The transanal/transperineal (ta/tp) endoscopic approach is the only approach from below that can be used in anorectal endoscopic surgery and is useful in cases of a narrow pelvis, bulky tumor, or obesity, where it is difficult to reach the deep pelvis via a transabdominal approach1,2,3. However, with the growing number of cases, there has been an increase in the number of reports of carbon dioxide embolism (CDE) as a complication when using this technique4,5,6. CDE is a potentially fatal complication but is rare in laparoscopic surgery, with a frequency of 0.0014–0.6%7. On the other hand, it has been suggested that the incidence of CDE is higher with the ta/tp endoscopic approach (0.39–3.8%) than with laparoscopic surgery8,9. However, the exact incidence of CDE when using the ta/tp approach and ways to prevent and respond to it remain unclear.

Clinically significant CDE has a mortality rate as high as 28%10. In the management of CDE, early diagnosis and treatment are important for improving prognosis. Transesophageal echocardiography (TEE) is useful for the early diagnosis of CDE and is reported to be able to detect asymptomatic CDE by confirming bubbles in the right heart system. Studies using TEE in laparoscopic surgery have detected intracardiac bubbles in 69% of cholecystectomies, 17.1% of radical prostatectomies, and 100% of hysterectomies11,12,13. However, there are still few relevant reports on CDE using the ta/tp approach. The aims of this study were to investigate the frequency of subclinical CDE detected by monitoring with TEE during anorectal surgery and to find opportunities for early diagnosis of CDE.

Patients and methods

Patients who underwent surgery via the ta/tp approach for anorectal disease and who consented to participate in the study were included. Patients who were deemed unsuitable for TEE or who did not provide consent were excluded. Surgery was performed using the ta/tp approach as previously reported3,14,15,16. Briefly, total mesorectal excision was performed simultaneously (two-team approach) with the laparoscopic approach in the head-down position. Using the ta/tp approach, CO2 was insufflated using an AirSeal system (ConMed, Utica, NY) with high-flow smoke evacuation settings, and surgery was mainly performed using electrocautery. Ultrasonic incision devices and vessel sealers were used for larger vessels. Intraoperative TEE was used to observe the right ventricular system in a four-chamber view at the following time points: (1) before and after induction of anesthesia,(2) immediately after the start of insufflation,(3) immediately after starting the ta/tp approach,(4) when the insufflation pressure of the ta/tp approach was increased,(5) when EtCO2 decreased by > 3 mmHg within 5 min,(6) when significant bleeding occurred, and (7) at the end of insufflation. Blood gas tests were also performed at each of these time points. Changes in end-tidal carbon dioxide (EtCO2), oxygen saturation (SpO2), and blood pressure over time were taken from the anesthesia records. The grade of CDE was assessed by (1) presence of bubbles in the right atrium and (2) changes in EtCO2 and vital signs, as reported by Feigl et al. (Table 1)17.

Table 1 Tübingen Venous Air Embolism Grading Scale.

The study was approved by the Fukuoka University Medical Ethics Committee (approval number U19-10-006). Informed consent was obtained from all patients who participated in the study. The study was conducted in accordance with the Declaration of Helsinki and followed the recommendations of the CONSORT Statement.


Twenty-one patients were enrolled between November 2018 and June 2021. The patient background data are shown in Table 2. The indications for surgery were rectal cancer in 19 cases, chronicgranulomatosis in 1, and local recurrence of cervical cancer in 1. The transanal approach was used in 4 cases and the transperineal approach in 17. The surgical procedures were low anterior resection in 2 cases, intersphincteric resection in 2, abdominoperineal resection in 13, and total pelvic exenteration in 4. The median maximum insufflation pressure when the ta/tp approach was used was 13.512,13,14,15,16,17,18 mmHg.

Table 2 Background patient characteristics.

The results of Intraoperative TEE monitoring are shown in Table 3. One case (4.8%) was observed to have a bubble in the right atrium; in this case, there was no change in blood pressure, SpO2, or pulse rate but there was a decrease in EtCO2 from 39 to 35 mmHg, and the patient was classified as grade 3. When the anesthesiologist confirmed the bubble by TEE (Supplemental Movie), the insufflation pressure was decreased from 15 to 12 mmHg to 10 mmHg until the surgeon identified the bleeding site, which was found to be a branch of the internal pudendal vein. Immediately after the bleeding was stopped, TEE confirmed that the bubble had disappeared. EtCO2 returned to its original value in 15 min and remained stable thereafter (Fig.1). There were no findings of right heart loading, even in the case where a bubble was observed in the right ventricular system by TEE. The surgery was continued using the two-team approach with TEE monitoring and was completed with no further complications.

Table 3 Operative parameters.
Figure 1
figure 1

Intraoperative hemodynamic monitoring; TEE was performed immediately after a decrease in EtCO2 (arrow).

No bubble was found in the right atrium on TEE in any other cases. Among all patients, the range of variation in intraoperative EtCO2 was 5–22 mmHg (median 16.5). A decrease in intraoperative EtCO2 by > 3 mmHg within 5 min was observed in 90.5% (19/21) of patients (median 5 mmHg in 1–10 times) but TEE revealed a bubble in the atrium in only 1 case.


The incidence of CDE when performing anorectal surgery via the ta/tp approach has been increasingly reported in recent years8. At our institution, symptomatic CDE has occurred in 2 (1.6%) of 120 cases, which is similar to the 3 (4%) of 80 cases reported by Harnsberger et al9. On the other hand, occult (asymptomatic) CDE was observed in 1 (4.8%) of our 21 cases in this study, which was lower than the rate at which TEE detected asymptomatic CDE in various organs11,12,13.

The incidence of CDE is reported to be higher with the ta/tp approach than with routine laparoscopic surgery. Possible reasons for this are as follows: (1) ta/tp requires a high insufflation pressure to secure the operative field18; (2) the veins around the perineum are relatively large and flow into the systemic circulation; and 3) the high pelvic position results in low central venous pressure, which promotes uptake of CO2 easily when a small hole is formed.

A sudden decrease in SpO2 or blood pressure accompanied by a rapid decrease in EtCO2 should raise suspicion for CO2 embolism, which is diagnosed using TEE or transthoracic echocardiography19,20. However, the decrease in EtCO2 can fluctuate depending on various factors, such as ventilator settings. In our study, a decrease in EtCO2 by > 3 mmHg within 5 min, which is an indicator used to grade CDE, was observed a total of 77 times in 19 cases (1–10 times) without any sign of bubbles on TEE. Therefore, comprehensive judgment is necessary rather than relying on only changes in EtCO2 as an indicator of CDE. When performing operations in areas of the perineum with a well-developed venous plexus, it is important to be vigilant for CDE and changes in EtCO2, SpO2, or vital signs. A visible venous lumen in the absence of bleeding means that the insufflation pressure is higher than the central venous pressure, which increases the risk of CDE. If respiratory and circulatory changes are detected, CDE should be suspected and TEE used for a definitive diagnosis while promptly discontinuing delivery of air.

We have previously had a case of CDE during tp-TPE in which the patient's condition worsened due to delayed diagnosis of CDE and conversion to laparotomy was required21. Early detection and management of CDE is important to prevent a deteriorating situation. Early diagnosis of CDE and lowering of the insufflation pressure in the patient with asymptomatic CDE in our present study resulted in prompt disappearance of the bubble without serious complications, and surgery could be continued more safely with monitoring by TEE.

This study has several limitations. First, it included only 21 cases. Second, TEE was not performed continuously, so the possibility that some bubbles were missed cannot be excluded. Lack of continuous monitoring might be the reason for wide variation in the detection rate of bubbles across studies (17.1–100%)11,12,13. Finally, although TEE monitoring may contribute to early detection of CDE, it is not realistic to use this method routinely in the clinical setting because of the cost, equipment, and personnel required, and the need for prolonged placement in the esophagus. However, our experience is that if CDE is suspected, it can be diagnosed immediately by TEE, and if vital signs are stable after insufflation is stopped, surgery can be continued more safely with ta/tp by monitoring with TEE. Both surgeons and anesthesiologists should keep CDE in mind and collaborate closely to detect and manage this complication19,20.


In this study, TEE monitoring in patients undergoing anorectal surgery via a ta/tp approach detected asymptomatic CDE in 1 of 21 cases (4.8%). Clinicians should be aware of this rare but potentially lethal complication in view of the ever-increasing popularity of the ta/tp approach.