There were significant heterogeneities that appeared
according to the I² estimates in the comparisons of direct
evidence; some of them existed between the two RCTs
that came from the same study with identical trial design
and patients’ characteristics, and some of the heterogeneity
might have come from the different setting of the trials.
We investigated possible sources and performed
subgroup analyses if necessary, but that still failed to
explain this heterogeneity. The recent research of heterogeneity
suggested that I² estimates need to be interpreted
with caution when the meta-analysis only includes a limited
number of events or trials.64 Furthermore, observations
in clinical practice tend to suggest the intrinsic
heterogeneity nature of the overall GERD population.
For instance, it is known that Helicobacter pylori infection
has been reported to influence the healing of EO by
PPI.65 Also, the impact of the availability of over-thecounter
of PPIs on the recent study population remains
unclear. Theoretically, patients who failed to obtain relief
from self-treatment with over-the-counter PPIs might be
more likely to seek medical consultation and hence participate
in the recent clinical trials