Statistical significance meta-analysis

When the CAST collaborative group performed a meta-analysis of 1ST, CAST, and MAST-I, the trend seen in CAST and 1ST toward a beneficial effect of aspirin on the rate of death or dependency reached the threshold for statistical significance. Early aspirin therapy (160-300 mg/day) conferred an absolute reduction in the rate of recurrent ischemic stroke by 0.7% (7 per 1000 patients treated) (p < 0.001) and reduced the rate of death or dependency by 1.3% (13 per 1000 patients treated) (2p = 0.007). Aspirin caused about 2 hemorrhagic strokes among every 1000 patients treated, but prevented about 11 other strokes or deaths in hospital. 

A recent meta-analysis evaluated patients at doses of 15 and 30 mg daily. Patients were analyzed for periods of 2 to 24 weeks. The majority of patients enrolled were female, and more than 80% of the patients evaluated received adjunctive modification in lifestyle. An average weight loss of 3.6 kg (7.9 lbs) was demonstrated for patients treated with phentermine compared with placebo. Although modest in amount, this value was statistically significant.37... 

One meta-analysis reviewed patients receiving doses of 75 mg daily during periods of 6 to 52 weeks. Similar to study characteristics for phentermine, the majority of patients enrolled were female, and all patients implemented adjunctive lifestyle modifications. The average additional weight loss observed was 3 kg (6.6 lbs) compared with diet and exercise alone, resulting in a borderline statistically significant difference.37... 

In Tables 10 and 11, meta-analyses based on 576 cases gleaned by Goodman from a number of reports, there are only moderate variations in mortality among cases grouped by route of administration or by age. Using chi-square analysis, Goodman found no statistically significant age-related differences in susceptibility to the lethal effects of atropine. 

Intermittent claudication appears to benefit from ginkgo therapy many studies demonstrate improved walking distance and decreased pain. One meta-analysis of eight studies documented statistically significant improvement but questioned its clinical relevance. In some studies, the high doses (240 mg) appeared more effective. 

Another meta-analysis of placebo-controlled trials in depression published between 1980 and 2000 showed an increase in the response rates in the placebo arms of trials with a variety of antidepressants (Walsh et al.y 2002). Responses to placebo increased significantly in recent years, as shown by the high positive correlation with the year of publication. The association between response rate and year of publication was more statistically robust for placebo than for active medication. The change in placebo response rate did not appear to be explained directly by changes in study characteristics such as patient age, placebo lead-in or minimum required Hamilton Rating Scale for Depression score. A potential explanation could be the changing awareness of patients and the fact that many patients in recent clinical trials had been exposed to several previous treatments and thus expected to improve (see Box 5.4). 

With meta-analysis the statistical significance of the combined results can be overwhelming when all the differences are in the same direction. For example, when the authors performed a meta-analysis on the probability that maintenance antipsychotics produced a lower relapse rate in schizophrenics than placebo (53% relapsed on placebo and 20% on maintenance antipsychotics), the difference was significant to 10 Typically, when multiple studies have the same outcome, the results of a meta-analysis will be markedly statistically significant. By contrast, p values of 0.05 or 0.01 are very difficult to interpret because an artifact from a single study could produce such nonsignificant significant levels. 

Using meta-analytic techniques based on the means and the standard errors presented graphically in the poster, we estimated pooled data of the four effective dosages of quetiapine both for the BPRS and the CGI severity of illness change scores from baseline to endpoint. Quetiapine produced an improvement of 0.43 effect-size units in comparison with placebo, a difference that was highly statistically significant and about the same improvement as haloperidol. Thus, based on the BPRS or PANSS, quetiapine was similar to neuroleptics in efficacy (i.e., differences were nonsignificant). Based on our meta-analysis, quetiapine is clearly superior to... 

Controlled trials. Table 10-4 summarizes the results from four of the five well-controlled, albeit small, trials comparing lithium with an antipsychotic in classic manic patients. These studies presented their data in a way that allowed for inclusion in a meta-analysis. Each study was a well-controlled, double-blind design, finding lithium superior to an antipsychotic, and the meta-anaiysis of the combined studies demonstrated this difference to be highly statistically significant. 

In addition, Platman (92) studied 13 patients on lithium and 10 on chlorpromazine (CPZ), finding lithium consistently superior overall, but not statistically significant on any individual rating scale. The general state of patients on lithium was markedly superior to those on CPZ because the majority were discharged with no other treatment, whereas all of the patients on CPZ required additional concurrent drugs. Because the author did not present data on individual patients, these results could not be included in our meta-analysis, but his results are also consistent with the outcome in Table 10-4. Due to the small sample size, the results should be interpreted cautiously. 

The meta-analysis of all studies comparing clomipramine or SRIs with placebo for the treatment of OCD found that the active drug produced a better result in every trial. We then calculated the effect size and the statistical significance for clomipramine alone and fluvoxamine alone. Their results showed a highly significant effect for both drugs (Table 13-10 and Table 13-11). There were also several studies with sertraline, fluoxetine, and paroxetine that demonstrated similar results (219, 220, 221, 222. 223,. 224, 225 and 226). 

In controlled trials, SRIs are better than placebo. All studies found these agents consistently and significantly more effective than standard antidepressants but each had a small sample size, and the results were not striking. Because of the consistency across studies, however, our meta-analysis with clomipramine was highly statistically significant. 

A meta-analysis of studies of the effect of thyroxine in TSH-suppressive doses on measurements of bone mineral density has suggested a small but statistically significant adverse effect on bone mineral density, especially in postmenopausal women (39). 

Intranasal calcitonin is associated with fewer adverse effects than parenteral formulations, probably because of low systemic availability. However, a meta-analysis has confirmed that adverse events are poorly reported in clinical trials (21). The pooled relative risk for rhinitis from four trials (n = 1663) was 1.72, but this did not reach statistical significance. 

The efficacy of chronic therapy with oral GPIIb/llla inhibitors has been assessed in five major randomized placebo-controlled trials (EXCITE, OPUS, SYMPHONY SYMPHONY II, and BRAVO) (63,64). These agents were associated with a statistically significant increase in mortality in three out of the five trials. A meta-analysis of these trials (n = 45,523) demonstrated a significant increase in mortality (2.8% vs. 2.1% for placebo, odds ratio 1.35, 95% confidence interval 1.15-1.61), mostly... 

---