Abstract
Introduction: The primary focus on predicting ability to wean from mechanical ventilatory support (MVS) has been on ventilatory capacity, lung pathology, diaphragmatic strength and nutrition. While this focus has not been incorrect, results have been disappointing. Clinicians continue to be challenged by patients requiring prolonged MVS. Left ventricular (LV) function is emerging as a poorly explored but potentially significant factor impacting ventilator dependence. Methods: A prospective, comparative design was used to follow 53 critically ill adults requiring MVS until completely weaned up to 4 weeks following initiation of MVS. Subjects were male (n=27, 51%), white (n=50, 94%) and aged 21 to 89 years (mean=63.9, SD=12.9). LV function was measured weekly using surrogate indicators of preload [pulmonary artery occlusion pressure (PAOP) and atrial natriuretic peptide levels (ANP)] and afterload (plasma norepinephrine). Cox proportional hazards regression was used to develop predictive models of MVS duration, treating the longitudinally measured indicators of LV function as time-dependent covariates. Hypotheses were tested using likelihood-ratio statistics. Results: Most subjects (n=40, 75.5%) exhibited compromised LV function based on either PAOP (mean= 16.6 mmHg, SD=5.2) or plasma norepinephrine (median=957pg/mL, SQR=562) levels at MVS initiation. Within 4 weeks, 69.8% (n=37) had weaned completely from MVS. Examination of the surrogate indicators of LV function revealed that subjects having a relative decrease in plasma norepinephrine levels since MVS initiation [χ2(1)=2.73, p<.05] and having lower absolute plasma norepinephrine levels [χ2(1)=3.65, p<.05] were more likely to wean; however, neither absolute or relative change since MVS initiation in either PAOP or ANP predicted MVS duration. Conclusions: These findings suggest LV function may play a role in ventilator dependence and that decreasing afterload may shorten duration of MVS.
