In an empirical evaluation of the funnel plot, researchers could not visually identify publication bias

GSTM1 and GSTT1 polymorphisms and nasopharyngeal cancer risk: an evidence-based meta-analysis

BACKGROUND

Previous evidence implicates polymorphisms of GSTM1 and GSTT1, candidates of phase II enzymes, as risk factors for various cancers. A number of studies have conducted on the association of GSTM1 and GSTT1 polymorphism with susceptibility to nasopharyngeal carcinoma (NPC). However, inconsistent and inconclusive results have been obtained. In the present study, we aimed to assess the possible associations of NPC risk with GSTM1 and GSTM1 null genotype, respectively.

METHODS

The associated literature was acquired through deliberate searching and selected based on the established inclusion criteria for publications, then the extracted data were further analyzed using systematic meta-analyses.

RESULTS

A total of 85 articles were identified, of which eight case-control studies concerning NPC were selected. The results showed that the overall OR was 1.42 (95%CI = 1.21-1.66) for GSTM1 polymorphism. While for GSTT1 polymorphism, the overall OR was 1.12 (95% CI = 0.93-1.34).

CONCLUSION

The data were proven stable via sensitivity analyses. The results suggest GSTM1 deletion as a risk factor for NPC and failed to suggest a marked correlation of GSTT1 polymorphisms with NPC risk.

Polymorphisms of CYP1A1 and GSTM1 and laryngeal cancer risk: evidence-based meta-analyses

PURPOSE

Previous evidence implicates CYP1A1 and GSTM1 polymorphisms as risk factors for various cancers. A number of studies have been devoted to the association of CYP1A1 or GSTM1 polymorphism with susceptibility to laryngeal carcinoma, with the results inconsistent and inconclusive. The aim of the present study was to assess the possible associations of laryngeal cancer risk with CYP1A1 genetic variation and GSTM1 null genotype respectively.

METHODS

The associated literature was acquired through deliberate searching and selected based on the established inclusion criteria for publications, then the extracted data were further analyzed using systematic meta-analyses.

RESULTS

The results showed that the overall odds ratio (OR) was 1.32 (95% CI = 1.08-1.61) for CYP1A1 Mspl polymorphism. Using subgroup analysis, the pooled ORs were 1.38 (95% CI = 0.98-1.95) in Asians and 1.29 (95% CI = 1.01-1.65) in Caucasians. For CYP1A1 exon7 polymorphism, the overall OR was 1.38 (95% CI = 0.98-1.95). The overall OR was 1.24 (95% CI = 1.03-1.49) for GSTM1 polymorphism and the pooled ORs were 1.36 (95% CI = 0.75-2.48) in Asians, 1.16 (95% CI = 0.94-1.44) in Caucasians and 1.52 (95% CI = 1.05-2.19) in Turkey population.

CONCLUSIONS

The data suggest CPY1A1 MspI polymorphism as a risk factor for laryngeal cancer in Caucasians but not in Asians. However, the results suggest a marked correlation of GSTM1 polymorphism with laryngeal cancer risk in Turkey population but not Caucasians and Asians.

More than numbers: the power of graphs in meta-analysis

In meta-analysis, the assessment of graphs is widely used in an attempt to identify or rule out heterogeneity and publication bias. A variety of graphs are available for this purpose. To date, however, there has been no comparative evaluation of the performance of these graphs. With the objective of assessing the reproducibility and validity of graph ratings, the authors simulated 100 meta-analyses from 4 scenarios that covered situations with and without heterogeneity and publication bias. From each meta-analysis, the authors produced 11 types of graphs (box plot, weighted box plot, standardized residual histogram, normal quantile plot, forest plot, 3 kinds of funnel plots, trim-and-fill plot, Galbraith plot, and L'Abbé plot), and 3 reviewers assessed the resulting 1,100 plots. The intraclass correlation coefficients (ICCs) for reproducibility of the graph ratings ranged from poor (ICC = 0.34) to high (ICC = 0.91). Ratings of the forest plot and the standardized residual histogram were best associated with parameter heterogeneity. Association between graph ratings and publication bias (censorship of studies) was poor. Meta-analysts should be selective in the graphs they choose for the exploration of their data.

The effect of a quantitative resuscitation strategy on mortality in patients with sepsis: a meta-analysis

OBJECTIVE

Quantitative resuscitation consists of structured cardiovascular intervention targeting predefined hemodynamic end points. We sought to measure the treatment effect of quantitative resuscitation on mortality from sepsis.

DATA SOURCES

We conducted a systematic review of the Cochrane Library, MEDLINE, EMBASE, CINAHL, conference proceedings, clinical practice guidelines, and other sources using a comprehensive strategy.

STUDY SELECTION

We identified randomized control trials comparing quantitative resuscitation with standard resuscitation in adult patients who were diagnosed with sepsis using standard criteria. The primary outcome variable was mortality.

DATA ABSTRACTION

Three authors independently extracted data and assessed study quality using standardized instruments; consensus was reached by conference. Preplanned subgroup analysis required studies to be categorized based on early (at the time of diagnosis) vs. late resuscitation implementation. We used the chi-square test and I to assess for statistical heterogeneity (p < 0.10, I > 25%). The primary analysis was based on the random effects model to produce pooled odds ratios with 95% confidence intervals.

RESULTS

The search yielded 29 potential publications; nine studies were included in the final analysis, providing a sample of 1001 patients. The combined results demonstrate a decrease in mortality (odds ratio 0.64, 95% confidence interval 0.43-0.96); however, there was statistically significant heterogeneity (p = 0.07, I = 45%). Among the early quantitative resuscitation studies (n = 6) there was minimal heterogeneity (p = 0.40, I = 2.4%) and a significant decrease in mortality (odds ratio 0.50, 95% confidence interval 0.37-0.69). The late quantitative resuscitation studies (n = 3) demonstrated no significant effect on mortality (odds ratio 1.16, 95% confidence interval 0.60-2.22).

CONCLUSION

This meta-analysis found that applying an early quantitative resuscitation strategy to patients with sepsis imparts a significant reduction in mortality.

Interpretation of tests of heterogeneity and bias in meta-analysis

Statistical tests of heterogeneity and bias, in particular publication bias, are very popular in meta-analyses. These tests use statistical approaches whose limitations are often not recognized. Moreover, it is often implied with inappropriate confidence that these tests can provide reliable answers to questions that in essence are not of statistical nature. Statistical heterogeneity is only a correlate of clinical and pragmatic heterogeneity and the correlation may sometimes be weak. Similarly, statistical signals may hint to bias, but seen in isolation they cannot fully prove or disprove bias in general, let alone specific causes of bias, such as publication bias in particular. Both false-positive and false-negative signals of heterogeneity and bias can be common and their prevalence may be anticipated based on some rational considerations. Here I discuss the major common challenges and flaws that emerge in using and interpreting statistical tests of heterogeneity and bias in meta-analyses. I discuss misinterpretations that can occur at the level of statistical inference, clinical/pragmatic inference and specific cause attribution. Suggestions are made on how to avoid these flaws, use these tests properly and learn from them.

Use of Isotonic Sodium Bicarbonate to Prevent Radiocontrast Nephropathy in Patients with Mild Pre-Existing Renal Impairment: A Meta-Analysis

Acute renal dysfunction after radiocontrast in patients with pre-existing renal impairment is not uncommon and is associated with significant morbidity and mortality. Isotonic sodium bicarbonate solution was first reported to reduce radiocontrast nephropathy in 2004. This first study was, however, limited by its small sample size and as such, the use of isotonic sodium bicarbonate to prevent radiocontrast nephropathy is still not widely used by many anaesthetists and intensivists. We meta-analysed relevant randomised controlled studies sourced from the Cochrane Controlled Trial Register (2007 issue 4), EMBASE and MEDLINE databases (1966 to April 15, 2008) without any language restriction. The use of isotonic sodium bicarbonate was associated with a significant reduction in risk of an incremental rise in serum creatinine concentration 25% above baseline (relative risk 0.22, 95% confidence interval [CI]: 0.11 to 0.44, P <0.0001; I 2 =0%>) and had a protective effect on the absolute change in serum creatinine concentration (weighted-mean-difference -9.4 μmol/l, 95% CI: -17.2 to -1.7, P=0.02; I 2 =0%>) and creatinine clearance (weighted-mean-difference 3.7 ml/min, 95% CI: 0.55 to 6.80, P=0.02; I 2 =57.1%) after radiocontrast. The incidence of acute renal failure requiring dialysis was low (1.4%) and was not significantly different after the use of isotonic sodium bicarbonate (relative risk 0.59, 95% CI: 0.15 to 2.42, P=0.47; I 2 =0%). With the limited data available, isotonic sodium bicarbonate appears to be safe and very effective in reducing radiocontrast nephropathy in patients with mild pre-existing renal impairment. A large randomised controlled study is needed to confirm whether isotonic bicarbonate can improve patient centred clinical outcomes.

Assessment of publication bias for the surgeon scientist

Background

Publication bias occurs when statistically non-significant (negative) findings are not published. It can profoundly affect the results of systematic reviews and meta-analyses.

Methods

Qualitative and quantitative methods of detecting publication bias are described, including their advantages and disadvantages.

Results and conclusion

Accepted quality standards for the reporting of meta-analyses recommend assessment of publication bias, but currently there is no uniform standard for reporting. Quantitative methods are being used with increasing frequency. Authors should take steps to minimize publication bias, and use both qualitative and quantitative assessment methods to determine whether it is present.

Contour-enhanced meta-analysis funnel plots help distinguish publication bias from other causes of asymmetry

OBJECTIVES

To present the contour-enhanced funnel plot as an aid to differentiating asymmetry due to publication bias from that due to other factors.

STUDY DESIGN AND SETTING

An enhancement to the usual funnel plot is proposed that allows the statistical significance of study estimates to be considered. Contour lines indicating conventional milestones in levels of statistical significance (e.g., <0.01, <0.05, <0.1) are added to funnel plots.

RESULTS

This contour overlay aids the interpretation of the funnel plot. For example, if studies appear to be missing in areas of statistical nonsignificance, then this adds credence to the possibility that the asymmetry is due to publication bias. Conversely, if the supposed missing studies are in areas of higher statistical significance, this would suggest the cause of the asymmetry may be more likely to be due to factors other than publication bias, such as variable study quality.

CONCLUSIONS

We believe this enhancement to funnel plots (i) is simple to implement, (ii) is widely applicable, (iii) greatly improves interpretability, and (iv) should be used routinely.

Systematic review of emergency physician-performed ultrasonography for lower-extremity deep vein thrombosis

OBJECTIVES

The authors performed a systematic review to evaluate published literature on diagnostic performance of emergency physician-performed ultrasonography (EPPU) for the diagnosis and exclusion of deep venous thrombosis (DVT).

METHODS

Structured search criteria were used to query MEDLINE and EMBASE, followed by a hand search of published bibliographies. Relevance and inclusion criteria required prospective investigation of emergency department (ED) outpatients with suspected DVT; diagnostic evaluations had to consist of EPPU followed by criterion standard (radiology-performed) imaging. Two authors independently extracted data from included studies; study quality was assessed utilizing a validated tool for quality assessment of diagnostic accuracy studies (QUADAS). Pooled data were analyzed using an unweighted summary receiver-operating-characteristic (SROC) curve; sensitivity and specificity were estimated using a random effects model.

RESULTS

The initial search yielded 1,162 publications. Relevance screening and selection yielded six articles including 936 patients. Four of the six studies reported adequate blinding but a number of other methodologic flaws were identified. A random effects model yielded an overall sensitivity of 0.95 (95% confidence interval [CI] = 0.87 to 0.99) and specificity of 0.96 (95% CI = 0.87 to 0.99).

CONCLUSIONS

Systematic review of six studies suggests that EPPU may be accurate for the diagnosis of DVT compared with radiology-performed ultrasound (US). However, given the methodologic limitations identified among the primary studies, the estimates of diagnostic test performance may be overly optimistic. Further research into EPPU for suspected DVT is needed before it can be adopted into routine clinical practice.

Comparative benefits and harms of competing medications for adults with attention-deficit hyperactivity disorder: a systematic review and indirect comparison meta-analysis

RATIONALE

Recommended medication prescribing hierarchies for adult attention-deficit hyperactivity disorder (ADHD) vary between different guideline committees. Few trials directly compare competing ADHD medications in adults and provide little insight for clinicians making treatment choices.

OBJECTIVE

The objective of this study was to assess comparative benefits and harms of competing medications for adult ADHD using indirect comparison meta-analysis.

MATERIALS AND METHODS

Eligible studies were English-language publications of randomized controlled trials comparing ADHD drugs to placebo. Data sources were electronic bibliographic databases, Drugs@FDA, manufacturer data, and reference lists. Two reviewers independently abstracted data on design, internal validity, population, and results. Benefits and harms were compared between drug types using indirect comparison meta-regression (ratio of relative risks).

RESULTS

Twenty-two placebo-controlled trials were included (n = 2,203). Relative benefit of clinical response for shorter-acting stimulants, primarily immediate release methylphenidate, was 3.26 times greater than for patients taking longer-acting stimulants (95% CI 2.03, 5.22) and 2.24 times greater than for patients taking longer-acting forms of bupropion (95% CI 1.23, 4.08). Immediate release methylphenidate is also the only drug shown to reduce ADHD symptoms in adults with substance abuse disorders. Neither non-stimulants nor longer-acting stimulants reduced adverse effects compared to shorter-acting stimulants. Key gaps in evidence were academic, occupational, social functioning, cardiovascular toxicity, and longer-term outcomes, influences of ADHD subtype and/or comorbidities, and misuse/diversion of the drugs.

CONCLUSIONS

Current best evidence supports using immediate release methylphenidate as first-line treatment for most adults with ADHD.

Recent developments in meta-analysis

The art and science of meta-analysis, the combination of results from multiple independent studies, is now more than a century old. In the last 30 years, however, as the need for medical research and clinical practice to be based on the totality of relevant and sound evidence has been increasingly recognized, the impact of meta-analysis has grown enormously. In this paper, we review highlights of recent developments in meta-analysis in medical research. We outline in particular how emphasis has been placed on (i) heterogeneity and random-effects analyses; (ii) special consideration in different areas of application; (iii) assessing bias within and across studies; and (iv) extension of ideas to complex evidence synthesis. We conclude the paper with some remarks on ongoing challenges and possible directions for the future.

Interferon treatment in hemodialysis patients with chronic hepatitis C virus infection: a systematic review of the literature and meta-analysis of treatment efficacy and harms

BACKGROUND

Hepatitis C virus (HCV) infection is prevalent in patients undergoing hemodialysis and is associated with greater mortality. We determined the efficacy and harms of interferon (IFN) and pegylated IFN (PEG-IFN) treatment of hemodialysis patients with chronic HCV infection and identified factors associated with these outcomes.

STUDY DESIGN

Meta-analysis and meta-regression of randomized controlled trials, uncontrolled trials, and prospective observational studies.

SETTING & POPULATION

Hemodialysis patients with chronic HCV infection.

SELECTION CRITERIA FOR STUDIES

MEDLINE indexed studies since 1966, sample size greater than 10.

INTERVENTION

IFN-based treatment, including PEG-IFN with and without ribavirin.

OUTCOMES

Sustained virological response (SVR) 6 months after treatment, rate of treatment discontinuation caused by adverse events, and factors associated with these outcomes.

RESULTS

20 studies of 459 IFN-treated patients, 3 studies of 38 PEG-IFN-treated patients, and 2 studies of 49 PEG-IFN and ribavirin-treated patients met inclusion criteria. The overall SVR rate was 41% (95% confidence interval [CI], 33 to 49) for IFN and 37% (95% CI, 9 to 77) for PEG-IFN. Treatment discontinuation rates were 26% (95% CI, 20 to 34) for IFN and 28% (95% CI, 12 to 53) for PEG-IFN. SVR was higher with 3 million units (MU) or higher of IFN 3 times weekly, with lower mean HCV RNA, and with lower rates of cirrhosis, HCV genotype 1 or elevated transaminase, but these findings were not statistically significant. Treatment discontinuation rates were greater in studies using larger doses.

LIMITATIONS

Publication bias, few randomized controlled trials, and limitations in generalizability to all hemodialysis patients.

CONCLUSION

IFN treatment of hemodialysis patients results in an SVR rate of 41%. Higher dose, lower mean HCV RNA level, and lower rates of cirrhosis, transaminase level increase, and HCV genotype 1 may be associated with greater SVR rates, but additional studies using individual patient data are needed.

Staging investigations for oesophageal cancer: a meta-analysis

The aim of the study was to compare the diagnostic performance of endoscopic ultrasonography (EUS), computed tomography (CT), and ¹⁸F-fluoro-2-deoxy-D-glucose positron emission tomography (FDG-PET) in staging of oesophageal cancer. PubMed was searched to identify English-language articles published before January 2006 and reporting on diagnostic performance of EUS, CT, and/or FDG-PET in oesophageal cancer patients. Articles were included if absolute numbers of true-positive, false-negative, false-positive, and true-negative test results were available or derivable for regional, celiac, and abdominal lymph node metastases and/or distant metastases. Sensitivities and specificities were pooled using a random effects model. Summary receiver operating characteristic analysis was performed to study potential effects of study and patient characteristics. Random effects pooled sensitivities of EUS, CT, and FDG-PET for regional lymph node metastases were 0.80 (95% confidence interval 0.75–0.84), 0.50 (0.41–0.60), and 0.57 (0.43–0.70), respectively, and specificities were 0.70 (0.65–0.75), 0.83 (0.77–0.89), and 0.85 (0.76–0.95), respectively. Diagnostic performance did not differ significantly across these tests. For detection of celiac lymph node metastases by EUS, sensitivity and specificity were 0.85 (0.72–0.99) and 0.96 (0.92–1.00), respectively. For abdominal lymph node metastases by CT, these values were 0.42 (0.29–0.54) and 0.93 (0.86–1.00), respectively. For distant metastases, sensitivity and specificity were 0.71 (0.62–0.79) and 0.93 (0.89–0.97) for FDG-PET and 0.52 (0.33–0.71) and 0.91 (0.86–0.96) for CT, respectively. Diagnostic performance of FDG-PET for distant metastases was significantly higher than that of CT, which was not significantly affected by study and patient characteristics. The results suggest that EUS, CT, and FDG-PET each play a distinctive role in the detection of metastases in oesophageal cancer patients. For the detection of regional lymph node metastases, EUS is most sensitive, whereas CT and FDG-PET are more specific tests. For the evaluation of distant metastases, FDG-PET has probably a higher sensitivity than CT. Its combined use could however be of clinical value, with FDG-PET detecting possible metastases and CT confirming or excluding their presence and precisely determining the location(s).

Differences between systematic reviews/meta-analyses of hyaluronic acid/hyaluronan/hylan in osteoarthritis of the knee

OBJECTIVE

To explore reasons for discrepant results between systematic reviews (SR)/meta-analyses (MA) of the efficacy and safety of hyaluronic acid/hyaluronan/hylan (HA) therapy in the treatment of osteoarthritis (OA) of the knee.

METHODS

A decision algorithm was utilised to identify reasons for discordance among six SR. Sources of discordance such as clinical question, trial selection and inclusion, data extraction, assessment of study quality, assessment of the ability to combine trials, and statistical methods for data synthesis were examined.

RESULTS

A similar question was asked in all six SR. Different trials were selected for inclusion in the reviews mainly because of differences in the search strategies and selection criteria. Although similar methods for data extraction were utilised, differences were found both in the outcome measures and time-points selected for extraction. Methodological quality was not always formally assessed. Different statistical methods for data synthesis resulted in conflicting estimates of therapeutic effect.

CONCLUSIONS

Reasons for the inconsistency of results reported in the six SR were identified. Using the principles of the GRADE approach for estimating the therapeutic effect of HA in the treatment of OA of the knee, there is moderate evidence suggesting that further research is unlikely to change our confidence in the estimate of the effect. In the balance of benefit to harm, the trade-off is probable benefit with respect to pain reduction and physical function improvement with low risk of harm.

The prognostic importance of a small acute decrement in kidney function in hospitalized patients: a systematic review and meta-analysis

BACKGROUND

Recently, acute kidney injury defined by small changes in serum creatinine levels was associated with worse short-term outcomes; however, the precision and variability of this association was not fully explored.

STUDY DESIGN

Systematic review and meta-analysis.

SETTING & PARTICIPANTS

Hospitalized patients.

SELECTION CRITERIA FOR STUDIES

MEDLINE and EMBASE databases were searched for observational cohort studies and randomized controlled trials published from 1990 through February 2007 that provided information for small changes in serum creatinine levels.

PREDICTOR

Small acute changes in serum creatinine levels by absolute and percentage of changes in serum creatinine levels (lower threshold for increase in serum creatinine <0.5 mg/dL or <25%).

OUTCOME

Short-term mortality (<or=30 days).

RESULTS

Compared with controls, patients with a 10% to 24% increase in creatinine levels had a relative risk (RR) of death of 1.8 (95% confidence interval [CI], 1.3 to 2.5). By comparison, subjects with a 25% to 49% acute change in creatinine levels had an RR of death of 3.0 (95% CI, 1.6 to 5.8), and those with the largest change (>or=50%) had the greatest RR of death (RR, 6.9; 95% CI, 2.0 to 24.5). Results were similar when absolute changes in creatinine levels were considered and when pooled estimates of adjusted RR were used.

LIMITATIONS

Individual patient data were unavailable; thus, only group-level data were pooled for meta-analysis. Results showed a significant degree of statistical heterogeneity that was only partially ameliorated by separating studies into subsets based on clinical setting.

CONCLUSIONS

Short-term mortality and acute decreases in renal function are associated through a graded relationship such that even mild changes in serum creatinine levels portend worse outcome in a variety of clinical settings and patient-types.

Methods for meta-analysis in genetic association studies: a review of their potential and pitfalls

Meta-analysis offers the opportunity to combine evidence from retrospectively accumulated or prospectively generated data. Meta-analyses may provide summary estimates and can help in detecting and addressing potential inconsistency between the combined datasets. Application of meta-analysis in genetic associations presents considerable potential and several pitfalls. In this review, we present basic principles of meta-analytic methods, adapted for human genome epidemiology. We describe issues that arise in the retrospective or the prospective collection of relevant data through various sources, common traps to consider in the appraisal of evidence and potential biases that may interfere. We describe the relative merits and caveats for common methods used to trace inconsistency across studies along with possible reasons for non-replication of proposed associations. Different statistical models may be employed to combine data and some common misconceptions may arise in the process. Several meta-analysis diagnostics are often applied or misapplied in the literature, and we comment on their use and limitations. An alternative to overcome limitations arising from retrospective combination of data from published studies is to create networks of research teams working in the same field and perform collaborative meta-analyses of individual participant data, ideally on a prospective basis. We discuss the advantages and the challenges inherent in such collaborative approaches. Meta-analysis can be a useful tool in dissecting the genetics of complex diseases and traits, provided its methods are properly applied and interpreted.

A systematic review of BNP as a predictor of prognosis in persons with coronary artery disease

OBJECTIVE

This systematic review was conducted to examine whether B-type natriuretic peptide (BNP) can predict mortality and other cardiac endpoints in persons diagnosed with coronary artery disease (CAD).

DESIGN AND METHODS

Databases were searched from 1989 to February 2005 for primary studies that measured BNP for the purpose of diagnosis, prognosis, and monitoring treatment.

RESULTS

In 18 studies, concentrations of BNP were found to have consistent positive associations with poorer prognoses for persons with CAD. The overall range of effect (95% confidence interval) was 2.31 to 5.02, measured via a random effects meta-analysis on studies reporting an odds ratio. Prognostic ability was similar for mortality and non-fatal outcomes. Ranges of estimated measures of effect (i.e., odds ratio, relative risk, hazard ratio) were concentrated between 1.33 to 2.94 for mortality and 1.01 to 3.03 for non-fatal outcomes.

CONCLUSIONS

Further research is needed to assess whether prognostic ability differs by comorbidity or concomitant treatment. As well, the importance and selection of cut points remains unresolved. Until greater clarity is given to these matters, it would be prudent for clinicians to employ caution when using concentrations of BNP to predict the prognosis of persons with CAD.

Atypical antipsychotics in bipolar disorder: systematic review of randomised trials

BACKGROUND

Atypical antipsychotics are increasingly used for treatment of mental illnesses like schizophrenia and bipolar disorder, and considered to have fewer extrapyramidal effects than older antipsychotics.

METHODS

We examined efficacy in randomised trials of bipolar disorder where the presenting episode was either depression, or manic/mixed, comparing atypical antipsychotic with placebo or active comparator, examined withdrawals for any cause, or due to lack of efficacy or adverse events, and combined all phases for adverse event analysis. Studies were found through systematic search (PubMed, EMBASE, Cochrane Library), and data combined for analysis where there was clinical homogeneity, with a special reference to trial duration.

RESULTS

In five trials (2,206 patients) participants presented with a depressive episode, and in 25 trials (6,174 patients) the presenting episode was manic or mixed. In 8-week studies presenting with depression, quetiapine and olanzapine produced significantly better rates of response and symptomatic remission than placebo, with NNTs of 5-6, but more adverse event withdrawals (NNH 12). With mania or mixed presentation atypical antipsychotics produced significantly better rates of response and symptomatic remission than placebo, with NNTs of about 5 up to six weeks, and 4 at 6-12 weeks, but more adverse event withdrawals (NNH of about 22) in studies of 6-12 weeks. In comparisons with established treatments, atypical antipsychotics had similar efficacy, but significantly fewer adverse event withdrawals (NNT to prevent one withdrawal about 10). In maintenance trials atypical antipsychotics had significantly fewer relapses to depression or mania than placebo or active comparator. In placebo-controlled trials, atypical antipsychotics were associated with higher rates of weight gain of >or=7% (mainly olanzapine trials), somnolence, and extrapyramidal symptoms. In active controlled trials, atypical antipsychotics were associated with lower rates of extrapyramidal symptoms, but higher rates of weight gain and somnolence.

CONCLUSION

Atypical antipsychotics are effective in treating both phases of bipolar disorder compared with placebo, and as effective as established drug therapies. Atypical antipsychotics produce fewer extrapyramidal symptoms, but weight gain is more common (with olanzapine). There is insufficient data confidently to distinguish between different atypical antipsychotics.

Procalcitonin Test in the Diagnosis of Bacteremia: A Meta-analysis

STUDY OBJECTIVE

We seek to evaluate the diagnostic performance of the procalcitonin test for the diagnosis of bacteremia in the emergency department (ED) population.

METHODS

We conducted a search of MEDLINE, bibliographies of previous systemic reviews, and pertinent national meeting research abstracts. We included studies that assessed the diagnostic accuracy of procalcitonin for bacteremia, with blood culture as the reference standard. We included prospective investigations of adults and children with suspected infection studied in the ED or at admission. Two authors independently extracted data and assessed study quality; consensus was reached by conference. The analysis was based on the I2 statistic for heterogeneity, unweighted summary receiver-operating characteristic curve, and random-effects pooled sensitivity and specificity across studies using the same test threshold.

RESULTS

The search yielded 348 publications and 1 unpublished study. Seventeen studies met the inclusion criteria and provided a sample of 2,008 subjects. There was a substantial degree of inconsistency (I2=64%). The unweighted summary receiver-operating characteristic curve provided the best overall estimate of test performance, with an area under the curve of 0.84 (95% confidence interval [CI] 0.75 to 0.90). Sensitivity analysis based on study quality did not significantly change the results. Subgroup analysis including only studies that used a test threshold of 0.5 or 0.4 ng/mL yielded pooled estimates for sensitivity and specificity of 76% (95% CI 0.66 to 0.84) and 70% (95% CI 0.60 to 0.79), respectively.

CONCLUSION

We found the diagnostic performance of the procalcitonin test for identifying bacteremia in ED patients to be moderate. Future research designed to determine the utility of the procalcitonin test as a diagnostic tool used in isolation for detecting bacteremia in ambulatory patients is needed before widespread clinical use.

The appropriateness of asymmetry tests for publication bias in meta-analyses: a large survey

BACKGROUND

Statistical tests for funnel-plot asymmetry are common in meta-analyses. Inappropriate application can generate misleading inferences about publication bias. We aimed to measure, in a survey of meta-analyses, how frequently the application of these tests would be not meaningful or inappropriate.

METHODS

We evaluated all meta-analyses of binary outcomes with é 3 studies in the Cochrane Database of Systematic Reviews (2003, issue 2). A separate, restricted analysis was confined to the largest meta-analysis in each of the review articles. In each meta-analysis, we assessed whether criteria to apply asymmetry tests were met: no significant heterogeneity, I2 < 50%, é 10 studies (with statistically significant results in at least 1) and ratio of the maximal to minimal variance across studies > 4. We performed a correlation and 2 regression asymmetry tests and evaluated their concordance. Finally, we sampled 60 meta-analyses from print journals in 2005 that cited use of the standard regression test.

RESULTS

A total of 366 of 6873 (5%) and 98 of 846 meta-analyses (12%) in the wider and restricted Cochrane data set, respectively, would have qualified for use of asymmetry tests. Asymmetry test results were significant in 7%-18% of the meta-analyses. Concordance between the 3 tests was modest (estimated k 0.33-0.66). Of the 60 journal meta-analyses, 7 (12%) would qualify for asymmetry tests; all 11 claims for identification of publication bias were made in the face of large and significant heterogeneity.

INTERPRETATION

Statistical conditions for employing asymmetry tests for publication bias are absent from most meta-analyses; yet, in medical journals these tests are performed often and interpreted erroneously.

Use of intravenous magnesium to treat acute onset atrial fibrillation: a meta-analysis

OBJECTIVES

To assess the effects of intravenous magnesium on converting acute onset atrial fibrillation to sinus rhythm, reducing ventricular response and risk of bradycardia.

DESIGN AND DATA SOURCES

Randomised controlled trials evaluating intravenous magnesium to treat acute onset atrial fibrillation from MEDLINE (1966 to 2006), EMBASE (1990 to 2006) and Cochrane Controlled Trials Register without language restrictions.

REVIEW METHODS

Two researchers independently performed the literature search and data extraction.

RESULTS

10 randomised controlled trials, including a total of 515 patients with acute onset atrial fibrillation, were considered. Intravenous magnesium was not effective in converting acute onset atrial fibrillation to sinus rhythm when compared to placebo or an alternative antiarrhythmic drug. When compared to placebo, adding intravenous magnesium to digoxin increased the proportion of patients with a ventricular response <100 beats/min (58.8% vs 32.6%; OR 3.2, 95% CI 1.93 to 5.42; p<0.001). When compared to calcium antagonists or amiodarone, intravenous magnesium was less effective in reducing the ventricular response (21.4% vs 58.5%; OR 0.19, 95% CI 0.09 to 0.44; p<0.001) but also less likely to induce significant bradycardia or atrioventricular block (0% vs 9.2%; OR 0.13, 95% CI 0.02 to 0.76; p = 0.02). The use of intravenous magnesium was associated with transient minor symptoms of flushing, tingling and dizziness in about 17% of the patients (OR 14.5, 95% CI 3.7 to 56.7; p<0.001).

CONCLUSIONS

Adding intravenous magnesium to digoxin reduces fast ventricular response in acute onset atrial fibrillation. The effect of intravenous magnesium on the ventricular rate and its cardiovascular side effects are less significant than other calcium antagonists or amiodarone. Intravenous magnesium can be considered as a safe adjunct to digoxin in controlling the ventricular response in atrial fibrillation.

Application and investigation of a bound for outcome reporting bias

Background

Direct empirical evidence for the existence of outcome reporting bias is accumulating and this source of bias is recognised as a potential threat to the validity of meta-analysis of randomised clinical trials.

Methods

A method for calculating the maximum bias in a meta-analysis due to publication bias is adapted for the setting where within-study selective non-reporting of outcomes is suspected, and compared to the alternative approach of missing data imputation. The properties of both methods are investigated in realistic small sample situations.

Results

The results suggest that the adapted Copas and Jackson approach is the preferred method for reviewers to apply as an initial assessment of robustness to within-study selective non-reporting.

Conclusion

The Copas and Jackson approach is a useful method for systematic reviewers to apply to assess robustness to outcome reporting bias.

The case of the misleading funnel plot

Evidence based medicine insists on rigorous standards to appraise clinical interventions. Failure to apply the same rules to its own tools could be equally damaging

Meta-analysis of frusemide to prevent or treat acute renal failure

OBJECTIVE

To investigate the potential beneficial and adverse effects of frusemide to prevent or treat acute renal failure in adults.

DESIGN

Meta-analysis of randomised controlled trials.

DATA SOURCES

Cochrane controlled trials register (2005 issue 4), Embase, and Medline (1966 to 1 February 2006), without language restrictions.

REVIEW METHODS

Two reviewers checked the quality of the studies and independently extracted data.

RESULTS

Nine randomised controlled trials totalling 849 patients with or at risk of acute renal failure were included. Outcome measures not significantly different after frusemide treatment were in-hospital mortality (relative risk 1.11, 95% confidence interval 0.92 to 1.33), risk for requiring renal replacement therapy or dialysis (0.99, 0.80 to 1.22), number of dialysis sessions required (weight mean difference--0.48 sessions, -1.45 to 0.50), and proportion of patients with persistent oliguria (urine output < 500 ml/day: 0.54, 0.18 to 1.61). Stratifying studies that used frusemide to prevent or treat acute renal failure did not change the results on mortality (relative risk ratio 2.10, 95% confidence interval 0.67 to 6.63) and the risk for requiring dialysis (4.12, 0.46 to 37.2). Evidence suggested an increased risk of temporary deafness and tinnitus in patients treated with high doses of frusemide (relative risk 3.97, 95% confidence interval 1.00 to 15.78).

CONCLUSIONS

Frusemide is not associated with any significant clinical benefits in the prevention and treatment of acute renal failure in adults. High doses may be associated with an increased risk of ototoxicity.

Preventing post-organ transplantation cytomegalovirus disease with ganciclovir: a meta-analysis comparing prophylactic and preemptive therapies

BACKGROUND

Cytomegalovirus (CMV) causes significant morbidity and mortality in transplant recipients, but there is no consensus regarding the most appropriate prevention method. The goal of this meta-analysis was to compare the efficacy of universal prophylaxis and preemption using ganciclovir.

METHODS

Literature searches for randomized and nonrandomized controlled trials of ganciclovir prophylaxis and preemption were conducted. Because of the lack of head-to-head trials, indirect comparisons of meta-analyses of the prevention strategies were performed. Meta-analyses were conducted using a random effects model to estimate the overall risk ratios for various clinical outcomes. We assessed the event rates for control groups across the trials for comparability.

RESULTS

Literature searches identified 17 universal prophylaxis trials and 9 preemption trials with 1560 and 457 subjects, respectively. Overall event rates for CMV disease in control groups across the studies were similar (approximately 26%). The relative risk of CMV disease in prophylaxis trials was 0.34 (95% confidence interval, 0.24-0.48) when trials of patients with prophylaxis of short duration and trials that only evaluated patients with high-risk serostatus were excluded. The relative risk of CMV disease for study subjects in all preemption trials was 0.30 (95% confidence interval, 0.15-0.60), compared with that for control subjects. There was no statistically significant difference in CMV disease between prevention strategies. Similarly, no differences between strategies were found for all-cause mortality or rejection. There were insufficient data to adequately evaluate graft loss and opportunistic infection.

CONCLUSIONS

On the basis of indirect comparisons of meta-analyses of prevention strategies, universal prophylaxis and preemption are equally effective in reducing the incidence of CMV disease.

Epidural pain relief versus systemic opioid-based pain relief for abdominal aortic surgery

BACKGROUND

Epidural analgesia offers greater pain relief compared to systemic opioid-based medications, but its effect on morbidity and mortality is unclear.

OBJECTIVES

To assess the benefits and harms of postoperative epidural analgesia in comparison with postoperative systemic opioid-based pain relief for adult patients who underwent elective abdominal aortic surgery.

SEARCH STRATEGY

We searched the Cochrane Central Register of Controlled Trials via OVID (CENTRAL) (The Cochrane Library, Issue 3, 2004); OVID MEDLINE (1966 to July 2004); and EMBASE (1980 to June 2004). We assessed non-English language reports and contacted researchers in the field. We did not seek unpublished data.

SELECTION CRITERIA

We included all randomized controlled trials comparing postoperative epidural analgesia and postoperative systemic opioid-based analgesia for adult patients who underwent elective open abdominal aortic surgery.

DATA COLLECTION AND ANALYSIS

Two authors independently assessed trial quality and extracted data. We contacted study authors for additional information and data.

MAIN RESULTS

Thirteen studies involving 1224 patients met our inclusion criteria; 597 patients received epidural analgesia and 627 received systemic opioid analgesia. The epidural analgesia group showed significantly lower visual analogue scale for pain on movement (up to postoperative day three), regardless of the site of epidural catheter and epidural formulation. Postoperative duration of tracheal intubation and mechanical ventilation was significantly shorter by about 20% in the epidural analgesia group. The overall incidence of cardiovascular complication; myocardial infarction; acute respiratory failure (defined as an extended need for mechanical ventilation); gastrointestinal complication; and renal insufficiency was significantly lower in the epidural analgesia group, especially in trials that used thoracic epidural analgesia.

AUTHORS' CONCLUSIONS

Epidural analgesia provides better pain relief (especially during movement) for up to three postoperative days. It reduces the duration of postoperative tracheal intubation by roughly 20%. The occurrence of prolonged postoperative mechanical ventilation, overall cardiac complication, myocardial infarction, gastric complication and renal complication was also reduced by epidural analgesia, especially thoracic. However, current evidence does not confirm the beneficial effect of epidural analgesia on postoperative mortality and other types of complications.

Benzodiazepines for neuroleptic-induced tardive dyskinesia

BACKGROUND

Tardive dyskinesia (TD) is a disfiguring movement disorder, often of the orofacial region, frequently caused by the use of neuroleptic drugs. A wide range of strategies have been used to help manage tardive dyskinesia, and for those who are unable to have their antipsychotic medication stopped or substantially changed, the benzodiazepine group of drugs have been suggested as a useful adjunctive treatment.

OBJECTIVES

To determine the effects of benzodiazepines for neuroleptic-induced tardive dyskinesia in people with schizophrenia or other chronic mental illnesses.

SEARCH STRATEGY

1. Electronic searches. For the update of 2006, we searched The Cochrane Schizophrenia Group Trials Register (November 2005). For the previous two updates (1996, 2002) the review authors searched Biological Abstracts (1982-2002), the Cochrane Schizophrenia Group's Register of trials (February 2002), EMBASE (1980-2002), LILACS (1982-2002), MEDLINE (1966-2002), PsycLIT (1974-2002), SCISEARCH (2002), hand searched references of all included/excluded studies and contacted the first author of each included trial.

SELECTION CRITERIA

We included all randomised clinical studies focusing on people with schizophrenia (or other chronic mental illnesses) and neuroleptic-induced tardive dyskinesia that compared benzodiazepines with placebo or no intervention.

DATA COLLECTION AND ANALYSIS

We independently extracted data from the studies and ensured that they were reliably selected, and quality assessed. For homogenous dichotomous data we calculated random effects, relative risk (RR), 95% confidence intervals (CI) and, where appropriate, numbers needed to treat (NNT) on an intention-to-treat basis. We synthesised continuous data from valid scales by using a weighted mean difference (WMD). For continuous outcomes we preferred endpoint data to change data.

MAIN RESULTS

We identified three trials (total N=56, one additional trial since 2002, n=24). Using benzodiazepines as an adjunctive treatment did not result in any clear changes for a series of tardive dyskinesia medium-term outcomes (n=30, 2 RCTs, RR not improved to clinically important extent 1.08 CI 0.57 to 2.05). One trial (n=24) found end point abnormal movement scores to be better for those receiving adjunct benzodiazepines(WMD AIMS -3.22 CI -4.63 to -1.81 ). Less than 10% in both groups left these studies before completion and none of the studies reported clear adverse effects.

AUTHORS' CONCLUSIONS

One small study reports some preliminary evidence that benzodiazepines may have some effect in neuroleptic induced tardive dyskinesia. Inconclusive results from other studies means routine clinical use is not indicated and these treatments remain experimental.

Imiquimod for Actinic Keratosis: Systematic Review and Meta-Analysis

Benefit and harm associated with treating actinic keratosis (AK) with the immune response modifier imiquimod was assessed using published randomized-controlled trials. Five randomized double-blind trials lasted 12-16 weeks and treated 1,293 patients. Complete clearance occurred in 50% of patients treated with imiquimod, compared to 5% treated with vehicle, and the number needed to treat (NNT) for one patient to have their keratosis completely cleared after 12-16 weeks was 2.2 (95% confidence interval 2.0-2.5). For partial (>/=75%) clearance the NNT was 1.8 (1.7-2.0). The proportion of patients with any adverse event, any local adverse event, or any treatment-related adverse event was substantially higher with imiquimod than with vehicle, and numbers needed to harm for one additional adverse event with imiquimod over 12-16 weeks ranged from 3.2 to 5.9. Particular local adverse events with imiquimod included erythema (27%), scabbing or crusting (21%), flaking (9%), erosion (6%), edema (4%), and weeping (3%). Imiquimod 5% cream was effective in the treatment of AK, preventing potential development of squamous cell carcinoma. Future investigation might be aimed at elucidating optimal dosing to minimize adverse events without detriment to efficacy, and evaluating long-term recurrence.

Viscosupplementation for the treatment of osteoarthritis of the knee

BACKGROUND

Osteoarthritis (OA) is the most prevalent chronic joint disorder worldwide and is associated with significant pain and disability.

OBJECTIVES

To assess the effects of viscosupplementation in the treatment of OA of the knee. The products were hyaluronan and hylan derivatives (Adant, Arthrum H, Artz (Artzal, Supartz), BioHy (Arthrease, Euflexxa, Nuflexxa), Durolane, Fermathron, Go-On, Hyalgan, Hylan G-F 20 (Synvisc Hylan G-F 20), Hyruan, NRD-101 (Suvenyl), Orthovisc, Ostenil, Replasyn, SLM-10, Suplasyn, Synject and Zeel compositum).

SEARCH STRATEGY

MEDLINE (up to January (week 1) 2006 for update), EMBASE, PREMEDLINE, Current Contents up to July 2003, and the Cochrane Central Register of Controlled Trials (CENTRAL) were searched. Specialised journals and reference lists of identified randomised controlled trials (RCTs) and pertinent review articles up to December 2005 were handsearched.

SELECTION CRITERIA

RCTs of viscosupplementation for the treatment of people with a diagnosis of OA of the knee were eligible. Single and double-blinded studies, placebo-based and comparative studies were eligible. At least one of the four OMERACT III core set outcome measures had to be reported (Bellamy 1997).

DATA COLLECTION AND ANALYSIS

Each trial was assessed independently by two reviewers for its methodological quality using a validated tool. All data were extracted by one reviewer and verified by a second reviewer . Continuous outcome measures were analysed as weighted mean differences (WMD) with 95% confidence intervals (CI). However, where different scales were used to measure the same outcome, standardized mean differences (SMD) were used. Dichotomous outcomes were analyzed by relative risk (RR).

MAIN RESULTS

Seventy-six trials with a median quality score of 3 (range 1 to 5) were identified. Follow-up periods varied between day of last injection and eighteen months. Forty trials included comparisons of hyaluronan/hylan and placebo (saline or arthrocentesis), ten trials included comparisons of intra-articular (IA) corticosteroids, six trials included comparisons of nonsteroidal anti-inflammatory drugs (NSAIDs), three trials included comparisons of physical therapy, two trials included comparisons of exercise, two trials included comparisons of arthroscopy, two trials included comparisons of conventional treatment, and fifteen trials included comparisons of other hyaluronans/hylan. The pooled analyses of the effects of viscosupplements against 'placebo' controls generally supported the efficacy of this class of intervention. In these same analyses, differential efficacy effects were observed for different products on different variables and at different timepoints. Of note is the 5 to 13 week post injection period which showed a percent improvement from baseline of 28 to 54% for pain and 9 to 32% for function. In general, comparable efficacy was noted against NSAIDs and longer-term benefits were noted in comparisons against IA corticosteroids. In general, few adverse events were reported in the hyaluronan/hylan trials included in these analyses.

AUTHORS' CONCLUSIONS

Based on the aforementioned analyses, viscosupplementation is an effective treatment for OA of the knee with beneficial effects: on pain, function and patient global assessment; and at different post injection periods but especially at the 5 to 13 week post injection period. It is of note that the magnitude of the clinical effect, as expressed by the WMD and standardised mean difference (SMD) from the RevMan 4.2 output, is different for different products, comparisons, timepoints, variables and trial designs. However, there are few randomised head-to-head comparisons of different viscosupplements and readers should be cautious, therefore, in drawing conclusions regarding the relative value of different products. The clinical effect for some products, against placebo, on some variables at some timepoints is in the moderate to large effect-size range. Readers should refer to relevant tables to review specific detail given the heterogeneity in effects across the product class and some discrepancies observed between the RevMan 4.2 analyses and the original publications. Overall, the analyses performed are positive for the HA class and particularly positive for some products with respect to certain variables and timepoints, such as pain on weight bearing at 5 to 13 weeks postinjection. In general, sample-size restrictions preclude any definitive comment on the safety of the HA class of products; however, within the constraints of the trial designs employed no major safety issues were detected. In some analyses viscosupplements were comparable in efficacy to systemic forms of active intervention, with more local reactions but fewer systemic adverse events. In other analyses HA products had more prolonged effects than IA corticosteroids. Overall, the aforementioned analyses support the use of the HA class of products in the treatment of knee OA.

Systematic review and meta-analysis of randomised trials and cohort studies of mycophenolate mofetil in lupus nephritis

Mycophenolate mofetil (MMF) is an immunosuppressant drug being used for induction and maintenance of remission of lupus nephritis in systemic lupus erythematosus. Evidence about its use was sought from full publications and abstracts of randomised trials and cohort studies by using a variety of search strategies. Efficacy and adverse event outcomes were sought. Five randomised trials enrolled patients with World Health Organization (WHO) class III, IV, or V (mostly IV) lupus nephritis, predominantly comparing MMF (1 to 3 g daily) with cyclophosphamide and steroid. Complete response and complete or partial response was significantly more frequent with MMF than with cyclophosphamide, with numbers needed to treat of 8 (95% confidence interval 4.3 to 60) to induce one additional complete or partial response, with wide confidence intervals. Death was reported less frequently with MMF (0.7%, 1 death in 152 patients) than with cyclophosphamide (7.8%, 12 deaths in 154 patients), with a number needed to treat to prevent (NNTp) one death of 14 (8 to 48). Hospital admission was also lower with MMF (1.7% versus 15%; NNTp 7.4 [4.8 to 16]). Serious infections, leucopaenia, amenorrhoea, and hair loss were all significantly less frequent with MMF than with cyclophosphamide, but diarrhoea was significantly more common with MMF. Ten of 18 cohort studies enrolled only patients with lupus nephritis (author-defined or WHO class III to V). Seven of these 10 reported that complete or partial response with MMF (mostly 1 or 2 g daily) with steroid occurred in 121/151 (80%) and that treatment failure or no response occurred in 30/151 (20%). Adverse events were generally similar in cohort studies with and without only patients with lupus nephritis. In all 18 cohorts, gastrointestinal adverse events (diarrhoea, nausea, vomiting) occurred in 30%, infection in 23%, and serious infection in 4.3%. Adverse event discontinuations occurred in 14% and lack of efficacy occurred in 10%. There was a single death with MMF, a mortality rate over the course of 1 year of approximately 0.2%. The results form a basis on which to plan future studies and provide a guide for the use of MMF in lupus nephritis until results of larger studies are available. At least one such study is under way.

[The contribution of alcohol to nutrition: addition or substitution according to cultural origins]

The relation of alcohol intake to diet is analyzed through multiple linear regression for a sample of 475 males living in Geneva. Control variables are age, relative weight index, marital status and employment status. Whereas alcohol is associated with higher dietary intake for people of mediterranean origin, it tends to replace food calories for natives of german speaking areas.