Can a Pain Monitor Predict When Postoperative Intensive Care Unit Patients Are Ready to Breathe on Their Own? A Pilot Study

<Background and Rationale>

Delayed liberation from invasive mechanical ventilation after surgery is associated with increased risk of ventilator-associated pneumonia, prolonged sedation and analgesic exposure, ICU-acquired weakness, delirium, and longer ICU and hospital stays. Identifying patients at risk of extubation failure before or during a spontaneous breathing trial (SBT) could allow clinicians to better time weaning attempts and reduce these complications.

The autonomic nervous system plays a broader physiological role than heart rate control alone, contributing to respiratory muscle regulation, the cholinergic anti-inflammatory pathway, and the stress response. Heart rate variability (HRV) is a validated, non-invasive marker of autonomic function, and reduced HRV has been associated with multiorgan dysfunction and mortality in critically ill patients. The weaning process, particularly the SBT, imposes substantial autonomic and cardiorespiratory stress through abrupt changes in intrathoracic pressure, preload and afterload, and oxygen demand, resembling a physiological stress test.

Several studies have examined HRV around the time of SBT and its association with weaning or extubation outcomes, generally showing that reduced total power or altered low-frequency/high-frequency ratios are associated with SBT failure. However, prior studies have relied on offline, retrospective HRV processing methods that are not readily applicable at the bedside, and none have evaluated HRV parameters derived from a real-time, point-of-care Analgesia Nociception Index (ANI) monitor in relation to extubation outcomes. In addition, no prior study has evaluated whether autonomic function measured at ICU admission, prior to any weaning attempt, predicts subsequent extubation success, nor has this question been examined specifically in postoperative surgical ICU patients.

The ANI monitor (MDoloris Medical Systems, France) is a commercially available, non-invasive device that derives HRV-based indices from surface ECG in real time, producing an instantaneous index (ANIi), a smoothed mean index (ANIm), and a signal quality/magnitude index (Energy, based on SDNN). ANI monitoring is an established, nationally approved technology for intraoperative nociception monitoring in Korea since 2021, but its use as a physiologic prognostic tool in the ICU, specifically for predicting extubation outcomes, remains unexplored.

<Study Design>

This is a single-center, prospective, observational pilot cohort study conducted in the surgical ICU of a tertiary academic hospital. As a pilot study, no formal sample size calculation was performed; findings are intended to generate preliminary effect estimates and inform the sample size calculation for a subsequent multicenter study.

To minimize confounding from sedative agents with differing autonomic effects, enrollment is restricted to patients sedated with dexmedetomidine as the primary agent, since dexmedetomidine acts directly on alpha-2 receptors to suppress sympathetic activity, in contrast to GABA-mediated agents such as propofol or benzodiazepines, which have been shown to affect HRV parameters differently.

<Measurement Procedures>

ANI-derived HRV parameters (ANIi, ANIm, Energy) are measured non-invasively via chest-surface ECG patches for 5 minutes at each of four standardized time points: (1) baseline, on the day of ICU admission after confirming stable sedation (RASS -2 to -5 maintained for at least 30 minutes); (2) immediately before the SBT, at transition to pressure support ventilation, using the pre-existing ventilator settings; (3) during the SBT, 10-15 minutes after transition to pressure support ventilation (recommended settings: pressure support 8 cmH2O, PEEP 5 cmH2O, adjustable at clinician discretion, with actual settings recorded); and (4) immediately after extubation, regardless of subsequent success or failure. If ICU admission and extubation occur on the same day, baseline and pre-SBT measurements are obtained separately.

The ANI monitor display is shielded from the treating clinical team throughout the study period so that monitoring values do not influence clinical decision-making. The decision to proceed with SBT and extubation remains entirely at the discretion of the treating physicians and follows standardized institutional SBT initiation/termination criteria (based on oxygen saturation, respiratory rate, heart rate change from baseline, blood pressure, work of breathing, and mental status), which are recorded for use as covariates but do not constitute a study intervention.

Only the SBT and extubation attempt performed on the initial intubation episode (i.e., the final weaning attempt prior to the first extubation) are included in the primary analysis; SBTs or extubation attempts occurring after reintubation are excluded from primary analysis.

<Data Collection>

In addition to the ANI parameters, the following data are collected via electronic medical record review: demographic and baseline characteristics (age, sex, height, weight, surgery type, elective/emergency status, comorbidities); severity scores at ICU admission (SOFA, APACHE II); sedation and analgesia data (agent, dose, RASS, use of patient-controlled analgesia); hemodynamic and ventilatory parameters (vasopressor use, ventilator settings including FiO2, PEEP, respiratory rate, tidal volume, and actual SBT settings); inflammatory and hematologic laboratory values (CRP, procalcitonin, lactate, complete blood count); and clinical outcomes including length of ICU/hospital stay, mortality, and delirium status (CAM-ICU). No additional invasive procedures, blood draws, or interventions beyond standard ICU care are performed for research purposes; ANI monitoring uses only chest-surface ECG patches, similar to routine cardiac monitoring.

<Statistical Analysis Plan>

Continuous variables will be assessed for normality using the Shapiro-Wilk test and summarized as mean (SD) or median (IQR) accordingly; categorical variables will be summarized as frequencies and percentages, with Fisher's exact test applied when expected cell counts are below 5.

Associations between each ANI parameter at each time point and the primary outcome will be assessed using univariable logistic regression, reported as odds ratios with 95% confidence intervals. A multivariable model adjusting for SOFA score and cumulative opioid dose (fentanyl equivalents) will be performed; these two covariates were selected a priori based on their established association with weaning outcomes in the literature, with the number of covariates deliberately limited given the small pilot sample size, following the events-per-variable principle to avoid model overfitting.

Because three time points (baseline, pre-SBT, and during-SBT ANI) are each tested as primary predictors, this multiplicity is explicitly acknowledged as exploratory analysis, and formal multiplicity correction (e.g., Bonferroni) is applied exploratorily rather than as a primary inferential adjustment, consistent with the pilot nature of the study. For parameters showing a significant association, receiver operating characteristic curve analysis will be performed to explore the area under the curve and candidate cut-off values.

Secondary analyses include univariable logistic regression for the association between post-extubation ANI and 48-hour reintubation, and for the association between the within-day change in ANI (pre-SBT to during-SBT, ΔANI) and successful extubation. Length of stay will be analyzed using Spearman correlation after assessing normality. All statistical tests are two-sided, with significance defined as p < 0.05.

Statistical analysis will be performed after completion of enrollment for the pilot cohort (target: up to approximately 50 participants over a 6-month enrollment period), with a total study duration of 12 months from IRB approval, including follow-up medical record review.