Search

Objective Although using meta-analysis to combine evidence from a number of studies should reduce both bias and uncertainty, it is sometimes not the case, because published studies represent a biased selection of the evidence. Copas proposed a selection model to assess the sensitivity of meta-analysis conclusions to possible selection bias. However, this relatively complex model awaits both reliable software and an empirical evaluation. This article reports work addressing both these issues. Study Design and Setting We took 157 meta-analyses with binary outcomes, analyzed each one using the Copas selection model, and evaluated each analysis using a prespecified protocol. The evaluation aimed to assess the usefulness of the Copas selection model to a typical Cochrane reviewer. Results In approximately 80% of meta-analyses, the overall interpretation of the Copas selection model was clear, with better results among the 22 with evidence of selection bias. However, as with the “Trim and Fill” method, allowing for selection bias can result in smaller standard errors for the treatment estimate. Conclusion When a reliable test for selection bias is significant, we recommend systematic reviewers to try the Copas selection model, although the results should be interpreted cautiously.

Addition of the anti-CD20 monoclonal antibody rituximab to chemotherapy (R-chemo) has been shown to improve response rates and progression-free survival in patients with indolent or mantle cell lymphoma. However, the impact of R-chemo on overall survival is unclear. We performed a comprehensive systematic review and meta-analysis to examine the efficacy of combined immunochemotherapy using R-chemo compared with the identical chemotherapy alone with respect to overall survival in patients with advanced indolent lymphoma or mantle cell lymphoma.Medical databases and conference proceedings were searched for randomized controlled trials published from January 1990 through December 2005 that compared R-chemo with chemotherapy alone in patients with newly diagnosed or relapsed indolent lymphoma or mantle cell lymphoma. We included full-text and abstract publications. Endpoints were overall survival, disease control, overall response, and toxicity. A fixed-effects model was assumed in all meta-analyses. For binary data, the relative risk was used as an indicator of treatment effect, and the Mantel-Haenszel method was used to pool relative risks. Statistical tests for heterogeneity were one-sided; statistical tests for effect estimates were two-sided.Seven randomized controlled trials involving 1943 patients with follicular lymphoma, mantle cell lymphoma, or other indolent lymphomas were included in the meta-analysis. Five studies were published as full-text articles, and two were in abstract form. Patients treated with R-chemo had better overall survival (hazard ratio [HR] for mortality = 0.65; 95% confidence interval [CI] = 0.54 to 0.78), overall response (relative risk of tumor response = 1.21; 95% CI = 1.16 to 1.27), and disease control (HR of disease event = 0.62; 95% CI = 0.55 to 0.71) than patients treated with chemotherapy alone. R-chemo improved overall survival in patients with follicular lymphoma (HR for mortality = 0.63; 95% CI = 0.51 to 0.79) and in patients with mantle cell lymphoma (HR for mortality = 0.60; 95% CI = 0.37 to 0.98). However, in the latter case, there was heterogeneity among the trials (P = .07), making the survival benefit less reliable.In patients with indolent or mantle cell lymphoma, R-chemo is superior to chemotherapy alone with respect to overall survival.

BACKGROUND: Abstracts of presentations at scientific meetings are usually available only in conference proceedings. If subsequent full publication of results reported in these abstracts is based on the magnitude or direction of the results, publication bias may result. Publication bias creates problems for those conducting systematic reviews or relying on the published literature for evidence about health and social care. OBJECTIVES: To systematically review reports of studies that have examined the proportion of meeting abstracts and other summaries that are subsequently published in full, the time between meeting presentation and full publication, and factors associated with full publication. SEARCH METHODS: We searched MEDLINE, Embase, the Cochrane Library, Science Citation Index, reference lists, and author files. The most recent search was done in February 2016 for this substantial update to our earlier Cochrane Methodology Review (published in 2007). SELECTION CRITERIA: We included reports of methodology research that examined the proportion of biomedical results initially presented as abstracts or in summary form that were subsequently published. Searches for full publications had to be at least two years after meeting presentation. DATA COLLECTION AND ANALYSIS: Two review authors extracted data and assessed risk of bias. We calculated the proportion of abstracts published in full using a random‐effects model. Dichotomous variables were analyzed using risk ratio (RR), with multivariable models taking into account various characteristics of the reports. We assessed time to publication using Kaplan‐Meier survival analyses. MAIN RESULTS: Combining data from 425 reports (307,028 abstracts) resulted in an overall full publication proportion of 37.3% (95% confidence interval (CI), 35.3% to 39.3%) with varying lengths of follow‐up. This is significantly lower than that found in our 2007 review (44.5%. 95% CI, 43.9% to 45.1%). Using a survival analyses to estimate the proportion of abstracts that would be published in full by 10 years produced proportions of 46.4% for all studies; 68.7% for randomized and controlled trials and 44.9% for other studies. Three hundred and fifty‐three reports were at high risk of bias on one or more items, but only 32 reports were considered at high risk of bias overall. Forty‐five reports (15,783 abstracts) with 'positive' results (defined as any 'significant' result) showed an association with full publication (RR = 1.31; 95% CI 1.23 to 1.40), as did 'positive' results defined as a result favoring the experimental treatment (RR =1.17; 95% CI 1.07 to 1.28) in 34 reports (8794 abstracts). Results emanating from randomized or controlled trials showed the same pattern for both definitions (RR = 1.21; 95% CI 1.10 to 1.32 (15 reports and 2616 abstracts) and RR = 1.17; 95% CI, 1.04 to 1.32 (13 reports and 2307 abstracts), respectively. Other factors associated with full publication include oral presentation (RR = 1.46; 95% CI 1.40 to 1.52; studied in 143 reports with 115,910 abstracts); acceptance for meeting presentation (RR = 1.65; 95% CI 1.48 to 1.85; 22 reports with 22,319 abstracts); randomized trial design (RR = 1.51; 95% CI 1.36 to 1.67; 47 reports with 28,928 abstracts); and basic research (RR = 0.78; 95% CI 0.74 to 0.82; 92 reports with 97,372 abstracts). Abstracts originating at an academic setting were associated with full publication (RR = 1.60; 95% CI 1.34 to 1.92; 34 reports with 16,913 abstracts), as were those considered to be of higher quality (RR = 1.46; 95% CI 1.23 to 1.73; 12 reports with 3364 abstracts), or having high impact (RR = 1.60; 95% CI 1.41 to 1.82; 11 reports with 6982 abstracts). Sensitivity analyses excluding reports that were abstracts themselves or classified as having a high risk of bias did not change these findings in any important way. In considering the reports of the methodology research that we included in this review, we found that reports published in English or from a native English‐speaking country found significantly higher proportions of studies published in full, but that there was no association with year of report publication. The findings correspond to a proportion of abstracts published in full of 31.9% for all reports, 40.5% for reports in English, 42.9% for reports from native English‐speaking countries, and 52.2% for both these covariates combined. AUTHORS' CONCLUSIONS: More than half of results from abstracts, and almost a third of randomized trial results initially presented as abstracts fail to be published in full and this problem does not appear to be decreasing over time. Publication bias is present in that 'positive' results were more frequently published than 'not positive' results. Reports of methodology research written in English showed that a higher proportion of abstracts had been published in full, as did those from native English‐speaking countries, suggesting that studies from non‐native English‐speaking countries may be underrepresented in the scientific literature. After the considerable work involved in adding in the more than 300 additional studies found by the February 2016 searches, we chose not to update the search again because additional searches are unlikely to change these overall conclusions in any important way.

BACKGROUND: Anaemia associated with cancer and cancer therapy is an important clinical factor in the treatment of malignant diseases. Therapeutic alternatives are recombinant human erythropoiesis stimulating agents (ESAs) and red blood cell transfusions. OBJECTIVES: To assess the effects of ESAs to either prevent or treat anaemia in cancer patients. SEARCH METHODS: This is an update of a Cochrane review first published in 2004. We searched the Central Register of Controlled Trials (CENTRAL), MEDLINE and EMBASE and other databases. Searches were done for the periods 01/1985 to 12/2001 for the first review, 1/2002 to 04/2005 for the first update and to November 2011 for the current update. We also contacted experts in the field and pharmaceutical companies. SELECTION CRITERIA: Randomised controlled trials on managing anaemia in cancer patients receiving or not receiving anti‐cancer therapy that compared the use of ESAs (plus transfusion if needed). DATA COLLECTION AND ANALYSIS: Several review authors assessed trial quality and extracted data. One review author assessed quality assessment and extracted data, a second review author checked for correctness. MAIN RESULTS: This update of the systematic review includes a total of 91 trials with 20,102 participants. Use of ESAs significantly reduced the relative risk of red blood cell transfusions (risk ratio (RR) 0.65; 95% confidence interval (CI) 0.62 to 0.68, 70 trials, N = 16,093). On average, participants in the ESAs group received one unit of blood less than the control group (mean difference (MD) ‐0.98; 95% CI ‐1.17 to ‐0.78, 19 trials, N = 4,715). Haematological response was observed more often in participants receiving ESAs (RR 3.93; 95% CI 3.10 to 3.71, 31 trials, N = 6,413). There was suggestive evidence that ESAs may improve Quality of Life (QoL). There was strong evidence that ESAs increase mortality during active study period (hazard ratio (HR) 1.17; 95% CI 1.06 to 1.29, 70 trials, N = 15,935) and some evidence that ESAs decrease overall survival (HR 1.05; 95% CI 1.00 to 1.11, 78 trials, N = 19,003). The risk ratio for thromboembolic complications was increased in patients receiving ESAs compared to controls (RR 1.52, 95% CI 1.34 to 1.74; 57 trials, N = 15,498). ESAs may also increase the risk for hypertension (fixed‐effect model: RR 1.30; 95% CI 1.08 to 1.56; random‐effects model: RR 1.12; 95% CI 0.94 to 1.33, 31 trials, N = 7,228) and thrombocytopenia/haemorrhage (RR 1.21; 95% CI 1.04 to 1.42; 21 trials, N = 4,507). There was insufficient evidence to support an effect of ESA on tumour response (fixed‐effect RR 1.02; 95% CI 0.98 to 1.06, 15 trials, N = 5,012). AUTHORS' CONCLUSIONS: ESAs reduce the need for red blood cell transfusions but increase the risk for thromboembolic events and deaths. There is suggestive evidence that ESAs may improve QoL. Whether and how ESAs affects tumour control remains uncertain. The increased risk of death and thromboembolic events should be balanced against the potential benefits of ESA treatment taking into account each patient’s clinical circumstances and preferences. More data are needed for the effect of these drugs on quality of life and tumour progression. Further research is needed to clarify cellular and molecular mechanisms and pathways of the effects of ESAs on thrombogenesis and their potential effects on tumour growth.

A new test for the detection of publication bias in meta-analysis with sparse binary data is proposed. The test statistic is based on observed and expected cell frequencies, and the variance of the observed cell frequencies. These quantities are utilized in a rank correlation test. Type I error rate and power of the test are evaluated in simulations; results are compared to those of two other commonly used test procedures. Sample sizes were generated according to findings in a survey of eight German medical journals. Simulation results indicate that, in contrast to existing test procedures, the new test holds the prescribed significance level when data are sparse. However, the power of all tests is low in many situations of practical importance.

Cancer and cancer therapy-associated anemia may have an impact on tumor response and overall survival. Additionally, anemia represents an important economic factor. Therefore, therapeutic alternatives such as erythropoietin (EPO) and red blood cell transfusions have to be evaluated systematically. The effectiveness of recombinant human EPO to prevent or alleviate anemia in patients with malignant disease was determined. Randomized controlled trials comparing prophylaxis or treatment of anemia with EPO plus red blood cell transfusion (RBCT) or RBCT only in patients with malignant disease undergoing antineoplastic therapy were included. The endpoints needed for RBCT were hematological response (hemoglobin increase of 2g/dL or hematocrit increase of 6%), tumor response, and overall survival. Medical databases (Cochrane Library, MEDLINE, EMBASE) and conference proceedings were searched (1985-2001). Full-text and abstract publications were included as well as unpublished data. Data extraction and quality assessment were done in duplicate. All authors were contacted to obtain missing data. Out of 33 eligible studies, 27 trials with 3,287 randomized patients were included.

Background Granulopoiesis-stimulating factors, such as granulocyte-colony-stimulating factor (G-CSF) and granulocyte-macrophage-colony-stimulating factor (GM-CSF), are being used to prevent febrile neutropenia and infection in patients undergoing treatment for malignant lymphoma. The question of whether G-CSF and GM-CSF improve dose intensity, tumour response, and overall survival in this patient population has not been answered yet. Since the results from single studies are inconclusive, a systematic review was undertaken. Objectives To determine the effectiveness of G-CSF and GM-CSF in patients with malignant lymphoma with respect to preventing neutropenia, febrile neutropenia and infection; improving quality of life, adherence to treatment protocol, tumour response, freedom from treatment failure (FFTF) and overall survival (OS); and adverse effects. Search strategy We searched The Cochrane Library, MEDLINE, EMBASE, CancerLit, and other relevant literature databases; Internet databases of ongoing trials; and conference proceedings of the American Society of Clinical Oncology and the American Society of Hematology (1980 - 2007). We included full-text and abstract publications as well as unpublished data. Selection criteria Randomised controlled trials comparing prophylaxis with G-CSF or GM-CSF versus placebo/no prophylaxis in adult patients with malignant lymphoma undergoing chemotherapy were included for review. Both study arms had to receive identical chemotherapy and supportive care. Data collection and analysis Trial eligibility and quality assessment, data extraction and analysis were done by two reviewers independently. Authors were contacted to obtain missing data. Main results We included 13 eligible randomised controlled trials with 2607 randomised patients. Compared with no prophylaxis, both G-CSF and GM-CSF did not improve overall survival (hazard ratio 0.97; 95% CI 0.87 to 1.09) or FFTF (hazard ratio 1.11; 95% CI 0.91 to 1.35). Prophylaxis significantly reduced the relative risk (RR) for severe neutropenia (RR 0.67; 95% confidence interval (CI) 0.60 to 0.73), febrile neutropenia (RR 0.74; 95% CI 0.62 to 0.89) and infection (RR 0.74; 95% CI 0.64 to 0.85). There was no evidence that either G-CSF or GM-CSF reduced the number of patients requiring intravenous antibiotics (RR 0.82; 95%CI 0.57 to 1.18); lowered infection related mortality (RR 0.93; 95% CI 0.51 to 1.71); or improved complete tumour response (RR 1.03; 95% CI 0.95 to 1.10).One study evaluated quality of life parameters and found no differences between the treatment groups. Authors' conclusions G-CSF and GM-CSF, when used as a prophylaxis in patients with malignant lymphoma undergoing conventional chemotherapy, reduce the risk of neutropenia, febrile neutropenia and infection. However, based on the randomised trials currently available, there is no evidence that either G-CSF or GM-CSF provide a significant advantage in terms of complete tumour response, FFTF or OS.

Background Ketotifen is an antihistamine which may be used to treat asthma. Since administering inhaled therapy to younger children can be difficult, an oral agent such as ketotifen offers potential advantages. Objectives The objective of this review is to determine, whether ketotifen alone or in combination with other co-interventions results in better control of asthma in children with asthma and/or wheezing and examine its safety profile. Search strategy We searched the Cochrane Airways Group register of trials (based on MEDLINE, EMBASE, CINAHL and handsearched respiratory journals) and reference lists of articles. The latest search was carried out in October 2002. Selection criteria Clinical studies had to be randomised-controlled and double-blinded, comparing oral ketotifen with placebo in children with asthma and/or wheeze for at least eight weeks at a dose not less than one mg daily. Data collection and analysis Two reviewers independently performed selection of trials, quality assessment and data extraction; a third reviewer was included in the consensus process if necessary. Main results A total of 26 relevant studies involving 1826 participants were included in this review. Children's age ranged from 4 months to 18 years and ketotifen was given between 10 and 32 weeks. The proportion of children able to reduce or stop their bronchodilator use within 12 to 16 weeks of treatment was significantly higher in the ketotifen group (relative risk 2.39, 95% CI 1.64 to 3.48) based on four trials; this result was statistically significant in a subgroup of two trials with well described and adequate method of blinding. Statistically significant beneficial effects of ketotifen were also observed in the following secondary outcomes: efficacy evaluated by physician (10 trials) and parents/patients (7 trials), asthma symptom score (4 trials), asthma exacerbations (2 trials), and reduction in use of oral steroids (4 trials). However, sub-group analyses of trials with well described and adequate method of blinding was only significant for the outcome asthma symptom score and non-significant for the remaining secondary outcomes. Reported side effects were more frequent in the ketotifen group (sedation: 21%, weight gain: 27%) than in the placebo group (sedation: 12%, weight gain: 17%). Reviewer's conclusions Evidence from randomised controlled trials indicates that ketotifen alone or in combination with other co-interventions improves control of asthma and wheezing in children with mild and moderate asthma. However due to the high proportion of children with atopy in some trials the results cannot necessarily be generalised to all asthmatic children. The benefit is obtained at the cost of minor side effects, namely sedation and weight gain. The validity of this conclusion is limited by the low reported, methodological quality of included trials.

Summary Objectives: In this paper three statistical methods [logistic regression, classification and regression tree (CART), and fuzzy inference] for the prediction of lymph node metastasis in carcinoma of the tongue are compared. Methods: A retrospective collection of data in 75 patients treated for tongue cancer was carried out at the Clinic and Policlinic for Oral and Maxillo-facial Surgery at the University Hospital of Freiburg in Germany between January 1990 and December 1999; biopsy material was used for laboratory evaluations. Statistical methods for the prediction of lymph node metastasis were compared using ROC curves and accuracy rates. Results: All three methods show similar results for the prediction of lymph node metastasis with slightly superior results for fuzzy inference and CART. A great overlap is apparent in the ROC curves. The best result observed for fuzzy inference and CART was a sensitivity of 79.2% [95% confidence interval: (57.8%; 92.9%)] and a specificity of 86.3% (73.7%; 94.3%); the best result for predictions based on the logistic regression was a sensitivity of 66.7% (44.7%; 84.4%) and a specificity of 80.4% (66.9%; 90.2%). Accuracy rates of fuzzy method and CART were higher [accuracy rate for fuzzy method and CART: 84% (73.7%; 91.4%), for logistic regression method: 73.3%, 95%-CI: (61.9%; 82.9%)]. Conclusions: From a clinical point of view, the predictive ability of the three methods is not sufficiently large to justify use of these methods in daily practice. Other factors probably on the molecular level are needed for the prediction of lymph node metastasis.

Objective:In hospital operating rooms (ORs), specially conditioned air is supplied to protect patients from airborne agents that may cause infections. This study investigated whether it is hygienically safe to shut down the air supply at night if measures are taken to ensure a timely restart before surgery is performed.Design:Experimental study.Setting:Neurosurgical OR of a German university hospital.Methods:The ventilation system was switched off and restarted after 10 hours. Particles suspended in the air near the operating table were counted, OR temperature was measured, and settle plates were exposed and incubated.Results:In 13 investigations, a median of 1.3 × 104 particles 0.5 μm/m3 or greater (range, 5.8 × 103 to 1.1 × 105) were documented immediately after restart in the morning. After 10 minutes and subsequently, no test showed a particle count exceeding the threshold limit of 1.0 × 104 particles 0.5 μm/m3 or greater recommended by the German Society of Hygiene and Microbiology. Only a few colony-forming units (CFU) were detected per settle plate (median, 0 CFU/60 cm2; range, 0 to 8) and OR temperatures quickly reached normal levels.Conclusions:Shutting down OR ventilation during off-duty periods does not appear to result in an unacceptably high particle count or microbial contamination of the OR air shortly after the system is restarted. Because substantial energy and cost savings are likely, this should be considered in hygienically safe heating, ventilation, and air conditioning systems. However, normal ventilation should be established at least 30 minutes before surgical activity.

The granulopoiesis-stimulating factors, granulocyte colony-stimulating factor (G-CSF) and granulocyte-macrophage colony-stimulating factor (GM-CSF), are used to prevent neutropenia and febrile neutropenia in patients with malignant lymphoma. The question as to whether G-CSF/GM-CSF improves dose-intensity, tumour response and overall survival (OS) has not yet been answered. As the results from single studies are inconclusive, a systematic review was performed to determine the effectiveness of G-CSF/GM-CSF. Randomized controlled trials comparing prophylaxis with G-CSF/GM-CSF versus no prophylaxis in adult patients with malignant lymphoma undergoing conventional chemotherapy were included. Medical databases (Cochrane Library, Medline, Embase) and conference proceedings were searched. All authors were contacted to obtain missing data. We included 11 studies making a total of 1434 patients. Compared with no prophylaxis, G-CSF/GM-CSF significantly reduced the relative risk (RR) for neutropenia [RR 0.64 [95% confidence interval (CI) 0.55-0.75]] febrile neutropenia (RR 0.74 [95% CI 0.62-0.89]) and infection (RR 0.74 [95% CI 0.64-0.85]). G-CSF/GM-CSF did not decrease infection-related mortality (RR 2.07 [95% CI 0.81-5.34]), improve complete remission (CR) (RR 1.06 [95% CI 0.96-1.16]) or OS (HR 0.98 [95% CI 0.81-1.18]). In conclusion, G-CSF/GM-CSF given during conventional chemotherapy in malignant lymphoma patients reduced the RR of neutropenia, febrile neutropenia and infection. However, there is no evidence that G-CSF/GM-CSF improved CR and OS in this clinical setting.

BACKGROUND: Malignant biliary obstruction, which requires endoscopic stenting as palliative therapy, is often complicated by clogging of the stent with subsequent jaundice and/or cholangitis. Stent clogging may be caused by microbiological adhesion and biliary stasis. Therefore, antibiotics and choleretic agents like ursodeoxycholic acid (UDCA) have been investigated to see whether they prolong stent patency. OBJECTIVES: To evaluate if UDCA and/or antibiotics may prolong stent patency and survival in patients with strictures of the biliary tract and endoscopically inserted stents. SEARCH METHODS: The Cochrane Hepato‐Biliary Group Controlled Trials Register, The Cochrane Library, MEDLINE, Current Contents, EMBASE, and CancerLit were searched until June 2001. Reference lists of the identified articles were checked for further trials. SELECTION CRITERIA: All randomised or quasi‐randomised clinical trials investigating UDCA and/or antibiotics in patients with biliary stents were considered for inclusion, regardless of blinding, language, and publication status. DATA COLLECTION AND ANALYSIS: Trial inclusion, quality assessment, and data extraction were performed independently by two reviewers. Principal investigators were contacted for further information. Survival data were combined by using hazard ratios (with 95% confidence interval (95% CI)). MAIN RESULTS: Five non‐blinded randomised trials with 258 patients with malignant strictures treated with polyethylene stents were included. Three trials, including 152 patients, investigated a combination of UDCA and antibiotics versus no treatment. The meta‐analysis of these three trials does not show a significant treatment effect on the duration of stent patency (hazard ratio (random effects model) 0.58, 95% CI 0.22 to 1.54) or mortality (hazard ratio (fixed effect model) 0.99, 95% CI 0.68 to 1.43). Two trials with 106 patients compared antibiotics with no treatment, one of these trials used a combination of antibiotics and rowachol (an 'alternative' drug of the 'mint' family). The pooled results of these two trials do not show significant effects of antibiotics on the duration of stent patency (hazard ratio (fixed effect model) 0.69 (95% CI 0.37 to 1.30)) or mortality (hazard ratio (fixed effect model) 1.23 (95% CI 0.72 to 2.08). Data concerning duration of hospital stay, frequency of cholangitis, and rate of infectious complications due to selection of antibiotic resistant bacteria strains were not available. AUTHORS' CONCLUSIONS: Treatment with UDCA and/or antibiotics to prevent clogging of biliary stents in patients with malignant stricture of the biliary tract cannot be recommended routinely on the basis of the existing randomised clinical trials. Further trials are needed with rigorous methodology and sufficient statistical power.

The application of artificial neural networks (ANNs), especially feed-forward neural networks (FFNNs), has become very popular for diagnosis and prognosis in clinical medicine, often accompanied by exaggerated statements of their potential. The excitement stems mainly from the fact that ANNs were developed as attempts to model the decision process of the human brain. Traditionally, logistic regression models and proportional hazard regression models have been used in these applications. In this article, FFNNs are introduced as flexible, nonlinear regression models and necessary precautions for their use are discussed. Furthermore, the results of a literature survey of applications of ANNs in prostate cancer published between 1999 and 2001 are described; most applications suffer from methodologic deficiencies. It is concluded that there is so far no evidence that the application of ANNs provide real progress in the field of diagnosis and prognosis in prostate cancer.

Background Anaemia associated with cancer and cancer therapy is an important clinical factor in the treatment of malignant diseases. Therapeutic alternatives are recombinant human erythropoietin (EPO) and red blood cell transfusions. Objectives The aim of this systematic review was to assess the effect of erythropoietin to either prevent or treat anaemia in cancer patients. Search strategy We searched the Central Register of Controlled Trials, MEDLINE (01/1985 to 12/2001), EMBASE (01/1985 to 12/2001), other databases and reference lists of articles. We also contacted experts in the field and pharmaceutical companies. Selection criteria Randomised controlled trials comparing the use of recombinant human erythropoietin (plus transfusion if needed) with red blood cell transfusions alone for the treatment or prevention of anaemia in cancer patients. Data collection and analysis Two reviewers independently assessed trial quality and extracted data. All authors from included studies were contacted for additional information. Main results Twenty seven trials with 3,287 adults were included. Use of erythropoietin significantly reduced the relative risk of red blood cell transfusions (RR 0.67; 95% CI 0.62 to 0.73, 25 trials, n = 3,069). On average participants in the erythropoietin group received one unit of blood less than the control group (WMD -1.00; 95% CI-1.31 to -0.70, 13 trials, n = 2,056). For participants with baseline haemoglobin below 10 g/dL haematological response was observed more often in participants receiving EPO (RR 3.60; 95% CI 3.07 to 4.23, 14 trials, n = 2,347). There was inconclusive evidence whether EPO improves tumour response (fixed effect RR 1.36; 95% CI 1.07 to 1.72, seven trials, n = 1,150; random effects: RR 1.21; 95% CI 0.92 to 1.59) and overall survival (adjusted data: HR 0.81; 95% CI 0.67 to 0.99; unadjusted data: HR 0.84; 95% CI 0.69 to 1.02, 19 trials, n = 2,865). There were no statistically significant adverse effects. Evidence was inconclusive with respect to quality of life and fatigue. Reviewers' conclusions There is consistent evidence that the administration of erythropoietin reduces the risk for blood transfusions and the number of units transfused in cancer patients. For patients with baseline haemoglobin below 10 g/dL there is strong evidence that erythropoietin improves haematological response. There is inconclusive evidence whether erythropoietin improves tumour response and overall survival. Research on side effects is inconclusive.

Randomisierte klinische Studien sind ein wichtiges Mittel fur die Erkenntnisgewinnung in der medizinischen Forschung; ihre Ergebnisse bilden die Grundlage fur eine evidenzbasierte Medizin (Evidence-Based Medicine; EBM). Dabei spielt die Zusammenfassung der Ergebnisse mehrerer Studien zur gleichen Fragestellung im Rahmen von Meta-Analysen eine immer wichtigere Rolle. Bezieht sich eine solche Meta-Analyse nur auf publizierte Studien, so kann ihr Ergebnis sehr stark davon abhangen, inwieweit alle zur Fragestellung durchgefuhrten Studien auch tatsachlich publiziert wurden. Dieses, als Publikations-Bias bezeichnete Phanomen muss bei der kritischen Bewertung der Ergebnisse von Meta-Analysen unbedingt beachtet werden. In diesem Kapitel wird diese Problematik in leicht verstandlicher Form und am Beispiel von konkreten Studien diskutiert. Gleichzeitig wird auf das Prinzip der evidenzbasierten Medizin und die Rolle der Cochrane Collaboration hingewiesen.

In early 1996 a hospital-wide methicillin-resistant Staphylococcus aureus (MRSA) epidemic was recognized in a 900-bed university hospital. In order to investigate hospital-specific transmission routes, a case-control study was carried out. Cases and controls were matched for age (+/- 10 years), sex, admission date (+/- 10 days) and clinical department on admission. Data on potential risk factors, were retrieved by chart review. Between June 1996 and February 1997, 67 patients with hospital-acquired MRSA were identified. Molecular typing showed that 85% of the cases carried an indistinguishable strain. The average time at risk for cases and controls was 17.3 and 23.7 days, respectively (P= 0.01). Seventeen patients (25.4%) developed infection. Conditional multivariate regression analysis showed that intensity of care (P= 0.002), number of transfers (P= 0.019), and fluoroquinolone therapy (P= 0.025) were independently associated with acquisition of MRSA. Intensity of care can be considered as a surrogate marker for a number of manipulations which represent the main risk factors for MRSA transmission. Frequent transfers within the hospital hinder, not only the epidemiological analyses, but also efforts to bring an outbreak under control. Our findings give epidemiological support to recent molecular studies which suggest that fluoroquinolone use may increase the transmissibility of MRSA in hospitals.

In the present study (Aapro et al, 2009), results of an updated meta-analysis of 12 randomised controlled trials with 2297 patients are presented on the effects of Epoetin β on survival, tumour progression and thromboembolic events (TEEs) in cancer patients. Special emphasis is given to the impact of different haemoglobin (Hb) levels at initiation of Epoetin β treatment. Overall, the risk for death was increased by a factor of 1.13 (95% CI 0.87–1.46) and that for thromboembolic events was increased by a factor of 1.62 (95% CI 1.13–2.31) in patients receiving Epoetin β compared with controls. When the analysis was limited to patients receiving Epoetin β at Hb⩽10 g dl−1, the authors identified no evidence for a negative impact on survival (Epoetin β vs control HR 0.99, 95% CI 0.70–1.40) and a favourable effect for disease progression (HR 0.73, 95% CI 0.57–0.94). However, the authors describe an increased risk for death in patients with Hb levels of > 11 g dl−1 at initiation of Epoetin β therapy. To evaluate the effect of baseline haemoglobin level on overall survival and other outcomes, Aapro et al. stratified the data by entry Hb and compared the treatment estimates in such created subgroups. Based on this analysis, the authors conclude that no detrimental effect of Epoetin β on survival or tumour progression is apparent when initiated at Hb levels up to 11 g dl−1. However, no statistical test for heterogeneity between subgroups was conducted. Accordingly, it is debatable whether the observed differences in treatment effects between entry Hb subgroups is due to entry Hb acting as an effect modifier or is just a play of chance. For two subgroups, a Z test on heterogeneity (Borenstein et al, 2009) can be calculated from estimated treatment effects and width of confidence intervals in the respective subgroups. The Z test on heterogeneity is based on the difference in the two estimated treatment effects divided by a pooled standard error. Information to conduct this test is given in Figure 2 of Aapro et al. for overall survival and other outcomes. For overall survival, a Z test for heterogeneity comparing the subgroups entry Hb⩽11 g dl−1 and entry Hb>11 g dl−1 would yield a P-value of 0.65. This test would also be non-significant for progression-free survival (P=0.36) and time to TEE (P=0.90). Bohlius et al (2009) also evaluated baseline Hb level as a potential effect modifier of treatment with erythropoiesis-stimulating agents in a larger dataset with 13 407 cancer patients using a Cox regression model stratified by study. The analysis with five Hb subgroups yielded a P-value of 0.75. The authors consider target haemoglobin level as a potential effect modifier. Target Hb level was defined as the maximum Hb level actually achieved up to 28 days after the end of treatment, which is time-dependent information. Using the maximum Hb level during follow-up for stratification is problematic for several reasons. Patients are divided into subgroups based on future information, which requires a specialised methodology in survival data with censored observations. Furthermore, the maximum Hb level during follow-up is dependent on the Epoetin β dosage, as well as other factors such as disease severity and number of blood transfusions. Accordingly, as correctly pointed out by the authors in the discussion, results of the analysis of target Hb level need to be interpreted with caution because of methodological limitations and potential confounding (see also Altman and de Stavola, 1994). In addition, interpretation of the results for maximum Hb level achieved is difficult, as it is unclear which effect is actually estimated in subgroups that are created by stratifying on information that is not available before the follow-up ends, either because the patient may die or is lost to follow-up. A more interpretable analysis of target Hb levels could be based on a Cox regression model containing both baseline Hb level and time-dependent Hb level as covariates (Parmar and Machin, 1995). Furthermore, in such Cox regression models, time-dependent information is modelled appropriately and the effect of future Hb values in addition to the baseline Hb level can be quantified and tested against the null hypothesis that future Hb values do not contain any additional information. Evaluating the impact of haemoglobin levels on survival and other outcomes in cancer patients treated with Epoetin β is an important research topic. The authors give some indications on the safety of Epoetin β when used within the Hb intervention and target levels as recommended in the revised European label. However, in order to get a more definite answer on the impact of initial and target haemoglobin levels on survival outcomes, analysis in a larger dataset using appropriate statistical methods for time-dependent information would be essential.

Background Systematic reviews and meta-analyses of pre-clinical studies, in vivo animal experiments in particular, can influence clinical care. Publication bias is one of the major threats of validity in systematic reviews and meta-analyses. Previous empirical studies suggested that systematic reviews and meta-analyses have become more prevalent until 2010 and found evidence for compromised methodological rigor with a trend towards improvement. We aim to comprehensively summarize and update the evidence base on systematic reviews and meta-analyses of animal studies, their methodological quality and assessment of publication bias in particular. Methods/Design The objectives of this systematic review are as follows: To investigate the epidemiology of published systematic reviews of animal studies until present. To examine methodological features of systematic reviews and meta-analyses of animal studies with special attention to the assessment of publication bias. To investigate the influence of systematic reviews of animal studies on clinical research by examining citations of the systematic reviews by clinical studies. Eligible studies for this systematic review constitute systematic reviews and meta-analyses that summarize in vivo animal experiments with the purpose of reviewing animal evidence to inform human health. We will exclude genome-wide association studies and animal experiments with the main purpose to learn more about fundamental biology, physical functioning or behavior. In addition to the inclusion of systematic reviews and meta-analyses identified by other empirical studies, we will systematically search Ovid Medline, Embase, ToxNet, and ScienceDirect from 2009 to January 2013 for further eligible studies without language restrictions. Two reviewers working independently will assess titles, abstracts, and full texts for eligibility and extract relevant data from included studies. Data reporting will involve a descriptive summary of meta-analyses and systematic reviews. Discussion Results are expected to be publicly available later in 2013 and may form the basis for recommendations to improve the quality of systematic reviews and meta-analyses of animal studies and their use with respect to clinical care.

Background: Erythropoiesis stimulating agents (ESAs) consistently have been shown to decrease transfusions in anemic oncology patients. However, whether they increase mortality in cancer patients is under debate. Results from individual studies conflict, and results from literature based meta-analyses are inconclusive. We conducted a meta-analysis based on individual patient data (IPD) from all available randomized controlled trials (RCTs). Meta-analyses with individual patient data offer several advantages over study-level analysis, including the ability to gain statistical power and increase validity using time-to-event analyses, to adjust for prognostic variables that may have confounded the original treatment comparisons and to investigate subgroups in which treatment may be either more or less effective or harmful. Methods: An international collaboration conducted an individual patent data meta-analysis of ESA effects on mortality in cancer patients. With guidance from an independent steering committee of international experts in hematology, oncology, radiotherapy, epidemiology, medical statistics and a consumer representative, we developed a detailed protocol and statistical analysis plan. Independent RCT investigators and representatives from pharmaceutical companies who submitted data provided additional input through the project’s advisory board. IPD from RCTs of ESA plus red blood cell transfusion (RBCT) (as needed) versus placebo or no ESA plus RBCT (as needed), for prophylaxis or treatment of anemia in cancer patients with or without concurrent antineoplastic therapy, were included. Hazard ratios and 95% confidence intervals (CIs) were calculated per study and meta-analyzed with fixed-effects and random-effects models. Primary endpoints were overall survival (during active study phase and for the longest follow-up available) for patients receiving chemotherapy, and for all cancer patients regardless of anticancer treatment. Stratified multivariable Cox-regression analyses were conducted to assess the impact of baseline imbalances and to identify potential effect modifiers. Duplicate main statistical analyses were conducted independently at two academic statistical departments. Results: Data on 13933 patients enrolled in 53 studies were included in the analysis. Data were provided by the companies Amgen Inc., Johnson & Johnson Pharmaceutical Research & Development, L.L.C., and F. Hoffmann-La Roche Ltd.; and by five independent investigators. Results are currently undergoing internal verification and final evaluation and will be presented at the meeting. Conclusion: Final conclusions will be presented at the meeting. Future analyses using IPD will be conducted to estimate the risks (clots, tumor progression) and potential benefits (transfusion needs and quality of life/fatigue) from other outcomes.

Background: Erythropoiesis-stimulating agents (ESAs) reduce anemia in cancer patients; however, there are concerns that ESAs increase mortality in some patients. Within the framework of an international collaboration we thus conducted a patient-level meta-analysis to assess the effects of ESAs on mortality in cancer patients. Methods: We performed a meta-analysis, based on the intention-to-treat principle, of cancer patients enrolled in randomized controlled trials comparing epoetin alfa, epoetin beta or darbepoetin alfa plus red blood cell transfusions as needed versus transfusion alone, for prophylaxis or treatment of anemia while or after receiving anticancer therapy. Patient-level data were obtained and analyzed by independent statisticians at two academic departments. Primary endpoints were on-study mortality and overall survival in patients receiving chemotherapy, defined as all patients from studies in which =/> 70% of study population received chemotherapy, and in all cancer patients regardless of anticancer therapy. On-study mortality was defined as death from any cause between date of randomization and 28 days after end of active study phase. Overall survival was defined as death from any cause between date of randomization and longest follow up available. Hazard ratios (HR) and 95% confidence intervals (CIs) were calculated per study and meta-analyzed with fixed-effects and random-effects models. Stratified multivariable Cox-regression analyses were conducted to assess the impact of baseline imbalances. Tests for interactions were used to identify differences in effect across pre-specified subgroups. All analyses were pre-specified in a peer-reviewed protocol (Bohlius et al. 2008). An independent steering committee consisting of clinicians and methodologists agreed on all analyses and interpretations. Independent investigators and representatives from manufacturers contributing data offered advice, but had no decision-making authority. Results: Data from 13,933 cancer patients enrolled in 53 studies were included in the analysis; 38 trials including 10,441 patients used mainly chemotherapy. Including all cancer patients ESAs increased on-study mortality by 17% (HR 1.17; 95% CI 1.06–1.30), with little evidence for a difference between chemotherapy and other trials (p for interaction=0.42), and worsened overall survival by 6% (HR 1.06; 95% CI 1.00–1.12). In the chemotherapy population on-study mortality was increased by 10% (HR 1.10, 95% CI 0.98–1.24) and overall survival was worsened by 4% (HR 1.04; 95% CI 0.97–1.11). Adjusting for known prognostic factors had little effect on the overall estimates. There was no conclusive evidence for effect modification by patient level characteristics such as age, sex, Hb and Hct at baseline, Hb ceiling, type and stage of tumor or study level characteristics (anticancer treatment, ESA treatment schedules, study design and quality) for the outcomes tested. Conclusion: ESA treatment in cancer patients increased on-study mortality by 17% and worsened overall survival by 6%. For patients undergoing chemotherapy the increase was less pronounced, but could not be excluded. In clinical practice, risks of ESAs must be balanced against benefits of ESAs depending on the clinical circumstances of the individual patient.

Background Rituximab (R) has shown impressive response and prolonged progression free survival (PFS) in patients with indolent lymphoma when combined with CHOP. Randomized phase III trials adding rituximab to a variety of different regimen confirmed this benefit in both previously treated and untreated patients with advanced indolent lymphoma. Furthermore these trials indicating a trend towards improved overall survival (OS) for a combined immunochemotherapy with R. Here we report updated results of a comprehensive systematic review in this group of patients comparing R and chemotherapy with chemotherapy alone with respect to OS, disease control, overall response (OR) and toxicity. Methods Only randomized controlled trials (RCT) comparing R-chemo with chemotherapy alone in patients with newly diagnosed or relapsed indolent lymphoma and mantle cell lymphoma (MCL) were included. Medical databases (Cochrane Library, MEDLINE, EMBASE) and conference proceedings were searched (1990–2005). We included full-text and abstract publications. Number needed to treat (NNT) were calculated to facilitate interpretation. Results We included seven eligible RCTs involving a total of 1943 patients with follicular lymphoma (FL), MCL and other indolent lymphoma. Studies were published as full text (5), and in abstract form (2). OS was statistically significant improved in the R-chemo group when compared to chemotherapy alone (HR; hazard ratio: 0,65; 95% CI 0,54 – 0,78). OR (RR; relative risk: 1.21; 95% CI 1.16–1.27) and disease control (HR: 0.62; 95% CI 0.55–0.71) were also significantly superior after R-chemo. The RR for developing fever and leukocytopenia was significantly higher with R-chemo, but not associated with an increased risk of infection. Conclusion The systematic review demonstrated improved OS, OR and disease control for patients with indolent lymphoma and in the subgroups of follicular and mantle cell lymphoma when treated with R-chemo compared to chemotherapy alone.

Background: Epoetin (EPO), darbepoetin (DARB) and red blood cell transfusions (RBCT) are therapeutic alternatives to treat anemia associated with cancer and cancer therapy. Results of randomized controlled trials (RCTs) conflict, and some question the safety of EPO and DARB. Previously, we systematically reviewed this topic (Bohlius et al, JNCI 2005). Since then many new studies became available necessitating an update the prior systematic review. Objectives: To determine the effectiveness and safety of recombinant human erythropoietin and darbepoetin to prevent or alleviate anemia in cancer patients (pts), and to compare current and previous results of systematic review. Methods: Included RCTs compared EPO or DARB plus RBCT if needed with observation plus RBCT for prophylaxis or treatment of anemia in cancer patients receiving or not receiving antineoplastic therapy. Patients had solid tumors or hematological malignancies including MDS. Endpoints were rates of RBCT, hematological response (defined as transfusion free Hb increase of 2 g/dL), tumor response, overall survival and thrombo embolic events. Medical databases (Cochrane Library, MEDLINE, EMBASE) and conference proceedings were searched (first review:1985–2001, update up to 04/2005). We included full-text and abstract publications plus unpublished data. Data extraction and quality assessment were in duplicate. Results: The update included 57 trials with 9,353 patients. Use of EPO or DARB significantly reduced the relative risk (RR) of RBCTs (RR 0.64; 95%-CI 0.60–0.68; 42 trials, n = 6,510). On average, participants in the EPO/DARB groups received one unit of blood less than controls (weighted mean difference −1.05; 95% CI− 1.32 −0.78; 14 trials, n = 2,353). Hb response (baseline Hb < 12 g/dL) was more likely in the EPO/DARB group (RR 3.43; 95%-CI 3.07–3.84; 22 trials, n = 4,307). Thrombo embolic complications were more frequent in patients receiving EPO/DARB: RR 1.67 (95%-CI 1.35–2.06; 35 trials, n = 6,769). Uncertainties remain whether and how EPO/DARB affects tumor response (fixed effect RR 1.12; 95%-CI 1.01–1.23; random effects: RR 1.09; 95%-CI 0.94–1.26; 13 trials, n = 2,833) or overall survival (HR 1.08; 95%-CI 0.99–1.18; 42 trials, n = 8,167). Subgroup analyses comparing EPO and DARB did not identify clinically or statistically significant differences. This update confirms previous findings for transfusion rates and hematological response. In contrast, the update suggests EPO/DARB may reduce survival whereas the first review suggested possible benefits (HR 0.81; 95%-CI 0.67–0.99; 19 trials, n = 2,865). Factors possibly contributing to these conflicting results include mortality due to thrombo embolic complications and tumor progression but also methodological limitations such as baseline imbalances. Conclusion: EPO or DARB given to cancer pts reduces the relative risk for red blood cell transfusion and increases likelihood of hematological response. However, EPO/DARB also increases the risk of thrombo embolic complications. The impact on tumor response and survival is uncertain.

Objective Open angle glaucoma is one of the most common causes of blindness in industrialised nations. Treatments to lower ocular pressure are widely used in glaucoma prevention and treatment, despite conflicting evidence. Design We performed meta-analyses to reassess the effectiveness of pressure lowering treatment to delay the development of glaucoma in ocular hypertension, as well as progression of manifest open angle glaucoma. Data sources Medline, Embase, and the Cochrane Library. Selection of studies Eligible studies were randomised controlled trials with a concurrent untreated control group and information on time to glaucomatous changes to visual field and optic disc. Trial reports were reviewed independently by two investigators in an unblinded standardised manner. Results Meta-analysis of trials in ocular hypertension showed a significant preventive effect of reducing intraocular pressure on progression to glaucoma (hazard ratio 0.56, 95% confidence interval 0.39 to 0.81, P = 0.01; number needed to treat 12). Pooled data of studies in manifest glaucoma showed a significant delay of visual field deterioration (0.65, 0.49 to 0.87, P = 0.003; NNT = 7), with subgroup analysis showing a larger effect in patients with raised pressure and a reduced effect in normal tension glaucoma (subgroup comparison: not significant). Conclusions Lowering intraocular pressure in patients with ocular hypertension or manifest glaucoma is beneficial in reducing the risk of visual field loss in the long term.

Background: High-dose chemotherapy with autologous stem cell support (HDT) has been proven effective in relapsed aggressive Non-Hodgkin lymphoma (NHL). However, conflicting results of HDT as part of first-line treatment have been reported in randomized controlled trials (RCTs). Here, we report our updated meta-analysis to better define the role of HDT in these patients. Methods: RCTs were identified by computerized search and handsearching of conference proceedings. Data extraction and quality assessment was performed independently by two reviewers. First authors were contacted to request individual patient data. Eight investigators provided us with individual patient data, for five trials data were extracted from survival curves. The hazard ratio (HR) was used as a measure of treatment effect; the inverse variance method (fixed effect model) was used for pooling. The relative risk was determined for binary data. Results: 15 RCTs including 3079 patients were eligible for this meta-analysis. Overall treatment-related mortality was 6.0% in the HDT group and not significantly different compared to conventional chemotherapy (RR 1.33, p=0.59). Analysis of 13 studies including 2018 patients showed significantly higher CR rates in the group receiving HDT (RR 1.10, p=0.004). However, HDT did not have an effect on OS, when compared to conventional chemotherapy. The pooled HR was 1.04 (p=0.58). There was no statistical heterogeneity among the trials and sensitivity analyses underscored the robustness of these results. Subgroup analysis of prognostic groups according to IPI did not show any survival difference between HDT and controls in 12 trials (low and low-intermediate risk IPI: HR 1.41, high-intermediate and high risk IPI: HR 0.97). Event-free survival (EFS) also showed no significant difference between HDT and CT (HR 0.93, p=0.31). We incorporated several additional variables to possibly identify other risk factors such as the proportions of diffuse large cell lymphoma, protocol adherence, the HDT strategy used, response status of patients before HDT, the conditioning regimen used, and methodological issues. However, our analyses demonstrate that the results described here are not related on either of these factors. Conclusion: Despite higher CR rates, there is no benefit for HDT in patients with aggressive NHL when incorporated in first-line treatment.

Background The heterogeneity statistic I2, interpreted as the percentage of variability due to heterogeneity between studies rather than sampling error, depends on precision, that is, the size of the studies included. Methods Based on a real meta-analysis, we simulate artificially 'inflating' the sample size under the random effects model. For a given inflation factor M = 1, 2, 3,... and for each trial i, we create a M-inflated trial by drawing a treatment effect estimate from the random effects model, using si2 MathType@MTEF@5@5@+=feaagaart1ev2aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacPC6xNi=xH8viVGI8Gi=hEeeu0xXdbba9frFj0xb9qqpG0dXdb9aspeI8k8fiI+fsY=rqGqVepae9pg0db9vqaiVgFr0xfr=xfr=xc9adbaqaaeGaciGaaiaabeqaaeqabiWaaaGcbaGaem4Cam3aa0baaSqaaiabdMgaPbqaaiabikdaYaaaaaa@2FBE@/M as within-trial sampling variance. Results As precision increases, while estimates of the heterogeneity variance τ2 remain unchanged on average, estimates of I2 increase rapidly to nearly 100%. A similar phenomenon is apparent in a sample of 157 meta-analyses. Conclusion When deciding whether or not to pool treatment estimates in a meta-analysis, the yard-stick should be the clinical relevance of any heterogeneity present. τ2, rather than I2, is the appropriate measure for this purpose.

In addition to vitamin C (and other vitamins/antioxidants), clinical ecologists (functional medicine) recommend selenium supplementation as a fundamental therapeutic remedy for the treatment of environment associated health disorders. This recommendation is based on the postulation that the trace element selenium inhibits oxidative stress generated during endogenous detoxification of xenobiotics (phase 1) by increasing selenium-dependent glutathione peroxidase activity, and that it counteracts heavy metal toxicity by forming inert metal complexes. The objective of this review was to investigate whether there are any valid studies providing reliable evidence of the therapeutic benefits of selenium supplementation in potentially environment associated health disorders. A systematic review was conducted based on the rigorous and well-defined methods developed by the Cochrane Collaboration. To achieve the demanding standards for systematic review set by the Cochrane Collaboration, study selection, quality assessment and data abstraction were performed independently and in duplicate using a standardized protocol. Overall, 1290 studies were identified as being eligible for inclusion. Twelve of these met the inclusion criteria and their quality was evaluated individually. None of the studies included in the analysis provided evidence of the therapeutic benefits of selenium supplementation in environment associated health disorders.

Background Network meta-analysis estimates all relative effects between competing treatments and can produce a treatment hierarchy from the most to the least desirable option according to a health outcome. While about half of the published network meta-analyses present such a hierarchy, it is rarely the case that it is related to a clinically relevant decision question. Methods We first define treatment hierarchy and treatment ranking in a network meta-analysis and suggest a simulation method to estimate the probability of each possible hierarchy to occur. We then propose a stepwise approach to express clinically relevant decision questions as hierarchy questions and quantify the uncertainty of the criteria that constitute them. The steps of the approach are summarized as follows: a) a question of clinical relevance is defined, b) the hierarchies that satisfy the defined question are collected and c) the frequencies of the respective hierarchies are added; the resulted sum expresses the certainty of the defined set of criteria to hold. We then show how the frequencies of all possible hierarchies relate to common ranking metrics. Results We exemplify the method and its implementation using two networks. The first is a network of four treatments for chronic obstructive pulmonary disease where the most probable hierarchy has a frequency of 28%. The second is a network of 18 antidepressants, among which Vortioxetine, Bupropion and Escitalopram occupy the first three ranks with frequency 19%. Conclusions The developed method offers a generalised approach of producing treatment hierarchies in network meta-analysis, which moves towards attaching treatment ranking to a clear decision question, relevant to all or a subset of competing treatments.

BACKGROUND: Recent trials suggest improved response rates for purine antagonists compared to alkylator‐based regimens in the treatment of B‐CLL. However, none was able to show a survival advantage. OBJECTIVES: To determine if there is any advantage of purine antagonists compared to alkylating agents (alone or in combination) in the treatment of patients with previously untreated B‐CLL. SEARCH METHODS: Medical databases (Cochrane Library, MEDLINE, EMBASE), conference proceedings and internet‐based trial registers were searched electronically and/or by hand (1990 to 2003). All references were checked for further trial information. We also contacted experts in the field and pharmaceutical companies. SELECTION CRITERIA: Randomised controlled trials comparing purine antagonists as single agents with alkylator‐based regimens in patients with previously untreated B‐CLL were included. We included full‐text and abstract publications as well as unpublished data. DATA COLLECTION AND ANALYSIS: Data extraction and quality assessment were done in duplicate by two independent reviewers. Missing data were obtained from original authors. Endpoints included overall survival, overall response rate, rate of complete remissions, progression‐free survival, treatment‐related morbidity and mortality. MAIN RESULTS: Five trials with 1838 randomised patients were included. There is some evidence for improved overall survival after treatment with purine antagonists compared to alkylators, but statistical significance was not reached (HR 0.89 [95% CI 0.78 to 1.01], 4 trials, N = 1638). However, the relative risk for achieving an overall response (RR 1.22 [95% CI 1.13 to 1.31], 5 trials, N = 1751) and complete remission (RR 1.94 [95% CI 1.65 to 2.28], 5 trials, N = 1751) was significantly higher, resulting in a longer progression‐free survival (HR 0.70 [95% CI 0.61 to 0.82], 4 trials, N = 1638). Incidence of grade III/IV infections was significantly higher in patients receiving treatment with purine antagonists (RR 1.83 [95% 1.30 to 2.58], 4 trials, N = 1620). There was no significant difference concerning the relative risk for grade III/IV neutropenia (RR 1.14 [95% CI 0.98‐1.34], 4 trials, N = 1620) and therapy‐related mortality (RR 0.94 [95% CI 0.45 to 1.95]). Overall incidence of haemolytic anaemia was low, but significantly increased in the purine antagonist group (RR 3.36 [95% CI 1.27 to 8.91], 3 trials, N = 1258). AUTHORS' CONCLUSIONS: Despite significantly increased overall response and complete remission rates and longer progression‐free survival with first‐line treatment of B‐CLL patients with single‐agent purine antagonists, we were not able to detect a statistically significant improvement of overall survival compared to alkylator‐based regimens. Furthermore, the use of purine antagonists also augments the risk for grade III/IV infections and haemolytic anaemia.

The application of artificial neural networks (ANNs) for prognostic and diagnostic classification in clinical medicine has become very popular. In particular, feed-forward neural networks have been used extensively, often accompanied by exaggerated statements of their potential. In this paper, the essentials of feed-forward neural networks and their statistical counterparts (that is, logistic regression models) are reviewed. We point out that the uncritical use of ANNs may lead to serious problems, such as the fitting of implausible functions to describe the probability of class membership and the underestimation of misclassification probabilities. In applications of ANNs to survival data, further difficulties arise. Finally, the results of a search in the medical literature from 1991 to 1995 on applications of ANNs in oncology and some important common mistakes are reported. It is concluded that there is no evidence so far that application of ANNs represents real progress in the field of diagnosis and prognosis in oncology.

Background Health professionals and policymakers aspire to make healthcare decisions based on the entire relevant research evidence. This, however, can rarely be achieved because a considerable amount of research findings are not published, especially in case of ‘negative’ results - a phenomenon widely recognized as publication bias. Different methods of detecting, quantifying and adjusting for publication bias in meta-analyses have been described in the literature, such as graphical approaches and formal statistical tests to detect publication bias, and statistical approaches to modify effect sizes to adjust a pooled estimate when the presence of publication bias is suspected. An up-to-date systematic review of the existing methods is lacking. Methods/design The objectives of this systematic review are as follows: • To systematically review methodological articles which focus on non-publication of studies and to describe methods of detecting and/or quantifying and/or adjusting for publication bias in meta-analyses. • To appraise strengths and weaknesses of methods, the resources they require, and the conditions under which the method could be used, based on findings of included studies. We will systematically search Web of Science, Medline, and the Cochrane Library for methodological articles that describe at least one method of detecting and/or quantifying and/or adjusting for publication bias in meta-analyses. A dedicated data extraction form is developed and pilot-tested. Working in teams of two, we will independently extract relevant information from each eligible article. As this will be a qualitative systematic review, data reporting will involve a descriptive summary. Discussion Results are expected to be publicly available in mid 2013. This systematic review together with the results of other systematic reviews of the OPEN project (To Overcome Failure to Publish Negative Findings) will serve as a basis for the development of future policies and guidelines regarding the assessment and handling of publication bias in meta-analyses.