Meta-analysis definition

When our most recent - and most definitive - meta-analysis was published, the headlines in many newspapers blazoned that antidepressants don t work .1 The Daily Telegraph headline phrased it more specifically, clarifying that antidepressants are no better than dummy pills ,2 but even this headline was not entirely accurate. What our analyses actually showed was that antidepressants work statistically better than placebos, but that this statistical difference was not clinically meaningful. It was too small a difference to be of much importance in the life of a severely depressed person. 

In simple terms, meta-analysis is the practice of using statistical methods to combine and quantify the outcomes of a series of studies in a single, pooled analysis. What is crucial in this definition is the emphasis on the use of statistical methods. In most biomedical research, the scientific review has a lengthy history and is still widely used. However, insofar that it does not utilise statistical methods for pooling results, and tends to summarise more in qualitative rather than quantitative terms it cannot be regarded as meta-analysis. 

The largest meta-analysis is the Meta-Analysis of Chemotherapy in Head and Neck Cancer (MACH-NC) study evaluating 63 trials with a total of 10,741 patients (Table 2) (69). MACH-NC assessed individual data rather than literature-based data with the inclusion of updated data and unpublished trials. For two-thirds of the trials, individual data were updated to a median follow-up of 6.8 yr. The meta-analysis was subcategorized into locoregional treatment with and without concomitant chemotherapy, induction/adju-vant chemotherapy, and laryngeal preservation with induction chemotherapy rather than definitive treatment for laryngeal and hypopharyngeal tumors. 

These particular points relate to each individual trial, but equally there will be similar considerations needed at the level of the development plan. In order for the overall, ordered programme of clinical trials to be scientifically sound there needs to be a substantial amount of commonality across the trials in terms of endpoints, definitions of analysis sets, recording of covariates and so on. This will facilitate the use of integrated summaries and meta-analysis for the evaluation and presentation of the complete programme or distinct parts of that programme, and outside of that, will allow a consistency of approach to the evaluation of the different trials. 

In this context, Lilly (474) reported a meta-analysis of three controlled studies of patients with TD who were treated with olanzapine. These authors found an 11-fold decrease in TD on olanzapine versus haloperidol based on the AIMS scale. There were a few patients who developed TD in the first 6 weeks of olanzapine, but this could have been from previous drug exposure, now not suppressed by the neuroleptic. Interestingly, there were no new cases (0/375) of TD developing in patients on long-term olanzapine treatment, whereas there were three of 83 cases on haloperidol. It is very difficult to arrive at definitive evidence about TD because most patients have received previous neuroleptic therapy and because TD-like symptoms occur spontaneously, providing an alternative explanation. It is clear that it is difficult to prove that olanzapine causes TD but equally difficult to prove that it does not. The 11-fold decreased incidence, however, is strong evidence that at least it produces much less TD. 

In a meta-analysis of the metabolic and psychosocial impact of pumps, 52 studies were found 22 were published before 1987 and 13 after 1993, the year in which the results of the DCCT were published (225). The authors stated that therefore conclusions about efficacy are not definitive. All pump malfunctions were reported before 1988. All types of changes were reported when the frequency and severity of hypoglycemia were compared with prepump times. Infection and skin irritation were expressed in different ways in the various studies. The risk of diabetic ketoacidosis fell after 1993. Most users preferred to continue pump treatment, mainly because of more flexibility, greater freedom, and improved glycemic control. 

Scottish Medical Journal 36 76-82 Easton JD, Albers GW, Gaplan LR et al. (2004). Discussion Reconsideration of TLA terminology and definitions. Neurology 62 S29-S34 Elliott WJ (1998). Circadian variation in the timing of stroke onset a meta-analysis. 

FIGURE 18.14 Dose intensity meta-analysis for doxorubicin in patients with osteosarcoma. Each bubble represents a separate clinical trial and the size of the bubble is proportional to the number of patients treated during the trial. Doxorubicin was administered prior to definitive surgical resection and effect is >90% necrosis of tumor in the resected specimen. Dose intensity or dose rate is measured in mg/m /week. (Reproduced with permission from Smith MA, Ungerleider RS, Horowitz ME, Simon R. J Natl Cancer Inst 1991 83 1460-70.)... 

However, the validity of these findings has been questioned, because some papers chosen for meta-analysis by the Cochrane group were suggested to have been incorrectly included (14,15). The results obtained from the meta-analysis have also been challenged by another meta-analysis of albumin administration in critically ill patients, which showed no increased risk in mortality (16). This illustrates the need for high-quality, randomized, controlled trials to generate definitive evidence. 

Several small clinical trials have suggested that total intravenous anesthesia with propofol reduces the incidence of postoperative nausea and vomiting and results in shorter emergence times. However, a systematic review (52) and a meta-analysis (53) have shown that most studies were small, did not have follow-up for more than 6 hours postoperatively, and were sponsored by industry. The results were difficult to combine, owing to heterogeneous definitions of postoperative nausea and vomiting. 

A meta-analysis of 24 studies showed that cimetidine and ranitidine, but not famotidine or nizatidine, caused small increases in blood alcohol concentrations (19). The mechanism is unclear, but one possibility is inhibition of gastric alcohol dehydrogenase. However, relative to accepted legal definitions of intoxication the effect of any histamine receptor antagonist on blood alcohol concentrations is unlikely to be clinically significant. 

In a careful meta-analysis, 29 randomized, controlled studies of the incidence of transient radicular irritation were identified (243). Lidocaine and mepivacaine were identified as the two local anesthetics that most commonly cause transient radicular irritation, while prilo-caine, bupivacaine, and ropivacaine had the lowest incidences. Owing to insufficient data, definitive statements could not be made about the effects of the baricity of the local anesthetic, the concentration, and the effect of vasoconstrictors, although all these factors seemed not to be relevant. With regard to intrathecal ropivacaine, the incidence in the formal studies was zero. However, there has been one previous report after intrathecal administration, and one report of transient radicular irritation following epidural anesthesia with ropivacaine the symptoms resolved within 24 hours (244). 

Systematic reviews are recent additions to the medical literature. In contrast to traditional narrative reviews, these reviews aim to answer a precisely defined clinical question and to do so in a way that is transparent and designed to minimize bias. Some of the defining features of systematic reviews are (1) a clear definition of the clinical question to be addressed, (2) an extensive and explicit strategy to find ail studies (published or unpublished) that may be eligible for inclusion in the review, (3) criteria by which studies are included and excluded, (4) a mechanism to assess the quahty of each study and, in some cases, (5) synthesis of results by use of statistical techniques of meta-analysis. By contrast, traditional reviews are subjective, are rarely well focused on a clinical question, lack explicit criteria for selection of studies to be reviewed, do not indicate criteria to assess the quality of included studies, and rarely can use meta-analysis. 

Meta-analysis is a term coined by RosenthaP for pooling data from several studies. A useful definition of meta-analysis was suggested by Huque, namely, ... a statistical analysis which combines or integrates the results of several independent clinical trials, considered by the analyst to be combinable . ... 

About 75% of the 400 million people with chronic HBV live in Asia. Herbal medication use is a common therapeutic modality in many parts of the world and has been studied extensively in China. A meta-analysis of over 500 papers, including randomized controlled trials, concluded that existing studies were of poor quality, limiting the definitive interpretation of results. However, the meta-analysis did identify that bufotoxin and kurorinone were associated with increased seroconversion of HBeAg and clearance of HBV DNA. Further evaluation of these active components as a possible therapeutic alternative is warranted, but they are not currently being recommended for routine use. 

An important aspect that contributes substantially to the quality and strength of evidence is the availability of patient-level data and individual study protocols for each study in the meta-analysis. Such availability allows evaluation of each study s quality and eligibility for inclusion in the meta-analysis. It allows for more precise outcome definition and ascertainment. The patient-level data permit time-to-event and subgroups analyses, including dose-response analyses. These issues are discussed further in the following section below (see Section 13.6). 

In order to study the possible association antidepressant drugs and adverse suicidal events in adult patients, FDA planned and conducted a meta-analysis study of randomized trials of antidepressants. The meta-analysis had several key features that supported its quality and utility for regulatory actions (1) hypotheses generated from previous and independent evidence provided the meta-analysis objectives (2) the meta-analysis was based on well-defined inclusion criteria and an exhaustive set of trials with patient-level data available (3) the meta-analysis employed rigorous and consistent outcome definitions across trials and patients and (4) the meta-analysis was based on prespecified statistical analysis plan. 

---