Propofol Compared to Midazolam Sedation and to General Anesthesia for Percutaneous Microwave Ablation in Patients with Hepatic Malignancies: A Single-Center Comparative Analysis of Three Historical Cohorts

The role of anesthetic techniques in percutaneous tumor ablation procedures is a highly debated topic worldwide since it may have impact on pain, anxiety and intraprocedural patient’s movements, thereby achieving an adequate, complete ablation zone (ideally a > 5 mm circumferential safety margin) [1‐8] Several anesthetic methods can be used, such as general anesthesia, spinal anesthesia, and sedation using midazolam/fentanyl (hereafter: midazolam) or propofol ( ± intravenous analgesia) [1, 3]. The choice of anesthetic technique differs among institutes and is often based on the clinician’s and patient’s preferences and local availability. General and spinal anesthesia are invasive techniques which require specialized actions and are associated with higher systemic complication rates compared to sedation [1, 9]. Midazolam and propofol sedation are known for their short time to onset of action and short time to clearance [10, 11]. Moderate conscious sedation with midazolam was prospectively reported to be safe during biliary, tunneled catheter, diagnostic and vascular interventional procedures [12]. However, midazolam sedation tends to be associated with agitation, irregular breathing, respiratory depression and thoracic movement, which might lead to inadequate needle placement, needle tracking and creation of an insufficient tumor-free ablation margin [13, 14].

Over the past 15 years, propofol has become the drug of choice for many outpatient and short procedures, mainly due to its favorable pharmaceutical properties [3]. Guidelines for diagnostic and therapeutic purposes in gastrointestinal endoscopy were the first to describe clear consensus on sedation management with propofol [15‐18]. Also in pediatric diagnostic imaging studies, compared to midazolam sedation, propofol is preferred in order to reduce undesired motion artifacts [19].

However, to our knowledge, there is no consensus which anesthetic technique should be used for an image-guided percutaneous liver ablation procedure, since there are no comparative studies evaluating the impact of anesthetic technique on local disease control and oncological outcomes. The aim of this study was to retrospectively analyze safety, efficacy and periprocedural perception of pain following percutaneous microwave ablation (MWA) for hepatocellular carcinomas (HCCs) or colorectal liver metastases (CRLMs), of the three most-used techniques in current-day clinical practice: general anesthesia, midazolam and propofol sedation.

Due to the expanding role of interventional radiology in liver cancer treatment, the amount and complexity of thermal ablation procedures have raised the demand for safe anesthetic management in the ambulatory interventional oncology suite [3]. This comparative analysis of three historical cohorts described the outcomes of patients undergoing percutaneous liver tumor MWA for CRLM or HCC to identify potential differences between general anesthesia, midazolam and propofol sedation.

Anesthetic technique was the most significant predictor of LTPFS in the Cox regression model in favor of propofol versus midazolam sedation (HR 7.94; P < 0.001) and in favor of general anesthesia versus midazolam sedation (HR 6.33; P = 0.014). This result suggests that patients who underwent a percutaneous procedure under general anesthesia or propofol sedation had an equally reduced risk of developing LTP compared to patients treated under midazolam sedation. These outcomes imply that propofol sedation results in fewer patient movements, better control of breathing and less pain compared to midazolam sedation. General anesthesia and deep sedation with propofol apparently lead to more controlled ablative procedures with superior precision in needle placement and needle tracking, presumably creating wider and more accurate ablation zones. General anesthesia is the ideal technique due to the fact that one can request apnea at any time with completely controllable respiration. Propofol appears to be equivalent regarding local control, although it is theoretically possible that the continued respiration during probe placement contributed to the two cases of probe-induced hepatic hemorrhage.

The efficacy following percutaneous MWA under propofol sedation and general anesthesia in this series is comparable to the per-lesion LTPFS reported in the most recent surgical series following open MWA for similar sized liver tumors [26]. This may indicate equipoise has been reached between the open and percutaneous approach.

Although the CIRSE quality improvement guidelines mention that thermal liver ablation can be performed under intravenous sedation and general anesthesia, to the best of our knowledge, this is the first study that has compared anesthesia techniques for liver tumor ablation [1]. Kim et al. retrospectively compared general anesthesia to midazolam/fentanyl sedation in a small number of renal cell carcinoma patients treated with percutaneous radiofrequency ablation [27]. The authors also reported a significantly higher LTP rate in the midazolam group, mainly caused by insufficient pain control and breath holding during the procedure leading to incomplete ablations.

Midazolam sedation is traditionally being used for interventional procedures because of its reported safety [12]. From a pharmacodynamics point of view, midazolam differs widely from propofol, which is known to achieve a more profound sedation level and shorter recovery time [28]. Several series compared midazolam sedation to propofol in interventional procedures. One outdated trial included 40 patients with intracranial vascular disease and randomized between the two [29]. No differences were found with regard to complications (pain, inappropriate movements and respiratory changes) and both patient’s and physician’s satisfaction score. However, another randomized study concluded that propofol sedation was associated with superior physician satisfaction (P < 0.05) and less respiratory depression and anxiety compared to midazolam for equivalent sedation levels in patients undergoing a percutaneous transluminal angioplasty (P < 0.05) [14].

In other medical fields, propofol is being used extensively for various procedures. In gastro-intestinal endoscopy, one meta-analysis of 22 studies reported that propofol sedation was associated with shorter recovery and discharge time and that patients were more likely to cooperate compared to traditional sedative agents [28]. One recently published, double-blind, randomized trial revealed that significantly fewer patients who received propofol remembered being awake during outpatient colonoscopy compared to midazolam sedation (respectively 2% vs. 17%, P < 0.0001) [30]. More patients who received propofol were “very satisfied” with their level of consciousness compared to midazolam (86.3% vs. 74%, P = 0.0005). Twenty-six percent of midazolam procedures were rated as “difficult” by the treating physician compared to 4.3% for propofol (P < 0.0001). Anesthesia-related complications were fewer in the propofol group (2.7% vs. 11.7%, P < 0.0001). Another randomized trial also reported less pain perception (P < 0.001) and greater patient and endoscopist satisfaction during colonoscopy in case of propofol-based sedation (n = 126) compared with midazolam/fentanyl (n = 136) [31].

Interestingly, several in vitro studies describe another potential advantage of propofol—that it may contribute to immune modulation, anti-inflammation and inhibition of cancer cell proliferation and invasion [32, 33].

This study has several limitations. First, the three groups are retrospectively analyzed; in other words, the anesthetic technique was not randomly allocated. As such, the possibility of selection bias is not negligible. Though all procedures were analyzed consecutively from a prospective registry database and even though univariate and multivariate analysis was performed to correct for potential biases, there are no guarantees that exclude residual confounding. Because general anesthesia was often chosen for patients with contraindications for both sedation techniques, assessing patient-based oncological endpoints such as overall or cancer-specific survival was considered untrustworthy. Intraprocedural pain perception contains subjective measurements which may have introduced recall bias. Whenever possible, data were reviewed separately by two researchers (RSP and VZP). Since periprocedural parameters, such as pain, were digitally reported by the anesthesiologist and nurse anesthetist, these factors were presumably being more extensively documented in the general anesthesia and propofol groups. In addition, monitoring and administration of midazolam/fentanyl were performed by the interventional radiologist, who, even though specifically trained and certified for this procedure, had limited knowledge of the systemic effects, while general anesthesia and propofol sedation were always administered by a specialized anesthetic assistant under direct supervision of an anesthesiologist.

Despite the fact that propofol administration should be reserved for anesthesia providers, a recently published survey showed that anesthesia providers are not uniformly available during interventional procedures [3]. This could result in situations where interventional radiologists are increasingly being involved in administering sedative drugs and managing complications, as was the case in our institution. Another limitation was the unequal median follow-up duration between the groups (P < 0.001); however, since the majority of LTPs appeared within the first 6 months post-treatment (Fig. 2, numbers at risk), the likelihood of developing LTP decreases over time (plateau curve). Although overall survival is generally considered the most relevant oncological endpoint, the efficacy of closely related ablation techniques to eradicate tumors can best be elucidated by comparing the time to LTP. Although multiple lesions in one patient cannot be considered independent, the per-patient analysis (counting LTP of one of the ablated lesions in a single patient as an event) showed equal differences between the three groups.

To conclude, propofol sedation represents a valid alternative to general anesthesia for percutaneous liver tumor ablation, and midazolam sedation does not. Midazolam sedation was inferior to both general anesthesia and to propofol with regard to local tumor control. Compared to midazolam sedation, propofol reduced the periprocedural perception of anxiety and pain, decreased patient movements and resulted in better control of breathing. This probably contributed to more precise needle placements and tracking with higher ablation accuracy, which is reflected by the superior LTPFS. Propofol-based sedation reduces the number of repeat procedures and should be favored over midazolam sedation in percutaneous liver tumor ablation. Future research should focus on the added value of innovative techniques such as one lung and high-frequency jet ventilation.