Exploring population pharmacokinetic models in patients treated with vancomycin during continuous venovenous haemodiafiltration (CVVHDF).

BACKGROUND: Therapeutic antibiotic dose monitoring can be particularly challenging in septic patients requiring renal replacement therapy. Our aim was to conduct an exploratory population pharmacokinetic (PK) analysis on PK of vancomycin following intermittent infusion in critically ill patients receiving continuous venovenous haemodiafiltration (CVVHDF); focussing on the influence of dialysis-related covariates. METHODS: This was a retrospective single-centre tertiary level intensive care unit (ICU) study, which included patients treated concurrently with vancomycin and CVVHDF between January 2015 and July 2016. We extracted clinical, laboratory and dialysis data from the electronic healthcare record (EHR), using strict inclusion criteria. A population PK analysis was conducted with a one-compartment model using the PMetrics population PK modelling package. A base structural model was developed, with further analyses including clinical and dialysis-related data to improve model prediction through covariate inclusion. The final selected model simulated patient concentrations using probability of target attainment (PTA) plots to investigate the probability of different dosing regimens achieving target therapeutic concentrations. RESULTS: A total of 106 vancomycin dosing intervals (155 levels) in 24 patients were examined. An acceptable 1-compartment base model was produced (Plots of observed vs. population predicted concentrations (Obs-Pred) R2 = 0.78). No continuous covariates explored resulted in a clear improvement over the base model. Inclusion of anticoagulation modality and vasopressor use as categorical covariates resulted in similar PK parameter estimates, with a trend towards lower parameter estimate variability when using regional citrate anti-coagulation or without vasopressor use. Simulations using PTA plots suggested that a 2 g loading dose followed by 750 mg 12 hourly as maintenance dose, commencing 12 h after loading, is required to achieve adequate early target trough concentrations of at least 15 mg/L. CONCLUSIONS: PTA simulations suggest that acceptable trough vancomycin concentrations can be achieved early in treatment with a 2 g loading dose and maintenance dose of 750 mg 12 hourly for critically ill patients on CVVHDF.

The diagnostic yield of transthoracic echocardiography in the intensive care unit: A retrospective observational analysis.

BACKGROUND: Transthoracic echocardiography (TTE) is a commonly requested ICU investigation. Despite this, limited data exist regarding the diagnostic yield of unselected TTEs in a heterogenous ICU population. METHODS: A retrospective, cross-sectional, single-center study was performed. All ICU patients admitted from January 2018 to February 2019 were included. AIMS: The primary aim was to define the indications for, and diagnostic yield of, TTEs performed in the ICU. We also investigated the association between major abnormalities identified on TTE and mortality. RESULTS: There were 358 patients admitted to the ICU during the study period. Of these patients, 115 (32%) had a TTE performed during their ICU stay. The primary indication was to assess left ventricular function. Just under two-thirds of TTEs (65%) were normal or had minor abnormalities. Compared to the rest of the ICU population in our study (including both patients without a TTE performed and patients with a normal TTE), patients with an abnormal TTE had higher ICU (35.9% vs 21.3%, Odds Ratio [OR], 2.06; 95% CI, 1.02-4.19, P = .04) and in-hospital (43.6% vs 30.3%, OR, 2.64; 95% CI, 1.33-5.25, P = .01) mortality. CONCLUSIONS: A formal TTE was performed in one-third of patients during their ICU admission. Major abnormalities were identified in over one-third of these TTEs. ICU and in-hospital mortality were higher in patients with an abnormal TTE.

YEARS Algorithm Versus Wells' Score: Predictive Accuracies in Pulmonary Embolism Based on the Gold Standard CT Pulmonary Angiography.

OBJECTIVES: This study retrospectively applied Wells' score and YEARS algorithm to the same sample of patients to evaluate the predictive performance of each when compared with the gold standard CT pulmonary angiography. DESIGN: Retrospective analytical study. SETTING: A tertiary University Hospital in Ireland. PATIENTS: Data from 794 patients who underwent CT pulmonary angiography to rule out pulmonary embolism. INTERVENTIONS: Patients were analyzed using retrospective application of both Wells' score and YEARS algorithm. Sensitivity, specificity, and diagnostic odds ratio were calculated and compared. MEASUREMENTS AND MAIN RESULTS: Of 794 scans, 78 (9.8%) were positive for pulmonary embolism. The YEARS algorithm was more sensitive than the Wells' score (97.44% vs 74.36%) but was less specific (13.97% vs 33.94%). Furthermore, the diagnostic odds ratio of YEARS was higher than Wells' score (6.27 vs 1.48). YEARS provides better negative predictive value (98% vs 92.4%), and both scores have poor positive predictive value (10.9%). CONCLUSIONS: Both scores successfully exclude pulmonary embolism, although YEARS has a better negative predictive value. Both exhibit poor positive predictive value.

Early warning scores and critical care transfer - patient heterogeneity, low sensitivity, high mortality.

BACKGROUND: Emergency warning systems (EWS) are becoming a standard of care, but have unproven screening value in early critical illness. Similarly, emergency response team (ERT) care is of uncertain value. These questions are most controversial in mixed patient populations, where screening performance might vary, and intensivist-led ERT care might divert resources from existing patients. AIMS: To examine triggering events, disposition and outcome data for an intensivist-staffed EWS-ERT system. METHODS: We analysed process and outcome data over three years, classing EWS-triggered patients into three categories (non-escalated, escalated ward care and critical care transfer). The relationships between EWS data, pre-triggering clinical data, and patient disposition and outcome were examined. RESULTS: There were 1675 calls in 1190 patients. Most occurred later during admission, with critical care transfer in a minority; the rest were followed by escalated or non-escalated ward care. Patients transferred to critical care had high mortality (40.3%); less than half of patient transfers occurred following triggering EWS score predicted overall hospital mortality, but not mortality after critical care. CONCLUSIONS: In a diverse hospital population, most triggering patients did not receive critical care and most critical care transfers occurred without triggering. Triggering was an insensitive screening measure for critical illness, followed by poor outcome. Higher scores predicted higher probability of transfer, but not later mortality, suggesting that EWS is being used as a decision aid but is not a true severity of illness score. Other, non-EWS data are needed for earlier detection and for prioritizing access to critical care.

Post-Intensive Care COVID Survivorship Clinic: A Single-Center Experience.

Patients discharged from the ICU post-COVID-19 pneumonitis may experience long-term morbidity related to their critical illness, the treatment for this and the ICU environment. The aim of this study was to characterize the cognitive, psychologic, and physical consequences of COVID-19 in patients admitted to the ICU and discharged alive. DESIGN: Prospective cohort study. SETTING: Post-intensive care syndrome (PICS) follow-up clinic at Tallaght University Hospital, a tertiary referral center with a 16-bed mixed medical-surgical ICU, including critical care physicians, a psychologist, a physiotherapist, and a research nurse. PATIENTS: Patients who had been admitted to the ICU in our tertiary referral center with COVID-19 pneumonitis 6 months earlier. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: A total of 22 patients attended the 6-month PICS follow-up clinic following admission to ICU with COVID-19 pneumonitis. Mean grip strength was low at the 6-month follow-up at 24.1 pounds (sd 9.8) with a minimally active median metabolic equivalent (MET) of 970 METs/wk (interquartile range, 0-7,794 METs/wk). Only 59% of patients were independent with regard to their activities of daily living. Eight of 14 patients (57%) had returned to work by 6 months post-ICU discharge. Their mean Intensive Care Psychological Assessment Tool (IPAT) score was 6.6 (sd 4.6) with a Post-Traumatic Stress Disorder Checklist for Diagnostic and Statistical Manual of Mental Disorders-5th Edition (PCL-5) score of 21.1 (sd 17.5) and a mean Montreal Cognitive Assessment (MoCA) score of 24 (sd 8.4); suggestive of mild cognitive impairment. In a multivariable regression model, only Acute Physiology and Chronic Health Evaluation II score was significantly independently associated with MoCA score as a cognitive PICS outcome (beta-coefficient, -1.6; se, 0.6; p = 0.04). None of the predictor variables were significantly independently associated with IPAT and PCL-5 as psychologic outcomes, nor with International Physical Activity Questionnaire-Short Form as a physical PICS outcome. CONCLUSIONS: In this single-center prospective cohort study, we found that patients have a high burden of physical and psychologic impairment at 6 months following ICU discharge post-COVID-19 pneumonitis; in many cases requiring specialist referrals for long-term input. We advocate for increased resources for this much needed follow-up multidisciplinary intervention for an ever-growing population of patients.

An evaluation of ciprofloxacin pharmacokinetics in critically ill patients undergoing continuous veno-venous haemodiafiltration.

BACKGROUND: The study aimed to investigate the pharmacokinetics of intravenous ciprofloxacin and the adequacy of 400 mg every 12 hours in critically ill Intensive Care Unit (ICU) patients on continuous veno-venous haemodiafiltration (CVVHDF) with particular reference to the effect of achieved flow rates on drug clearance. METHODS: This was an open prospective study conducted in the intensive care unit and research unit of a university teaching hospital. The study population was seven critically ill patients with sepsis requiring CVVHDF. Blood and ultrafiltrate samples were collected and assayed for ciprofloxacin by High Performance Liquid Chromatography (HPLC) to calculate the model independent pharmacokinetic parameters; total body clearance (TBC), half-life (t1/2) and volume of distribution (Vd). CVVHDF was performed at prescribed dialysate rates of 1 or 2 L/hr and ultrafiltration rate of 2 L/hr. The blood flow rate was 200 ml/min, achieved using a Gambro blood pump and Hospal AN69HF haemofilter. RESULTS: Seventeen profiles were obtained. CVVHDF resulted in a median ciprofloxacin t1/2 of 13.8 (range 5.15-39.4) hr, median TBC of 9.90 (range 3.10-13.2) L/hr, a median Vdss of 125 (range 79.5-554) L, a CVVHDF clearance of 2.47+/-0.29 L/hr and a clearance of creatinine (Clcr) of 2.66+/-0.25 L/hr. Thus CVVHDF, at an average flow rate of ~3.5 L/hr, was responsible for removing 26% of ciprofloxacin cleared. At the dose rate of 400 mg every 12 hr, the median estimated Cpmax/MIC and AUC0-24/MIC ratios were 10.3 and 161 respectively (for a MIC of 0.5 mg/L) and exceed the proposed criteria of >10 for Cpmax/MIC and > 100 for AUC0-24/MIC. There was a suggestion towards increased ciprofloxacin clearance by CVVHDF with increasing effluent flow rate. CONCLUSIONS: Given the growing microbial resistance to ciprofloxacin our results suggest that a dose rate of 400 mg every 12 hr, may be necessary to achieve the desired pharmacokinetic-pharmacodynamic (PK-PD) goals in patients on CVVHDF, however an extended interval may be required if there is concomitant hepatic impairment. A correlation between ciprofloxacin clearance due to CVVHDF and creatinine clearance by the filter was observed (r2 = 0.76), providing a useful clinical surrogate marker for ciprofloxacin clearance within the range studied. TRIAL REGISTRATION: Current Controlled Trials ISRCTN52722850.

Memory improvement in aging as a function of exposure to mood-matching music.

This study examined the effect of matching musical emotion and the mood of the listener on working memory and free recall in normal aging. Memory measures were taken at baseline in healthy young and older participants, and, following a happy or sad mood induction, again after exposure to both mood-matching and -mismatching music in a counterbalanced repeated measures design. Compared to baseline, [i] recall was greater following mood-matching than mood-mismatching music in both groups, and was reduced following mood-mismatching music in older adults, [ii] working memory was greater in the mood-matching condition, but did not differ from baseline in the mismatching condition. The results have significant implications for the increasingly popular forms of intervention involving music used with older populations experiencing cognitive decline.

Anaesthetic considerations of adults with Morquio's syndrome - a case report.

BACKGROUND: The anaesthetic management of patients with Morquio syndrome is complicated by a number of factors including odontoid hypoplasia, atlantoaxial instability, thoracic kyphosis, and deposition of mucopolysaccharides in the soft tissue of the oropharnyx. CASE PRESENTATION: Herein we describe the anaesthetic considerations and management of a 26 year old adult with Morquio syndrome, who presented for an elective hip replacement. CONCLUSION: This report details an awake fiberoptic intubation in an adult with Morquio syndrome. We recommend that this approach be considered in patients with Morquio syndrome undergoing general anaesthesia.

An outbreak of colonization with linezolid-resistant Staphylococcus epidermidis in an intensive therapy unit.

OBJECTIVES: To report an outbreak of colonization with linezolid-resistant Staphylococcus epidermidis in an intensive therapy unit (ITU). METHODS: An outbreak of colonization with linezolid-resistant S. epidermidis affecting 16 patients in an ITU was investigated using PFGE. Environmental and staff screening was carried out as part of the investigation. Usage of linezolid in the hospital and in the ITU was reviewed. Resistant strains were screened for the presence of the G2576T mutation using PCR-RFLP genotyping. The interventions made to control the outbreak were restriction of linezolid prescription and specific infection control measures, including isolation of colonized patients and increased environmental cleaning. RESULTS: Linezolid-resistant S. epidermidis strains from the 16 colonized patients were genetically related. The same strain was also cultured from environmental samples in the ITU. An increase in linezolid usage in the hospital and in the ITU occurred in the 6 months prior to the emergence of the resistant strain. Infection control measures and restriction of linezolid prescription controlled the outbreak. All resistant isolates contained the G2576T mutation. CONCLUSIONS: An outbreak of colonization with linezolid-resistant S. epidermidis occurred in the ITU in our institution. The resistant strain colonized the environment and probably spread from patient to patient. The outbreak was associated with an increase in the linezolid usage in the ITU and in the institution as a whole. Restriction of linezolid usage and infection control measures were introduced to control the outbreak. The emergence of linezolid resistance in S. epidermidis has implications for the use of linezolid as a therapeutic agent.

Mortality rates of patients undergoing emergency laparotomy in an Irish university teaching hospital.

INTRODUCTION: Emergency laparotomy (EL) is a commonly performed operation with increased morbidity and mortality. Currently, there is a lack of published outcomes following emergency laparotomy within an Irish population. The aim of this study was to assess our outcomes and compare these to predefined outcomes from NELA. METHODS: A review of a prospectively maintained database of all patients who underwent an emergency laparotomy between January 1st 2015 and October 31st 2016 was performed. Patient demographics, operative indication and procedures, preoperative lactate, time of surgery, admission to high dependency unit (HDU) and mortality (30- and 90-day mortality) were included. Statistical analysis was performed using Minitab V18 with p < 0.05 considered significant. RESULTS: One hundred twenty-four emergency operations were performed on 120 patients. The median age was 60 years. Indications for surgery included the following (%): peritonitis (32.45%), obstruction (22.5%), complicated hernia (19.1%), mesenteric ischaemia (15%), trauma (4.1%), and acute haemorrhage (3.3%). A consultant surgeon and consultant anaesthetist were present at 79 and 78% of EL carried out, respectively. Reported 30- and 90-day mortality were 6.66 and 11.6%, respectively. Thirty-day mortality was 20% in octogenarians. Index lactate did not correlate with mortality (p = 0.43). A significant proportion of mortalities had procedures carried out between 6 pm and midnight with the highest mortality rate in patients with mesenteric ischaemia (p < 0.05). CONCLUSION: EL is associated with high mortality rates. Our mortality figures compare favourably with the NELA data. We support the development of a national database to facilitate improvements in the quality of care delivered to this high-risk cohort.

Previous cocaine self-administration disrupts reward expectancy encoding in ventral striatum.

The nucleus accumbens core (NAc) is important for integrating and providing information to downstream areas about the timing and value of anticipated reward. Although NAc is one of the first brain regions to be affected by drugs of abuse, we still do not know how neural correlates related to reward expectancy are affected by previous cocaine self-administration. To address this issue, we recorded from single neurons in the NAc of rats that had previously self-administered cocaine or sucrose (control). Neural recordings were then taken while rats performed an odor-guided decision-making task in which we independently manipulated value of expected reward by changing the delay to or size of reward across a series of trial blocks. We found that previous cocaine self-administration made rats more impulsive, biasing choice behavior toward more immediate reward. Further, compared to controls, cocaine-exposed rats showed significantly fewer neurons in the NAc that were responsive during odor cues and reward delivery, and in the reward-responsive neurons that remained, diminished directional and value encoding was observed. Lastly, we found that after cocaine exposure, reward-related firing during longer delays was reduced compared to controls. These results demonstrate that prior cocaine self-administration alters reward-expectancy encoding in NAc, which could contribute to poor decision making observed after chronic cocaine use.

Temporal distribution of porcine circovirus 2 genogroups recovered from postweaning multisystemic wasting syndrome affected and nonaffected farms in Ireland and Northern Ireland.

Porcine circovirus type 2 (PCV2) is now recognized as the essential infectious component of porcine postweaning multisystemic wasting syndrome (PMWS). PMWS was first recognized in high-status, specific pathogen-free pigs in Canada in 1991 and is now an economically important disease that affects the swine industry around the world. Recently, reports of genomic studies on PCV2 viruses indicated that 2 distinctive genogroups of PCV2 exist.4,10 This report involves the results of a study on the distribution of predominant PCV2 genogroups recovered from samples taken from PMWS-affected and PMWS-nonaffected farms on the island of Ireland over a 9-year period and the results of a study on PCV2 genogroup recovery from fecal samples taken from a farm in Northern Ireland from 2003 to 2005 that was first diagnosed as PMWS positive in August 2005. The results indicate that, although at least 2 distinct genogroups of PCV2 have been circulating on pig farms on the island of Ireland, there does not appear to be a direct relationship between infection with these different genogroups of PCV2 and the development of PMWS.

Structural Analysis of All the Nickel 14-Membered Tetraaza Macrocycles in the Cambridge Structural Database.

One hundred and thirty-nine nickel complexes containing the 1,4,8,11-tetraazacyclotetradecane backbone were found in the Cambridge Structural Database. Cluster analysis was used to separate the structures according to their macrocyclic configurations, using their conformations as a separation criteria. The most commonly found configuration was the trans III configuration. The amount of trans I macrocycles relative to trans III macrocycles was found to be higher for square-planar complexes than octahedral ones. Molecular mechanical conformational searches were done for select structures found in the Cambridge Structural Database.

Pharmacokinetics of amphotericin B lipid complex in critically ill patients undergoing continuous venovenous haemodiafiltration.

The objective of this study was to examine the effect of continuous venovenous haemodiafiltration (CVVHDF) on the pharmacokinetics of amphotericin B (AmB) in critically ill patients following administration of amphotericin B lipid complex (ABLC). Plasma and ultrafiltrate (UF) samples were collected from patients administered ABLC and either receiving or not receiving CVVHDF. Pharmacokinetic (PK) analysis was performed on eight profiles from patients receiving CVVHDF and six profiles from patients not receiving CVVHDF. For patients receiving CVVHDF, the following median PK data were calculated: area under the concentration-time curve (AUC) = 13.9 h·µg/mL, volume of distribution at steady state (V(ss)) = 1476L and drug clearance (CL) = 27.4 L/h; for patients not receiving CVVHDF, the corresponding median PK data were 11.5 h µg/mL, 2048 L and 43.7 L/h, respectively. The median half-lives calculated during the dosage interval (t(1/2int)) were 30.9 h and 32.5 h on and off CVVHDF, respectively, and the total range of t(1/2int) values was 15.6-180.4 h. Observed median peak concentrations on Day 1 were 0.563 µg/mL and 0.468 µg/mL in patients on and off CVVHDF, respectively. From AmB present in the UF, clearance via CVVHDF contributed<1% of total plasma clearance. The AmB concentration-time profiles for patients administered ABLC on and off CVVHDF were compared and no statistically significant differences in AUC, CL, t(1/2int) and V(ss) were observed. In conclusion, CVVHDF had no clinically significant effect on the pharmacokinetics of AmB following administration of ABLC.

How is intensive care reimbursed? A review of eight European countries.

Reimbursement schemes in intensive care are more complex than in other areas of healthcare, due to special procedures and high care needs. Knowledge regarding the principles of functioning in other countries can lead to increased understanding and awareness of potential for improvement. This can be achieved through mutual exchange of solutions found in other countries. In this review, experts from eight European countries explain their respective intensive care unit reimbursement schemes. Important conclusions include the apparent differences in the countries' reimbursement schemes-despite all of them originating from a DRG system-, the high degree of complexity found, and the difficulties faced in several countries when collecting the data for this collaborative work. This review has been designed to assist the intensivist clinician and researcher in understanding neighbouring countries' approaches and in putting research into the context of a European perspective. In addition, steering committees and decision makers might find this a valuable source to compare different reimbursement schemes.

An open prospective study of amikacin pharmacokinetics in critically ill patients during treatment with continuous venovenous haemodiafiltration.

BACKGROUND: The objectives of the current study were to determine amikacin pharmacokinetics in patients undergoing treatment with continuous venovenous haemodiafiltration (CVVHDF) in an Intensive Care Unit (ICU), and to determine whether peak and trough concentration data could be used to predict pharmacokinetic parameters. An open prospective study was undertaken, comprising five critically ill patients with sepsis requiring CVVHDF. METHODS: Peak and trough plasma concentrations and multiple serum levels in a dosage interval were measured and the latter fitted to both a one- and two-compartment model. Blood and ultrafiltrate samples were collected and assayed for amikacin to calculate the pharmacokinetic parameters; total body clearance (TBC), elimination rate constant (k) and volume of distribution (Vd). The concentration of amikacin in ultrafiltrate was used to determine the clearance via CVVHDF. CVVHDF was performed at prescribed dialysate rates of 1-2l h-1 and ultrafiltration rate of 2l h-1. Blood was pumped at 200ml/min using a Gambro blood pump and Hospal AN69HF haemofilter. Amikacin dosing was according to routine clinical practice in the Intensive Care Unit. RESULTS: The multi serum level study indicated that the one compartment model was adequate to characterize the pharmacokinetics in these patients suggesting that peak and trough plasma level data may be used to estimate individual patient pharmacokinetic parameters and to optimise individual patient dosing during treatment with CVVHDF. CVVHDF resulted in an amikacin k of 0.109+/-0.025 h, t1/2 of 6.74 +/- 1.69h, TBC of 3.39+/-0.817 h-1, and Vd of 31.4 +/- 3.27. The mean clearance due to CVVHDF of 2.86 l h-1 is similar to the creatinine clearance of 2.74 +/-0.4 lh-1. Amikacin was significantly cleared by CVVHDF, and its half life in patients on CVVHDF was approximately 2-3 times that reported in subjects without renal impairment and not undergoing haemodiafiltration for any reason. CONCLUSIONS: CVVHDF contributes significantly to total clearance of amikacin. The use of pharmacokinetic parameter estimates obtained from two steady state serum-drug concentrations (peak and trough) can be used to guide individualised dosing of critically ill patients treated with CVVHDF. This is considered a useful strategy in this patient cohort, particularly in avoiding the risk of underdosing.

Rhabdomyolysis associated with polydipsia induced hyponatraemia.

A 41-year-old white male with a history of alcoholism and depression was brought from prison into the emergency department (ED) after having had a witnessed tonic-clonic seizure lasting approximately 5 min. During the 24 h prior to admission, the patient's cell mate reported that he was restless and had consumed 11 litres of water. The patient had also been taking regular escitalopram for his depression. On arrival to the ED, the patient was found to have a sodium level of 112 mmol/l. After correction of his hyponatraemia the patient developed rhabdomyolisis with a creatine kinase level of 65 064 IU/l. To prevent an acute kidney injury a high volume alkaline diuresis protocol was started. Once corrected, his sodium level remained normal and he was discharged home after making a full neurological recovery. Rhabdomyolysis has rarely been associated with the correction of hyponatraemia.

Determining the economic cost of ICU treatment: a prospective "micro-costing" study.

OBJECTIVE: To prospectively assess the cost of patients in an adult intensive care unit (ICU) using bottom-up costing methodology and evaluate the usefulness of "severity of illness" scores in estimating ICU cost. METHODS AND DESIGN: A prospective study costing 64 consecutive admissions over a 2-month period in a mixed medical/surgical ICU. RESULTS: The median daily ICU cost (interquartile range, IQR) was 2,205 euro (1,932 euro-3,073 euro), and the median total ICU cost (IQR) was 10,916 euro (4,294 euro-24,091 euro). ICU survivors had a lower median daily ICU cost at 2,164 per day, compared with 3,496 euro per day for ICU non-survivors (P = 0.08). The requirements for continuous haemodiafiltration, blood products and anti-fungal agents were associated with higher daily and overall ICU costs (P = 0.002). Each point increase in SAPS3 was associated with a 305 euro (95% CI 31 euro-579 euro) increase in total ICU cost (P = 0.029). However, SAPS3 accounted for a small proportion of the variance in this model (R (2) = 0.08), limiting its usefulness as a stand-alone predictor of cost in clinical practice. A model including haemodiafiltration, blood products and anti-fungal agents explained 54% of the variance in total ICU cost. CONCLUSION: This bottom-up costing study highlighted the considerable individual variation in costs between ICU patients and identified the major factors contributing to cost. As the requirement for expensive interventions was the main driver for ICU cost, "severity of illness" scores may not be useful as stand-alone predictors of cost in the ICU.