Problematic meta-analyses: Bayesian and frequentist perspectives on combining randomized controlled trials and non-randomized studies

PURPOSE

In the literature, the propriety of the meta-analytic treatment-effect produced by combining randomized controlled trials (RCT) and non-randomized studies (NRS) is questioned, given the inherent confounding in NRS that may bias the meta-analysis. The current study compared an implicitly principled pooled Bayesian meta-analytic treatment-effect with that of frequentist pooling of RCT and NRS to determine how well each approach handled the NRS bias.

MATERIALS & METHODS

Binary outcome Critical-Care meta-analyses, reflecting the importance of such outcomes in Critical-Care practice, combining RCT and NRS were identified electronically. Bayesian pooled treatment-effect and 95% credible-intervals (BCrI), posterior model probabilities indicating model plausibility and Bayes-factors (BF) were estimated using an informative heavy-tailed heterogeneity prior (half-Cauchy). Preference for pooling of RCT and NRS was indicated for Bayes-factors > 3 or < 0.333 for the converse. All pooled frequentist treatment-effects and 95% confidence intervals (FCI) were re-estimated using the popular DerSimonian-Laird (DSL) random effects model.

RESULTS

Fifty meta-analyses were identified (2009-2021), reporting pooled estimates in 44; 29 were pharmaceutical-therapeutic and 21 were non-pharmaceutical therapeutic. Re-computed pooled DSL FCI excluded the null (OR or RR = 1) in 86% (43/50). In 18 meta-analyses there was an agreement between FCI and BCrI in excluding the null. In 23 meta-analyses where FCI excluded the null, BCrI embraced the null. BF supported a pooled model in 27 meta-analyses and separate models in 4. The highest density of the posterior model probabilities for 0.333 < Bayes factor < 1 was 0.8.

CONCLUSIONS

In the current meta-analytic cohort, an integrated and multifaceted Bayesian approach gave support to including NRS in a pooled-estimate model. Conversely, caution should attend the reporting of naïve frequentist pooled, RCT and NRS, meta-analytic treatment effects.

Modelling of intensive care unit (ICU) length of stay as a quality measure: a problematic exercise

BACKGROUND

Intensive care unit (ICU) length of stay (LOS) and the risk adjusted equivalent (RALOS) have been used as quality metrics. The latter measures entail either ratio or difference formulations or ICU random effects (RE), which have not been previously compared.

METHODS

From calendar year 2016 data of an adult ICU registry-database (Australia & New Zealand Intensive Care Society (ANZICS) CORE), LOS predictive models were established using linear (LMM) and generalised linear (GLMM) mixed models. Model fixed effects quality-metric formulations were estimated as RALOSR for LMM (geometric mean derived from log(ICU LOS)) and GLMM (day) and observed minus expected ICU LOS (OMELOS from GLMM). Metric confidence intervals (95%CI) were estimated by bootstrapping; random effects (RE) were predicted for LMM and GLMM. Forest-plot displays of ranked quality-metric point-estimates (95%CI) were generated for ICU hospital classifications (metropolitan, private, rural/regional, and tertiary). Robust rank confidence sets (point estimate and 95%CI), both marginal (pertaining to a singular ICU) and simultaneous (pertaining to all ICU differences), were established.

RESULTS

The ICU cohort was of 94,361 patients from 125 ICUs (metropolitan 16.9%, private 32.8%, rural/regional 6.4%, tertiary 43.8%). Age (mean, SD) was 61.7 (17.5) years; 58.3% were male; APACHE III severity-of-illness score 54.6 (25.7); ICU annual patient volume 1192 (702) and ICU LOS 3.2 (4.9). There was no concordance of ICU ranked model predictions, GLMM versus LMM, nor for the quality metrics used, RALOSR, OMELOS and site-specific RE for each of the ICU hospital classifications. Furthermore, there was no concordance between ICU ranking confidence sets, marginal and simultaneous for models or quality metrics.

CONCLUSIONS

Inference regarding adjusted ICU LOS was dependent upon the statistical estimator and the quality index used to quantify any LOS differences across ICUs. That is, there was no "one best model"; thus, ICU "performance" is determined by model choice and any rankings thereupon should be circumspect.

Hospital-acquired complications: the relative importance of hospital- and patient-related factors

OBJECTIVE

To quantify the prevalence of hospital-acquired complications; to determine the relative influence of patient- and hospital-related factors on complication rates.

DESIGN, PARTICIPANTS

Retrospective analysis of administrative data (Integrated South Australian Activity Collection; Victorian Admitted Episodes Dataset) for multiple-day acute care episodes for adults in public hospitals.

SETTING

Thirty-eight major public hospitals in South Australia and Victoria, 2015-2018.

MAIN OUTCOME MEASURES

Hospital-acquired complication rates, overall and by complication class, by hospital and hospital type (tertiary referral, major metropolitan service, major regional service); variance in rates (intra-class correlation coefficient, ICC) at the patient, hospital, and hospital type levels as surrogate measures of their influence on rates.

RESULTS

Of 1 558 978 public hospital episodes (10 029 918 bed-days), 151 486 included a total of 214 286 hospital-acquired complications (9.72 [95% CI, 9.67-9.77] events per 100 episodes; 2.14 [95% CI, 2.13-2.15] events per 100 bed-days). Complication rates were highest in tertiary referral hospitals (12.7 [95% CI, 12.6-12.8] events per 100 episodes) and for episodes including intensive care components (37.1 [95% CI, 36.7-37.4] events per 100 episodes). For all complication classes, inter-hospital variation was determined more by patient factors (overall ICC, 0.55; 95% CI, 0.53-0.57) than by hospital factors (ICC, 0.04; 95% CI, 0.02-0.07) or hospital type (ICC, 0.01; 95% CI, 0.001-0.03).

CONCLUSIONS

Hospital-acquired complications were recorded for 9.7% of hospital episodes, but patient-related factors played a greater role in determining their prevalence than the treating hospital.

Long-term clinical outcomes in patients with a working diagnosis of myocardial infarction with non-obstructed coronary arteries (MINOCA) assessed by cardiovascular magnetic resonance imaging

BACKGROUND

Myocardial infarction with non-obstructed coronary arteries (MINOCA) is a distinct entity among patients presenting with troponin-positive acute chest pain. We have previously reported on the incremental diagnostic capability of cardiovascular magnetic resonance (CMR) in this cohort. There is paucity of evidence on the long-term (> 5 years) clinical outcomes of these patients as graded by their acute CMR diagnosis.

METHODS AND RESULTS

A total of 229 patients with a working diagnosis of MINOCA who underwent CMR assessment during the acute admission (2010-2017) were prospectively studied. The primary endpoint was major adverse cardiac events (MACE) defined as a composite of all-cause mortality and cardiovascular readmissions, identified from hospital and primary care records. CMR performed at a median of 6 days (IQR 2, 8) from presentation provided a diagnosis in 85% of the patients (38% myocarditis, 28% acute myocardial infarction and 19% Takotsubo cardiomyopathy). Over a median follow-up of 7.1 years (IQR 3.7, 8.2), 56 (24%) patients experienced a MACE. We found a strong association between CMR diagnosis and MACE (log rank 30.47, p < 0.001). In multivariate analysis, age (hazard ratio = 1.07; 95% confidence interval = 1.05, 1.10; p < 0.001) and CMR diagnosis of acute myocardial infarction (hazard ratio = 8.87; 95% confidence interval = 2.58, 30.4; p = 0.001) were independent predictors of MACE.

CONCLUSIONS

In a large cohort of patients with a working diagnosis of MINOCA, one in four suffer a MACE during long-term clinical follow-up. CMR diagnosis of acute myocardial infarction and age were significant predictors of MACE even in the absence of significant coronary artery obstruction.

Multivariate meta-analysis of critical care meta-analyses: a meta-epidemiological study

Background

Meta-analyses typically consider multiple outcomes and report univariate effect sizes considered as independent. Multivariate meta-analysis (MVMA) incorporates outcome correlation and synthesises direct evidence and related outcome estimates within a single analysis. In a series of meta-analyses from the critically ill literature, the current study contrasts multiple univariate effect estimates and their precision with those derived from MVMA.

Methods

A previous meta-epidemiological study was used to identify meta-analyses with either one or two secondary outcomes providing sufficient detail to structure bivariate or tri-variate MVMA, with mortality as primary outcome. Analysis was performed using a random effects model for both odds ratio (OR) and risk ratio (RR); borrowing of strength (BoS) between multivariate outcome estimates was reported. Estimate comparisons, β coefficients, standard errors (SE) and confidence interval (CI) width, univariate versus multivariate, were performed using Lin’s concordance correlation coefficient (CCC).

Results

In bivariate meta-analyses, for OR (n = 49) and RR (n = 48), there was substantial concordance (≥ 0.69) between estimates; but this was less so for tri-variate meta-analyses for both OR (n = 25; ≥ 0.38) and RR (≥ -0.10; n = 22). A variable change in the multivariate precision of primary mortality outcome estimates compared with univariate was present for both bivariate and tri-variate meta-analyses and for metrics. For second outcomes, precision tended to decrease and CI width increase for bivariate meta-analyses, but was variable in the tri-variate. For third outcomes, precision increased and CI width decreased. In bivariate meta-analyses, OR coefficient significance reversal, univariate versus MVMA, occurred once for mortality and 6 cases for second outcomes. RR coefficient significance reversal occurred in 4 cases; 2 were discordant with OR. For tri-variate OR meta-analyses reversal of coefficient estimate significance occurred in two cases for mortality, nine cases for second and 7 cases for third outcomes. In RR meta-analyses significance reversals occurred for mortality in 2 cases, 6 cases for second and 3 cases for third; there were 7 discordances with OR. BoS was greater in trivariate MVMAs compared with bivariate and for OR versus RR.

Conclusions

MVMA would appear to be the preferred solution to multiple univariate analyses; parameter significance changes may occur. Analytic metric appears to be a determinant.

Modelling hospital outcome: problems with endogeneity

Background

Mortality modelling in the critical care paradigm traditionally uses logistic regression, despite the availability of estimators commonly used in alternate disciplines. Little attention has been paid to covariate endogeneity and the status of non-randomized treatment assignment. Using a large registry database, various binary outcome modelling strategies and methods to account for covariate endogeneity were explored.

Methods

Patient mortality data was sourced from the Australian & New Zealand Intensive Society Adult Patient Database for 2016. Hospital mortality was modelled using logistic, probit and linear probability (LPM) models with intensive care (ICU) providers as fixed (FE) and random (RE) effects. Model comparison entailed indices of discrimination and calibration, information criteria (AIC and BIC) and binned residual analysis. Suspect covariate and ventilation treatment assignment endogeneity was identified by correlation between predictor variable and hospital mortality error terms, using the Stata™ “eprobit” estimator. Marginal effects were used to demonstrate effect estimate differences between probit and “eprobit” models.

Results

The cohort comprised 92,693 patients from 124 intensive care units (ICU) in calendar year 2016. Patients mean age was 61.8 (SD 17.5) years, 41.6% were female and APACHE III severity of illness score 54.5(25.6); 43.7% were ventilated. Of the models considered in predicting hospital mortality, logistic regression (with or without ICU FE) and RE logistic regression dominated, more so the latter using information criteria indices. The LPM suffered from many predictions outside the unit [0,1] interval and both poor discrimination and calibration. Error terms of hospital length of stay, an independent risk of death score and ventilation status were correlated with the mortality error term. Marked differences in the ventilation mortality marginal effect was demonstrated between the probit and the "eprobit" models which were scenario dependent. Endogeneity was not demonstrated for the APACHE III score.

Conclusions

Logistic regression accounting for provider effects was the preferred estimator for hospital mortality modelling. Endogeneity of covariates and treatment variables may be identified using appropriate modelling, but failure to do so yields problematic effect estimates.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12874-021-01251-8.

Multivariate Meta-Analysis of the Mortality Effect of Prone Positioning in the Acute Respiratory Distress Syndrome

BACKGROUND

The efficacy of prone positioning (PP) as therapy of the acute respiratory distress syndrome (ARDS) has varied in recent meta-analyses. The efficacy question was reviewed using a cohesive multivariate meta-analysis model incorporating all available common time-point data.

METHODS

Data from a core group of 8 randomized controlled trials (2001-2013) utilized in 8 current meta-analyses (2014-2017) was extracted for common time points. Multivariate meta-analysis and meta-regression models for prone-hours per day, mechanical ventilation tidal-volume and baseline patient PaO₂/FiO₂, considered as continuous and categorical predictors, determined the pooled relative risk (RR) of mortality for prone versus supine positioning.

RESULTS

Mortality RR at 28-30 days, 2-3 months and 6-months was not significant overall (P > 0.05). Meta-regression of categorical predictors indicated significant mortality reduction (P ≤ 0.001) for ≥ 12 prone-hours (versus < 12), lung protective ventilation (versus none) and moderate-severe ARDS (versus all ARDS). Meta-regressions of continuous predictors were also significant (P ≤ 0.021) and yielded treatment inflection points of efficacious therapy for ≥ 12 prone-hours per day, ≤ 8.5 mL/kg tidal volume and ≤ PaO₂/FiO₂ ratio of 130.

CONCLUSIONS

The mortality treatment effect of PP in ARDS, was not demonstrated in the unadjusted meta-analysis model. Moderator effects indicated consistent significant benefit of prone positioning. In the absence of individual patient data, multivariate models provide more decisive conclusions than individual time point analyses.

Safety of the endotracheal tube for prolonged mechanical ventilation

RATIONALE

The endotracheal tube (ETT) is the most common route for invasive mechanical ventilation (MV) yet controversy attends its long-term safety.

OBJECTIVE

Assess the safety of ETT compared with tracheostomy tube (TT) for MV support in the intensive care unit (ICU).

METHODS

Retrospective analysis of five year national dataset of 128,977 adults (age > 15-years) admitted for MV therapy with tracheostomy tube (TT; n = 4772) or without (ETT; n = 124,204), excluding those with neurological diagnoses or likely to require a surgical airway (n = 27,466), in 93 public health service ICUs across Australia, between July 2013-June 2018.

MEASUREMENTS

Hospital survival (including liberation from MV) for ETT Group compared with TT Group using a probit regression model adjusted for confounding using fixed, endogenous and non-random treatment assignment covariates, and their interactions; analysed and plotted as marginal effects by duration of MV.

RESULTS

Median duration of MV was 2 (IQR =1-4) days, predominantly via ETT (124,205; 96.3%), and 21,620 (16.7%) died. Temporal trend for ETT increased (OR = 1.06 per year, 95%CI =1.03-1.10) compared to TT, even for prolonged (>3 weeks) MV (38.1%). Higher risk-adjusted mortality was associated with longer duration of MV and after 9 days of MV with retention of ETT compared with TT - average (mortality) treatment effect 12.6% (95%CI =10.7-14.5). The latter was not significant after 30 days of MV.

CONCLUSIONS

The safety of ETT compared with TT beyond short-term MV (≤9-days) is uncertain and requires prospective evaluation with additional data.

ECMO, ARDS and meta-analyses: Bayes to the rescue?

PURPOSE

A recent meta-analysis by Munshi et al. (Lancet Respiratory Medicine, 2019) claimed mortality treatment efficacy for extra corporeal membrane oxygenation (ECMO) in the acute respitratory syndrome (ARDS) despite very low meta-analytic study numbers (n = 2 (RCTs), risk-ratio (RR) 0·73 (95%CI: 0·58-0·92); n = 5 (2 RCT, 3 observational), RR 0·69 (95%CI: 0·50-0·95)). We explore this efficacy claim by a comprehensive re-analysis of the data.

METHODS

Data were sourced from the two- and five-study meta-analyses, conducted using the Der-Simonian & Laird (DSL) method. A variety of frequentist (DSL, restricted maximum likelihood (REML), Paul-Mandel (PM), with/without Hartung-Knapp-Sidik-Jonkman variance correction), a beta-binomial model (BBN)) and Bayesian models (2 finite-mixture and several Markov-Chain-Monte-Carlo) were used to estimate treatment effects. Fragility-indices, the minimum patients changing mortality outcome needed to induce a conclusion change were also applied.

RESULTS

For the 2-study and 5-study meta-analysis only the uncorrected frequentist estimators (DSL, REML, PM) demonstrated significant RR. Except for the BBN model, which was significant for the 2-study meta-analysis, intervals for all other models included the null. Both meta-analyses demonstrated fragility.

CONCLUSIONS

Having canvassed the conduct of both meta-analyses presented by Munshi et al. and proffered alternative methods, we find no certainty regarding the efficacy of ECMO in ARDS.

Effect of angiotensin II receptor blocker and salt supplementation on short-term blood pressure variability in type 2 diabetes

High blood pressure variability (BPV) has been associated with increased cardiovascular (CV) risk. The effect of dietary salt and renin-angiotensin-aldosterone system (RAAS) activity on short-term BPV in type 2 diabetes mellitus (T2DM) is not well characterised. We aimed to determine the effect of dietary salt (sodium chloride, NaCl) supplementation on 24-h mean arterial BPV (24hBPV) during angiotensin II receptor blocker (telmisartan) use and to evaluate the effects of age, sex, plasma renin activity (PRA) and serum aldosterone on 24hBPV. In a randomised, double-blind, crossover study, patients with T2DM (n = 28), treated with telmisartan received NaCl (100 mmol/24 h) or placebo capsules during 2 weeks of telmisartan. Following a 6-week washout, the protocol was repeated in reverse. 24hBPV was evaluated as a co-efficient of variation [CV (%) = mean/standard deviation] × 100). Twenty-four hour urinary sodium excretion, ambulatory BP and biochemical tests were performed at each phase. Results were analysed using a linear mixed model to generate predicted values for 24hBPV. Predicted 24hBPV was higher with telmisartan vs baseline (p = 0.01), with a trend towards reduced 24hBPV with salt (p = 0.052). Predicted 24hBPV was lower in females (p = 0.017), increasing age (p = 0.001) and increasing PRA (p = 0.011). In patients with T2DM, predicted 24hBPV increased from baseline with telmisartan, but there was no additional increase in predicted 24hBPV with salt supplementation. This suggests that in the short-term, salt supplementation has no apparent deleterious effects on 24hBPV. Long-term studies are required to evaluate the effect of 24hBPV on CV outcomes in patients with T2DM.

Comparison of ultrasound-guided transmuscular quadratus lumborum block catheter technique with surgical pre-peritoneal catheter for postoperative analgesia in abdominal surgery: a randomised controlled trial

Following abdominal surgery, the provision of postoperative analgesia with local anaesthetic infusion through both transmuscular quadratus lumborum block and pre-peritoneal catheter have been described. This study compared these two methods of postoperative analgesia following laparotomy. Eighty-two patients 18-85 years of age scheduled to undergo elective surgery were randomly allocated to receive either transmuscular quadratus lumborum block or pre-peritoneal catheter block. In the transmuscular quadratus lumborum group, an 18-gauge Tuohy needle was passed through the quadratus lumborum muscle under ultrasound guidance to reach its anterior aspect. A 20-ml bolus of ropivacaine 0.375% was administered and catheters placed bilaterally. In the pre-peritoneal catheter group, 20 ml of ropivacaine 0.375% was infiltrated at each of three subcutaneous sub-fascial levels, and pre-peritoneal plane catheters were placed bilaterally. Both groups received an infusion of ropivacaine 0.2% at 5 ml.h^-1 , continued up to 48 h along with a multimodal analgesic regime that included regular paracetamol and patient-controlled analgesia with fentanyl. The primary end-point was postoperative pain score on coughing, assessed using a numerical rating score (0-10). Secondary outcomes were pain score at rest, fentanyl usage until 48 h post-operation, satisfaction scores and costs. There was no treatment difference between the two groups for pain score on coughing (p = 0.24). In the transmuscular quadratus lumborum group, there was a reduction in numerical rating score at rest (p = 0.036) and satisfaction scores on days 1 and 30 (p = 0.004, p = 0.006, respectively), but fentanyl usage was similar. In the transmuscular quadratus lumborum group, the highest and lowest blocks observed in the recovery area were T4 and L1, respectively. The transmuscular quadratus lumborum technique cost 574.64 Australian dollars more per patient than the pre-peritoneal catheter technique.

First evidence of Kunjin strain of West Nile virus associated with saltwater crocodile (Crocodylus porosus) skin lesions

Recently, the Kunjin strain of West Nile virus (WNV_KUN ) has been detected using qRT-PCR in belly skin lesions of farmed juvenile saltwater crocodiles. This follows an established association between similar lesions and West Nile virus in American alligators. The lesions present as cutaneous lymphohistiocytic aggregates in the dermal layers of both species. While these lesion do not create an obvious defect on the live crocodile, upon tanning the lesion area collapses and does not uptake the dye evenly, thus reducing its aesthetic appeal. As a result, skins are being rejected jeopardising the economic viability of the Australian crocodile industry. Over 50 skin lesions have since been confirmed as WNV_KUN -positive and preliminary evidence of lesion restructuring is presented. Horizontal transmission of WNV_KUN by mosquitoes is well-established but other transmission routes, such as ingestion and cloacal shedding, need further evaluation. An infection trial is currently underway to ensure WNV_KUN is the causative agent of these skin lesions.

Risk related therapy in meta-analyses of critical care interventions: Bayesian meta-regression analysis

PURPOSE

The relationship between treatment efficacy and patient risk is explored in a series of meta-analyses from the critical care domain, focusing on mortality outcome.

METHODS

Systematic reviews of randomized controlled trials were identified by electronic search over the period 2002 to July 2018. A Bayesian meta-regression model was employed, using the risk difference metric to estimate the relationship between mortality difference and control arm risk, and estimate the mortality difference with and without adjusting for control arm risk.

RESULTS

Of 780 initially identified published systematic reviews, 113 had appropriate mortality data comprising 123 analysable groups. The 123 meta-analyses were pharmaceutical therapeutic (59.3%), non-pharmaceutical therapeutic (24.4%) and nutritional (16.3%), with a 25% overall average control arm mortality. In 25/123 (20%) analyses, meta-regression indicated significant baseline risk (Bayesian 95% credible intervals excluding zero). In all analyses, the relationship between risk-difference and control arm risk was negative indicating a positive treatment effect with increasing control arm risk. Adjusted estimates identified six studies with significant positive treatment effects, not evident until after adjustment for control arm risk.

CONCLUSION

Underlying risk-related therapy is apparent in meta-analyses of the critically-ill and identification is of importance to both the conduct and interpretation of these meta-analyses.

Post Anaesthesia Care Unit Discharge: A Clinical Scoring System versus Traditional Time-based Criteria

This prospective cohort analysis compared the efficiency of time-based discharge criteria (Group 1) to a modified clinical scoring system (Group 2), incorporating the assessment of pain and temperature, in the post anaesthesia care unit (PACU). Two consecutive series of patients (n=292) were assessed following general anaesthesia for various surgical procedures. The time taken for patients to satisfy their respective discharge criteria was recorded as PACU length of stay (LOS). Patient group and other factors that may have influenced PACU-LOS were examined using time-to-event analysis. The raw PACU-LOS was not shown to be different between the two groups (log rank test, P=0.12). Covariate adjusted estimates were used to compare the two discharge criteria and also to identify other factors influencing PACU-LOS. The Cox regression model was poorly specified and a log-logistic accelerated failure time model was found to be the most parsimonious predictive model.

Predictors of decreased PACU-LOS were the treatment group (Group 2 versus Group 1) and the covariate recording anaesthetic airway choice (no endotracheal tube (ETT) versus ETT). Surgical time, as a linear function, intra- and postoperative opioid administration, as well as postoperative antiemetic use were predictors of increased PACU-LOS. Patient age, gender, urgency of surgery, and ASA classification were not predictive of PACU-LOS. Using covariate adjusted estimates, the new PACU discharge criteria, based on the Aldrete's scoring system, was associated with a significantly reduced PACU-LOS in comparison with time-based criteria.

Assessment of Outcome over a 10-year Period of Patients Admitted to a Multidisciplinary Adult Intensive Care Unit with Haematological and Solid Tumours

The risk factors for time to mortality, censored at 30 days, of patients admitted to an adult teaching hospital ICU with haematological and solid malignancies were assessed in a retrospective cohort study.

Patients, demographics and daily ICU patient data, from admission to day 8, were identified from a prospective computerized database and casenote review in consecutive admissions to ICU with haematological and solid tumours over a 10-year period (1989–99). The cohort, 108 ICU admissions in 89 patients was of mean age (±SD) 55±14 years; 43% were female. Patient diagnoses were leukaemia (35%), lymphoma (38%) and solid tumours (27%). Median time from hospital to ICU admission was five days (range 0–67). On ICU admission, 50% had septic shock and first day APACHE II score was 28±9. Forty-six per cent of patients were ventilated. ICU and 30-day mortality were 39% and 54% respectively. Multivariate Cox model predictors (P<0.05), using only ICU admission day data were: Charlson comorbidity index (CCI), time to ICU admission (days) and mechanical ventilation. For daily data (admission through day 8), predictors were: cohort effect (2nd vs 1st five-year period); CCI; time to ICU admission (days); APACHE II score and mechanical ventilation.

Outcomes were considered appropriate for severity of illness and demonstrated improvement over time. Ventilation was an independent outcome determinant. Controlling for other factors, mortality has improved over time (1st vs 2nd five year period). Analysis restricted to admission data alone may be insensitive to particular covariate effects.

Modelling Thirty-day Mortality in the Acute Respiratory Distress Syndrome (ARDS) in an Adult ICU

Variables predicting thirty-day outcome from Acute Respiratory Distress Syndrome (ARDS) were analysed using Cox regression structured for time-varying covariates. Over a three-year period, 1996–1998, consecutive patients with ARDS (bilateral chest X-ray opacities, PaO2/FiO2 ratio of <200 and an acute precipitating event) were identified using a prospective computerized data base in a university teaching hospital ICU.

The cohort, 106 mechanically ventilated patients, was of mean (SD) age 63.5 (15.5) years and 37% were female. Primary lung injury occurred in 45% and 24% were postoperative. ICU-admission day APACHE II score was 25 (8); ARDS onset time from ICU admission was 1 day (median: range 0-16) and 30 day mortality was 41% (95% CI: 33%-51%). At ARDS onset, PaO2/FiO2 ratio was 92 (31), 81% had four-quadrant chest X-ray opacification and lung injury score was 2.75 (0.45). Average mechanical ventilator tidal volume was 10.3 ml/ predicted kg weight. Cox model mortality predictors (hazard ratio, 95% CI) were: APACHE II score, 1.15 (1.09-1.21); ARDS lag time (days), 0.72 (0.58-0.89); direct versus indirect injury, 2.89 (1.45-5.76); PaO2/FiO2 ratio, 0.98 (0.97-0.99); operative versus non-operative category, 0.24 (0.09-0.63). Time-varying effects were evident for PaO2/FiO2 ratio, operative versus non-operative category and ventilator tidal volume assessed as a categorical predictor with a cut-point of 8 ml/kg predicted weight (mean tidal volumes, 7.1 (1.9) vs 10.7 (1.6) ml/kg predicted weight). Thirty-day survival was improved for patients ventilated with lower tidal volumes.

Survival predictors in ARDS were multifactorial and related to patient-injury-time interaction and level of mechanical ventilator tidal volume.

The search for biomarkers in the critically ill: a cautionary tale

The search for biomarkers has been described as a dismal patchwork of fragmented research. We review biomarkers in sepsis in the critically ill in terms of conventional single circulating proteins. Despite sepsis biomarker publications trebling over the past 6 years, currently only one, procalcitonin, has materialised promise. We survey genomic biomarker initiatives, single nucleotide polymorphisms (SNPs) and gene signatures. Despite many SNP associations with sepsis susceptibility and a limited number of genome-wide association studies, the status of these associations is that of genomic signposts only. The standing of gene signatures in the paradigmatic discipline, breast cancer, is described. Uncertainties in the understanding of the sepsis process are documented - the dissociation between blood and tissue element activity, or compartmentalisation. The paradox of the active search for gene signatures to refine the sepsis phenotype and discover target subtypes for new therapies in the absence of such therapies is presented.

Genome-wide common and rare variant analysis provides novel insights into clozapine-associated neutropenia

This corrects the article DOI: 10.1038/mp.2016.97.

Reconsidering lactate as a sepsis risk biomarker

Objectives

There has been renewed interest in lactate as a risk biomarker in sepsis and septic shock. However, the ability of the odds ratio (OR) and change in the area under the receiver operator characteristic curve (AUC-ROC) to assess biomarker added-value has been questioned.

Design, setting and participants

A sepsis cohort was identified from the ICU database of an Australian tertiary referral hospital using APACHE III diagnostic codes. Demographic information, APACHE III scores, 24-hour post-admission patient lactate levels, and hospital mortality were accessed.

Measurements and main results

Hospital mortality was modelled using a base predictive logistic regression model and sequential addition of admission lactate, lactate clearance ([lactate^admission—lactate^final]/lactate^admission), and area under the lactate-time curve (LTC). Added-value was assessed using lactate index OR; AUC-ROC difference (base-model versus lactate index addition); net (mortality) reclassification index (NRI; range -2 to +2); and net benefit (NB), the number of true positives per patient adjusted for the number of false positives. The data set comprised 717 patients with mean(SD) age and APACHE III score 61.1(16.5) years and 68.3(28.2) respectively; 59.2% were male. Admission lactate was 2.3(2.5) mmol/l; with lactate of ≥ 4 mmol/L (37% hospital mortality) in 17% and patients with lactate < 4 mmol/L having 18% hospital mortality. The admission base-model had an AUC-ROC = 0.81 with admission lactate OR = 1.127 (95%CI: 1.038, 1.224), AUC-ROC difference of 0.0032 (-0.0037, 0.01615; P = 0.61), and NRI 0.240(0.030, 0.464). The over-time model had an AUC-ROC = 0.86 with (i) clearance OR = 0.771, 95%CI: 0.578, 1.030; P = 0.08; AUC-ROC difference 0.001 (-0.003, 0.014; P = 0.78), and NRI 0.109(-0.193, 0.425) and (ii) LTC OR = 0.997, 95%CI: 0.989, 1.005, P = 0.49; AUC-ROC difference 0.004 (-0.002, 0.004; P = 0.34), and NRI 0.111(-0.222, 0.403). NB was not incremented by any lactate index.

Conclusions

Lactate added-value assessment is dependent upon the performance of the underlying predictive model and should incorporate risk reclassification and net benefit measures.

Continuous transversus abdominis plane block vs intermittent bolus for analgesia after abdominal surgery: a randomized trial

BACKGROUND

Continuous and intermittent bolus techniques of transversus abdominis plane (TAP) blocks have been used for analgesia after abdominal surgery. Although both are effective, there are no studies comparing them. The aim of this study is to compare analgesia and cost-effectiveness between these groups.

METHODS

After obtaining ethical approval, 20 American Society of Anesthesiologists ASA grade I to III patients undergoing elective abdominal surgery were recruited with 10 patients allocated to each arm. Bilateral ultrasound-guided TAP blocks were performed with an initial bolus of 0.5% ropivacaine 20 mL per side, followed by catheter insertion. After surgery, the continuous infusion group received 0.2% ropivacaine 8 mL/hour on each side and the intermittent bolus group received doses of 0.2% ropivacaine 20 mL per side every 8 hours for 48 hours. Both groups received intravenous fentanyl patient-controlled analgesia and regular oral paracetamol. Parameters recorded included numerical rating scores for pain and post-operative analgesic consumption at baseline (time 0) and at 1 hour, 1 day and 2 days post-operatively. The duration of catheter insertion, complications, patient satisfaction and information regarding costs were also recorded. Patient satisfaction was assessed utilizing a 4-point "Likert" scale on day 2 and on day 30. Pain and Likert scores were analysed by non-parametric sum rank test and all two-sampled t-tests assumed unequal variances.

RESULTS

There was no difference between duration of TAP block, anesthetic and surgical technique and length of stay (p=0.23). Primary outcomes: pain scores at rest and cough were not significantly different (p=0.20) between the groups. Satisfaction scores were similar at day 2 and 30 (p=0.77). However, the bolus group was more cost-effective (AU$347.98 vs AU$429.43).

CONCLUSION

Continuous or bolus TAP blocks are effective analgesic techniques in abdominal surgery, with bolus technique being more economical.

Evidence for genetic heterogeneity between clinical subtypes of bipolar disorder

We performed a genome-wide association study of 6447 bipolar disorder (BD) cases and 12 639 controls from the International Cohort Collection for Bipolar Disorder (ICCBD). Meta-analysis was performed with prior results from the Psychiatric Genomics Consortium Bipolar Disorder Working Group for a combined sample of 13 902 cases and 19 279 controls. We identified eight genome-wide significant, associated regions, including a novel associated region on chromosome 10 (rs10884920; P=3.28 × 10-8) that includes the brain-enriched cytoskeleton protein adducin 3 (ADD3), a non-coding RNA, and a neuropeptide-specific aminopeptidase P (XPNPEP1). Our large sample size allowed us to test the heritability and genetic correlation of BD subtypes and investigate their genetic overlap with schizophrenia and major depressive disorder. We found a significant difference in heritability of the two most common forms of BD (BD I SNP-h2=0.35; BD II SNP-h2=0.25; P=0.02). The genetic correlation between BD I and BD II was 0.78, whereas the genetic correlation was 0.97 when BD cohorts containing both types were compared. In addition, we demonstrated a significantly greater load of polygenic risk alleles for schizophrenia and BD in patients with BD I compared with patients with BD II, and a greater load of schizophrenia risk alleles in patients with the bipolar type of schizoaffective disorder compared with patients with either BD I or BD II. These results point to a partial difference in the genetic architecture of BD subtypes as currently defined.

Genome-wide association study identifies SESTD1 as a novel risk gene for lithium-responsive bipolar disorder

This corrects the article DOI: 10.1038/mp.2015.165.

Genome-wide common and rare variant analysis provides novel insights into clozapine-associated neutropenia

This corrects the article DOI: 10.1038/mp.2016.97.

New initiatives in critical care: distinguishing hype from hope

Recent viewpoints on critical care have expressed frustration at the slow development of new therapeutic agents and the failure of investigator-initiated trials. Several new directions have been proposed: personalised medicine and the embracing of "omic" technologies, resolving the heterogeneity of treatment effects, and adaptive trial designs. We examine these approaches in the context of analysis of randomised controlled trials (RCTs). The curse of treatment effect heterogeneity is found not only in critical care but also in cancer oncology. We find the uncritical appeal to personalised medicine to be misplaced because such treatments are not identified at the personal level, but at the group or stratified level. The analysis of RCTs has foundered over the problem of accounting for the centre effect and rejecting the random effects approach. Enthusiasm for adaptive trial designs has been articulated at the rhetorical, not the substantive, level.

Nonlinear electrophoresis in the presence of dielectric decrement

The nonlinear phenomena that occur in the electric double layer (EDL) that forms at charged surfaces strongly influence electrokinetic effects, including electro-osmosis and electrophoresis. In particular, saturation effects due to either dielectric decrement or ion crowding effects are of paramount importance. Dielectric decrement significantly influences the ionic concentration in the EDL at high ζ potential, leading to the formation of a condensed layer near the particle's surface. In this article, we present a model incorporating both steric effects due to the finite size of ions and dielectric decrement to describe the physics in the electric double layer. The model remains valid in both weakly and strongly nonlinear regimes, as long as the electric double layer remains in quasiequilibrium. We apply this model to the study of two archetypal problems in electrokinetics, namely the electrophoresis of particles with fixed surface charges and the electrophoresis of ideally polarizable particles.

Calorie delivery and clinical outcomes in the critically ill: a systematic review and meta-analysis

OBJECTIVES

To determine the effect of calorie delivery on hospital mortality among critically ill adults receiving enteral nutrition (EN). Secondary outcomes included the effect of calorie delivery on intensive care unit and hospital length of stay (LOS), duration of mechanical ventilation (MV) and incidence of new-onset pneumonia.

METHODS

We identified randomised clinical trials of EN, with or without supplemental parenteral nutrition (PN), involving adult ICU patients for whom mortality data were available, and when there was a significant difference in calorie supplementation between intervention arms (P < 0.05). We searched English language electronic databases (1946-2014), bibliographies of nutrition society guidelines and high-impact nutrition and critical care journals. We calculated summary odds ratio (OR) estimates and 95% confidence intervals using a random effects estimator, and used meta-regression to assess the effect on mortality of average calories delivered.

RESULTS

Of 1545 articles identified, 16 eligible studies involving 3473 patients were included. Five studies involved supplemental PN. Mean calorie delivery ranged from 126 kcal/day (SD, 115 kcal/day) to 2086 kcal/day (SD, 460 kcal/day). Mortality was 26.0% in the lower calorie delivery group and 26.5% in the higher calorie delivery group. There was no effect of increased calorie delivery on mortality (OR, 1.02; 95% CI , 0.85-1.24; P = 0.27; I2 = 16.3%). ICU and hospital LOS and incidence of newonset pneumonia did not differ between groups. Duration of MV was decreased with lower calorie delivery (weighted mean difference, 2.92 days; 95% CI, -4.49 to -1.35 days; P < 0.001; I2 = 14.7%). Meta-regression analysis did not show an overall effect on mortality of average calories delivered (P = 0.73; I2 = 40.8%).

CONCLUSION

Delivery of increased calories via the enteral route, with or without supplemental PN, was not associated with a survival benefit.

Exome arrays capture polygenic rare variant contributions to schizophrenia

Schizophrenia is a highly heritable disorder. Genome-wide association studies based largely on common alleles have identified over 100 schizophrenia risk loci, but it is also evident from studies of copy number variants (CNVs) and from exome-sequencing studies that rare alleles are also involved. Full characterization of the contribution of rare alleles to the disorder awaits the deployment of sequencing technology in very large sample sizes, meanwhile, as an interim measure, exome arrays allow rare non-synonymous variants to be sampled at a fraction of the cost. In an analysis of exome array data from 13 688 individuals (5585 cases and 8103 controls) from the UK, we found that rare (minor allele frequency < 0.1%) variant association signal was enriched among genes that map to autosomal loci that are genome-wide significant (GWS) in common variant studies of schizophrenia genome-wide association study (P_GWAS = 0.01) as well as gene sets known to be enriched for rare variants in sequencing studies (P_RARE = 0.026). We also identified the gene-wise equivalent of GWS support for WDR88 (WD repeat-containing protein 88), a gene of unknown function (P = 6.5 × 10⁻⁷). Rare alleles represented on exome chip arrays contribute to the genetic architecture of schizophrenia, but as is the case for GWAS, very large studies are required to reveal additional susceptibility alleles for the disorder.

Assessing changes over time in healthcare provider performance: addressing regression to the mean over multiple time points

In recent years, the evaluation of healthcare provider performance has become standard for governments, insurance companies, and other stakeholders. Often, performance is compared across providers using indicators in one time period, for example a year. However it is often important to assess changes in the performance of individual providers over time. Such analyses can be used to determine if any providers have significant improvements, deteriorations, unusual patterns or systematic changes in performance. Studies which monitor healthcare provider performance in this way have to date typically been limited to comparing performance in the most recent period with performance in a previous period. It is also important to consider a longer-term view of performance and assess changes over more than two periods. In this paper, we develop test statistics that account for variable numbers of prior performance indicators, and show that these are particularly useful for assessing consecutive improvements or deteriorations in performance. We apply the tests to coronary artery bypass graft mortality rates in New York State hospitals, and mortality data from Australian and New Zealand intensive care units. Although our applications are to medical data, the new tests have broad application in other areas.

Clinical features of organophosphate poisoning: A review of different classification systems and approaches

PURPOSE

The typical toxidrome in organophosphate (OP) poisoning comprises of the Salivation, Lacrimation, Urination, Defecation, Gastric cramps, Emesis (SLUDGE) symptoms. However, several other manifestations are described. We review the spectrum of symptoms and signs in OP poisoning as well as the different approaches to clinical features in these patients.

MATERIALS AND METHODS

Articles were obtained by electronic search of PubMed(®) between 1966 and April 2014 using the search terms organophosphorus compounds or phosphoric acid esters AND poison or poisoning AND manifestations.

RESULTS

Of the 5026 articles on OP poisoning, 2584 articles pertained to human poisoning; 452 articles focusing on clinical manifestations in human OP poisoning were retrieved for detailed evaluation. In addition to the traditional approach of symptoms and signs of OP poisoning as peripheral (muscarinic, nicotinic) and central nervous system receptor stimulation, symptoms were alternatively approached using a time-based classification. In this, symptom onset was categorized as acute (within 24-h), delayed (24-h to 2-week) or late (beyond 2-week). Although most symptoms occur with minutes or hours following acute exposure, delayed onset symptoms occurring after a period of minimal or mild symptoms, may impact treatment and timing of the discharge following acute exposure. Symptoms and signs were also viewed as an organ specific as cardiovascular, respiratory or neurological manifestations. An organ specific approach enables focused management of individual organ dysfunction that may vary with different OP compounds.

CONCLUSIONS

Different approaches to the symptoms and signs in OP poisoning may better our understanding of the underlying mechanism that in turn may assist with the management of acutely poisoned patients.

Use of a concentrated enteral nutrition solution to increase calorie delivery to critically ill patients: a randomized, double-blind, clinical trial

BACKGROUND

Critically ill patients typically receive ∼60% of estimated calorie requirements.

OBJECTIVES

We aimed to determine whether the substitution of a 1.5-kcal/mL enteral nutrition solution for a 1.0-kcal/mL solution resulted in greater calorie delivery to critically ill patients and establish the feasibility of conducting a multicenter, double-blind, randomized trial to evaluate the effect of an increased calorie delivery on clinical outcomes.

DESIGN

A prospective, randomized, double-blind, parallel-group, multicenter study was conducted in 5 Australian intensive care units. One hundred twelve mechanically ventilated patients expected to receive enteral nutrition for ≥2 d were randomly assigned to receive 1.5 (n = 57) or 1.0 (n = 55) kcal/mL enteral nutrition solution at a rate of 1 mL/kg ideal body weight per hour for 10 d. Protein and fiber contents in the 2 solutions were equivalent.

RESULTS

The 2 groups had similar baseline characteristics (1.5 compared with 1.0 kcal/mL). The mean (±SD) age was 56.4 ± 16.8 compared with 56.5 ± 16.1 y, 74% compared with 75% were men, and the Acute Physiology and Chronic Health Evaluation II score was 23 ± 9.1 compared with 22 ± 8.9. The groups received similar volumes of enteral nutrition solution [1221 mL/d (95% CI: 1120, 1322 mL/d) compared with 1259 mL/d (95% CI: 1143, 1374 mL/d); P = 0.628], which led to a 46% increase in daily calories in the group given the 1.5-kcal/mL solution [1832 kcal/d (95% CI: 1681, 1984 kcal/d) compared with 1259 kcal/d (95% CI: 1143, 1374 kcal/d); P < 0.001]. The 1.5-kcal/mL solution was not associated with larger gastric residual volumes or diarrhea. In this feasibility study, there was a trend to a reduced 90-d mortality in patients given 1.5 kcal/mL [11 patients (20%) compared with 20 patients (37%); P = 0.057].

CONCLUSIONS

The substitution of a 1.0- with a 1.5-kcal/mL enteral nutrition solution administered at the same rate resulted in a 46% greater calorie delivery without adverse effects. The results support the conduct of a large-scale trial to evaluate the effect of increased calorie delivery on clinically important outcomes in the critically ill.

Fixed effects modelling for provider mortality outcomes: Analysis of the Australia and New Zealand Intensive Care Society (ANZICS) Adult Patient Data-base

BACKGROUND

Risk adjusted mortality for intensive care units (ICU) is usually estimated via logistic regression. Random effects (RE) or hierarchical models have been advocated to estimate provider risk-adjusted mortality on the basis that standard estimators increase false outlier classification. The utility of fixed effects (FE) estimators (separate ICU-specific intercepts) has not been fully explored.

METHODS

Using a cohort from the Australian and New Zealand Intensive Care Society Adult Patient Database, 2009-2010, the model fit of different logistic estimators (FE, random-intercept and random-coefficient) was characterised: Bayesian Information Criterion (BIC; lower values better), receiver-operator characteristic curve area (AUC) and Hosmer-Lemeshow (H-L) statistic. ICU standardised hospital mortality ratios (SMR) and 95%CI were compared between models. ICU site performance (FE), relative to the grand observation-weighted mean (GO-WM) on odds ratio (OR), risk ratio (RR) and probability scales were assessed using model-based average marginal effects (AME).

RESULTS

The data set consisted of 145355 patients in 128 ICUs, years 2009 (47.5%) & 2010 (52.5%), with mean(SD) age 60.9(18.8) years, 56% male and ICU and hospital mortalities of 7.0% and 10.9% respectively. The FE model had a BIC = 64058, AUC = 0.90 and an H-L statistic P-value = 0.22. The best-fitting random-intercept model had a BIC = 64457, AUC = 0.90 and H-L statistic P-value = 0.32 and random-coefficient model, BIC = 64556, AUC = 0.90 and H-L statistic P-value = 0.28. Across ICUs and over years no outliers (SMR 95% CI excluding null-value = 1) were identified and no model difference in SMR spread or 95%CI span was demonstrated. Using AME (OR and RR scale), ICU site-specific estimates diverged from the GO-WM, and the effect spread decreased over calendar years. On the probability scale, a majority of ICUs demonstrated calendar year decrease, but in the for-profit sector, this trend was reversed.

CONCLUSIONS

The FE estimator had model advantage compared with conventional RE models. Using AME, between and over-year ICU site-effects were easily characterised.

Copy number variation in schizophrenia in Sweden

Schizophrenia (SCZ) is a highly heritable neuropsychiatric disorder of complex genetic etiology. Previous genome-wide surveys have revealed a greater burden of large, rare copy number variations (CNVs) in SCZ cases and identified multiple rare recurrent CNVs that increase risk of SCZ although with incomplete penetrance and pleiotropic effects. Identification of additional recurrent CNVs and biological pathways enriched for SCZ CNVs requires greater sample sizes. We conducted a genome-wide survey for CNVs associated with SCZ using a Swedish national sample (4719 cases and 5917 controls). High-confidence CNV calls were generated using genotyping array intensity data, and their effect on risk of SCZ was measured. Our data confirm increased burden of large, rare CNVs in SCZ cases as well as significant associations for recurrent 16p11.2 duplications, 22q11.2 deletions and 3q29 deletions. We report a novel association for 17q12 duplications (odds ratio=4.16, P=0.018), previously associated with autism and mental retardation but not SCZ. Intriguingly, gene set association analyses implicate biological pathways previously associated with SCZ through common variation and exome sequencing (calcium channel signaling and binding partners of the fragile X mental retardation protein). We found significantly increased burden of the largest CNVs (>500 kb) in genes present in the postsynaptic density, in genomic regions implicated via SCZ genome-wide association studies and in gene products localized to mitochondria and cytoplasm. Our findings suggest that multiple lines of genomic inquiry--genome-wide screens for CNVs, common variation and exonic variation--are converging on similar sets of pathways and/or genes.

The role of non-invasive positive pressure ventilation in post-extubation respiratory failure: An evaluation using meta-analytic techniques

Background:

The use of non-invasive positive pressure ventilation (NIPPV) in post-extubation respiratory failure is not well-established. Meta-analytic techniques were used to assess the effects of prophylactic application of NIPPV (prior to the development of respiratory failure) and therapeutic application of NIPPV (subsequent to the development of respiratory failure).

Materials and Methods:

Randomized controlled trials (RCTs) from 1966 to May 2010 were identified using electronic databases. RCTs, which reported the use of NIPPV in post-extubation respiratory failure with defined assessable endpoints: reintubation, mortality and length of stay, were included.

Results:

Reintubation was the primary outcome, mortality and lengths of stay were the secondary outcomes. Risk ratios (RR) were calculated for discrete outcomes and weighted mean differences (WMD) for continuous measures. There were 13 trials with 1420 patients; 9 prophylactic with 861 patients and 4 therapeutic with 559 patients. In the prophylactic group, NIPPV was associated with lower rates of reintubation: RR 0.53 (95% confidence interval [CI], 0.28-0.98), P = 0.04. In the therapeutic group, NIPPV showed a null effect on reintubation: RR 0.79 (95% CI, 0.50-1.25), P = 0.31. The analysis on the secondary outcomes suggested significant reduction of hospital mortality with prophylactic application of NIPPV: RR 0.62 (95% CI 0.4-0.97), P = 0.03, with no effect on the other outcomes. Therapeutic application of NIPPV reduced intensive care unit length of stay: WMD −1.17 (95% CI −2.82 to −0.33), P = 0.006, but no effect on the other secondary outcomes.

Conclusions:

The results of this review suggested prophylactic NIPPV was beneficial with respect to reintubation and the therapeutic use of NIPPV showed a null effect.

Performance of clinical scoring systems in acute organophosphate poisoning

INTRODUCTION

Clinical scoring systems are used to predict mortality rate in hospitalized patients. Their utility in organophosphate (OP) poisoning has not been well studied.

METHODS

In this retrospective study of 396 patients, we evaluated the performance of the Acute Physiology and Chronic Health Evaluation (APACHE) II score, the Simplified Acute Physiology Score (SAPS) II, Mortality Prediction Model (MPM) II, and the Poisoning Severity Score (PSS). Demographic, laboratory, and survival data were recorded. Receiver operating characteristic (ROC) curves were generated, and the area under the curve (AUC) was calculated to study the relationship between individual scores and mortality rate.

RESULTS

The mean (standard deviation) age of the patients was 31.4 (12.7) years, and at admission, their pseudocholinesterase (median, interquartile) level was 317 (222-635) U/L. Mechanical ventilation was required in 65.7% of the patients and the overall mortality rate was 13.1%. The mean (95% confidence interval) scores were as follows: APACHE-II score, 16.4 (15.5-17.3); SAPS-II, 34.4 (32.5-36.2); MPM-II score, 28.6 (25.7-31.5); and PSS, 2.4 (2.3-2.5). Overall, the AUC for mortality was significantly higher for APACHE-II (0.77) and SAPS-II (0.77) than the PSS (0.67). When patients were categorized, the AUCs were better for WHO Class II (0.71-0.82) than that for Class I compounds (0.60-0.66). For individual compounds, the AUC for APACHE-II was highest in quinalphos (0.93, n = 46) and chlorpyrifos (0.86, n = 38) and lowest in monocrotophos (0.60, n = 63). AUCs for SAPS-II and MPM-II were marginally but not significantly lower than those for APACHE-II. The PSS was generally a poorer discriminator compared to the other scoring systems across all categories.

CONCLUSIONS

In acute OP poisoning, the generic scoring systems APACHE-II and SAPS-II outperform the PSS. These tools may be used to predict the mortality rate in OP poisoning.

Statistical process control of mortality series in the Australian and New Zealand Intensive Care Society (ANZICS) adult patient database: implications of the data generating process

Background

Statistical process control (SPC), an industrial sphere initiative, has recently been applied in health care and public health surveillance. SPC methods assume independent observations and process autocorrelation has been associated with increase in false alarm frequency.

Methods

Monthly mean raw mortality (at hospital discharge) time series, 1995–2009, at the individual Intensive Care unit (ICU) level, were generated from the Australia and New Zealand Intensive Care Society adult patient database. Evidence for series (i) autocorrelation and seasonality was demonstrated using (partial)-autocorrelation ((P)ACF) function displays and classical series decomposition and (ii) “in-control” status was sought using risk-adjusted (RA) exponentially weighted moving average (EWMA) control limits (3 sigma). Risk adjustment was achieved using a random coefficient (intercept as ICU site and slope as APACHE III score) logistic regression model, generating an expected mortality series. Application of time-series to an exemplar complete ICU series (1995-(end)2009) was via Box-Jenkins methodology: autoregressive moving average (ARMA) and (G)ARCH ((Generalised) Autoregressive Conditional Heteroscedasticity) models, the latter addressing volatility of the series variance.

Results

The overall data set, 1995-2009, consisted of 491324 records from 137 ICU sites; average raw mortality was 14.07%; average(SD) raw and expected mortalities ranged from 0.012(0.113) and 0.013(0.045) to 0.296(0.457) and 0.278(0.247) respectively. For the raw mortality series: 71 sites had continuous data for assessment up to or beyond lag₄₀ and 35% had autocorrelation through to lag₄₀; and of 36 sites with continuous data for ≥ 72 months, all demonstrated marked seasonality. Similar numbers and percentages were seen with the expected series. Out-of-control signalling was evident for the raw mortality series with respect to RA-EWMA control limits; a seasonal ARMA model, with GARCH effects, displayed white-noise residuals which were in-control with respect to EWMA control limits and one-step prediction error limits (3SE). The expected series was modelled with a multiplicative seasonal autoregressive model.

Conclusions

The data generating process of monthly raw mortality series at the ICU level displayed autocorrelation, seasonality and volatility. False-positive signalling of the raw mortality series was evident with respect to RA-EWMA control limits. A time series approach using residual control charts resolved these issues.

Evaluating the performance of Australian and New Zealand intensive care units in 2009 and 2010

The Australian and New Zealand Intensive Care Society Adult Patient Database (ANZICS APD) is one of the largest databases of its kind in the world and collects individual admissions' data from intensive care units (ICUs) around Australia and New Zealand. Use of this database for monitoring and comparing the performance of ICUs, quantified by the standardised mortality ratio, poses several theoretical and computational challenges, which are addressed in this paper. In particular, the expected number of deaths must be appropriately estimated, the ICU casemix adjustment must be adequate, statistical variation must be fully accounted for, and appropriate adjustment for multiple comparisons must be made. Typically, one or more of these issues have been neglected in ICU comparison studies. Our approach to the analysis proceeds by fitting a random coefficient hierarchical logistic regression model for the inhospital death of each patient, with patients clustered within ICUs. We anticipate the majority of ICUs will be estimated as performing 'usually' after adjusting for important clinical covariates. We take as a starting point the ideas in Ohlssen et al and estimate an appropriate null model that we expect these ICUs to follow, taking a frequentist rather than a Bayesian approach. This methodology allows us to rigorously account for the aforementioned statistical issues and to determine if there are any ICUs contributing to the Australian and New Zealand Intensive Care Society database that have comparatively unusual performance. In addition to investigating the yearly performance of the ICUs, we also estimate changes in individual ICU performance between 2009 and 2010 by adjusting for regression-to-the-mean.

Genome-wide association study in a Swedish population yields support for greater CNV and MHC involvement in schizophrenia compared with bipolar disorder

Schizophrenia (SCZ) and bipolar disorder (BD) are highly heritable psychiatric disorders with overlapping susceptibility loci and symptomatology. We conducted a genome-wide association study (GWAS) of these disorders in a large Swedish sample. We report a new and independent case-control analysis of 1507 SCZ cases, 836 BD cases and 2093 controls. No single-nucleotide polymorphisms (SNPs) achieved significance in these new samples; however, combining new and previously reported SCZ samples (2111 SCZ and 2535 controls) revealed a genome-wide significant association in the major histocompatibility complex (MHC) region (rs886424, P=4.54 × 10(-8)). Imputation using multiple reference panels and meta-analysis with the Psychiatric Genomics Consortium SCZ results underscored the broad, significant association in the MHC region in the full SCZ sample. We evaluated the role of copy number variants (CNVs) in these subjects. As in prior reports, deletions were enriched in SCZ, but not BD cases compared with controls. Singleton deletions were more frequent in both case groups compared with controls (SCZ: P=0.003, BD: P=0.013), whereas the largest CNVs (>500 kb) were significantly enriched only in SCZ cases (P=0.0035). Two CNVs with previously reported SCZ associations were also overrepresented in this SCZ sample: 16p11.2 duplications (P=0.0035) and 22q11 deletions (P=0.03). These results reinforce prior reports of significant MHC and CNV associations in SCZ, but not BD.

Assessing the performance of a continuous infusion for potassium supplementation in the critically ill

Hypokalaemia is a common problem in critically ill patients, which if untreated, can result in dysrhythmia or another adverse outcome. We assessed the safety and efficacy of a continuous infusion of potassium chloride versus an existing intermittent infusion regimen. In this open-label randomised parallel-arm active-controlled pilot study, critically ill adults with plasma potassium concentration between 2.5 and 3.8 mmol/l were randomised to receive either a continuous infusion or intermittent infusions of potassium chloride for establishment and maintenance of normokalaemia. The primary outcome was the mean difference in plasma potassium concentration over time between the two study arms as assessed by a linear mixed-effects model. Although a statistically significant difference was observed (0.22 mmol/l; 95% confidence interval 0.17, 0.27; P <0.0001), this did not reach the pre-determined level indicative of a treatment effect (0.5 mmol/l). The continuous group demonstrated less variance in (mean) plasma potassium as reflected in narrower confidence intervals in a prediction-by-time model. The incidence rate ratio of dysrhythmia, assessed by a mixed-effects Poisson model, was similar in each group (0.62; 95% confidence interval 0.32, 1.21; P=0.16). We recorded no adverse events directly attributable to infusion of potassium chloride in either study arm. Although titrated continuous infusion did not demonstrate a clinically important difference by comparison with intermittent infusions for the maintenance of normokalaemia, there was more consistent control of plasma potassium with no observed complications or adverse events. Therefore, this trial showed an acceptable efficacy and safety profile for the continuous infusion regimen, suggesting scope for further study.

Enteral nutrition in Australian and New Zealand intensive care units: a point-prevalence study of prescription practices

BACKGROUND

Enteral nutrition (EN) is widely accepted as the preferred method for providing nutrition therapy to critically ill patients. However, optimal energy goals and the best way to achieve those goals are ill defined.

OBJECTIVE

To determine the type and energy concentration of commonly prescribed EN formulations and whether energy-dense formulations (> 1 kcal/mL) are used.

DESIGN

Prospective, observational, multicentre, single-day, point-prevalence study.

PARTICIPANTS AND SETTING

All patients present in 38 Australian and New Zealand intensive care units at 10:00 on 17 November 2010.

MAIN OUTCOME MEASURES

Demographic data, admission diagnosis and information on EN administration were collected.

RESULTS

522 patients were enrolled. Mean age was 58.7 (SD, 17.3) years, 65% were male and 79% were mechanically ventilated. On study day, 220/522 patients received EN (43%; 95% CI, 39%-48%). ICU admission source, Acute Physiology and Chronic Health Evaluation (APACHE) III diagnostic category, APACHE II score and ventilation on study day predicted receipt of EN. Of those receiving EN, 111/220 (51%; 95% CI, 44%-57%) received a 1 kcal/mL formulation and the remainder received an energy-dense formulation - 2 kcal/mL, 39/220 (18%; 95% CI, 13%-23%); and 1.5 kcal/mL, 32/220 (15%; 95% CI, 10%-20%). There were no significant predictors for receipt of energy-dense versus 1 kcal/mL EN.

CONCLUSIONS

1 kcal/mL and energy-dense formulations are administered with about equal frequency in Australian and New Zealand ICUs. This finding supports future research into the evaluation of optimal nutritional delivery amounts using EN formulations with differing energy concentrations.

A review of statistical estimators for risk-adjusted length of stay: analysis of the Australian and new Zealand Intensive Care Adult Patient Data-Base, 2008-2009

BACKGROUND

For the analysis of length-of-stay (LOS) data, which is characteristically right-skewed, a number of statistical estimators have been proposed as alternatives to the traditional ordinary least squares (OLS) regression with log dependent variable.

METHODS

Using a cohort of patients identified in the Australian and New Zealand Intensive Care Society Adult Patient Database, 2008-2009, 12 different methods were used for estimation of intensive care (ICU) length of stay. These encompassed risk-adjusted regression analysis of firstly: log LOS using OLS, linear mixed model [LMM], treatment effects, skew-normal and skew-t models; and secondly: unmodified (raw) LOS via OLS, generalised linear models [GLMs] with log-link and 4 different distributions [Poisson, gamma, negative binomial and inverse-Gaussian], extended estimating equations [EEE] and a finite mixture model including a gamma distribution. A fixed covariate list and ICU-site clustering with robust variance were utilised for model fitting with split-sample determination (80%) and validation (20%) data sets, and model simulation was undertaken to establish over-fitting (Copas test). Indices of model specification using Bayesian information criterion [BIC: lower values preferred] and residual analysis as well as predictive performance (R2, concordance correlation coefficient (CCC), mean absolute error [MAE]) were established for each estimator.

RESULTS

The data-set consisted of 111663 patients from 131 ICUs; with mean(SD) age 60.6(18.8) years, 43.0% were female, 40.7% were mechanically ventilated and ICU mortality was 7.8%. ICU length-of-stay was 3.4(5.1) (median 1.8, range (0.17-60)) days and demonstrated marked kurtosis and right skew (29.4 and 4.4 respectively). BIC showed considerable spread, from a maximum of 509801 (OLS-raw scale) to a minimum of 210286 (LMM). R2 ranged from 0.22 (LMM) to 0.17 and the CCC from 0.334 (LMM) to 0.149, with MAE 2.2-2.4. Superior residual behaviour was established for the log-scale estimators. There was a general tendency for over-prediction (negative residuals) and for over-fitting, the exception being the GLM negative binomial estimator. The mean-variance function was best approximated by a quadratic function, consistent with log-scale estimation; the link function was estimated (EEE) as 0.152(0.019, 0.285), consistent with a fractional-root function.

CONCLUSIONS

For ICU length of stay, log-scale estimation, in particular the LMM, appeared to be the most consistently performing estimator(s). Neither the GLM variants nor the skew-regression estimators dominated.

Hyperfiltration in type 1 diabetes: does it exist and does it matter for nephropathy?

AIMS/HYPOTHESIS

Hyperfiltration is widely regarded as a contributing factor to the development of microalbuminuria and progressive nephropathy in type 1 diabetes. However, recent studies have questioned this conclusion.

METHODS

To address this conflicting evidence, we examined the association between hyperfiltration and progression to microalbuminuria in 2,318 adults with type 1 diabetes. We also compared the estimated GFR in our diabetic patients with rates observed in 6,247 adults from the Finnish general population, using age- and sex-specific z scores.

RESULTS

The distribution of estimated GFR in adults with type 1 diabetes and normoalbuminuria was not significantly different from that expected in the general population (p = 0.51, Mann-Whitney test). Type 1 diabetic patients with a higher estimated GFR were also no more likely to develop microalbuminuria over a median of 5.2 years of follow-up than those with normal estimated GFR. This was the case regardless of whether hyperfiltration was defined by an absolute threshold, deciles of estimated GFR or a z score, using creatinine- or cystatin-based clearance formulas in men or in women.

CONCLUSIONS/INTERPRETATION

Together with other studies, these data suggest that creatinine- or cystatin-based estimates of GFR do not predict the development of microalbuminuria in patients with type 1 diabetes. Moreover, in the absence of incipient or overt nephropathy, conventionally determined renal function in patients with type 1 diabetes appears no different from that in the general population. This is hardly surprising, given that these individuals, by all definitions, do not have kidney disease.

Robust meta-analytic conclusions mandate the provision of prediction intervals in meta-analysis summaries

OBJECTIVES

Results of meta-analyses typically conclude that future large studies may be mandated. However, the predictive ability of these estimates is deficient. We explored meta-analytic prediction intervals as means for providing a clear and appropriate future treatment summary reflecting current estimates.

STUDY DESIGN

A meta-epidemiological study of binary outcome critical care meta-analyses published between 2002 and 2010. Computation of 95% DerSimonian-Laird and Bayesian random-effects meta-analytic confidence intervals (CI) and 95% credible intervals (CrI), respectively, and frequentist (PI) and Bayesian (PrI) prediction intervals for odds ratio (OR) and risk ratio (RR) were undertaken. Bayesian calculations included the probability that the OR and RR point estimates ≥1.

RESULTS

Seventy-two meta-analyses from 70 articles were identified, containing between three and 80 studies each, with median nine studies. For both frequentist and Bayesian settings, 49-69% of the meta-analyses excluded the null. All significant CrI had high probabilities of efficacy/harm. The number of PI vs. PrI excluding 1 was 25% vs. 3% (OR), 26% vs. 3% (RR) of the total meta-analyses. Unsurprisingly, PI/PrI width was greater than CI/CrI width and increased with increasing heterogeneity and combination of fewer studies.

CONCLUSION

Robust meta-analytic conclusions and determination of studies warranting new large trials may be more appropriately signaled by consideration of initial interval estimates with prediction intervals. Substantial heterogeneity results in exceedingly wide PIs. More caution should be exercised regarding the conclusions of a meta-analysis.

Soluble receptor for AGE (RAGE) is a novel independent predictor of all-cause and cardiovascular mortality in type 1 diabetes

AIMS/HYPOTHESIS

Activation of the receptor for AGE (RAGE) is implicated in the development and progression of vascular complications of diabetes. In this study, we explore factors and mortality outcomes associated with soluble RAGE (sRAGE) in a multicentre nationwide cohort of Finnish adults with type 1 diabetes.

METHODS

Baseline sRAGE concentrations were estimated in 3,100 adults with type 1 diabetes. Clinical and biological variables independently associated with sRAGE were identified using multivariate regression analysis. Independent predictors of mortality were determined using Cox and Fine-Gray proportional-hazards models.

RESULTS

The main independent determinants of sRAGE concentrations were estimated glomerular filtration rate, albuminuria, body mass index, age, duration of diabetes, HbA(1c) and insulin dose (all p < 0.05). During a median of 9.1 years of follow-up there were 202 deaths (7.4 per 1,000 patient years). sRAGE was independently associated with all-cause (Cox model: HR 1.03) and cardiovascular mortality (Fine-Gray competing risks model: HR 1.06) such that patients with the highest sRAGE concentrations had the greatest risk of mortality, after adjusting for age, sex, macrovascular disease, HDL-cholesterol, HbA(1c), triacylglycerol, high-sensitivity C-reactive protein (hsCRP) and the presence and severity of chronic kidney disease. Although polymorphisms in the gene coding for RAGE were significantly associated with sRAGE concentrations, none were associated with mortality outcomes.

CONCLUSIONS/INTERPRETATION

Increased concentrations of sRAGE are associated with increased all-cause and cardiovascular mortality in type 1 diabetes, potentially reflecting the activation and production of RAGE in the context of accelerated vascular disease. These novel findings highlight the importance of the RAGE activation in the prevention and management of diabetic complications.

Variability in management of early severe sepsis

OBJECTIVE

A study was undertaken to characterise how doctors in emergency medicine (EM), acute medicine (AM) and critical care (ICU) in the UK, USA and Australia and New Zealand (ANZ) approach the initial resuscitative care of patients with severe sepsis.

METHODS

In 2007, members on the mailing lists of UK, US and ANZ EM, ICU and AM specialist organisations were invited to answer an anonymous scenario-based online survey. Respondents described their management of a patient with pneumonia and signs of sepsis. Multiple-choice questions were based on the Surviving Sepsis Campaign (SSC) 6-hour resuscitation bundle guidelines while avoiding the specific terms "sepsis" and "SSC guidelines".

RESULTS

The response rate was 21% (2461/11 795). Only two respondents (0.1%) complied with all SSC resuscitation recommendations. Inter-specialty and inter-country variations included differences in reporting initial lactate measurement (ranging from 30% in US-EM to 79% in UK-EM), fluid resuscitation targeting a central venous pressure of 8-12 mm Hg (from 15% in ANZ-ICU to 60% in UK-EM), blood transfusion for a central venous oxygen saturation <70% and haematocrit <30% (from 15% in ANZ-ICU to 70% in US-EM and UK-EM) and insertion of invasive monitoring (intra-arterial catheter: 89% in UK-ICU vs 20% in US-EM; central venous catheter: 83% in UK-ICU vs 44% in US-EM). 81% of respondents identified at least one reason why they did not implement all the recommendations; the reasons varied by region and specialty.

CONCLUSIONS

Reported management of early sepsis varies between specialities and countries, and the responses do not follow SSC guidelines. Concerns relate to knowledge, attitudes and resources.