Warm and humidified insufflation gas during gynecologic laparoscopic surgery reduces postoperative pain in predisposed patients—a randomized, controlled multi-arm trial

The study was designed as a monocentric, prospective, randomized, double-blinded controlled trial with three parallel intervention arms. Before trial commencement, the study design was changed to be a single-blinded trial because the surgeons and study staff could not be effectively blinded with respect to the devices used during the laparoscopic procedure. All patients and ward staff were not aware of the method used during laparoscopy.

The methods and trial are described in detail in supplement 2.

The study included 150 participants with an indication to a laparoscopic gynecological surgery. It was conducted at the Department of Anesthesiology and the Department of Gynecology and Obstetrics, University Hospital Aachen, Germany between July 2016 and September 2018.

The participants were randomized in 3 groups of 50 subjects each. In group “AIR” (control group “AIR”), a forced air warming blanket and cold and dry insufflation gas was used during surgery. In group “HUMI”, insufflation was performed with warm and humidified insufflation gas and no warming blanket was used (“HUMI”). Group “HUMI +” was treated with a combination of a forced air warming blanket and warm, humidified gas (“HUMI +”).

Eligible patients were female, aged between 18 and 69 years with a body mass index under 35, admitted to the hospital for laparoscopic surgery with a planned duration of more than 60 min.

Exclusion criteria were patients who were pregnant or not using sufficient contraception, who were breastfeeding, who were engaged in alcohol or drug abuse, who were either expected not to comply with instructions or with limited ability to comply with instructions for this study, who were unwilling or unable to give informed consent, who participated in another interventional study within the last 3 months, who are committed to an institution and/ or penitentiary by judicial or official order, and who are employees of the investigator cooperation companies.

The primary endpoint was postoperative pain recorded by the visual analogue scale upon arrival in the recovery room, before transfer to the ward, at 8 p.m. on the day of surgery, at 8 a.m. and 8 p.m. on postoperative days 1 to 7 specifically for abdominal pain, pain in the shoulders, pain upon movement, and pain upon coughing. Secondary endpoints were analgesic consumption, the duration of epidural anesthesia, postoperative nausea and vomiting, differences in activities of daily living (ADL), the length of stay in post-anesthesia care unit (PACU), and the total length of the hospital stay.

This study was designed to address heating capabilities and pain reduction. Three groups were constructed, and sample size and statistical power were calculated to detect a difference in core body temperature. In a Cohen's delta effect size power analysis, a sample size of 50 in each of the treatment groups would give a power of 0.8 to detect a difference of at least 0.2 °C between groups in a balanced design. The effects on body temperature have not been published yet.

After patients were enrolled by the study team and written informed consent was obtained, study participants were randomized with equal allocation ratios to the three interventions using permuted block randomization (block size 6) stratified by endometriosis (Yes/No). Computer-generated sequences were used. To maintain allocation concealment, the randomization sequence and the block size were concealed from the investigators and the study team until database lock and the assignment to study participants was carried out with a web-based application maintained by the Institute of Medical Informatics, RWTH Aachen University.

Outcome variables were described within each treatment group using standard descriptive statistics (frequency, minimum, maximum, quartiles, mean, and standard deviation). Descriptive statistics for pain scores were also calculated separately for each measurement timepoint. Analyses of pain scores were performed on the maximum pain score calculated as the maximum of the abdominal-, movement-, upon coughing-, and shoulder pain scores for each subject at each measurement occasion. A linear mixed effects model was used to model the pain score [25, 26]. Estimated treatment effects at each measurement occasion were calculated from the model along with nominal 95% confidence intervals. Explorative tests for the treatment effect on activities of daily living (ADL) scores and the frequency of nausea severity levels on day 1 and on discharge day were conducted with Kruskal–Wallis rank sum test and Pearson’s Chi-squared tests, respectively. The length of the stay in PACU was analyzed using a general linear model with gamma-distributed errors. Analyses were conducted using R [27]. Mixed models were fitted with lme4 [28].

A total of 208 patients with an indication to a laparoscopic gynecologic surgery were assessed for eligibility between July 2016 and September 2018. The trial ended after the planned 150 interventions were completed. 58 subjects were not included, as they did not meet inclusion criteria, declined to participate, participated in another study, or for other reasons. The 150 patients who were included in the study were randomized either to the control group or to one of the two intervention groups, stratified by endometriosis. Seventy-four women declared to suffer from endometriosis and 76 did not have a diagnosis of endometriosis. Four of these patients were newly diagnosed with endometriosis during this study.

One patient randomized to the control group accidently received no warming blanket, but warm humidified insufflation gas instead of the allocated intervention with forced air warming blanket alone. One patient randomized into group “HUMI” did not receive surgery and consequently the allocated intervention, because of a preoperative spontaneous rupture of the ovarian cyst that was the indication for laparoscopy. Three patients randomized into group “HUMI +” did not receive the allocated intervention: in two cases because the planned surgery was not performed due to different reasons and the other because the patient received a forced air warming blanket only instead of the allocated intervention.

In two cases the intervention was discontinued: one patient from group “AIR” because of the intraoperative indication to a conversion of the laparoscopic procedure into laparotomy and one patient from group “HUMI” because of the lowering of core body temperature below 35 °C during the intervention with the necessity to use in addition a forced air warming blanket.

Follow-up data were available for 146 of the 150 randomized patients. In one case from group “AIR” and one case from group “HUMI +”, follow-up data were missing, because the VAS questionnaires were not available. There were also no follow-up data from the two patients without surgery after randomization.

Primary intention-to-treat analysis was performed on the full set of follow-up data (Fig. 1).

Bild vergrößern

Table 1 shows the baseline characteristics of all patients who received surgery. Demographic data, risk factors for cardiovascular complications, patient’s medical history, intraoperative medication, IV fluids, and insufflated CO₂, as well as the type and length of operative procedures were recorded. Significant differences between the groups were seen in ASA classification and intraoperative use of paracetamol. No other differences were recorded.

Table 3 shows a descriptive analysis of the secondary endpoints, like postoperative analgesic consumption, flow rate and duration of epidural anesthetics, length of stay in PACU, and length of the hospital stay. Compared to group “AIR”, group “HUMI +” showed significantly less consumption of Ibuprofen at day 2 (− 0.5871 g ± 0.258; p-value = 0.0471) (Table 4). No differences were found in the other secondary endpoints. Not shown are data of nausea and vomiting and ADL scores: there was no difference in occurrence of nausea and vomiting on day 1 (p-value 0.989) or on discharge day (p-value 0.6362). ADL scores were the same in each treatment group separately on day 1 (p-value 0.45) and on discharge day (p-value 0.2117).

Additionally, exploratory post hoc analysis of postoperative pain scores was performed. We analyzed the treatment effect on postoperative pain scores of the two intervention groups combined, both received warm and humidified gas, and in comparison to the control group, which received cold and dry gas: no significant differences in this model were detected.

Results stratified by endometriosis are shown in Fig. 4A and B. Figure 4A shows pain scores stratified by endometriosis with higher pain scores in patients suffering from endometriosis nearly in all groups over the whole period. The differences between groups “HUMI” vs “AIR” and “HUMI +” vs “AIR” are shown in Fig. 4B. There were significantly lower pain score levels of subjects suffering from endometriosis in group “HUMI” and group “HUMI +” compared to group “AIR” at different timepoints.

Bild vergrößern

Analysis of data stratified by abdominal surgery showed significantly less pain scores in patients who did not undergo previous abdominal surgery in both groups “HUMI” and “HUMI +” compared to group “AIR” at different timepoints (Fig. 5A and B).

Bild vergrößern

A total of 109 (74%) study objects experienced adverse events during study intervention. All adverse events were classified as mild and they were not, or highly unlikely to be, related to the study (see Supplement, Table S1).

Our post hoc analysis revealed two important results. First, patients suffering from endometriosis showed higher pain score levels than non-endometriosis patients nearly over the whole period during our observation (Fig. 4A). However, we detected significantly less pain in both intervention groups compared to the control group if endometriosis was present, which lasted for several days (Fig. 4B). This is a novel finding, indicating that female patients suffering from endometriosis may especially profit from the use of humidified and warm insufflation gas, while generally in higher pain. This can have several possible explanations. First, patients with endometriosis may be suffering from chronic pain, which often causes structural and functional changes in the nociceptive system. Endometriosis induces inflammation in the tissue surrounding it [35] and can be found in 70% of patients suffering from chronic pelvic pain [36]. Additionally, the surgery characteristics may also influence the results of the study. For instance, surgical treatment of endometriosis, which is, according to our clinic standards, performed via excision of the affected areas, results in the stripping of peritoneum and exposure of relatively large surfaces of underlying tissue to the insufflation gas, possibly facilitating tissue drying and other local reactions. This differs from other gynecologic surgeries, like hysterectomy, myomectomy, or ovarian cystectomy, where the exposed surface can be smaller or covered by coagulation area or surgical sutures.

When grouping our patients according to prior or non-prior abdominal surgery, we measured a higher pain level in patients without pre-existent surgery (Fig. 5A), which may be due to the fact, that in pre-operated patients with a likelihood for adhesions in 40–63% of gynecological or obstetric patients, [37] pain relief by adhesiolysis may have been beneficial per se [38, 39]. Nonetheless, the use of WHCD leads to a significant reduction of pain in the patient group facing higher pain levels. Therefore, patient groups with the risk of higher postoperative pain (in our study endometriosis patients and those without prior surgery) may potentially benefit from using warm humidified insufflation gas, which prevents the drying of the wound surfaces during surgery that may contribute to intraoperative pain [7]. In rodent experiments, the use of humidified gas protected against both adhesions and the surface reaction of mesothelium and peritoneum was ameliorated compared to dry gas [11, 12]. In porcine models, the authors describe an increase in peritoneal damage simply using dry gas; these studies demonstrated further that HIF-1α, a string indicator for tissue hypoxia but also a modulator in pain regulation, was enhanced [15, 40]. Additionally, HIF-1α disturbances have been revealed to be present in endometriosis formation and reaction [41]. We did not measure HIF-1α in tissue or abdominal fluid, etc. to proof this theory, but it may be of interest for further studies investigating effects on abdominal pain by insufflation gas.

Our study has several strengths and limitations that need to be addressed. First, the study design had a clear randomization, single blinding, a pre-defined surgical team, and a variety of measures to assess “pain” in a well-defined homogeneous population.

Since the study was designed to address heating capabilities and pain reduction, three groups were constructed, although for the research question, if humidified and warm insufflation gas reduces postoperative pain, a two-sided model would have been sufficient. In our analysis, we remained in the group “HUMI”/ “AIR” and “HUMI +”/ “AIR” design; one model was calculated with both groups pooled against group “AIR” without changing significance levels and under consideration of statistical balancing. Because we did not see any differences, we remained in the three-group model to follow the investigation plan properly.

This study was powered to detect a difference in intraoperative core body temperature. As a limitation, we must mention that no power was calculated for questioning pain score differences.

The unequal distribution of patients stratified by previous abdominal surgery is worth mentioning. While the endometriosis/ non-endometriosis patients are nearly equally distributed, there are 66% with previous abdominal surgery and only 1/3 without. Therefore, we must assume that the results stratified by previous abdominal surgery are perhaps less meaningful.

It is of interest whether the results of this monocentric study can be extrapolated to other patient populations (e.g., urologic) or reproduced in other hospital settings.

Further studies on the pathogenetic mechanisms of the observed differences, involving inflammatory pathways and HIF-1α, especially in the subgroup of patients with endometriosis, are needed.

The adverse events we recorded were overall mild and not related to the study and unlikely explainable simply by the use of differently warm and humid insufflation gas.