Comparison of carbon dioxide and normal saline for uterine distension in outpatient hysteroscopy

FERTILITY AND STERILITY Copyright 1996 American Society for Reproductive Medicine Printed on acid-free paper in U. S. A. Comparison of carbon dioxide and normal saline for uterine distension in outpatient hysteroscopy Fritz Nagele, M.D.* Nikolas Bournas, M.D. Hugh O'Connor, M.RC.O.G. Mark Broadbent, M.R.C.O.G. Robert Richardson, M.RC.O.G. Adam Magos, M.RC.O.G. University Department of Obstetrics and Gynaecology, Minimally Invasive Therapy Unit and Endoscopy Training Centre, The Royal Free Hospital, London, United Kingdom Objective: To evaluate patient acceptance and the clinical feasibility of carbon dioxide compared with normal saline for uterine distension in outpatient hysteroscopy. Design: Prospective, randomized clinical trial. Setting: Outpatient hysteroscopy clinic in a university hospital. Patients: One hundred fifty-seven patients undergoing outpatient hysteroscopy. Interventions: Outpatient hysteroscopy was performed with carbon dioxide or normal saline with endometrial biopsy when indicated. Main Outcome Measures: Need for local anesthesia, cervical dilatation, view of uterine cavity, need to change from carbon dioxide to normal saline distension, procedure time, patient discomfort (lower abdominal pain, shoulder tip pain, nausea) and complications. Results: Carbon dioxide was used for uterine distension in 79 women and normal saline was used in 78. Cervical dilatation was required more often with carbon dioxide hysteroscopy, although there was no increased requirement for local anesthesia. Hysteroscopic vision was similar between the two media, but eight carbon dioxide cases had be to converted to liquid distension. Procedure times were significantly longer for carbon dioxide hysteroscopy as was the occurrence of bubbles during the procedure. Lower abdominal pain and shoulder tip pain were significantly worse with carbon dioxide distension. Although the incidence of nausea and vomiting was higher with the use of carbon dioxide, the differences did not achieve statistical significance. Conclusion: The use of normal saline for uterine distension had no adverse affects on the hysteroscopic view. It provided a shorter operating time and was well accepted by patients. Because of its easy availability and low cost, normal saline is an excellent alternative to carbon dioxide in women undergoing outpatient hysteroscopy. Fertil Steril 1996;65:305-9 Key Words: Outpatient hysteroscopy, distension media, patient acceptance Received May 18, 1995; revised and accepted July 27, 1995. * Reprint requests: Fritz Nagele, M.D., University Department of Obstetrics and Gynaecology, Minimally Invasive Therapy Unit & Endoscopy Training Centre, The Royal Free Hospital, Hampstead, London NW3 2QG, United Kingdom (FAX: 39-171- 431-1321). In recent years diagnostic hysteroscopy has been shown to be superior to conventional dilatation and curettage (D&C) in evaluating abnormal uterine bleeding. Panoramic hysteroscopy allows a complete visual inspection of the uterine cavity and the endocervical canal, directed endometrial biopsy, and, thus, more appropriate management (1-3). In addition, hysteroscopy can be performed as an outpatient procedure in the majority of cases even without the need for local anesthesia (4, 5). Panoramic hysteroscopy requires distension of the uterine cavity with a gaseous (C0 2) or liquid (Hyskon, normal saline, 5% dextrose [Pharmacia Ltd., Knowlhill, Milton Keynes, Buckinghamshire, United Kingdom]) medium. Carbon dioxide has the same refractive index as air and provides optimal clarity (6). As a gas, it is easy to infuse, although success depends on a secure seal between the hysteroscope in the cervix and care in avoiding endometrial contact, which may cause bleeding. The main disadvantage with CO2 is the need for an expensive insuf- Nagele et at. Distension media in outpatient hysteroscopy 305

Table 1 Main Indications for Outpatient Hysteroscopy Indication Menorrhagia Plus irregular bleeding Plus intermenstrual bleeding Plus dysmenorrhea Irregular bleeding Postmenopausal bleeding Otherst Normal saline (n = 78) 30 (38.5)* 19 (24.4) 7 (8.9) 3 (3.8) 4 (5.1) 12 (15.4) 3 (3.8) * Values in parentheses are percentages. t Subfertility, cervical polyps, postcoital bleeding. CO2 (n = 79) 22 (27.8) 21 (26.6) 7 (8.9) 6 (7.6) 8 (10.1) 5 (6.3) 10 (12.7) Total (n = 157) 52 (33.1) 40 (25.5) 14 (8.9) 9 (5.7) 12 (7.6) 17 (10.8) 13 (8.3) flation apparatus to control flow rate and intrauterine pressure to achieve optimal distension of the uterine cavity (7, 8). In contrast to the wide experience with CO2, there is limited objective data on the use of liquids for outpatient hysteroscopy, particularly low viscosity fluids such as normal saline (0.9% sodium chloride, Viaflex container; Baxter Healthcare Ltd., Thetford, Norfolk, United Kingdom) (9, 10). Obvious advantages of such a solution are its ready availability, low cost, insufflation by a simple pressure bag, and rapid reabsorption from the peritoneal cavity in case oftranstuballeakage. Blood clots, viscous secretions, and debris can be washed out, with the result that hysteroscopy can be carried out during menstruation. Theoretical disadvantages of liquid distension for diagnostic hysteroscopy include the potential need for a greater cervical diameter to allow for circulation of the fluid within the uterine cavity and the inconvenience of external leakage (11). Although liquids such as normal saline appear an attractive alternative, CO2 and, to a lesser extent, Hyskon remain the most popular choice of distension media used for diagnostic hysteroscopy. As there is no comparative data, we carried out a prospective randomized study of CO2 and normal saline for uterine distension for outpatient diagnostic hysteroscopy with particular emphasis on the view obtained at hysteroscopy, the need for cervical dilatation, analgesia requirements, the duration of the procedure, patient comfort, and complications. MATERIALS AND METHODS Between February 1992 and December 1993, 157 women referred for outpatient diagnostic hysteroscopy were included in this study. The main indication for referral was abnormal uterine bleeding (Table 1), and all patients had a pelvic examination and, in most cases, a pelvic ultrasound scan. Bookings for the outpatient clinic were made several weeks in advance and the women were told to come irrespective of any vaginal bleeding. The plabned procedure 306 Nagele et al. Distension media in outpatient hysteroscopy was explained to them and a small descriptive booklet also was provided. Full informed consent for the study, which was supported by the local medical ethics committee, was obtained from all subjects. Hysteroscopy was performed by a small group of experienced operators according to an established protocol used by the clinic over several years. Patients were placed in a modified lithotomy position, and the vulva, vagina, and cervix were cleansed with warm antiseptic solution followed by a bimanual pelvic examination. The cervical canal was gently probed with a 3-mm diameter uterine sound and, if the endocervical canal was felt to be tight, the cervix was dilated to 5 to 6 mm. Premedication or analgesia was not administered routinely but, if the procedure was uncomfortable at any stage, local anesthesia was performed using a dental syringe fitted with a 27 French gauge needle and 2.2 ml of 2% lignocaine containing 1:80,000 adrenaline was injected intracervically up to the level of the internal os at 1, 5, 7, and 11 o'clock. Hysteroscopy was performed using a standard 4 mm hysteroscope with a 30 foreoblique lens and a 5.5 mm diagnostic sheath. Illumination was provided by a 250 W high-intensity cold light source via a fiber optic lead. The images were viewed on a high resolution color monitor using a chip camera and unusual lesions were recorded by video. Target biopsies and minor surgical procedures were performed using a 7-mm operative sheath with a operating channel fitted to the hysteroscope through which flexible scissors and grasping forceps could be inserted (all equipment manufactured by Karl Storz GmbH, Tuttlingen, Germany). The nature of randomization was to alternate the distension medium between CO 2 and normal saline on a weekly basis. Hysteroscopic procedures were divided equally among the surgeons irrespective of the distension medium. An electronic HAMOU hysteroflator (Karl Storz GmbH) providing a variable flow rate of up to 120 mumin at a pressure of 100 mmhg was used in the CO2 group. Normal saline was installed from a 500-mL bag wrapped in a pressure bag connected to a manometer pumped to 150 to 250 mmhg. The hysteroscope was maneuvered slowly through the endocervical canal into the uterine cavity under direct vision. The endometrial surface then was inspected systematically and the tubal ostia identified by rotation of the hysteroscope. The hysteroscope then was pulled back toward the internal os to obtain a panoramic view of the whole cavity. Finally, the endocervical canal was inspected during withdrawal of the hysteroscope. Endometrial biopsy or other operative procedures were performed as indicated, the latter often necessitating dilatation of the cervix to 7 to 8 mm to allow passage of the larger Fertility and Sterility

operative sheath into the uterine cavity. Operative procedures generally were carried out using normal saline distension and, similarly, if the view was poor in the CO 2 group because of uterine bleeding, the distension medium was changed to the liquid. The view obtained during hysteroscopy was ranked by the surgeon on a five-point scale (0 = excellent, 1 = good, 2 = adequate, 3 = poor, 4 = very poor); the hysteroscopic view in the CO 2 group, in whom the distension medium had to be changed to normal saline, was scored as "4." The presence of bubbles during the procedure was recorded as was the need to change distension media, The duration of the hysteroscopy was timed by measuring the interval between insertion and removal of the hysteroscope. At the end of the procedure each woman was asked to rank symptoms of lower abdominal pain, shoulder tip pain, and nausea on a five-point scale (0 = none, 1 = mild, 2 = moderate, 3 = severe, 4 = very severe). Any other adverse symptoms such as vomiting or feeling faint also were recorded. The data were analyzed using the X 2 test (including Yates correction when appropriate) and Mann Whitney U-test, and a result of P < 0.05 was considered as statistically significant. Odds ratios were calculated using the program EPI-INFO (CDC, Atlanta, Georgia), with exact confidence limits. RESULTS Seventy-nine patients were randomized to CO2 and 78 were randomized to normal saline hysteroscopy. The primary indication for hysteroscopy was similar for the two groups (Table 1). The mean age in the CO 2 group was 42.3 years (range 21 to 71 years) and 43.4 years for the normal saline group (range 24 to 73 years). There were 23 and 20 nulliparous women in the CO 2 and normal saline groups, respectively, and 6 (7.6%) of the CO 2 group, 1 with a cervical polyp, and 12 (15.4%) of the normal saline group were postmenopausal. The distribution of patients who had regular menses was similar in both groups with respect to the phase of menstrual cycle (follicular: CO 2 = 39, normal saline = 34; luteal: CO2 = 27, normal saline = 28). Significantly more patients of the CO2 group had undergone prior surgery such as D&C, conization, or endometrial ablation (C0 2: n = 13 [16.5%] versus normal saline: n = 4 [5.1 %], X2 [Yates correction] = 4.11, P < 0.05). The distribution of intrauterine and intracervical pathology evident at hysteroscopy were comparable in the two groups, with over half the patients having positive findings, mainly fibroids (C0 2 : n = 20 [25.3%] versus normal saline: n = 16 [20.5%]) and endometrial polyps (C0 2 : n = 11 [13.9%] versus normal saline: n = 20 [25.6%]). The operative sheath was used Table 2 Procedure Statistics in Patients Undergoing Outpatient IIysteroscopy Nonnal saline CO2 Total Procedure (n = 78) (n = 79) (n = 157) Cervical dilatationt 14 (17.9)* 28 (35.4) 42 (26.8) Local anesthesia 16 (20.5) 17 (21.5) 33 (21.0) Endometrial biopsy 65 (83.3) 59 (74.7) 124 (79.0) Operative sheath 5 (6.3) 3 (3.8) 8 (5.1) Procedure timet (min) 5.8 ::':: 3.4 7.2::':: 3.7 6.4::':: 3.7 * Values in parentheses are percentages. t CO 2 group is significantly different from normal saline group. t Values are means::':: SD. CO 2 group significantly longer than normal saline group. in eight patients for the purposes of endometrial polypectomy (n = 5) and target biopsy (n = 3). Significantly more patients randomized to CO2 hysteroscopy needed cervical dilatation for successful hysteroscopy (C0 2: n = 28 [35.4%] versus normal saline: n = 14 [17.9%], X2 [Yates correction] = 5.27, P < 0.05) (Table 2). In contrast, there were no significant differences between the groups with regard to the need for local anesthesia or the use of endometrial biopsy. Eight (10.1 %) patients who started with CO 2 had the distension medium changed to normal saline because of inadequate view secondary to insufficient distension due to cervical leakage (n = 5) or bleeding (n = 3). Intrauterine bubbles occurred significantly more often with CO2 than with normal saline (C0 2: n = 35 [44.3%] versus normal saline: n = 9 [11.5%], X2 [Yates correction] = 19.29, P < 0.001). However, overall there were no significant differences in the scores for vision between the two groups (poor vision or very poor vision occurred for 11 women in the CO 2 group and 6 women in the normal saline group; odds ratio 1.94, 95% confidence interval [Cll 0.61 to 6.74) and the phase of the menstrual cycle did not affect vision (Fig. 1). The total operating time was significantly longer for patients with gas compared with liquid distension (C02 : 7.18 ::!:: 3.67 minutes versus normal saline: 5.76 ::!:: 3.41 minutes, mean::!:: SD; P = 0.03) (Table 2). The difference in mean procedure time between the two groups was 1.42 minutes (95% CI -0.27 to 3.11 minutes). Hysteroscopy was unsuccessful in six (3.8%) women (C0 2 : n = 4, normal saline: n = 2) because of inability to visualize the entire uterine cavity including both tubal ostia (n = 5) and excessive pain despite local anesthesia (n = 1). Two patients from the CO 2 group and one after normal saline distension vomited after the procedure. Vasovagal attacks occurred in five patients (6.3%), all of whom had CO 2 hysteroscopy, but this difference failed to reach statistical significance. There were no other compli- Nagele et al. Distension media in outpatient hysteroscopy 307

Lower abdominal pain Shoulder tip pain Nausea Hysteroscopic vision 1.5 4 Figure 1 Mean ± SD scores for lower abdominal pain, shoulder tip pain, nausea, and hysteroscopic vision. Scores for lower abdominal pain, shoulder tip pain, and nausea: 0 = none, 1 = mild, 2 = moderate, 3 = severe, 4 = very severe. Scores for hysteroscopic vision: 0 = excellent, 1 = good, 2 = adequate, 3 = poor, 4 = very poor. Lower abdominal pain and shoulder tip pain were significantly worse with CO 2 compared with normal saline (lower abdominal pain: CO 2 1.44 ± 0.94 versus normal saline 0.92 ± 0.92; shoulder tip pain: CO2 0.47 ± 0.88 versus normal saline 0.03 ± 0.22). The differences with regard to nausea and hysteroscopic vision obtained during hysteroscopy were not statistically significant (nausea: CO 2 0.24 ± 0.64 versus normal saline 0.13 ± 0.40; hysteroscopic vision: CO2 1.06 ± 1.32 versus normal saline 0.88 ± 1.05). cations and no patients required inpatient admission. The patients responses to the questionnaire regarding discomfort they experienced during hysteroscopy are shown in Figure 1. Most women experienced only slight discomfort, but scores for lower abdominal pain and shoulder tip pain were significantly higher with CO2 distension; 4 patients in the normal saline group and 11 in the CO2 group experienced severe lower abdominal pain (odds ratio 0.33, 95% CI 0.08 to 1.20); only 1 woman experienced any shoulder tip pain in the normal saline group compared with 21 in the CO 2 group (odds ratio 0.04,95% CI 0.01 to 0.24). Although there was a tendency for more nausea after gaseous distension (13 patients experienced nausea in the CO2 group compared with 8 in the normal saline group; odds ratio 0.85, 95% CI 0.20 to 1.62), this failed to reach statistical significance. DISCUSSION Basic research done by Lindemann and others (12-14) concerning physiological, physical, and clinical properties of CO2 has established this as the standard medium for outpatient hysteroscopy. Although it has a long history of safety, severe complications have been reported associated with misuse of equipment and excessive flow rates and pressures (15, 16). Minor side effects with the use of CO2, as 308 Nagele et al. Distension media in outpatient hysteroscopy also demonstrated in the present investigation, include visual impairment by bubbles during the procedure, lower abdominal pain, shoulder tip pain, as well as nausea, vomiting, and vasovagal stimulation (17, 18). Our study, however, shows that normal saline is an equal and by some parameters a superior distension medium. Although overall there was no significant difference between CO2 and normal saline with regard to the quality of the view obtained at hysteroscopy, the presence of bubbles was significantly more frequent with the use of the gas, and 10% of CO 2 procedures had to be converted to liquid distension to obtain an adequate view of the uterine cavity. These two factors probably contributed to the 24% increase in procedure times associated with CO2 distension. Lower abdominal pain and shoulder tip pain were also significantly worse in patients undergoing CO 2 hysteroscopy, and indeed only one patient undergoing liquid hysteroscopy complained of shoulder pain. Equally important at a time of financial stringency in the health service, the equipment for liquid distension is far simpler and cheaper than that required for CO 2 hysteroscopy. Furthermore, low viscosity fluids such as normal saline do not suffer from the risks of anaphylaxis as with high viscosity solutions such as Dextran 70, and instrument care is not compromised (19). Minor operative procedures that may precipitate bleeding also are facilitated; low viscosity fluids are of course already widely used for hysteroscopic surgery such as myomectomy and endometrial ablation (20, 21). It is difficult to explain the apparent greater need for cervical dilatation with CO2 hysteroscopy, as the opposite was expected. The likeliest explanation for this finding is subtle differences in patient characteristics despite randomization, particularly regarding previous surgery and effects of parity. Although this might be taken as an indication of a more difficult hysteroscopy, the lack of any significant difference in the need for local anesthesia between the two groups suggests that any effect was slight. There was only 1 patient in 157 in whom the procedure had to be abandoned because of pain and, in another 5 cases, hysteroscopy was unsuccessful because of inability to visualize the entire uterine cavity adequately, giving a failure rate of 3.8%. De Jong et al. (22), who performed outpatient CO2 hysteroscopy under local anesthesia in 160 patients, reported that 5 women in their study described the pain as intolerable, and in 2 the procedure had to be abandoned, a failure rate of 4.4%. Nearly 20 years ago, Siegler et al. (23) performed diagnostic hysteroscopy under general anesthesia in 257 patients; failure to observe the uterine cavity adequately and to complete the examination occurred in 20 patients, and Fertility and Sterility

uterine perforation occurred in another 5 instances. Thus, the overall failure rate was approximately 10%. Although the technical equipment has improved, these data suggest that gynecologists have widely accepted hysteroscopy and have become more experienced in its technique. In our study the need for local anesthesia was the same for both CO 2 and normal saline, approximately 20%. We have reported previously that the majority of patients find the injection either more or as painful as the hysteroscopy itself (24), and therefore we do not feel that local anesthesia should be administered routinely. In summary, this study confirms the statement by LofIer (25), that an ideal medium for uterine distension does not exist. Hysteroscopists should be aware of the advantages and potential problems of each distension medium and its choice should be based on its capacity for diagnostic accuracy, patient acceptance, and cost. Our data suggest that normal saline provides comparable visualization to CO 2, with reduced procedure time and patient discomfort. Its low cost and simplicity makes normal saline a highly suitable alternative to CO2 as a uterine distension medium for outpatient hysteroscopy. REFERENCES 1. Goldrath MH, Shermann AI. Office hysteroscopy and suction curettage: can we eliminate the hospital diagnostic dilatation and curettage. Am J Obstet Gynecol 1985; 152:220-9 2. Gimpelson RJ, Rappold HO. A comparative study between panoramic hysteroscopy with directed biopsies and dilatation and curettage. Am J Obstet Gynecol 1988; 158:489-92 3. Loffer FD. Hysteroscopy with selective endometrial sampling compared with D&C for abnormal uterine bleeding: the value of a negative hysteroscopic view. Obstet Gyneco11989; 73:16-20 4. Mencaglia L, Perino A, Hamou J. Hysteroscopy in perimenopausal and post-menopausal women with abnormal uterine bleeding. J Reprod Med 1987;32:577-82 5. Hill NCW, Broadbent JAM, Magos AL, Baumann R, Lockwood GM. Local anaesthesia and cervical dilatation for outpatient diagnostic hysteroscopy. J Obstet Gynaecol 1992; 12:33-7 6. Van der Pas H, Vancaille T, editors. Gynecologic endoscopy. Vol. 2. Manual of hysteroscopy. Amsterdam: Elsevier Science Publishers, 1990:1-7 7. Rioux JE. Methods for uterine distension. In: Siegler AM, Lindem.ann HJ, editors. Hysteroscopy: principles and practice. Philadelphia: Lippincott, 1984:37-40 8. Soderstrom RM. Distending the uterus: what medium is best? Clin Obstet Gynecol 1992;35:225-8 9. Sugimoto O. Diagnostic and therapeutic hysteroscopy for traumatic intrauterine adhesions. Am J Obstet Gynecol 1978; 131:539-47 10. Quinones RG. Hysteroscopy: choosing distending media. Int J Fertil 1984;29:129-32 11. Baggish MS. Distending media for panoramic hysteroscopy. In: Baggish MS, Barbot J, Valle RF, editors. Diagnostic and operative hysteroscopy. Chicago: Year Book Medical Publishers, Inc., 1989:89-93 12. Lindemann HJ. Proceedings: CO 2 hysteroscopy, its development and current application possibilities. Geburtshilfe Frauenheilkd 1974;34:385-7 13. Lindemann HJ, Mohr J, Gallinat A, Buros M. Experimental and clinical investigations on the effects of carbon dioxide gas during hysteroscopy. Geburtshilfe Frauenheilkd 1976; 36:153-62 14. Lindemann HJ, Gallinat A. Physical and physiological principles of CO 2 hysteroscopy. Geburtshilfe Frauenheilkd 1976; 36:729-37 15. Obenhaus T, Maurer W. CO2 embolism during hysteroscopy. Anaesthesist 1990;39:243-6. 16. Ditton IN. Presumed carbon dioxide embolism associated with hysteroscopy. Anaesth Intensive Care 1992;20:123-4 17. La Scala GB, Sacchetti F, Dessanti L. Panoramic diagnostic microhysteroscopy: analysis of results obtained from 976 out patients. Acta Obstet Gynecol Scand 1987;144:1-94 18. Fraser IS. Personal techniques and results for outpatient diagnostic hysteroscopy. Gynaecol Endoscopy 1993;2:29-33 19. Siegler AM, Valle RF, Lindemann HJ, Mencaglia L, editors. Therapeutic hysteroscopy. St. Louis: CV Mosby Company, 1990. 20. Loffer FD. Removal of large symptomatic intrauterine growths by the hysteroscopic resectoscope. Obstet Gynecol 1990; 76:836-40 21. Magos AL, Baumann R, Lockwood GM, Turnbull AC. Experience with the first 250 endometrial resections for menorrhagia. Lancet 1991;337:1074-8 22. De Jong P, Doel F, Falconer A. Outpatient diagnostic hysteroscopy. Br J Obstet Gynaecol 1990;97:299-303 23. Siegler AM, Kemmann E, Gentile GP. Hysteroscopic procedures in 257 patients. Fertil SteriI1976;27:1267-73 24. Broadbent JAM, Hill NCW, Molnar BG, Rolfe KJ, Magos AL. Randomized placebo controlled trial to assess the role of intracervicallignocaine in outpatient hysteroscopy. Br J Obstet Gynaecol 1992; 99:777-80. 25. Loffer FD. Complications from uterine distension during hysteroscopy. In: Corfman RS, Diamond MP, DeCherney A, editors. Complications oflaparoscopy and hysteroscopy. Boston: Blackwell Scientific Publications, 1993:176-85. Nagele et al. Distension media in outpatient hysteroscopy 309