Limitations
The current study has several limitations. This study was retrospective
and observational in nature, and as such, may be subject to selection
bias. We did find higher rates of preoperative comorbidity and other
associated complications in patients that experienced GI complications,
which may bias some of the study’s findings. In order to account for
such possible confounding and selection bias, we did perform
multivariable modeling to adjust for such preoperative risk factors, as
well as performing propensity-matched sub-analysis to account for the
possibility of confounding baseline variables. However, it is possible
that there may be some unmeasured confounders that remain despite these
measures. The study findings also demonstrate an association between GI
complications and higher mortality but do not necessarily translate into
causality as the temporal relation of GI and other complications was not
evaluated, nor was the ultimate cause of death.