One assumes that all databases (such as SIDIAP) have a certain level of underreporting on variables of interest, although this should not differ between the two randomized study groups.
However, here we have significant heterogeneity of results both within the randomized trials (Stockholm trial [ 13] and HABITS trial [ 14], both conducted in Sweden) and between the observational and randomized study findings in the systematic review conducted by Col and coworkers [ 7].
This is the first randomized study between weekly low-dose of zoledronic acid and conventional dose of zoledronic acid investigating the prognostic and predictive effect of clinical and biochemical markers.
However, the results of the sensitivity analysis did not show any difference in the efficacy of the treatment in relation to the type and quality between the randomized and non-randomized studies, and importantly, the absence of statistical heterogeneity or publication bias suggests that the results are robust.
Yet FEAST is the first large and randomized study of the relationship between volume and composition of intravenous fluids and clinical outcome in acute severe sepsis.
This is a 10-year follow-up report of a prospective, randomized study between the Insall-Burstein II (Zimmer, Warsaw, IN) and the Press Fit Condylar Modular (Johnson & Johnson, Raynham, MA) knees.
However, it is noteworthy to mention that there are several differences between above randomized studies and our retrospective study.
Again we come back to the possibility of a skewed distribution between the randomized groups in that particular study.
A meta-analysis of 13 studies, 12 observational and 1 randomized study assessed the association between metformin and CRC.
One randomized study examined the relationship between dextrose administration after out-of-hospital cardiac arrest and neurological outcome and found no difference between the patients receiving 5% dextrose and those receiving half normal saline [ 45].
