Combining multiple imputation and meta‐analysis with individual participant data

Multiple imputation is a strategy for the analysis of incomplete data such that the impact of the missingness on the power and bias of estimates is mitigated. When data from multiple studies are collated, we can propose both within‐study and multilevel imputation models to impute missing data on covariates. It is not clear how to choose between imputation models or how to combine imputation and inverse‐variance weighted meta‐analysis methods. This is especially important as often different studies measure data on different variables, meaning that we may need to impute data on a variable which is systematically missing in a particular study. In this paper, we consider a simulation analysis of sporadically missing data in a single covariate with a linear analysis model and discuss how the results would be applicable to the case of systematically missing data. We find in this context that ensuring the congeniality of the imputation and analysis models is important to give correct standard errors and confidence intervals. For example, if the analysis model allows between‐study heterogeneity of a parameter, then we should incorporate this heterogeneity into the imputation model to maintain the congeniality of the two models. In an inverse‐variance weighted meta‐analysis, we should impute missing data and apply Rubin's rules at the study level prior to meta‐analysis, rather than meta‐analyzing each of the multiple imputations and then combining the meta‐analysis estimates using Rubin's rules. We illustrate the results using data from the Emerging Risk Factors Collaboration. © 2013 The Authors. Statistics in Medicine published by John Wiley & Sons Ltd.     

The varying size of exit wounds from center-fire rifles as a consequence of the temporary cavity

To study the question whether and how the size and position of the temporary cavity influence the morphology and especially the size of a bullet exit wound, test shots were fired to composite models consisting of gelatine and pig soft tissue covered with skin (at the exit site). The dimensions of the composite model were determined such that the exit planes were located either at the level of the narrow channel or within the temporary cavity or behind it. The chronological sequence of wound formation and its relation to the current position of the bullet were documented by means of a high-speed camera. Test shots were fired from a distance of 10 m using 5.56?×?45 mm cartridges with full metal-jacketed pointed bullets (v ₀?～?912 m/s, E ₀?～?1,663 J). The study proved that the extension and position of the temporary wound cavity was decisive for the size of the exit wound: An exit plane within the cavity resulted in particularly large skin lesions, whereas the wound diameters were much smaller if the exit plane was located in front or behind the cavity. The exit hole reaches its maximum size only after the bullet has left the target synchronous to the staggered expansion of the temporary cavity.     

Rapid osteolysis of the femoral neck: consequence of an insufficiency fracture of the hip?

Purpose

To describe the imaging and clinical features of rapid osteolysis of the femoral neck in an attempt to better understand this uncommon pathology.

Materials and methods

We retrospectively reviewed the files of 11 patients (six women and five men) aged 53–78 years diagnosed with rapid osteolysis of the femoral neck. Available imaging studies included radiographs, CT, MRI, and bone scintigraphy. Histopathological evaluations were available for seven cases.

Results

All patients presented with complaints of hip pain, six of whom had acute symptoms, while the rest had progressive symptoms and impairment. All but one case were found to have bone deposition in adjacent hip muscles. CT confirmed bone deposition in adjacent tissues and true osteolysis of the femoral neck with relative sparing of the articular surfaces. Bone scintigraphy and MRI were useful to exclude underlying neoplastic disease.

Conclusions

Rapid osteolysis of the femoral neck tends to occur in patients with underlying comorbidities leading to bone fragility and may actually represent a peculiar form of spontaneous insufficiency fracture. Recognition of its imaging features and clinical risk factors may help distinguish this process from other more concerning disorders such as infection or neoplasm.     

A ganglion of the patellar tendon in patellar tendon-lateral femoral condyle friction syndrome

Intratendinous ganglia are rare. We report the case of a sedentary woman with chronic mechanical anterolateral pain of the knee and an extensive ganglion of the patellar tendon as indicated on magnetic resonance (MR) and ultrasound (US) examinations. There was evidence of a high-riding patella, patellar malalignment and patellar tendon-lateral femoral condyle friction syndrome with significantly close contact between the patellar tendon and the lateral facet of the femoral trochlea. The ultrasound-guided aspiration of the ganglion enabled a localized injection of an anti-inflammatory drug (cortivazol) and the cytopathological examination of the fluid, which confirmed the diagnosis. Clinical improvement was maintained with knee rehabilitation and was satisfactory at follow-up after 1 year. To our knowledge, we report the first case of a ganglion of the patellar tendon subsequent to patellar tendon-lateral femoral condyle friction syndrome. We found that this case was illustrative of mucoid degeneration in connective tissue due to chronic repetitive microtraumas. Additionally, this case provided the opportunity to discuss the management of this condition in a sedentary individual with a high-riding patella and patellar malalignment.     

The elusive task of biomarkers of renal injury

Neutrophil gelatinase-associated lipocalin (NGAL) is one of the most promising candidate biomarkers of renal injury, with expression in renal tissue increasing dramatically after ischemia-reperfusion injury but not in the case of pure pre-renal failure. In a recent issue of Critical Care, Di Somma and colleagues reported that NGAL could improve the classification of acute kidney injury compared with clinical assessment and showed that NGAL was associated with poor prognosis. NGAL may therefore carry different information than biomarkers of renal function. This study finally provides additional evidence for the highly complex relationship between renal function and renal injury.Biomarkers are tools that should aid the physician in diagnosis, in risk stratification with prediction of outcome, and, at best, in clinical decision-making [1]. Because the usual renal biomarkers (for example, serum creatinine and/or urine output) may fail to detect acute kidney injury (AKI) at early onset, there is a current effort to search for and validate new diagnostic biomarkers. Neutrophil gelatinase-associated lipocalin (NGAL) is one of the most promising candidates because its level of expression in renal tissue increases dramatically after ischemia-reperfusion injury but not in the case of pure pre-renal failure [2]. Those data suggest the ability of NGAL to detect renal structural damage. Although first studies have highlighted that NGAL monitoring would perform better than standard biomarkers in detecting or quantifying AKI, other reports have yielded conflicting results and raised concerns regarding the accuracy of NGAL for this purpose [3].In this context, the study published by Di Somma and colleagues in Critical Care provides important additional insights [1]. The study included patients admitted to hospital from the emergency room, to assess the diagnosis and prognosis value of plasmatic NGAL using a point-of-care method. The authors attempted a multidimensional approach including the NGAL level combined with the initial clinical assessment of risk of presumed AKI by the caring physician. In a Net Reclassification Improvement analysis, NGAL improved classification of AKI by 32.4%. This improvement occurred mainly by moving patients from the AKI to the no-AKI subgroups. NGAL was confirmed to have a high predictive negative value in this population with low prevalence of AKI.At first glance, these results suggest that NGAL has an additional diagnostic value to clinical suspicion, which is a crucial requirement for a new biomarker to improve the predictive accuracy of the standard of care. Looking deeper, this study highlights concerns regarding the clinical validation of biomarkers of renal injury. Interestingly, the authors excluded acute renal dysfunctions from AKI, a classification that referred to a subset of disease characterized by transient decline of the glomerular filtration rate with presumed no or minor structural damage (similarly to pre-renal AKI) [4]. This study adds evidence to the recent study from Nejat and colleagues that these classifications based on clinical presumption may be erroneous, with increased serum levels of both biomarkers of renal function and injury in these subsets of patients [5]. Using classifications based on markers of glomerular function (Acute Kidney Injury Network/Risk Injury Failure Loss Endstage/Kidney Disease: Improving Global Outcomes) to validate biomarkers of renal injury introduces the risk of diverting such biomarkers from their true goal and significance, namely to detect organ damage. Using classifications based on a clinical assumption of pure pre-renal mechanism of renal failure obviously further exposes the patient to misclassification of outcome. Although an association exists between organ injury and the inherent loss of function, the relationship is obviously complex and may vary with respect to the causative process [6,7].The influence of comorbidities adds further complexity. Intriguingly, urine NGAL has been associated with long-term cardiovascular mortality in a cohort of older community-dwelling adults with no past history of clinical cardiovascular disease with a median level of 192 ng/ml, higher than the threshold proposed for detection of AKI in the study by Di Somma and colleagues (that is, 150 ng/ml), reflecting a high noise/signal ratio for the diagnosis of renal failure (which semantically better defines altered glomerular function) [8]. The expectation that a single biomarker (with a much hunted critical threshold) could at the same time capture these mechanisms of disease and could accurately predict the loss of function and/or stage of kidney injury is therefore obviously flawed. By analogy, who would expect increased serum troponin to accurately detect or predict heart failure? Other (bio)markers can do so.NGAL carries risk information beyond markers of renal function and clinical assessment. In the present study NGAL did not perform better than the estimated glomerular filtration rate or clinical judgment, which share common criteria with the clinical endpoint (namely serum creatinine and estimated glomerular filtration rate) - further suggesting that these biomarkers provide different information, including a prognostic value [9]. Is it time for intensivists to operate the paradigm shift at the bedside in the way we assess AKI by using NGAL to monitor renal damage? Further exploration of the association between biomarkers of renal function, biomarkers of renal injury and prognosis appears a crucial next step before doing so. The results of an ongoing large multicenter study assessing the association between NGAL and 1-year outcome in ICU patients will provide important insights into this issue (FROG-ICU trial, ClinicalTrials.gov:NCT01367093). Exploration of the links between hits, damage and driving forces of renal failure through the sources of different biomarkers [1], identification of the NGAL pathways (that is, systemic inflammation, chronic renal injury, acute renal injury, and so forth) [10] and increased specificity of biomarkers toward renal injury is then required [11,12] (Figure ​(Figure1).1). Most importantly, intensivists and emergency physicians should explore whether risk stratification using NGAL will translate in clinical decision-making.Open in a separate windowFigure 1Serum/urine neutrophil-gelatinase associated lipocalin increase might outweigh renal function biomarkers for prediction of adverse outcome. Several lines of evidence suggest that a rise of serum or urine level of neutrophil-gelatinase associated lipocalin (NGAL) might outweigh biomarkers of renal function for prediction of adverse outcome (that is, mortality). Further prospective studies should confirm these findings. Dashed lines enclose potential effectors linking biomarkers of renal function (for example, serum creatinine) and renal injury (for example, NGAL) to poor outcome. These factors can affect the serum level of these biomarkers (+) but can also influence outcome (-). Further research should explore the significance of these associations and will unveil the specificity of the biomarkers toward renal injury.In conclusion, current definitions of AKI are based on renal function biomarkers that carry different information than biomarkers of injury. NGAL can detect renal injury, and does so pretty well. Injury does not always translate into renal failure, however, and the converse holds true. Future research should aim at clarifying what we are really looking at with biomarkers of kidney injury, including NGAL, and what are the clinical implications. Pending these advancements, we certainly have to accept that injury is not function and that in most conditions trying to predict renal failure with biomarkers of injury will remain an elusive task.     

Incidence,risk factors and prediction of post-operative acute kidney injury following cardiac surgery for active infective endocarditis: an observational study

Introduction

Cardiac surgery is frequently needed in patients with infective endocarditis (IE). Acute kidney injury (AKI) often complicates IE and is associated with poor outcomes. The purpose of the study was to determine the risk factors for post-operative AKI in patients operated on for IE.

Methods

A retrospective, non-interventional study of prospectively collected data (2000–2010) included patients with IE and cardiac surgery with cardio-pulmonary bypass. The primary outcome was post-operative AKI, defined as the development of AKI or progression of AKI based on the acute kidney injury network (AKIN) definition. We used ensemble machine learning (“Super Learning”) to develop a predictor of AKI based on potential risk factors, and evaluated its performance using V-fold cross validation. We identified clinically important predictors among a set of risk factors using Targeted Maximum Likelihood Estimation.

Results

202 patients were included, of which 120 (59%) experienced a post-operative AKI. 65 (32.2%) patients presented an AKI before surgery while 91 (45%) presented a progression of AKI in the post-operative period. 20 patients (9.9%) required a renal replacement therapy during the post-operative ICU stay and 30 (14.8%) died during their hospital stay. The following variables were found to be significantly associated with renal function impairment, after adjustment for other risk factors: multiple surgery (OR: 4.16, 95% CI: 2.98-5.80, p<0.001), pre-operative anemia (OR: 1.89, 95% CI: 1.34-2.66, p<0.001), transfusion requirement during surgery (OR: 2.38, 95% CI: 1.55-3.63, p<0.001), and the use of vancomycin (OR: 2.63, 95% CI: 2.07-3.34, p<0.001), aminoglycosides (OR: 1.44, 95% CI: 1.13-1.83, p=0.004) or contrast iodine (OR: 1.70, 95% CI: 1.37-2.12, p<0.001). Post-operative but not pre-operative AKI was associated with hospital mortality.

Conclusions

Post-operative AKI following cardiopulmonary bypass for IE results from additive hits to the kidney. We identified several potentially modifiable risk factors such as treatment with vancomycin or aminoglycosides or pre-operative anemia.     

Noninvasive mechanical ventilation in patients having declined tracheal intubation

Purpose

Noninvasive ventilation (NIV) is a treatment option in patients with acute respiratory failure who are good candidates for intensive care but have declined tracheal intubation. The aim of our study was to report outcomes after NIV in patients with a do-not-intubate (DNI) order.

Methods

Prospective observational cohort study in all patients who received NIV for acute respiratory failure in 54 ICUs in France and Belgium, in 2010/2011.

Results

Goals of care, comfort, and vital status were assessed daily. On day 90, a telephone interview with patients and relatives recorded health-related quality of life (HRQOL), posttraumatic stress disorder-related symptoms, and symptoms of anxiety and depression. Post-ICU burden was compared between DNI patients and patients receiving NIV with no treatment-limitation decisions (TLD). Of 780 NIV patients, 574 received NIV with no TLD, and 134 had DNI orders. Hospital mortality was 44 % in DNI patients and 12 % in the no-TLD group. Mortality in the DNI group was lowest in COPD patients compared to other patients in the DNI group (34 vs. 51 %, P = 0.01). In the DNI group, HRQOL showed no significant decline on day 90 compared to baseline; day-90 data of patients and relatives did not differ from those in the no-TLD group.

Conclusions

Do-not-intubate status was present among one-fifth of ICU patients who received NIV. DNI patients who were alive on day 90 experienced no decrease in HRQOL compared to baseline. The prevalences of anxiety, depression, and PTSD-related symptoms in these patients and their relatives were similar to those seen after NIV was used as part of full-code management (clinicaltrial.govNCT01449331).     

Dyspnea and surface inspiratory electromyograms in mechanically ventilated patients

Context

Pressure support ventilation (PSV) must be tailored to the load capacity balance of the respiratory system. While "over assistance" generated hyperinflation and ineffective efforts, "under assistance" increased respiratory drive and causes dyspnea. Surface electromyograms (sEMGs) of extradiaphragmatic inspiratory muscles were responsive to respiratory loading/unloading.

Objectives

To determine if sEMGs of extradiaphragmatic inspiratory muscles vary with PSV settings and relate to the degree of discomfort and the intensity of dyspnea in acutely ill patients.

Design

Pathophysiological study, prospective inclusions of 12 intubated adult patients.

Interventions

Two PSV levels (high and low) and two expiratory trigger (ET) levels (high and low).

Measurements

Surface electromyograms of the scalene, parasternal, and Alae Nasi muscles (peak, EMG_max; area under the curve, EMG_AUC); dyspnea visual analogue scale (VAS); prevalence of ineffective triggering efforts.

Main results

For the three recorded muscles, EMG_max and EMG_AUC were significantly greater with low PS than high PS. The influence of ET was less important. A strong correlation was found between dyspnea and EMG_max. A significant inverse correlation was found between the prevalence of ineffective efforts and both dyspnea-VAS and EMG_min.

Conclusions

Surface electromyograms of extradiaphragmatic inspiratory muscles provides a simple, reliable and non-invasive indicator of respiratory muscle loading/unloading in mechanically ventilated patients. Because this EMG activity is strongly correlated to the intensity of dyspnea, it could be used as a surrogate of respiratory sensations in mechanically ventilated patients, and might, therefore, provide a monitoring tool in patients in whom detection and quantification of dyspnea is complex if not impossible.