour APACHE II score were 12.7 and 15.4, and the derived predicted mortality estimates were 15.5% and 19.3%, respectively. The actual hospital mortality was 16.3%. The overall discrimination ability, as measured by the area under the receiver operating characteristic curve, of the admission APACHE II model (83.8%, 95% confidence interval = 82.9-84.7) and the worst 24-hour APACHE II model (84.6%, 95% confidence interval = 83.7-85.5) was not significantly different (P = 1.00). Conclusion: Substitution of the worst 24-hour physiological variables with the admission physiological variables to calculate the admission APACHE II score maintains the overall discrimination ability of the traditional APACHE II model. The admission APACHE II model represents a potential alternative model

The clinical, epidemiological and microbiological features of an outbreak of infection and colonisation caused by gentamicin-resistant Acinetobacter baumanii (G RAB) in an 18-bed intensive care unit (ICU) of a 680-bed adult teaching hospital are described. A retrospective review of medical, laboratory and infection control records was followed by prospective surveillance. Typing of isolates was performed by restriction enzyme analysis (REA) of chromosomal DNA. The incidence of GRAB in the ICU increased from 1.26 cases per 1000 occupied bed days (OBDs) for January to June 1993, to 6.62 per 1000 OBDs for July to December 1993 (Chi square = 4.8, P < 0.05), confirming the existence of an outbreak. For the two year period, 1993 and 1994, a total of 45 cases of GRAB infection or colonisation was identified. Males and females were equally represented, with an age range of 16-79 years and a mean age of 51 years. Admitting diagnoses varied, with multiple trauma and head injury predominating (ten cases). For 35 of the 45 cases the initial site of GRAB isolation was sputum or other respiratory tract specimen. Specific treatment for GRAB was initiated in 23 patients, however no deaths were directly attributable to GRAB infection. The period of time between admission to the ICU and first isolation of GRAB ranged from three to 70 days with a median of nine days. Overall, ten (11%) of 91 staff hand samples and one of 37 (3%) environmental samples yielded GRAB. All GRAB isolates produced similar biochemical profiles and antibiotic resistance patterns, except for a group of five which were ciprofloxacin resistant. Thirty patient isolates, all ten staff hand isolates and the environmental isolate produced identical REA patterns. The remaining five patient isolates (all ciprofloxacin resistant) which were available for typing produced a different REA pattern. Our study has documented a moderate-sized outbreak of GRAB in an ICU setting. Typing of isolates using REA was useful in deline