Deep Sedation Using Propofol Target-controlled Infusion for Gastrointestinal Endoscopic Procedures

A Retrospective Cohort Study

María E. García Guzzo; María S. Fernandez; Delfina Sanchez Novas; Sandra S. Salgado; Sergio A. Terrasa; Gonzalo Domenech; Carlos A. Teijido


BMC Anesthesiol. 2020;20(195) 

In This Article


Few evidence-based studies have examined the safety outcomes of low-risk patients undergoing propofol TCI sedation for ambulatory GIEPs. Approximately 20,000 GIEPs are performed at our institution per year under sedation managed by anaesthetists, in accordance with institutional regulations. A safe anaesthetic technique with rapid turnover and discharge times is essential for this high-volume practice to remain efficient.

The study of our population demographics revealed that most patients scheduled for elective procedures on an outpatient basis had a low-risk profile (85% had ASA physical status classification scores I–II). Invasive diagnostic and therapeutic procedures were excluded from the investigation. The largest proportion of interventions consisted of colonoscopies alone, followed by colonoscopies combined with EGD; hence, we can state that this study involved a low-risk population undergoing low-risk procedures.

Regarding the administration of hypnotic drugs, Leslie et al. reported a mean total propofol dose of 200 mg in a study including more than 2000 patients undergoing GIEPs,[9] which is lower than the mean total propofol dose administered in the current investigation (291.2 mg). Although a smaller proportion of patients in the study by Leslie et al. received propofol in combination with opioids (fentanyl or alfentanyl), the weight-adjusted doses of fentanyl were similar in both studies (0.78 vs. 0.77 In addition, 37% of patients in the aforementioned study also received 2 mg of midazolam during anaesthesia induction, which has been proven to reduce propofol requirements.[15,16] Therefore, the lower administered doses of propofol in the study by Leslie et al. could probably be attributed to the addition of midazolam, and not opioids, to the anaesthetic regimen. Despite the fact that our patients received larger doses of propofol, this was not associated with an increased incidence of arterial hypotension events, which was reported to be up to 12% in both studies.[9]

When comparing our results with those from a study conducted by Chang et al.,[13] the administered propofol doses in our practice were higher, although dispensed using TCI in both investigations. Patients in the study by Chang et al. received propofol as well as 2–2.5 mg midazolam and a mean alfentanil dose of 493 ng. Again, combination of propofol TCI with benzodiazepines may account for the reduction in propofol requirements.

The procedure duration recorded in the present study (median time 25.07 ± 11 min) was similar to that reported by Leslie et al. However, we found differences in discharge times from the PACU; our mean time to hospital discharge was 42 min, IQR [37 to 48] compared with 60 min, IQR [33 to 82] in the previous study. Although it is difficult to compare the results without knowledge of the discharge criteria used in other studies, extended stay in the PACU may be related to the use of benzodiazepines. These drugs appear to prolong recovery time at the expense of similar hypotension rates.[17]

Arterial hypotension was one of the most frequently encountered adverse events, with an incidence of 12,64%. The administration of vasoactive drugs (19.6%) was more frequent than the occurrence of arterial hypotension, probably related to the fact that many anaesthesiologists selected to administer them pre-emptively to avoid hypotension events. Although these incidences may appear high, we did not encounter major cardiovascular complications and none of the patients included in our study required ACLS or died during the perioperative period.

Through multivariate analysis, colonoscopic procedures and higher propofol doses were found to be associated with a higher incidence of arterial hypotension events. As reported in the literature, bowel cleansing with sodium phosphate relates to significant orthostatic hypotension and increases in heart rate, probably due to intravascular volume contraction.[18,19] This scenario can lead to an increased propensity to arterial hypotension events when combined with administration of propofol for deep sedation.

With respect to respiratory adverse events, SaO2 < 95% was the most frequently encountered event (22.3%); this drop in pulse oximetry likely represents upper airway obstructions with no significant clinical impact. Events of SaO2 < 90% were encountered less often with an incidence of 6.9%. Only four (0.5%) patients required unplanned orotracheal intubation, suggesting that most SaO2 < 95 and < 90% events were resolved by non-invasive manoeuvres to unclog the airway including chin lift, jaw thrust, or insertion of oral/nasal cannulas. Considering the fact that the present study involved elective gastrointestinal procedures in relatively low-risk patients, the rates of significant hypoxia (SpO2 < 90% events) of 6.92% and unplanned orotracheal intubation of 0.5% were non-trivial. The authors believe that certain quality improvements could be incorporated to clinical practice, such as electroencephalographic monitoring and capnography tracing, which may allow for further analyses concerning the appropriate anaesthetic depth, recommended TCI target concentrations, and oxygenation/respiratory adequacy.

Multivariate analysis revealed a significant association between SaO2 < 90% events and obesity. Morbid obesity showed a 10.22 OR for these events, probably suggesting the need for alternate airway management and/or oxygen supplementation strategies under sedation for these patients. The use of a high-flow nasal cannula or continuous positive airway pressure via nasal mask (SuperN2va) have been proposed to reduce oxygen desaturation events for spontaneously breathing obese patients.[20,21]

Most of the limitations of this study are related to information bias due to its retrospective nature. Hemodynamic variables were automatically captured from multiparameter monitors and graphically displayed. Although vasoactive drug administration data were obtained from these anaesthetic charts, the exact administration time registered by anaesthetists may not have been accurate; hence, it is difficult to determine whether these interventions were therapeutic or pre-emptive. Moreover, even when the total propofol doses were accurately documented due to institutional regulations for drug control, the TCI models (Marsch vs Schnider) and target concentrations selected for each patient throughout the procedure were not recorded. We did not encounter major events during the course of the procedures or the immediate post-operative period, and all patients were discharged from the hospital on the day of the procedure. Therefore, it is difficult to evaluate whether intraprocedural hypotension events had any long-term, directly related, cardiovascular or neurovascular consequences. Regarding oxygen desaturation, the use of non-invasive manoeuvres destined to increase oxygen saturation in patients registering SpO2 values under 95% is not regularly recorded in the anaesthetic chart.