The Limitations of Systematic Reviews With Meta-Analyses in Sport Science. Editorial

Abstract

It is generally argued that systematic reviews are necessary because of the possibility of some subjective bias in narrative reviews. Thus, SRwMs are considered the gold standard of scientific evidence. However, “all that glitters is not gold,”2 as SRwMs are only as good as the articles they contain, not only from a risk-of-bias perspective2 but also when considering the characteristics of the samples, protocols, and outcomes included. Generalizations may erroneously pretend to be a one-size-fits-all solution for complex biological phenomena. When considering the factors and potential moderators of training interventions and subsequent performance and physiological adaptations, a suite of population characteristics (eg, training background, age, sex), training-regimen characteristics (eg, exercise type and mode, loading, timing), testing protocols and selected outcomes (eg, exercise type, performance parameter, timing), and other contextual factors should be included. Consideration of all these factors is important for a better characterization of any (acute or chronic) training effect on an individual athlete basis, as previously exemplified for postactivation performance-enhancement strategies.3 However, this approach is not always possible for most SRwMs and the included studies. This shortcoming means that any analysis from an SRwM may be limited because of the infinite combinations of all those potential factors and their moderators, which may result in suboptimal combinations.