Measurement of isoniazid in tuberculosis patients using finger sweat with creatinine normalisation: A controlled administration study.

BACKGROUND: Insufficient exposure and poor compliance with anti-tuberculosis (TB) medications are risk factors for treatment failure and the development of drug resistance. Measurement of drugs in biological samples, such as blood and saliva, can be used to assess adherence and make dose adjustments by therapeutic drug monitoring (TDM). Finger sweat testing is a convenient and non-invasive method to monitor patients. OBJECTIVES: To assess the feasibility of finger sweat testing for medication adherence and as a semi-quantitative tool for TDM analysis. METHODS: Ten patients provided finger sweat, blood and saliva samples following a controlled dose of isoniazid. Samples were analysed by liquid chromatography-mass spectrometry. RESULTS: Isoniazid can be detected in finger sweat 1-6 h following administration at typically prescribed dosages. The normalisation of isoniazid to creatinine increases the correlation between finger sweat and serum isoniazid concentration and provides a means to account for inconsistent sample volumes. CONCLUSION: We describe the time-course measurement of isoniazid (or drug-to-creatinine ratio) in finger sweat compared to the pharmacokinetic profile in blood for the first time. This technique, adaptable for other drugs, could reduce the burden on clinics and improve patient experience.

Imaging mass spectrometry: a new way to distinguish dermal contact from administration of cocaine, using a single fingerprint.

Here we show a new and significant application area for mass spectrometry imaging. The potential for fingerprints to reveal drug use has been widely reported, with potential applications in forensics and workplace drug testing. However, one unsolved issue is the inability to distinguish between drug administration and contamination by contact. Previous work using bulk mass spectrometry analysis has shown that this distinction can only be definitively made if the hands are washed prior to sample collection. Here, we illustrate how three mass spectrometry imaging approaches, desorption electrospray ionisation (DESI), matrix assisted laser desorption ionisation (MALDI) and time of flight secondary ion mass spectrometry (ToF-SIMS) can be used to visualise fingerprints at different pixel sizes, ranging from the whole fingerprint down to the pore structure. We show how each of these magnification scales can be used to distinguish between cocaine use and contact. We also demonstrate the first application of water cluster SIMS to a fingerprint sample, which was the sole method tested here that was capable of detecting excreted drug metabolites in fingerprints, while providing spatial resolution sufficient to resolve individual pore structure. We show that after administration of cocaine, lipids and salts in the fingerprint ridges spatially correlate with the cocaine metabolite, benzoylecgonine. In contrast after contact, we have observed that cocaine and its metabolite show a poor spatial correlation with the flow of the ridges.

On the relevance of cocaine detection in a fingerprint.

The finding that drugs and metabolites can be detected from fingerprints is of potential relevance to forensic science and as well as toxicology and clinical testing. However, discriminating between dermal contact and ingestion of drugs has never been verified experimentally. The inability to interpret the result of finding a drug or metabolite in a fingerprint has prevented widespread adoption of fingerprints in drug testing and limits the probative value of detecting drugs in fingermarks. A commonly held belief is that the detection of metabolites of drugs of abuse in fingerprints can be used to confirm a drug has been ingested. However, we show here that cocaine and its primary metabolite, benzoylecgonine, can be detected in fingerprints of non-drug users after contact with cocaine. Additionally, cocaine was found to persist above environmental levels for up to 48 hours after contact. Therefore the detection of cocaine and benzoylecgonine (BZE) in fingermarks can be forensically significant, but do not demonstrate that a person has ingested the substance. In contrast, the data here shows that a drug test from a fingerprint (where hands can be washed prior to donating a sample) CAN distinguish between contact and ingestion of cocaine. If hands were washed prior to giving a fingerprint, BZE was detected only after the administration of cocaine. Therefore BZE can be used to distinguish cocaine contact from cocaine ingestion, provided donors wash their hands prior to sampling. A test based on the detection of BZE in at least one of two donated fingerprint samples has accuracy 95%, sensitivity 90% and specificity of 100% (n = 86).

Borderline anaemia and postoperative outcome in women undergoing major abdominal surgery: a retrospective cohort study.

Pre-operative anaemia is typically diagnosed with a haemoglobin concentration < 120 g.l-1 for women and < 130 g.l-1 for men on the basis of limited evidence. This retrospective cohort study stratified women undergoing elective, major abdominal surgery based on pre-operative haemoglobin concentration: anaemic (< 120 g.l-1 ); borderline anaemic (120-129 g.l-1 ); and non-anaemic (> 130 g.l-1 ). Data from 1554 women were analysed. Women with borderline anaemia had a greater incidence of postoperative complications (55 (16%) vs. 110 (11%); p = 0.026), longer duration of hospital stay (median (IQR [range]) 3 (1-6 [0-69]) days vs. 2 (1-5 [0-80]) days; p = 0.017) and fewer days alive and out of hospital at postoperative day 30 (median (IQR [range]) 27 (23-29 [0-30]) vs. 28 (25-29 [0-30]) days; p = 0.017) compared with non-anaemic women. However, after matched cohort analysis, these outcome differences no longer remained statistically significant. After multivariable adjustment for procedure, Charlson comorbidity index and patient age, a negative relationship between logarithmic pre-operative haemoglobin concentration and duration of stay was found (parameter estimate (standard error) -0.006 (0.003) vs. 0.003 (0.003) for a haemoglobin concentration < 130 g.l-1 vs. > 130 g.l-1 , respectively; p = 0.03); the difference in duration of stay was approximately 50% greater for women with a haemoglobin concentration of 120 g.l-1 compared with those with a haemoglobin concentration of 130 g.l-1 . Although the contribution of borderline anaemia to the incidence of postoperative complications is uncertain, the current diagnostic criteria should be re-assessed.

Exploring a route to a selective and sensitive portable system for explosive detection- swab spray ionisation coupled to of high-field assisted waveform ion mobility spectrometry (FAIMS).

Paper spray mass spectrometry is a rapid and sensitive tool for explosives detection but has so far only been demonstrated using high resolution mass spectrometry, which bears too high a cost for many practical applications. Here we explore the potential for paper spray to be implemented in field applications with portable mass spectrometry. This involved (a) replacing the paper substrate with a swabbing material (which we call "swab spray") for compatibility with standard collection materials; (b) collection of explosives from surfaces; (c) an exploration of interferences within a ±â€¯0.5 m/z window; and (d) demonstration of the use of high-field assisted waveform ion mobility spectrometer (FAIMS) for enhanced selectivity. We show that paper and Nomex® are viable collection materials, with Nomex providing cleaner spectra and therefore greater potential for integration with portable mass spectrometers. We show that sensitive detection using swab spray will require a mass spectrometer with a mass resolving power of 4000 or more. We show that by coupling the swab spray ionisation source with FAIMS, it is possible to reduce background interferences, thereby facilitating the use of a low resolving power (e.g. quadrupole) mass spectrometer.

A multifaceted pharmacist intervention to improve antihypertensive adherence: a cluster-randomized, controlled trial (HAPPy trial).

WHAT IS KNOWN AND OBJECTIVES: About half of all patients taking antihypertensives discontinue treatment by 12 months. There is potential for substantial health gains at both individual and population levels through improved treatment adherence. The objective was to evaluate a community pharmacist intervention to improve adherence with antihypertensive medicines with a view to improving blood pressure (BP) control. DESIGN: prospective, non-blinded, cluster-randomized, controlled trial. PARTICIPANTS: adults with primary hypertension who obtained antihypertensives in the previous 6 months. Patients with poor refill adherence were preferentially identified with the help of a purpose-built software application. INTERVENTION: package comprising BP monitor; training on BP self-monitoring; motivational interviewing; medication use review; prescription refill reminders. FOLLOW-UP: six months. PRIMARY OUTCOME: change in proportion self-reporting medication adherence. Secondary outcome: BP changes. RESULTS: Participants (n = 395; intervention - 207; control - 188) had a mean age of 66.7 years; 51.1% were males. The proportion of adherent participants increased in both groups but was not significantly different between groups [57·2% to 63·6% (control) vs. 60·0% to 73·5% (intervention), P = 0·23]. The mean reduction in systolic BP was significantly greater in the intervention group (10·0 mmHg vs. 4·6 mmHg; P = 0·05). The proportion of patients who were non-adherent at baseline and adherent at 6 months was 22·6% (95%CI 5·1-40·0%) higher in the intervention group (61·8% vs. 39·2%, P = 0·007). Among participants with baseline BP above target, reduction of systolic BP was significantly greater in the intervention group [by 7·2 mmHg (95%CI 1·6-12·8 mmHg); (P = 0·01)]. Among participants non-adherent at baseline and above target BP, the proportion reporting adherence at 6 months was significantly greater in the intervention group [56·8% vs. 35·9%, P = 0·039). WHAT IS NEW AND CONCLUSION: This community pharmacist intervention resulted in improved adherence to antihypertensive medication and reduced systolic BP.

Diabetes is a major risk factor for mortality after lung transplantation.

Survival following lung transplant (LTx) remains significantly lower than after other solid organ transplants. Diabetes mellitus (DM) is a mortality risk factor not comprehensively studied in LTx recipients. Notably, neither the relation of time of DM onset to survival nor the actual causes of DM-associated excess mortality have been described. We determined DM status, DM diagnosis date and all-cause mortality in 386 consecutive adults who underwent LTx at our institution from January 1, 2001 to July 31, 2010. The relationship of DM to survival both as a categorical and time-dependent variable was studied. Fifty-three percent of patients had DM. Overall median survival was 5.2 (95% CI 3.8-6.6) years. At study end, 52% of patients had died, of whom 64% had DM. Estimated median survival was 10 years in patients without DM, 5.0 (3.3-6.8) years in patients with DM pre- and post-LTx and 4.3 (3.1-5.5) years in patients with new onset DM. As a time-dependent covariate, DM was the strongest risk factor for mortality, hazard ratio 3.96 (2.85-5.51). Bronchiolitis obliterans syndrome was the main cause of death in all patients surviving >90 days, but its incidence was not increased in patients with DM. Further studies are warranted to determine whether improved glycemic control could improve outcomes in LTx recipients.

Enhanced imaging of developed fingerprints using mass spectrometry imaging.

Latent fingermarks are invisible to the naked eye and normally require the application of a chemical developer followed by an optical imaging step in order to visualize the ridge detail. If the finger deposition is poor, or the fingermark is aged, it can sometimes be difficult to produce an image of sufficient quality for identification. In this work, we show for the first time how mass spectrometry imaging (in this case time-of-flight secondary ion mass spectrometry, ToF-SIMS) can be used to enhance the quality of partially recovered fingermarks. We show three examples of how chemical imaging can be used to obtain enhanced images of fingermarks deposited on aluminium foil, glass and the handle of a hand grenade compared with conventional development techniques.

Integrated Ion Beam Analysis (IBA) in Gunshot Residue (GSR) characterisation.

Gunshot Residue (GSR) is residual material from the discharge of a firearm, which frequently provides crucial information in criminal investigations. Changes in ammunition manufacturing are gradually phasing out the heavy metals on which current forensic GSR analysis is based, and the latest Heavy Metal Free (HMF) primers urgently demand new forensic solutions. Proton scanning microbeam Ion Beam Analysis (IBA), in conjunction with the Scanning Electron Microscope equipped with an Energy Dispersive X-ray Spectrometer (SEM-EDS), can be introduced into forensic analysis to solve both new and old problems, with a procedure entirely commensurate with current forensic practice. Six cartridges producing GSR particles known to be interesting in casework by both experience and the literature were selected for this study. A standard procedure to relocate the same particles previously analysed by SEM-EDS, based on both secondary electron (SE) and X-ray imaging was developed and tested. Elemental Particle Induced X-ray Emission (PIXE) mapping of the emitted X-rays allowed relocation in a scan of 10 µm × 10 µm of even a 1 µm GSR particle. The comparison between spectra from the same particle obtained by SEM-EDS and IBA-PIXE showed that the latter is much more sensitive at mid-high energies. Results that are very interesting in a forensic context were obtained with particles from a cartridge containing mercury fulminate in the primer. Particle-induced gamma-ray emission (PIGE) maps of a particles from HMF cartridges allowed identification of Boron and Sodium in particles from hands using the (10)B(p,α1γ)(7)Be, (11)B(p,p1γ)(11)B and (23)Na(p,p1γ)(23)Na reactions, which is extraordinary in a forensic context. The capability for quantitative analysis of elements within individual particles by IBA was also demonstrated, giving the opportunity to begin a new chapter in the research on GSR particles. The integrated procedure that was developed, which makes use of all the IBA signals, has unprecedented characterisation and discrimination power for GSR samples.

Evaluation of particle-induced X-ray emission and particle-induced γ-ray emission of quartz grains for forensic trace sediment analysis.

The independent verification in a forensics context of quartz grain morphological typing by scanning electron microscopy was demonstrated using particle-induced X-ray emission (PIXE) and particle-induced γ-ray emission (PIGE). Surface texture analysis by electron microscopy and high-sensitivity trace element mapping by PIXE and PIGE are independent analytical techniques for identifying the provenance of quartz in sediment samples in forensic investigations. Trace element profiling of the quartz grain matrix separately from the quartz grain inclusions served to differentiate grains of different provenance and indeed went some way toward discriminating between different quartz grain types identified in a single sample of one known forensic provenance. These results confirm the feasibility of independently verifying the provenance of critical samples from forensic cases.

Transfusion practice and guidelines in Australian and New Zealand intensive care units.

OBJECTIVE: To describe the relationship between clinical practice and national guidelines for the transfusion of red blood cells (RBCs), fresh frozen plasma (FFP), platelets, and cryoprecipitate in Australian and New Zealand intensive care units (ICUs). SETTING: Forty-seven ICUs over a 5-week period from August to September 2008. DESIGN: Prospective, observational, multicentre, cohort study. PATIENTS: A total of 874 patients receiving any type of blood transfusion. METHODS: All patients who were transfused at least one unit of any blood component were included. Patient-specific and blood-component specific data were gathered. Pre-transfusion haemoglobin, platelet count, international normalised ratio (INR), and fibrinogen levels were compared to national guidelines. RESULTS: Of all 874 patients, 757 received RBCs (86.6%), 231 (26.4%) received platelets, 340 (38.9%) received FFP, and 78 (8.9%) received cryoprecipitate. Bleeding was the reason for administration of RBCs in 46%, FFP in 55%, and platelets in 47% of transfusions. The mean (SD) pre-transfusion haemoglobin was 77.6 (9.5) g/l, while the geometric means (95% CI) for platelet count, INR, and fibrinogen were 67.0 (59.7-75.3) x 10(9)/l, 1.84 (1.76-1.93), and 1.4 (1.1-1.8) g/l, respectively. The proportions of transfusions not adherent to guidelines were 2% for RBC, but 53% for platelets, 29% for FFP, and 88% for cryoprecipitate (RBC vs. other transfusion p < 0.001 for all). CONCLUSIONS: Transfusion practice of RBCs in Australian and New Zealand ICUs is restrictive and is concordant with guidelines. However, the transfusion of other blood components is not.

A simple tool for mortality prediction in burns patients: APACHE III score and FTSA.

Prediction of outcome for patients with major thermal injury is important to inform clinical decision making, alleviate individual suffering and improve hospital resource allocation. Age and burn size are widely accepted as the two largest contributors of mortality amongst burns patients. The APACHE (Acute Physiology and Chronic Health Evaluation) III-j score, which incorporates patient age, is also useful for mortality prediction, of intensive care populations. Validation for the burns specific cohort is unclear. A retrospective cohort study was performed on patients admitted to the Intensive Care Unit (ICU) via the Victorian Adult Burns Service (VABS), to compare observed mortality with burns specific markers of illness severity and APACHE III-j score. Our primary aim was to develop a mortality prediction tool for the burns population. Between January 1, 2002 and December 31, 2008, 228 patients were admitted to the ICU at The Alfred with acute burns. The mean age was 45.6 years and 81% (n=184) were male. Patients had severe injuries: the average percent TBSA (total body surface area) was 28% (IQR 10-40) and percent FTSA (full thickness surface area) was 18% (IQR 10-25). 86% (n=197) had airway involvement. Overall mortality in the 7-year period was 12% (n=27). Non-survivors were older, had larger and deeper burns, a higher incidence of deliberate self-harm, higher APACHE III-j scores and spent less time in hospital (but similar time in ICU), compared with survivors. Independent risk factors for death were percent FTSA (OR 1.03, 95% CI 1.01-1.05, p=0.01) and APACHE III-j score (OR 1.04, 95% CI 1.02-1.07, p<0.001). Mortality prediction based on both of these variables in combination was more specific than either individual variable alone (AUROC 0.85, 95% CI 0.79-0.92). Likelihood of death for patients with severe thermal injury can be predicted with accuracy from APACHE III-j score and percent FTSA. Prospective validation of our model on different burn populations is necessary.

Focal cerebral oxygenation and neurological outcome with or without brain tissue oxygen-guided therapy in patients with traumatic brain injury.

BACKGROUND: In patients with severe traumatic brain injury (TBI), the depth and duration of cerebral hypoxia are independent predictors of outcome. This study aimed to evaluate the efficacy of brain oxygen-guided therapy in improving cerebral oxygenation and neurological outcome in severe TBI patients. METHODS: Thirty TBI patients had brain oxygen monitors placed contralateral to the side of mass lesions, or to the non-dominant side if injury was diffuse. The first 10 patients (Group 1, observational) had brain tissue oxygen (PbrO2) monitored, but not treated. The next 20 patients (Group 2, interventional) were treated according to brain tissue oxygen-guided algorithms aiming to improve cerebral oxygen availability. The 6-month neurological outcome of Group 2 patients was compared with that of Group 1 patients and with contemporary control patients (Group 3) treated without the use of brain oxygen monitoring. FINDINGS: The mean duration of brain hypoxic episodes (PbrO2 <15 mmHg) was 106 minutes in Group 1, and 34 minutes in Group 2 (p=0.01). Brain tissue oxygen was <15 mmHg for 10% of monitoring time in Group 1 and 2.8% in Group 2 (p=0.12). The peak incidence of cerebral hypoxic events in both groups occurred during post-injury day 5. The mean Injury Severity Score (ISS) of patients experiencing cerebral hypoxia was higher than that of patients without cerebral hypoxic episodes (33.7 vs 24.2, p=0.04). There was no statistically significant difference in neurological outcome between those patients treated with and those without brain oxygen-guided therapy. CONCLUSIONS: In TBI patients, brain tissue oxygen-guided therapy is associated with decreased duration of episodes of cerebral hypoxia. Larger studies are indicated to determine the effects of this therapy on neurological outcome.

Effect of pump prime on acidosis, strong-ion-difference and unmeasured ions during cardiopulmonary bypass.

We tested the hypothesis that a cardiopulmonary bypass prime with lactate would be associated with less acidosis than a prime with only chloride anions because of differences in the measured strong-ion-difference. We randomised 20 patients to a 1500 ml bypass prime with either a chloride-only solution (Ringer's Injection; anions: chloride 152 mmol/l) or a lactated solution (Hartmann's solution; anions: chloride 109 mmol/l, lactate 29 mmol/l). Arterial blood was sampled before bypass and then two, five, 15 and 30 minutes after initiating bypass. We used repeated measures analysis of variance to compare groups. In both groups, the base-excess and measured strong-ion-difference decreased markedly from baseline after two minutes of bypass. The chloride-only group had greater acidosis with lower base-excess and pH (P < 0.05), greatest after five minutes of bypass (C5). Contrary to our hypothesis, however, the difference between the groups was not due to a difference in the measured strong-ion-difference, P = 0.88. At C5 when the difference in standard base-excess between the groups was greatest, 1.9 mmol/l (95% confidence interval: 0.1 to 3.6 mmol/l, P < 0.05), the difference in the measured strong-ion-difference was only 0.2 mmol/l (95% confidence interval: -2.4 to 2.7 mmol/l, P > 0.05). There was, however a difference in the net-unmeasured-ions (strong-ion-gap). We conclude that acid-base changes with cardiopulmonary bypass may differ with the prime but that the early differences between chloride-only and lactated primes appear not to be due to differences in the measured strong-ion-difference. We suggest future studies examine other possible mechanisms including unmeasured ions.

Dynamic contrast-enhanced MRI for prostate cancer localization.

Radiotherapy dose escalation improves tumour control in prostate cancer but with increased toxicity. Boosting focal tumour only may allow dose escalation with acceptable toxicity. Intensity-modulated radiotherapy can deliver this, but visualization of the tumour remains limiting. CT or conventional MRI techniques are poor at localizing tumour, but dynamic contrast-enhanced MRI (DCE-MRI) may be superior. 18 patients with prostate cancer had T(2) weighted (T2W) and DCE-MRI prior to prostatectomy. The prostate was sectioned meticulously so as to achieve accurate correlation between imaging and pathology. The accuracy of DCE-MRI for cancer detection was calculated by a pixel-by-pixel correlation of quantitative DCE-MRI parameter maps and pathology. In addition, a radiologist interpreted the DCE-MRI and T2W images. The location of tumour on imaging was compared with histology, and the accuracy of DCE-MRI and T2W images was then compared. Pixel-by-pixel comparison of quantitative parameter maps showed a significant difference between the benign peripheral zone and tumour for the parameters K(trans), v(e) and k(ep). Calculation of areas under the receiver operating characteristic curve showed that the pharmacokinetic parameters were only "fair" discriminators between cancer and benign gland. Interpretation of DCE-MRI and T2W images by a radiologist showed DCE-MRI to be more sensitive than T2W images for tumour localization (50% vs 21%; p = 0.006) and similarly specific (85% vs 81%; p = 0.593). The superior sensitivity of DCE-MRI compared with T2W images, together with its high specificity, is arguably sufficient for its use in guiding radiotherapy boosts in prostate cancer.

Split-thickness skin graft donor sites: a comparative study of two absorbent dressings.

OBJECTIVE: To identify the optimal dressing for split-thickness skin graft (SSG) donor sites. METHOD: This prospective randomised controlled trial compared two dressings - a new absorbent form of a polyurethane film dressing (Tegaderm Absorbent, 3M) and our standard alginate dressing (Kaltostat, ConvaTec) - on SSG donor sites in 40 patients. Primary outcome measures were: reduced time to full healing; reduced postoperative pain; reduced leakage rates from the dressing. Secondary outcome measures related to acceptability of the dressings to the patient. RESULTS: On removal of the dressings at the first assessment, 79% of the Tegaderm Absorbent donor sites had healed completely, compared with 16% of the Kaltostat ones (p<0.001).A significantly greater median area had healed with Tegaderm Absorbent (100%), when compared with Kaltostat (89%) (p<0.001). Mean time to complete healing was also significantly faster for Tegaderm Absorbent than Kaltostat (14 versus 21 days) (p<0.001). Significantly fewer subjects experienced postoperative pain with Tegaderm Absorbent on both day 1 (21% versus 67%, p=0.006, NNT=3) and day 2 (17% versus 75%, p<0.001, NNT=2). Leakage rates reduced by 48% with Tegaderm Absorbent, with no leakage in the smaller donor sites. Tegaderm Absorbent was significantly easier to apply than Kaltostat (89% versus 27% found it'very easy') as was ease of removal (84% versus 11% found it'very easy') (p<0.0001). Patients found Tegaderm Absorbent dressings significantly more convenient to manage and bathe with. At one month post-surgery, Vancouver scar scores showed thatTegaderm Absorbent donor sites were less red, flatter, softer and less itchy. CONCLUSION: Tegaderm Absorbent provides a significant improvement in terms of donor-site pain, healing and ease of management.

After-hours discharge from intensive care increases the risk of readmission and death.

Despite reports showing night discharge from an intensive care unit (ICU) is associated with increased mortality, it is unknown if this has resulted in changes in practice in recent years. Our aim was to determine prevalence, trends and effect on patient outcome of discharge timing from ICU throughout Australia and New Zealand. Two datasets from the Australian and New Zealand Intensive Care Society Adult Patient Database (ANZICS APD) were examined: (1) All submissions to the APD from 1.1.2003 to 31.12.2004 to determine contemporary practices. (2) Forty hospitals which had submitted continuous data between 1.1.2000 and 31.12.2004 to determine trends in practice over time. Outcomes investigated were hospital mortality and ICU readmission rate. Between 1.1.2003 and 31.12.2004, the ANZICS APD reported 76,690 patients discharged alive from ICU; 13,968 (18.2%) were discharged after-hours (between 1800 and 0559 hours). After-hours discharges had a higher readmission rate (6.3% vs. 5.1%; P < or = 0.0001) and higher mortality (8.0% vs. 5.3%; P = < 0.0001). Peak readmission (8.6%) and mortality rates (9.7%) were seen in patients discharged between 0300 and 0400 hours. After-hours discharge was a predictor of mortality (odds ratio 1.42, 95% confidence interval 1.32-1.52; P= < 0.0001) in multivariate analysis. Between 2000 and 2004, after-hours discharges increased (P = 0.0015) with seasonal peaks during winter The risk of death increased as the proportion of patients discharged after-hours rose. After-hours discharge from ICU is associated with increased risk of death and readmission to ICU. It has become more frequent. The risk of death increases as more after-hours discharges occur.