- Email: [email protected]
How Safe Is the Intravasation Limit in Hysteroscopic Surgery?
Original Article
How Safe Is the Intravasation Limit in Hysteroscopic Surgery? B. M. P. Rademaker, MD, PhD*, P. J. M. van Kesteren, MD, PhD, P. de Haan, MD, PhD, D. Rademaker, and C. France From the Department of Anesthesiology (Drs. B. Rademaker and de Haan) and Department of Obstetrics and Gynaecology (Dr. van Kesteren), Onze Lieve Vrouwe Gasthuis, Vrije Universiteit Amsterdam (Ms. Rademaker), and the Universiteit van Amsterdam (Mr. France), Amsterdam, The Netherlands.
ABSTRACT Background: Transcervical resection of myomas (TCR-M) is considered a safe hysteroscopic procedure if intravasation is limited. Complications may occur if gas formation during myoma resection leads to gaseous embolism. However, the incidence of emboli during transcervical myoma resection is unknown. Therefore in this study the occurrence of physiological changes that indicate the formation of emboli was retrospectively determined in patients undergoing hysteroscopic myoma resection. In addition, these changes were related to the amount of fluid intravasation. Methods: The anesthesia records and operation files of 234 patients were screened for physiological changes that indicate embolism, as measured with standard intraoperative monitoring. These patients underwent surgery for intrauterine myomas with either a monopolar resectoscope with electrolyte-free distension fluid containing 3% sorbitol (limited to 1500-mL intravasation) or a bipolar resectoscope with normal saline solution (limited to 2500-mL intravasation). The patients were grouped according to the amount of fluid intravasation during the operation: Group 1: 500 mL or less, group 2: 500–1000 mL, group 3: 1000–1500 mL, and group 4: 1500–2500 mL. Results: Physiological changes that could be attributed to gaseous embolism were observed in 33% to 43% of patients with 1000 to 2500 mL fluid intravasation during transcervical myoma resection. Nearly half of those patients had cardiovascular disturbances that indicated the formation of emboli. Conclusion: During transcervical resection of myomas, physiological changes that could be attributed to gaseous embolism frequently occurred. Therefore cardiovascular disturbances that indicate gaseous embolism during transcervical resection of myomas may occur despite the limitation of intravasation according to current view. Journal of Minimally Invasive Gynecology (2011) 18, 355–361 Ó 2011 AAGL. All rights reserved. Keywords:
Hysteroscopy; Myoma; Embolism; Gas; Intravasation; Pathophysiology; Diathermia; St-segment changes; Hemoglobin oxygen saturation; Capnography
The hysteroscopic resection of submucosal myomas (TCR-M) is considered to be a low-risk procedure [1], but life-threatening complications may occur [2]. Complications may be the result of dilutional hyponatremia if electrolytefree distension fluid is used during monopolar TCR-M [3]. Severe hyponatremia can lead to complications such as encephalopathy and death [3–5]. To prevent the adverse effects of dilutional hyponatremia, it is advised to limit the intravasation of electrolyte free distension fluid to a maximum of 1500 mL [6]. If bipolar diathermia is used durThe authors have no commercial, proprietary, or financial interest in the products or companies described in this article. Corresponding author: B.M.P. Rademaker, MD, PhD, Department of Anesthesiology, Onze Lieve Vrouwe Gasthuis, Oosterpark 9, 1091 AC, Amsterdam, The Netherlands. E-mail: [email protected] Submitted September 15, 2010. Accepted for publication January 6, 2011. Available at www.sciencedirect.com and www.jmig.org 1553-4650/$ - see front matter Ó 2011 AAGL. All rights reserved. doi:10.1016/j.jmig.2011.01.010
ing TCR-M, normal saline solution is used for distension. Consequently, dilutional hyponatremia can be avoided, and up to 2500 mL of intravasation is considered to be safe [7]. Another complication of TCR-M is the occurrence of gaseous emboli [8]. We recently described severe clinical symptoms of gaseous embolism during TCR-M [9]. Moreover, we observed substantial cardiovascular disturbances that were believed to be associated with embolism in some of our patients. An example of the physiological changes that were seen in a patient is presented in Fig. 1. To investigate the frequency of physiological changes that could indicate gaseous embolism, patients scheduled for hysteroscopic myoma resection were retrospectively studied. It is likely that intrauterine gaseous embolism is correlated with the uptake of distension fluid. Therefore the incidence of the hemodynamic disturbances that might indicate gaseous embolism was related to the amount of distension fluid that was absorbed.
356
Fig. 1 This figure shows the anesthetic record and simultaneously recorded echocardiographic picture of a healthy 34-year-old patient who underwent a 45-minute procedure of bipolar transcervical myomectomy (TCR-M). Normal saline solution 1500 mL eventually intravasated during this procedure. The anesthesia record (upper graph) demonstrates the typical pattern of the physiological changes that indicate gaseous emboli: a decrease in etCO2, cardiovascular disturbances, ST-II segment depression, and SpO2 lowering. The thick red arrow indicates the start of TCR-m and the fine black arrow depicts the start of cardiovascular disturbances paralleled by the occurrence of gaseous emboli. (NIBDs 5 noninvasively measured blood pressure systolic; NIBDd 5 noninvasively measured blood pressure diastolic; HF-Puls 5 heart rate; ETCO2 5 etCO2; SpO2 5 peripheral oxygen saturation; ST-II 5 STII segment analysis). The lower picture shows a picture of a midesophageal echocardiography four-chamber view recording, demonstrating (solid red arrows) massive right-side atrial and ventricular embolization of gas and or debris during trans-cervical myomectomy.
Methods Anesthesia records and operation files of 234 patients scheduled for hysteroscopic myoma resection in the period between 2007 and March 2009 were studied. The data management system (MetaVision Anesthesia Information
Journal of Minimally Invasive Gynecology, Vol 18, No 3, May/June 2011
Management System; iMDsoft, Needham, MA) collected all hemodynamic (including ST-segment changes) ventilatory and respiratory data during surgery with a sample frequency of one per minute. Data were stored in the hospital’s central MSQL server. Patients’ data were analyzed offline by two trained independent observers. During anesthesia, standard monitoring (Philips IntelliVue MP90; Koninklijke Philips Electronics, Amsterdam, the Netherlands) was used: Three lead electrocardiography with ST-II segment analysis (STII), heart rate (HR), peripheral hemoglobin oxygen saturation (SpO2), noninvasively measured systolic arterial blood pressure (NIBDs), and diastolic arterial blood pressure (NIBDd) and end tidal CO2 (etCO2). All patients received general anesthesia with a propofol opioid-based intravenous technique. Either a laryngeal mask or an endotracheal tube was used to secure the airway. Patients were ventilated with oxygen in air (FiO2 .0.35), and normocapnia was maintained. Either monopolar or bipolar electrosurgery was used for trans-cervical myomectomy (monopolar or bipolar resectoscope from Olympus, Hamburg, Germany, connected to a VIO 300D generator, Erbe Electromedizin GmbH, T€ubingen, Germany). Normal saline solution was used as distention fluid during bipolar electrosurgery and electrolytefree solution containing 3% sorbitol was used as distention fluid during monopolar electrosurgery. Fluid management data were obtained with an Olympus Hysteromat, Hamburg, Germany, and inflow pressure was set at 80 mm Hg. Active suction was acquired by connecting the outflow channel to the standard vacuum available in the OR. Intravasation was calculated by subtracting the amount of fluid introduced via the resectoscope and the amount of fluid collected by fluid suction. The total amount of intravasated fluid was recorded in the patient’s personal operation file. Patients were divided into four groups according to the amount of fluid that was absorbed during the procedure: Group 1: ,500 mL; Group 2: 501–1000 mL; Group 3: 1001–1500 mL; Group 4: 1501–2500 mL. The anesthesia records were carefully screened for changes in intraoperative hemodynamic parameters that are suggestive of embolism [10]: a drop in etCO2 of 2 mmHg during the transcervical myomectomy, especially when accompanied by a decrease in blood pressure; signs of cardiovascular disturbance as measured by a drop in blood pressure (NIBDs and NIBDd decrease of .20%), not explained by hypovolemia, with or without extreme bradycardia or tachycardia; a 2% (or larger) decrease in SpO2 saturation; newly developed ST-II segment changes, especially in the inferior leads; postoperative decreases in SpO2-saturation. On the basis of these hemodynamic disturbances, the patients could be divided in three groups: (1) No apparent physiological changes that indicate the formation of emboli; (2) Physiological changes that might indicate the formation of emboli: (A) A decrease in etCO2 and a decrease in hemoglobin oxygen saturation without cardiovascular disturbances, or (B) a drop in etCO2 accompanied by newly developed ST-II segment changes without cardiovascular
Rademaker et al.
How Safe Is the Intravasation Limit in Hysteroscopic Surgery?
disturbances, (C) a period of postoperative hemoglobin oxygen desaturations; (D) cardiovascular disturbances during and after surgery. (3) Physiological changes that clearly indicate the formation of emboli: (A) A drop in etCO2 and cardiovascular disturbances; (B) a drop in etCO2 and a decrease in hemoglobin oxygen saturation accompanied with newly developed ST-II segment changes. Data Management Data are expressed as mean (95% CI) 6 STD for normally distributed continuous variables, median (range) for non-normally distributed variables. The proportion of patients with physiological changes that indicate embolism are presented as percentages. Data were analyzed by use of two-way analysis of variance, multivariate analysis and univariate analysis, c2 and t tests when appropriate. The p values ,.05 were considered significant. Results There were no differences among the four groups with respect to patient characteristics or obstetric history (Table 1). In group I significantly more patients underwent surgery for myoma classification 0 (p ,.05), and the opera-
357
tion time was significantly shorter (p ,.005). There was no difference in operation time between group II and groups III and IV. Significantly more patients underwent surgery by use of bipolar diathermia in groups III and IV. An example of the typical pattern of physiological changes suggestive for a substantial embolism in a patient is presented in Fig. 1. Physiological changes and cardiovascular variables that indicate the formation of emboli are presented as box plots in Fig. 2. Circulatory collapse requiring advanced life support was reported in one patient from group IV. No special care was needed in any other patient with apparent gas embolism. In this group of patients an expectative policy was used, with basic treatment and supportive measurements. Significant desaturation was observed in groups III and IV (p ,.0001). Groups III and IV differed significantly in this respect (p ,.0001) from groups I and II. In addition, the duration of desaturation lasted significantly (p ,.0001) longer in group III. There was a statistically significant difference observed with respect to the decrease of etCO2. In addition, the duration of etCO2 decreases lasted significantly longer in groups III and IV than in group I and II. Interindividual variation in the in etCO2 decreases during transcervical surgery were large, especially in groups III and IV. ST-segment analysis
Table 1 Data are mean (95% CI) 6 STD
Number of patients Age (years) Weight (kg) Length (cm) ASA (1/2/3) Para status 0 1 2 3 or more Myoma class 0 1 2 01112 Not specified Diathemia monopalar/bipolar Operation duration (min)
Intravasation Group I 0–500 mL
Group II 500–1000 mL
Group III 1000–1500 mL
Group IV 1500–2500 mL
78 45.9 (44.2–47.6) 6 7.58 70.4 (67.1–73.6) 6 14.36 168.3 (166.4–170.1) 6 8.2 57/20/0
33 46.3 (44.5–48.1) 6 5.07 70.1(65.6–74.5) 6 12.50 167.1 (164.4–169.9) 6 7.64 23/9/0
55 44.7 (42.4–46.9) 6 8.32 73.2 (68.4–77.9) 6 17.49 167.2 (165.1–169.3) 6 7.83 37/14/2
68 44.4 (42.8–46.1) 6 6.76 69.2 (66.5–71.9) 6 11.34 167.0 (165.1–168.8) 6 7.5 53/15/0
48 8 10 12
15 10 5 3
23 19 9 4
24 11 20 13
34* 26 8 10
7 14 8 4
14 10 22 9
23/55 28.3 (24.5–32.5) 6 17
7/26 36.3 (29.2–43.4) 6 20.1y
8/47y z 37.9 (32.6–41.4) 6 15.9y
17 20 18 11 2 0/68y x 37.8 (33.8–43.6) 6 20.4y
ASA 5 American Society of Anesthesiologists’ physical status classification. * Significantly different compared to group 3 and 4 (p ,.05). Significantly different from group 1 (p ,.005). z Significantly different compared with group 1 (p ,.001). x Significantly different compared with group 2 (p ,.005). y
358
Journal of Minimally Invasive Gynecology, Vol 18, No 3, May/June 2011
Fig. 2 Physiological changes and cardiovascular variables during transcervical myomectomy (TCR-m) and lowest value of postoperative hemoglobin oxygen saturation (SaO2). Data are presented as box-and-whisker plots. The box covers the interquartile range with the median indicated by the line within the box. The whiskers extend to the 10th and the 90th percentile values. Outliers are plotted individually. # Significantly different from groups I and II (p ,.05), ## significantly different from group I (p ,.0001).
revealed significantly more ST-segment changes in groups II and IV. After surgery, no difference was observed in the incidence of HbO2-saturations among the four groups. The proportion of patients having experienced either apparently no physiological changes that indicate the formation of emboli, physiological changes that might indicate the
formation of emboli, or physiological changes that clearly indicate the formation of emboli is presented in the pie diagrams of Fig. 3. Groups III and IV showed a significantly higher (p ,.0001) percentage of patients with physiological changes that can be attributed to emboli. The use of bipolar diathermia as compared with monopolar diathermia was
Rademaker et al.
How Safe Is the Intravasation Limit in Hysteroscopic Surgery?
Fig. 3 The percentage of patients in each group having experienced either: , Apparently no physiological changes that indicate the formation of emboli, Physiological changes that might indicate the formation of emboli, - Physiological changes that clearly indicate the formation of emboli. Groups III and IV showed a significantly (P , 0.0001) higher ratio of patients with physiological changes and cardiovascular disturbances that indicate the occurrence of embolisation.
not an independent risk factor for the development of cardiovascular disturbances that might indicate gaseous embolism (p 5 .124). Discussion In this study, physiological changes and cardiovascular disturbances, which could be attributed to gaseous embolism, frequently occurred during transcervical resection of myomas. In patients with intravasation between 1500 and 2500 mL, physiological changes that might imply a possible embolism appeared in 43%. Half of those patients had cardiovascular disturbances that clearly indicated the formation of emboli. Surprisingly, physiological changes that might indicate the formation of emboli could also be demonstrated in 33% of the group of patients who had intravasation between 1000 and 1500 mL, and 45% of these patients showed hemodynamic disturbances that clearly implied embolism. The formation of gaseous emboli during hysteroscopic surgery is observed frequently. In a recent study, gas bubbles were seen in the right atrium by transthoracic echocardiography in all patients [11], and a continuous flow of bubbles was observed in 20 of 23 patients. These bubbles consist of gases such as CO, CO2, and hydrogen [12–14], and, although soluble, they are apparently not dissolved before they reach the heart. The two fundamental factors determining the morbidity and mortality rates of gas embolism are directly related to the volume of gas entrainment and rate of accumulation [10]. One of our patients with apparent gas embolism received advanced life support. We cannot comment
359
on the percentage of patients with apparent gas embolism, who will need some kind of special care or additional support. In this retrospective study this percentage could not be determined, because of the fact that in the daily anesthesia practice correction of physiological derangements with minor hemodynamic or ventilatory interventions, including longer postoperative observation, are not systematically labelled as procedure related complications. The cut-off point between the occurrence of major catastrophic episode and subtle but unequivocally important symptoms is yet undefined. Our knowledge of the consequences of small-quantity gas emboli is imperfect [15]. Moreover, it is not clear why some patients show distinct clinical signs at the same level of intravasation while others do not. Further insight into the pathophysiology of the occurrence of gaseous emboli and subsequent physiological changes may identify patients at risk. It is well known that when a relatively large volume of gas enters the circulation rapidly, cardiovascular, pulmonary, and physiological changes occur [10]. We opted for the following changes as indications of a possible gaseous embolism: A drop in etCO2 of 2 mm Hg, a drop in blood pressure of more than 20%, decrease in SpO2 saturation R2%, and newly developed ST-II segment changes, especially in the inferior leads [6,10,16]. Of all clinical signs, a drop of 2 mm Hg etCO2 or more is considered to be the most important sign of intraoperative embolism, especially when accompanied by a decrease in blood pressure. In this study, a statistically significant difference with respect to the magnitude of the etCO2 decrease was observed, and the duration of the observed etCO2 decreases was significantly increased in groups III and IV. In contrast to etCO2 changes, hemoglobin oxygen desaturation is considered a relative late and nonspecific sign. Gas embolism can effect hemoglobin oxygen saturation by increasing ventilation/perfusion mismatch [10]. Despite the fact that our patients were mechanically ventilated with relatively high inspiratory admixtures of oxygen (FiO2 0.35–0.5), in groups III and IV SpO2 desaturations were observed during the procedure. We considered these decreases in SpO2 saturation most likely to be the result of gaseous emboli, particularly if the desaturation was accompanied by cardiovascular disturbances. Although alterations in the electrocardiogram rank low in sensitivity, gaseous embolism can result in electrocardiographic changes. ST-II segment depression occurs when right ventricular strain develops as a result of right ventricular load increase [11]. In addition, if gas emboli develop into paradoxical emboli, either through an open foramen ovale or through patent pulmonary pathways, they may disturb right coronary artery blood flow and also result in electrocardiography changes [9]. Indeed, a large percentage of our patients showed a combination of ST-II segment depression with the aforementioned physiological changes. Although the abovementioned changes may appear nonspecific, in our opinion the most likely explanation for them is the occurrence of gaseous emboli.
360
The cardiovascular disturbances that imply the formation of gaseous emboli became more obvious with an increased quantity of intravasation. This can be explained by distension fluid and gaseous products of electrosurgery entering the circulation through surgically induced open veins in the myometrium. An increase of intravasation leads to a concomitant increase of gaseous emboli into the circulation. Consequently, the occurrence of related cardiovascular disturbances will increase. It can be postulated that operation time correlates with apparent gas entry. Indeed, the operation time was shortest in group I, the group with the least fluid intravasation. However, although operation times were similar in groups II, III, and IV, significantly fewer cardiovascular disturbances were observed in group II as compared with groups III and IV. Therefore it seems logical to assume that operation time is not a decisive factor in the pathophysiology of the cardiovascular disturbances. It is not likely that the observed physiological changes are the result of volume overload, because volume loading is well tolerated in healthy individuals. Overhydration may be a plausible explanation for hemoglobin oxygen desaturation in the cardiovascular-compromised patient. However, this is improbable in ventilated, relatively healthy patients. Another explanation for the physiological changes is that they result from room air embolism during the process of removing and reinserting the scope, which then acts as a piston in the vaginal cylinder to force air in the open uterine veins. Corson et al [17] reported two patients (out of a series of five catastrophic hysteroscopic events) who died, apparently because room air entered the circulation through surgically exposed blood vessels. Indeed, during a recent TCRM, monitored with transesophageal echocardiography, we repeatedly observed emboli during introduction of the hysteroscope, whereas there was no electrosurgical activation of the instrument, and the resection had not yet started. Although it is clear that emboli develop during the use of diathermia [11], it is important to note that the occurrence of emboli may also be related to the introduction of the resectoscope. Experimental studies have shown that monopolar and bipolar diathermy may produce slightly different amounts of gas and insoluble particles of different sizes, but the composition of the gas appears to be the same [12–14]. A difference between monopolar and bipolar TCR-M could not be observed. However, in groups III and IV, only eight of 123 patients had monopolar energy, which is insufficient to draw conclusions. This study has several important limitations, the first of which is the retrospective design. During hysteroscopic resection of myomas, we observed cardiovascular disturbances in several of our patients. These changes were believed to be the result of gaseous emboli. Therefore a retrospective study was designed to substantiate the frequency of physiological changes and cardiovascular disturbances
Journal of Minimally Invasive Gynecology, Vol 18, No 3, May/June 2011
that are believed to be the result of gaseous emboli. Prospective studies with precordial Doppler scanning, transesophageal ultrasound echocardiography, or the newly developed multifrequency transcranial Doppler scanning (capable of differentiating gaseous from solid emboli) [18] are definitely needed to confirm our observations. Second, there is no actual proof of gaseous emboli and the physiological changes are nonspecific. Indeed, we did not use precordial Doppler scanning or visualize the intravascular emboli with transesophageal echocardiography. However, hysteroscopic resection of myomas is a circumscribed procedure in a healthy population. Therefore, if general anesthesia is provided, this should result in an uncomplicated anesthesia record. If cardiovascular disturbances are seen, they probably result from the procedure itself. Intravasation leading to volume loading is well tolerated in a healthy person, and therefore these physiological disturbances are most likely to result from gaseous emboli. Conclusion Our study suggests that an intravasation of 1000 mL or more in patients undergoing hysteroscopic myoma resection may result in cardiovascular disturbances that indicate the occurrence of gaseous emboli, independent of the type of diathermia or distension fluid used. Although this may be well tolerated by healthy young patients, it is potentially hazardous in an elderly, cardiopulmonary compromised patient. References 1. Brandner P, Neis KJ, Diebold P. Hysteroscopic resection of submucous myomas. Contrib Gynecol Obstet. 2000;20:81–90. 2. Cooper JM, Brady RM. Intraoperative and early postoperative complications of operative hysteroscopy. Obstet Gynecol Clin North Am. 2000;27:347–366. 3. Kim AH, Keltz MD, Arici A, Rosenberg M, Olive DL. Dilutional hyponatremia during hysteroscopic myomectomy with sorbitolmannitol distention medium. J Am Assoc Gynecol Laparosc. 1995;2: 237–242. 4. Arieff AI, Ayus JC. Endometrial ablation complicated by fatal hyponatremic encephalopathy. JAMA. 1993;270:1230–1232. 5. Istre O, Bjoennes J, Naess R, Hornbaek K, Forman A. Postoperative cerebral oedema after transcervical endometrial resection and uterine irrigation with 1.5% glycine. Lancet. 1994;344:1187–1189. 6. Groenman FA, Peters LW, Rademaker BM, Bakkum EA. Embolism of air and gas in hysteroscopic procedures: pathophysiology and implication for daily practice. J Minim Invasive Gynecol. 2008;15: 241–247. 7. Loffer FD, Bradley LD, Brill AI, Brooks PG, Cooper JM. Hysteroscopic fluid monitoring guidelines. The ad hoc committee on hysteroscopic training guidelines of the American Association of Gynecologic Laparoscopists. J Am Assoc Gynecol Laparosc. 2000;7: 167–168. 8. Bloomstone J, Chow CM, Isselbacher E, VanCott E, Isaacson KB. A pilot study examining the frequency and quantity of gas embolization during operative hysteroscopy using a monopolar resectoscope. J Am Assoc Gynecol Laparosc. 2002;9:9–14. 9. Rademaker BM, Groenman FA, van der Wouw PA, Bakkum EA. Paradoxical gas embolism by transpulmonary passage of venous emboli
Rademaker et al.
10. 11.
12.
13.
How Safe Is the Intravasation Limit in Hysteroscopic Surgery?
during hysteroscopic surgery: a case report and discussion. Br J Anaesth. 2008;101:230–233. Mirski MA, Lele AV, Fitzsimmons L, Toung TJ. Diagnosis and treatment of vascular air embolism. Anesthesiology. 2007;106:164–177. Leibowitz D, Benshalom N, Kaganov Y, Rott D, Hurwitz A, Hamani Y. The incidence and haemodynamic significance of gas emboli during operative hysteroscopy: a prospective echocardiographic study. Eur J Echocardiogr. 2010;11:429–431. Munro MG, Brill AI, Ryan T, Ciarrocca S. Electrosurgery-induced generation of gases: comparison of in vitro rates of production using bipolar and monopolar electrodes. J Am Assoc Gynecol Laparosc. 2003;10: 252–259. Munro MG, Weisberg M, Rubinstein E. Gas and air embolization during hysteroscopic electrosurgical vaporization: comparison of gas
14.
15. 16. 17. 18.
361 generation using bipolar and monopolar electrodes in an experimental model. J Am Assoc Gynecol Laparosc. 2001;8:488–494. Farrugia M, Hussain SY, Perrett D. Particulate matter generated during monopolar and bipolar hysteroscopic human uterine tissue vaporization. J Minim Invasive Gynecol. 2009;16:458–464. Barak M, Katz Y. Microbubbles: pathophysiology and clinical implications. Chest. 2005;128:2918–2932. Stoloff DR, Isenberg RA, Brill AI. Venous air and gas emboli in operative hysteroscopy. J Am Assoc Gynecol Laparosc. 2001;8:181–192. Corson SL, Brooks PG, Soderstrom RM. Gynecologic endoscopic gas embolism. Fertil Steril. 1996;65:529–533. Chen CI, Iguchi Y, Garami Z, Malkoff MD, Smalling RW, Campbell MS, et al. Analysis of emboli during carotid stenting with distal protection device. Cerebrovasc Dis. 2006;21:223–228.
How Safe Is the Intravasation Limit in Hysteroscopic Surgery? B. M. P. Rademaker, MD, PhD*, P. J. M. van Kesteren, MD, PhD, P. de Haan, MD, PhD, D. Rademaker, and C. France From the Department of Anesthesiology (Drs. B. Rademaker and de Haan) and Department of Obstetrics and Gynaecology (Dr. van Kesteren), Onze Lieve Vrouwe Gasthuis, Vrije Universiteit Amsterdam (Ms. Rademaker), and the Universiteit van Amsterdam (Mr. France), Amsterdam, The Netherlands.
ABSTRACT Background: Transcervical resection of myomas (TCR-M) is considered a safe hysteroscopic procedure if intravasation is limited. Complications may occur if gas formation during myoma resection leads to gaseous embolism. However, the incidence of emboli during transcervical myoma resection is unknown. Therefore in this study the occurrence of physiological changes that indicate the formation of emboli was retrospectively determined in patients undergoing hysteroscopic myoma resection. In addition, these changes were related to the amount of fluid intravasation. Methods: The anesthesia records and operation files of 234 patients were screened for physiological changes that indicate embolism, as measured with standard intraoperative monitoring. These patients underwent surgery for intrauterine myomas with either a monopolar resectoscope with electrolyte-free distension fluid containing 3% sorbitol (limited to 1500-mL intravasation) or a bipolar resectoscope with normal saline solution (limited to 2500-mL intravasation). The patients were grouped according to the amount of fluid intravasation during the operation: Group 1: 500 mL or less, group 2: 500–1000 mL, group 3: 1000–1500 mL, and group 4: 1500–2500 mL. Results: Physiological changes that could be attributed to gaseous embolism were observed in 33% to 43% of patients with 1000 to 2500 mL fluid intravasation during transcervical myoma resection. Nearly half of those patients had cardiovascular disturbances that indicated the formation of emboli. Conclusion: During transcervical resection of myomas, physiological changes that could be attributed to gaseous embolism frequently occurred. Therefore cardiovascular disturbances that indicate gaseous embolism during transcervical resection of myomas may occur despite the limitation of intravasation according to current view. Journal of Minimally Invasive Gynecology (2011) 18, 355–361 Ó 2011 AAGL. All rights reserved. Keywords:
Hysteroscopy; Myoma; Embolism; Gas; Intravasation; Pathophysiology; Diathermia; St-segment changes; Hemoglobin oxygen saturation; Capnography
The hysteroscopic resection of submucosal myomas (TCR-M) is considered to be a low-risk procedure [1], but life-threatening complications may occur [2]. Complications may be the result of dilutional hyponatremia if electrolytefree distension fluid is used during monopolar TCR-M [3]. Severe hyponatremia can lead to complications such as encephalopathy and death [3–5]. To prevent the adverse effects of dilutional hyponatremia, it is advised to limit the intravasation of electrolyte free distension fluid to a maximum of 1500 mL [6]. If bipolar diathermia is used durThe authors have no commercial, proprietary, or financial interest in the products or companies described in this article. Corresponding author: B.M.P. Rademaker, MD, PhD, Department of Anesthesiology, Onze Lieve Vrouwe Gasthuis, Oosterpark 9, 1091 AC, Amsterdam, The Netherlands. E-mail: [email protected] Submitted September 15, 2010. Accepted for publication January 6, 2011. Available at www.sciencedirect.com and www.jmig.org 1553-4650/$ - see front matter Ó 2011 AAGL. All rights reserved. doi:10.1016/j.jmig.2011.01.010
ing TCR-M, normal saline solution is used for distension. Consequently, dilutional hyponatremia can be avoided, and up to 2500 mL of intravasation is considered to be safe [7]. Another complication of TCR-M is the occurrence of gaseous emboli [8]. We recently described severe clinical symptoms of gaseous embolism during TCR-M [9]. Moreover, we observed substantial cardiovascular disturbances that were believed to be associated with embolism in some of our patients. An example of the physiological changes that were seen in a patient is presented in Fig. 1. To investigate the frequency of physiological changes that could indicate gaseous embolism, patients scheduled for hysteroscopic myoma resection were retrospectively studied. It is likely that intrauterine gaseous embolism is correlated with the uptake of distension fluid. Therefore the incidence of the hemodynamic disturbances that might indicate gaseous embolism was related to the amount of distension fluid that was absorbed.
356
Fig. 1 This figure shows the anesthetic record and simultaneously recorded echocardiographic picture of a healthy 34-year-old patient who underwent a 45-minute procedure of bipolar transcervical myomectomy (TCR-M). Normal saline solution 1500 mL eventually intravasated during this procedure. The anesthesia record (upper graph) demonstrates the typical pattern of the physiological changes that indicate gaseous emboli: a decrease in etCO2, cardiovascular disturbances, ST-II segment depression, and SpO2 lowering. The thick red arrow indicates the start of TCR-m and the fine black arrow depicts the start of cardiovascular disturbances paralleled by the occurrence of gaseous emboli. (NIBDs 5 noninvasively measured blood pressure systolic; NIBDd 5 noninvasively measured blood pressure diastolic; HF-Puls 5 heart rate; ETCO2 5 etCO2; SpO2 5 peripheral oxygen saturation; ST-II 5 STII segment analysis). The lower picture shows a picture of a midesophageal echocardiography four-chamber view recording, demonstrating (solid red arrows) massive right-side atrial and ventricular embolization of gas and or debris during trans-cervical myomectomy.
Methods Anesthesia records and operation files of 234 patients scheduled for hysteroscopic myoma resection in the period between 2007 and March 2009 were studied. The data management system (MetaVision Anesthesia Information
Journal of Minimally Invasive Gynecology, Vol 18, No 3, May/June 2011
Management System; iMDsoft, Needham, MA) collected all hemodynamic (including ST-segment changes) ventilatory and respiratory data during surgery with a sample frequency of one per minute. Data were stored in the hospital’s central MSQL server. Patients’ data were analyzed offline by two trained independent observers. During anesthesia, standard monitoring (Philips IntelliVue MP90; Koninklijke Philips Electronics, Amsterdam, the Netherlands) was used: Three lead electrocardiography with ST-II segment analysis (STII), heart rate (HR), peripheral hemoglobin oxygen saturation (SpO2), noninvasively measured systolic arterial blood pressure (NIBDs), and diastolic arterial blood pressure (NIBDd) and end tidal CO2 (etCO2). All patients received general anesthesia with a propofol opioid-based intravenous technique. Either a laryngeal mask or an endotracheal tube was used to secure the airway. Patients were ventilated with oxygen in air (FiO2 .0.35), and normocapnia was maintained. Either monopolar or bipolar electrosurgery was used for trans-cervical myomectomy (monopolar or bipolar resectoscope from Olympus, Hamburg, Germany, connected to a VIO 300D generator, Erbe Electromedizin GmbH, T€ubingen, Germany). Normal saline solution was used as distention fluid during bipolar electrosurgery and electrolytefree solution containing 3% sorbitol was used as distention fluid during monopolar electrosurgery. Fluid management data were obtained with an Olympus Hysteromat, Hamburg, Germany, and inflow pressure was set at 80 mm Hg. Active suction was acquired by connecting the outflow channel to the standard vacuum available in the OR. Intravasation was calculated by subtracting the amount of fluid introduced via the resectoscope and the amount of fluid collected by fluid suction. The total amount of intravasated fluid was recorded in the patient’s personal operation file. Patients were divided into four groups according to the amount of fluid that was absorbed during the procedure: Group 1: ,500 mL; Group 2: 501–1000 mL; Group 3: 1001–1500 mL; Group 4: 1501–2500 mL. The anesthesia records were carefully screened for changes in intraoperative hemodynamic parameters that are suggestive of embolism [10]: a drop in etCO2 of 2 mmHg during the transcervical myomectomy, especially when accompanied by a decrease in blood pressure; signs of cardiovascular disturbance as measured by a drop in blood pressure (NIBDs and NIBDd decrease of .20%), not explained by hypovolemia, with or without extreme bradycardia or tachycardia; a 2% (or larger) decrease in SpO2 saturation; newly developed ST-II segment changes, especially in the inferior leads; postoperative decreases in SpO2-saturation. On the basis of these hemodynamic disturbances, the patients could be divided in three groups: (1) No apparent physiological changes that indicate the formation of emboli; (2) Physiological changes that might indicate the formation of emboli: (A) A decrease in etCO2 and a decrease in hemoglobin oxygen saturation without cardiovascular disturbances, or (B) a drop in etCO2 accompanied by newly developed ST-II segment changes without cardiovascular
Rademaker et al.
How Safe Is the Intravasation Limit in Hysteroscopic Surgery?
disturbances, (C) a period of postoperative hemoglobin oxygen desaturations; (D) cardiovascular disturbances during and after surgery. (3) Physiological changes that clearly indicate the formation of emboli: (A) A drop in etCO2 and cardiovascular disturbances; (B) a drop in etCO2 and a decrease in hemoglobin oxygen saturation accompanied with newly developed ST-II segment changes. Data Management Data are expressed as mean (95% CI) 6 STD for normally distributed continuous variables, median (range) for non-normally distributed variables. The proportion of patients with physiological changes that indicate embolism are presented as percentages. Data were analyzed by use of two-way analysis of variance, multivariate analysis and univariate analysis, c2 and t tests when appropriate. The p values ,.05 were considered significant. Results There were no differences among the four groups with respect to patient characteristics or obstetric history (Table 1). In group I significantly more patients underwent surgery for myoma classification 0 (p ,.05), and the opera-
357
tion time was significantly shorter (p ,.005). There was no difference in operation time between group II and groups III and IV. Significantly more patients underwent surgery by use of bipolar diathermia in groups III and IV. An example of the typical pattern of physiological changes suggestive for a substantial embolism in a patient is presented in Fig. 1. Physiological changes and cardiovascular variables that indicate the formation of emboli are presented as box plots in Fig. 2. Circulatory collapse requiring advanced life support was reported in one patient from group IV. No special care was needed in any other patient with apparent gas embolism. In this group of patients an expectative policy was used, with basic treatment and supportive measurements. Significant desaturation was observed in groups III and IV (p ,.0001). Groups III and IV differed significantly in this respect (p ,.0001) from groups I and II. In addition, the duration of desaturation lasted significantly (p ,.0001) longer in group III. There was a statistically significant difference observed with respect to the decrease of etCO2. In addition, the duration of etCO2 decreases lasted significantly longer in groups III and IV than in group I and II. Interindividual variation in the in etCO2 decreases during transcervical surgery were large, especially in groups III and IV. ST-segment analysis
Table 1 Data are mean (95% CI) 6 STD
Number of patients Age (years) Weight (kg) Length (cm) ASA (1/2/3) Para status 0 1 2 3 or more Myoma class 0 1 2 01112 Not specified Diathemia monopalar/bipolar Operation duration (min)
Intravasation Group I 0–500 mL
Group II 500–1000 mL
Group III 1000–1500 mL
Group IV 1500–2500 mL
78 45.9 (44.2–47.6) 6 7.58 70.4 (67.1–73.6) 6 14.36 168.3 (166.4–170.1) 6 8.2 57/20/0
33 46.3 (44.5–48.1) 6 5.07 70.1(65.6–74.5) 6 12.50 167.1 (164.4–169.9) 6 7.64 23/9/0
55 44.7 (42.4–46.9) 6 8.32 73.2 (68.4–77.9) 6 17.49 167.2 (165.1–169.3) 6 7.83 37/14/2
68 44.4 (42.8–46.1) 6 6.76 69.2 (66.5–71.9) 6 11.34 167.0 (165.1–168.8) 6 7.5 53/15/0
48 8 10 12
15 10 5 3
23 19 9 4
24 11 20 13
34* 26 8 10
7 14 8 4
14 10 22 9
23/55 28.3 (24.5–32.5) 6 17
7/26 36.3 (29.2–43.4) 6 20.1y
8/47y z 37.9 (32.6–41.4) 6 15.9y
17 20 18 11 2 0/68y x 37.8 (33.8–43.6) 6 20.4y
ASA 5 American Society of Anesthesiologists’ physical status classification. * Significantly different compared to group 3 and 4 (p ,.05). Significantly different from group 1 (p ,.005). z Significantly different compared with group 1 (p ,.001). x Significantly different compared with group 2 (p ,.005). y
358
Journal of Minimally Invasive Gynecology, Vol 18, No 3, May/June 2011
Fig. 2 Physiological changes and cardiovascular variables during transcervical myomectomy (TCR-m) and lowest value of postoperative hemoglobin oxygen saturation (SaO2). Data are presented as box-and-whisker plots. The box covers the interquartile range with the median indicated by the line within the box. The whiskers extend to the 10th and the 90th percentile values. Outliers are plotted individually. # Significantly different from groups I and II (p ,.05), ## significantly different from group I (p ,.0001).
revealed significantly more ST-segment changes in groups II and IV. After surgery, no difference was observed in the incidence of HbO2-saturations among the four groups. The proportion of patients having experienced either apparently no physiological changes that indicate the formation of emboli, physiological changes that might indicate the
formation of emboli, or physiological changes that clearly indicate the formation of emboli is presented in the pie diagrams of Fig. 3. Groups III and IV showed a significantly higher (p ,.0001) percentage of patients with physiological changes that can be attributed to emboli. The use of bipolar diathermia as compared with monopolar diathermia was
Rademaker et al.
How Safe Is the Intravasation Limit in Hysteroscopic Surgery?
Fig. 3 The percentage of patients in each group having experienced either: , Apparently no physiological changes that indicate the formation of emboli, Physiological changes that might indicate the formation of emboli, - Physiological changes that clearly indicate the formation of emboli. Groups III and IV showed a significantly (P , 0.0001) higher ratio of patients with physiological changes and cardiovascular disturbances that indicate the occurrence of embolisation.
not an independent risk factor for the development of cardiovascular disturbances that might indicate gaseous embolism (p 5 .124). Discussion In this study, physiological changes and cardiovascular disturbances, which could be attributed to gaseous embolism, frequently occurred during transcervical resection of myomas. In patients with intravasation between 1500 and 2500 mL, physiological changes that might imply a possible embolism appeared in 43%. Half of those patients had cardiovascular disturbances that clearly indicated the formation of emboli. Surprisingly, physiological changes that might indicate the formation of emboli could also be demonstrated in 33% of the group of patients who had intravasation between 1000 and 1500 mL, and 45% of these patients showed hemodynamic disturbances that clearly implied embolism. The formation of gaseous emboli during hysteroscopic surgery is observed frequently. In a recent study, gas bubbles were seen in the right atrium by transthoracic echocardiography in all patients [11], and a continuous flow of bubbles was observed in 20 of 23 patients. These bubbles consist of gases such as CO, CO2, and hydrogen [12–14], and, although soluble, they are apparently not dissolved before they reach the heart. The two fundamental factors determining the morbidity and mortality rates of gas embolism are directly related to the volume of gas entrainment and rate of accumulation [10]. One of our patients with apparent gas embolism received advanced life support. We cannot comment
359
on the percentage of patients with apparent gas embolism, who will need some kind of special care or additional support. In this retrospective study this percentage could not be determined, because of the fact that in the daily anesthesia practice correction of physiological derangements with minor hemodynamic or ventilatory interventions, including longer postoperative observation, are not systematically labelled as procedure related complications. The cut-off point between the occurrence of major catastrophic episode and subtle but unequivocally important symptoms is yet undefined. Our knowledge of the consequences of small-quantity gas emboli is imperfect [15]. Moreover, it is not clear why some patients show distinct clinical signs at the same level of intravasation while others do not. Further insight into the pathophysiology of the occurrence of gaseous emboli and subsequent physiological changes may identify patients at risk. It is well known that when a relatively large volume of gas enters the circulation rapidly, cardiovascular, pulmonary, and physiological changes occur [10]. We opted for the following changes as indications of a possible gaseous embolism: A drop in etCO2 of 2 mm Hg, a drop in blood pressure of more than 20%, decrease in SpO2 saturation R2%, and newly developed ST-II segment changes, especially in the inferior leads [6,10,16]. Of all clinical signs, a drop of 2 mm Hg etCO2 or more is considered to be the most important sign of intraoperative embolism, especially when accompanied by a decrease in blood pressure. In this study, a statistically significant difference with respect to the magnitude of the etCO2 decrease was observed, and the duration of the observed etCO2 decreases was significantly increased in groups III and IV. In contrast to etCO2 changes, hemoglobin oxygen desaturation is considered a relative late and nonspecific sign. Gas embolism can effect hemoglobin oxygen saturation by increasing ventilation/perfusion mismatch [10]. Despite the fact that our patients were mechanically ventilated with relatively high inspiratory admixtures of oxygen (FiO2 0.35–0.5), in groups III and IV SpO2 desaturations were observed during the procedure. We considered these decreases in SpO2 saturation most likely to be the result of gaseous emboli, particularly if the desaturation was accompanied by cardiovascular disturbances. Although alterations in the electrocardiogram rank low in sensitivity, gaseous embolism can result in electrocardiographic changes. ST-II segment depression occurs when right ventricular strain develops as a result of right ventricular load increase [11]. In addition, if gas emboli develop into paradoxical emboli, either through an open foramen ovale or through patent pulmonary pathways, they may disturb right coronary artery blood flow and also result in electrocardiography changes [9]. Indeed, a large percentage of our patients showed a combination of ST-II segment depression with the aforementioned physiological changes. Although the abovementioned changes may appear nonspecific, in our opinion the most likely explanation for them is the occurrence of gaseous emboli.
360
The cardiovascular disturbances that imply the formation of gaseous emboli became more obvious with an increased quantity of intravasation. This can be explained by distension fluid and gaseous products of electrosurgery entering the circulation through surgically induced open veins in the myometrium. An increase of intravasation leads to a concomitant increase of gaseous emboli into the circulation. Consequently, the occurrence of related cardiovascular disturbances will increase. It can be postulated that operation time correlates with apparent gas entry. Indeed, the operation time was shortest in group I, the group with the least fluid intravasation. However, although operation times were similar in groups II, III, and IV, significantly fewer cardiovascular disturbances were observed in group II as compared with groups III and IV. Therefore it seems logical to assume that operation time is not a decisive factor in the pathophysiology of the cardiovascular disturbances. It is not likely that the observed physiological changes are the result of volume overload, because volume loading is well tolerated in healthy individuals. Overhydration may be a plausible explanation for hemoglobin oxygen desaturation in the cardiovascular-compromised patient. However, this is improbable in ventilated, relatively healthy patients. Another explanation for the physiological changes is that they result from room air embolism during the process of removing and reinserting the scope, which then acts as a piston in the vaginal cylinder to force air in the open uterine veins. Corson et al [17] reported two patients (out of a series of five catastrophic hysteroscopic events) who died, apparently because room air entered the circulation through surgically exposed blood vessels. Indeed, during a recent TCRM, monitored with transesophageal echocardiography, we repeatedly observed emboli during introduction of the hysteroscope, whereas there was no electrosurgical activation of the instrument, and the resection had not yet started. Although it is clear that emboli develop during the use of diathermia [11], it is important to note that the occurrence of emboli may also be related to the introduction of the resectoscope. Experimental studies have shown that monopolar and bipolar diathermy may produce slightly different amounts of gas and insoluble particles of different sizes, but the composition of the gas appears to be the same [12–14]. A difference between monopolar and bipolar TCR-M could not be observed. However, in groups III and IV, only eight of 123 patients had monopolar energy, which is insufficient to draw conclusions. This study has several important limitations, the first of which is the retrospective design. During hysteroscopic resection of myomas, we observed cardiovascular disturbances in several of our patients. These changes were believed to be the result of gaseous emboli. Therefore a retrospective study was designed to substantiate the frequency of physiological changes and cardiovascular disturbances
Journal of Minimally Invasive Gynecology, Vol 18, No 3, May/June 2011
that are believed to be the result of gaseous emboli. Prospective studies with precordial Doppler scanning, transesophageal ultrasound echocardiography, or the newly developed multifrequency transcranial Doppler scanning (capable of differentiating gaseous from solid emboli) [18] are definitely needed to confirm our observations. Second, there is no actual proof of gaseous emboli and the physiological changes are nonspecific. Indeed, we did not use precordial Doppler scanning or visualize the intravascular emboli with transesophageal echocardiography. However, hysteroscopic resection of myomas is a circumscribed procedure in a healthy population. Therefore, if general anesthesia is provided, this should result in an uncomplicated anesthesia record. If cardiovascular disturbances are seen, they probably result from the procedure itself. Intravasation leading to volume loading is well tolerated in a healthy person, and therefore these physiological disturbances are most likely to result from gaseous emboli. Conclusion Our study suggests that an intravasation of 1000 mL or more in patients undergoing hysteroscopic myoma resection may result in cardiovascular disturbances that indicate the occurrence of gaseous emboli, independent of the type of diathermia or distension fluid used. Although this may be well tolerated by healthy young patients, it is potentially hazardous in an elderly, cardiopulmonary compromised patient. References 1. Brandner P, Neis KJ, Diebold P. Hysteroscopic resection of submucous myomas. Contrib Gynecol Obstet. 2000;20:81–90. 2. Cooper JM, Brady RM. Intraoperative and early postoperative complications of operative hysteroscopy. Obstet Gynecol Clin North Am. 2000;27:347–366. 3. Kim AH, Keltz MD, Arici A, Rosenberg M, Olive DL. Dilutional hyponatremia during hysteroscopic myomectomy with sorbitolmannitol distention medium. J Am Assoc Gynecol Laparosc. 1995;2: 237–242. 4. Arieff AI, Ayus JC. Endometrial ablation complicated by fatal hyponatremic encephalopathy. JAMA. 1993;270:1230–1232. 5. Istre O, Bjoennes J, Naess R, Hornbaek K, Forman A. Postoperative cerebral oedema after transcervical endometrial resection and uterine irrigation with 1.5% glycine. Lancet. 1994;344:1187–1189. 6. Groenman FA, Peters LW, Rademaker BM, Bakkum EA. Embolism of air and gas in hysteroscopic procedures: pathophysiology and implication for daily practice. J Minim Invasive Gynecol. 2008;15: 241–247. 7. Loffer FD, Bradley LD, Brill AI, Brooks PG, Cooper JM. Hysteroscopic fluid monitoring guidelines. The ad hoc committee on hysteroscopic training guidelines of the American Association of Gynecologic Laparoscopists. J Am Assoc Gynecol Laparosc. 2000;7: 167–168. 8. Bloomstone J, Chow CM, Isselbacher E, VanCott E, Isaacson KB. A pilot study examining the frequency and quantity of gas embolization during operative hysteroscopy using a monopolar resectoscope. J Am Assoc Gynecol Laparosc. 2002;9:9–14. 9. Rademaker BM, Groenman FA, van der Wouw PA, Bakkum EA. Paradoxical gas embolism by transpulmonary passage of venous emboli
Rademaker et al.
10. 11.
12.
13.
How Safe Is the Intravasation Limit in Hysteroscopic Surgery?
during hysteroscopic surgery: a case report and discussion. Br J Anaesth. 2008;101:230–233. Mirski MA, Lele AV, Fitzsimmons L, Toung TJ. Diagnosis and treatment of vascular air embolism. Anesthesiology. 2007;106:164–177. Leibowitz D, Benshalom N, Kaganov Y, Rott D, Hurwitz A, Hamani Y. The incidence and haemodynamic significance of gas emboli during operative hysteroscopy: a prospective echocardiographic study. Eur J Echocardiogr. 2010;11:429–431. Munro MG, Brill AI, Ryan T, Ciarrocca S. Electrosurgery-induced generation of gases: comparison of in vitro rates of production using bipolar and monopolar electrodes. J Am Assoc Gynecol Laparosc. 2003;10: 252–259. Munro MG, Weisberg M, Rubinstein E. Gas and air embolization during hysteroscopic electrosurgical vaporization: comparison of gas
14.
15. 16. 17. 18.
361 generation using bipolar and monopolar electrodes in an experimental model. J Am Assoc Gynecol Laparosc. 2001;8:488–494. Farrugia M, Hussain SY, Perrett D. Particulate matter generated during monopolar and bipolar hysteroscopic human uterine tissue vaporization. J Minim Invasive Gynecol. 2009;16:458–464. Barak M, Katz Y. Microbubbles: pathophysiology and clinical implications. Chest. 2005;128:2918–2932. Stoloff DR, Isenberg RA, Brill AI. Venous air and gas emboli in operative hysteroscopy. J Am Assoc Gynecol Laparosc. 2001;8:181–192. Corson SL, Brooks PG, Soderstrom RM. Gynecologic endoscopic gas embolism. Fertil Steril. 1996;65:529–533. Chen CI, Iguchi Y, Garami Z, Malkoff MD, Smalling RW, Campbell MS, et al. Analysis of emboli during carotid stenting with distal protection device. Cerebrovasc Dis. 2006;21:223–228.