Paulson

cannot predict that even those with little rents in the tube won't get pregnant. I think we have women with intrauterine pregnancies even when they have something like that. At the time of a second-look laparoscopy, I might try to create a raw area over such a rent in the tube so it will reperitonealize. I was asked why we did second-look laparoscopies instead of hysterosalpingography on some of our pa-

August 1992 Am J Obstet Gynecol

tients. It was mostly patient choice. I am very aggressive and I feel that second-look laparoscopies are helpful because it will add to future fertility and predictive values. However, the majority of our patients did not have second-look laparoscopies and we have not analyzed our data on second-look laparoscopy versus hysterosalpingography, so I don't know how predictive it would be.

Intrauterine pressure and fluid absorption during continuous flow hysteroscopy Spyro P. Vulgaropulos, MD, Leigh C. Haley, MD, and Jaroslav F. Hulka, MD Chapel Hill, North Carolina OBJECTIVES: Our objectives were to document the causes of fluid absorption during continuous flow hysteroscopy and to determine under which operative conditions fluid overload may occur. STUDY DESIGN: Fifteen patients underwent operative hysteroscopy with 2% ethanol solution for uterine distention. Absorption of fluid was measured by blood alcohol, sodium, osmolarity, and hematocrit. Intrauterine pressures were measured with an obstetric pressure catheter. RESULTS: Alcohol absorption was noted in one patient during a myoma resection. Two additional patients, not in the study, had fluid absorption after partial perforations of the uterus. Under normal operative conditions there were no changes in sodium, osmolarity, or hematocrit. Intrauterine pressures ranged from 45 to 75 mm Hg. Experimental pressures of >200 mm Hg were not associated with fluid absorption. CONCLUSIONS: Intravasation of fluid may occur through open uterine venous channels with extensive resections and under low pressures in the presence of unrecognized perforations. (AM J OSSTET GVNECOL

1992;167:386-91.)

Key words: Hysteroscopy, fluid absorption, intrauterine pressure, ethanol, partial perforation Diagnostic and operative hysteroscopy procedures are done in a variety of settings with different distention media. The use of continuous flow distention with lowviscosity fluids such as 5% dextrose in water, saline solution, and glycine are popular for intrauterine laser, resections, coagulation, and ablations. However, there From the Department of Obstetrics and Gynecology, University of North Carolina School of Medicine. Supported in part by Bard Reproductive Systems, Tewksbury, Massachusetts. Presented as Invited Guest at the Fiftyjourth Annual Meeting of The South Atlantic Association of Obstetricians and Gynecologists, Palm Beach, Florida, January 26-29, I 992. Reprint requests:Jaroslav F. Hulka, MD, Department of Obstetrics and Gynecology,CB #7570 MacNider Building, University ofNorth Carolina School of Medicine, Chapel Hill, NC 27599-7570. 6/6/38059

386

is a risk of irrigation fluid absorption leading to volume overload and a rare risk of death subsequent to pulmonary edema or hyponatremia. 1 The mechanism of this intravasation is poorly understood but appears to be similar to complications of urologic transurethral prostate resection. The objective of this study is to measure some of the mechanisms involved in systemic absorption of irrigation fluid and to determine under which operative conditions volume overload can occur. Methods

Fifteen study participants were chosen from patients scheduled for diagnostic and operative hysteroscopy; they gave their full informed consent as approved by the Committee on the Protection of Human Subjects

Volume 167 Number 2

at the University of North Carolina Hospitals. Uterotubal cannulations, endometrial ablations, myoma resections, and septum resections were the indications for the procedures, which were mostly carried out with the patients under general inhalation anesthesia in the proliferative phase with no hormonal preparation. Pitressin was used during ablations and myoma resections only. A 24F Wolf resectoscope was used, and for myomas and resections the cervix was overdilated to 37F. Glycine, a nonelectrolyte solution commonly used for transurethral prostate resection by urologists, was used as an irrigant. Ethanol was added to create a 2% solution. Fluid absorption was then measured directly by blood alcohol levels. In addition, indirect estimates of fluid absorption were made by measuring the volume of outflow collected and by subtracting this volume from the amount of irrigant (inflow) used. Also, calculations of intravascular fluid absorption were made by measuring plasma sodium, osmolarity, and blood hematocrit. Blood samples were obtained and pressures were measured at the following times during a procedure: Time 1 Beginning of the procedure (baseline) Time 2 After 5 minutes with 8 feet of pump pressure Time 3 After 5 minutes with 14 feet of pump pressure or at 8 feet with outflow restricted Time 4 End of procedure Intrauterine pressures were measured directly with a standard obstetric pressure catheter placed alongside the hysteroscope inside the uterine cavity. A uterine model was also created for laboratory measures of flow rates and pressures with a variety of pump settings, pump types, and different hysteroscopes. The patients were used as their own controls, and one-way analysis of variance statistical measures were used to compare changes in plasma sodium, osmolarity, and blood hematocrit with a significance level at p < 0.05. Results

In vitro studies. The laboratory model used (Fig. 1) documented that a number of independent variables influenced intrauterine pressure and flow rates. These included the following: (1) height of fluid column or pump setting, (2) manufacturer of hysteroscope, (3) degree of cervical dilation, and (4) outflow valve closed or open. We first calibrated the pressure reading of our uterine model with a closed system (no cervical or valve outflow) in an operating room, simulating conditions of surgery. The three points measured were at 4, 6, and 8 feet of direct gravity flow, resulting in intrauterine pressures of 118,177, and 230 mm Hg, respectively. Higher direct gravity measures were not possible, because the operating room ceilings are not higher than 8 feet from the operating table surface. The Bard Hys-

Fluid absorption during hysteroscopy 387

tero-Flo (C. R. Bard, Tewksbury, Mass.) pump matched this calibration curve at 8 and 14 feet of pressure. We next measured the effect of opening the outflow valve and simulating cervical overdilation by allowing fluid to flow around the hysteroscope. Opening the outflow valve decreased pressure to about 50 to 60 mm Hg, and opening the cervix further decreased pressures to 40 to 50 mm Hg. Under these conditions there was minimal difference in intrauterine pressure with Bard pump settings of 8 or 14 feet of pump pressure. Urologists routinely use the outflow valve to regulate flow rates and pressures. Using the outflow valve to maintain the intrauterine pressure at either 50 or 80 mm Hg (Fig. 2), there was only a 12% difference in flow rates with Bard pump settings of 8 or 14 feet, and flow was approximately 375 to 450 mllmin. The Zimmer pump is designed with an upper limit of 80 mm of pressure, with a readout measured at the inflow valve (approximating intrauterine pressure). When 80 mm of measured intrauterine pressure was maintained, flow rates were reduced, even at the highest pump setting, to 150 mil min (Fig. 2). Fig. 3 shows intrauterine pressure and now rates with two different hysteroscopes. Surprisingly, the larger (24F) Wolf resectoscope had lower flow rates and higher resistant pressures than the smaller (22F) Olympus pediatric resectoscope. Clinical studies. Intrauterine pressure data from 14 patients thus far show a range of 45 to 75 mm Hg (with an average 63 mm Hg) under normal operative conditions (open cervix or outflow valve, no perforations). Closing overdilated cervices with a tenaculum for 5minute periods raised intrauterine pressure to an average of 122 mm Hg. In some patients, both the cervix and outflow valve were closed, creating intrauterine pressures of 230 and 280 mm Hg with 8 and 14 feet of pump pressure, respectively. No fluid absorption as measured by blood alcohol, volumetric discrepancy, or hyponatremia was seen during these in vivo pressure experiments. Blood studies on 15 patients revealed nonsignificant average declines in serum sodium from 139 to 136 (time 1 and time 4), p > 0.05. Average osmolarity was 285 to 289, and the mean hematocrits were 35 to 33. The mean measure of volumetric fluid deficit was 152 ml. One patient out of 15 had ethanol detected (37 mg/dl) at the end of the procedure. This patient had had a myoma resection and both the cervix and outflow valve had been closed for 3 minutes to increase visualization due to excessive bleeding. Through an oversight the valve had remained closed for nearly 10 more minutes. The pressure catheter was not functioning in this case. The volumetric measurement of absorption in this case was 900 ml, equivalent to the estimated volume required to reach a blood level of 37 mg/dl

388 Vulgaropulos, Haley, and Hulka

August 1992 Am J Obstet Gynecol

Irrigation Fluid

Cervix open (37 Fr) or Cervix closed (25 Fr)

t

Intra-uterine

Pressure Gauge

D Suction Drapes

Fig. 1. Experimental model. Variables of fluid dynamics include pump settings and types, type of hysteroscope, cervical dilation, outflow valve.

ethanol, given a 5 L blood volume and 2% ethanol solution. One other patient, not in the study, developed a lateral tear of the cervix after fibroid removal. The standard pressure settings, estimated at 75 mm Hg, were used for a 3-minute period to diagnose the perforation. During this time the patient had absorbed 1500 ml of fluid by volumetric estimation. This patient did not have an y clinical signs of volume overload. A third patient, not in the study because of severe diabetes and chronic renal insufficiency, developed acute pulmonary edema after a rapid (12 minute) endometrial ablation under local anesthesia. Serum sodium level was 129 mEq/L. There was an absorption, estimated by measuring input and output of the irrigation system, of 1300 ml. There was a suspected thinning and possible partial perforation of the fundus during the dilatation of the cervix during this procedure. There was good outflow from the cervix in this case, allowing an estimate of intrauterine pressure of 60 to 75 mrn Hg.

Comment

Hysteroscopic surgery with low-viscosity, continuousflow irrigation offers several advantages over other systerns! " These include better exposure in operative cases, continuous clearing of blood and tissue, and fewer reported complications.' Nevertheless, the risk of fluid overload and hyponatremia exists and is anal ogous to the complications with prostatic resection procedures. The finding of minimal to no absorption during hysteroscopy detected in most of these patients to date is encouraging. Hysteroscopy should be expected to have less than the 5% to 10% incidence during prostatic resection by the urologists. I. S This is due to the fact that prostatic tissue has an extensive plexus of venous sinuses that are opened up during resections, although paracervical venous plexi can be exposed during unsuspected cervical tears." We have drawn from studies in the urology literature that use ethanol added to the irrigation fluid to monitor absorption during prostatic resection.' Hahn" in 1988

Fluid absorption during hysteroscopy 389

Volu me 167 Num ber 2

500 400

-::

c: E 300

0 0

0 ;;::

~

Bard, 14 ft

-4-

Bard, 8 ft

--

..... ......

Zimmer

200 100

......

Intrauterine pressure maintained at:

t

t

50 mm Hg

80 mm Hg

0 Fig. 2. Factorsaffecting flow rates (in vitro). Maintaining intrauterine pressures with outflow valve.

OJ J:

80

Hystero-Flo Setting 14 It

-

E E 60

12 ft

Q)

L..

::J

C/) C/)

Q)

40

L..

c, Q)

c: 20

::> .~

Q)

0

0

200

400

600

Flow (cc/mln) Fig. 3. Uterine pressure versus flow. In vitro comparison of hysteroscopes. dem onstrated that absorbed glycine levels are in direct propo rtion to serum ethanol levels, whereas there was a proportional decline in th e serum sod iu m levels during prostati c resection s. Norlen" determined fluid absorption on the basis of th e osmolarity of th e irrigation Huid (glycine) and th e cha nge in serum sod iu m seen postop erati vely. The availab ility of ine xpensive breathalyzer units may mak e it possible to routinely monitor the expiratory gases continuously during surgery with a I % or 2% ethanol irrigant, to detect fluid absorption as it occurs during su rgery." In add ition to case rep orts, 10. II Morrison et al." ha ve the only other study in th e gyneco logy literature attemptin g to measure fluid absor ption during hysteroscopic surgery . In one patient of the 12 in th e study fluid ove rload develop ed after lase r ablation procedure in which saline solution was the irrigation fluid. Measureme nts of absorption show a rise in central venous pressure and serum chloride while the plasma albumin and hematocrit decreased. Intrauterine pressures were not included in the study.

The urology patient tends to be old er and more susceptib le to volume shifts. Another interesting difference between our patients and th e urology patients is ge nde r. In 1986 Ar ieff" reported 16 postoperative deaths ove r a I O-year peri od in health y women un dergoin g elective gen eral and gyneco logic surgery. T hey were found to have had fluid overload lead ing to the syndrome of inappropriate antid iuretic hormone resultin g in severe hyponat remi a. Fraser et al.!' also re po rts similar findings in II de ath s that he attributes to cent ra l diabetes mellitus and di abetes insipidus seconda ry to untreated sever e hypon atremia. Fraser and Ar ieff " have also sho wn in ra t stu dies th at the Na + -K + ad enosine 5'triphosphatase pump syste m is less effective in fem ale than in male brain syna pses. Progesterone can also decrease pump activity in myocardial and kidney tissue. I' These stud ies suggest that guarding aga inst iatrogenic hyponatremi a may be even more critical in female patients. How eve r, animal studies in our laboratories have failed to co nfirm th ese hypotheses to da te.

390 Vulgaropulos, Haley, and Hulka

This is a preliminary report of our studies so far, presented to report some clinically useful unexpected findings. Different pump settings and pumps (Figs. 2 and 3), result in wide differences in pressures and flow rates, as do different hysteroscopes (Fig. 3). Contrary to the current assumption that elevated pressures cause fluid absorption, pressures >200 mm Hg did not result in fluid absorption, yet a rapid procedure under normal pressure conditions did. We will continue to collect data on more patients, in collaboration with other hysteroscopists, to further determine which procedures are at risk and which are free of risk, of fluid absorption. We will also see if there is a pressure threshold above which fluid absorption occurs, keeping in mind the last patient cited, where symptomatic fluid overload occurred well within safe pressures. Our conclusions thus far are as follows: (I) Significant intravasation of fluid may occur with extensive resections which open uterine venous channels; (2) significant intravasation also can occur at low pressures in the presence of unrecognized perforations; (3) alcohol is a reliable indicator for measuring fluid absorption and may be used to monitor this risk; (4) there appears to be a minimal risk of absorption during hysteroscopy under normal operative conditions.

1.

2. 3. 4.

5.

6. 7.

8. 9.

10. 11.

REFERENCES Rhymer ]C, Bell T], Perry KC, Ward ]P. Hyponatremia following transurethral resection of the prostate. Br ] Urol 1985;57:450-2. Seki M, Mettler L, Semm K. Comparison of fluid hysteroscopy and CO, hysteroscopy. Endoscopy 1982;14:141-5. Quinones R. Hysteroscopy: choosing distending media. Int] FertiI1984;29:129-32. Peterson HB, Hulka]F, Phillips]M. American Association of Gynecologic Laparoscopists' 1988 membership survey on operative hysteroscopy. ] Reprod Med 1990;35: 590-1. Emmett ]F, Gilbaugh ]H, McLean P. Fluid absorption during transurethral resection: comparison of mortality and morbidity after irrigation with water and nonhemolytic solutions.] Urol 1969:101:884-9. Hulka ]F, Higgins G. Tears of the internal cervical os during dilation for routine curettage. AM] OBSTET GvNECOL 1961;82:913. Hulten]O,]orfeldtLS, Wictorisson YM.Monitoring fluid absorption during TURP by marking the irrigating fluid solution with ethanol. Scand] Urol NephroI1986;20:24551. Hahn RG. Ethanol monitoring of irrigating fluid absorption in transurethral prostatic surgery. Anesthesiology 1988;68:867-73. Norlen H. Isotonic solution of mannitol, sorbitol, glycine and distilled water as irrigating fluids during transurethral resection of the prostate and calculation of irrigating fluid influx. Scand] Urol Nephrol 1985;96(suppl): I-58. Van Boren M], Singelyn F, Donnez], Gribomon B. Dilutional hyponatremia associated with intrauterine endoscopic laser surgery. Anesthesiology 1989;71:449-50. Feinberg BI, Gimpleson R], Godier DE. Pulmonary edema after photocoagulation of the endometrium with the Nd:YAG laser. A case report. ] Reprod Med 1989;34:431-4.

August 1992

Am

J Obstet Gynecol

12. Morrison LLM, Davis], Sumner D. Absorption of irrigating fluid during laser photocoagulation of the endometrium in the treatment of menorrhagia. Br ] Obstet Gynaecol 1989;96:346-52. 13. Arieff AI. Hyponatremia, convulsions, respiratory arrest, and permanent brain damage after elective surgery in healthy women. N Engl] Med 1986;314: 1529-35. 14. Fraser CL, Kucharczyk], Arieff AI, Rollin C, Sarnacki P, Norman D. Sex differences result in increased morbidity from hyponatremia in female rats. Am ] Physiol 1989;256:R880-5. 15. Fraser CL, Arieff AI. Fetal central diabetes mellitus and insipidus resulting from untreated hyponatremia: a new syndrome. Ann Intern Med 1990;112:113-9. Editors' note: This manuscript was revised after these discussions were presented.

Discussion DR. RONALD L. BROWN, Charlotte, North Carolina. With increasing use of operative hysteroscopy as a surgical alternative for various gynecologic problems, there has become a heightened awareness of the potential complications associated with the procedure. The use of lasers and electocoagulation devices within the intrauterine cavity certainly poses a potential threat to the integrity of the uterine walls, which if violated can lead to a life-threatening situation. Fluid overload, as a result of excessive absorption of distention medium, is a potentially serious complication of hysteroscopy, the cause of which has never been fully understood. Goldrath was a pioneer of endometrial ablation with the Nd:YAG laser, and he recognized fluid overload in several of his patients. Goldrath et al. 1 hypothesized that insufficient dilatation of the cervix and prolonged operating time may be responsible for excessive fluid absorption. Others have suggested that if the fallopian tubes are not blocked that fluid will pass through them and into the peritoneal cavity and be absorbed into the circulation. In 1990 Garry' reported a study of 63 women undergoing endometrial ablation. In 30 of these women fluid was infused with a simple roller pump, and in 33 women the pump was connected to a pressure transducer that maintained intrauterine pressure at a preset level or less. The pressure-controlled group showed an 86% reduction in the volume absorbed, which was measured by inflow-outflow volumes. He suggested that damaged peripheral uterine arteries were the pathway for fluid absorption and if intrauterine pressure was maintained at the patient's mean blood pressure or less, there would be no retrograde flow into the systemic circulation and no fluid overload. Dr. Vulgaropulos has reported a clinical study of 12 patients in a well-designed protocol to directly measure fluid absorption and intrauterine pressure. This clinical study has been complemented by an experimental laboratory model that demonstrates the effects of several variables on intrauterine pressure and flow rates. It was shown very clearly in the experimental model that intrauterine pressure is primarily a function of the outflow resistance regardless of pump setting. If outflow

Volume 167 Number 2

was unobstructed, intrauterine pressure was never >60 mm Hg. Conversely, closure of the outflow valve and cervix resulted in pressures> 117 mm Hg at all pump settings. In the clinical study 11 patients underwent various types of operative hysteroscopy and were maintained at intrauterine pressures of 45 to 75 mm Hg with ethanol-labeled glycine as the distention medium. At an undesignated time during the surgery and for an unspecified length of time, outflow of glycine was obstructed in some of these patients, generating intrauterine pressure of 230 and 280 mm Hg at different pump settings. In none of the 11 patients was there any evidence of fluid absorption. In a twelfth patient there was evidence of fluid absorption as measured by blood alcohol level. It is unfortunate that no intrauterine pressures were measured on this patient, but it is safe to assume that normal operating pressures (i.e., pressures >75 mm Hg) were exceeded and probably >200 mm Hg because outflow was totally obstructed. In this patient there was apparent excessive intrauterine pressure for 13 minutes at a time in the surgery when there were obviously open, damaged uterine vessels, and she did have evidence of fluid absorption. In the other 11 patients in the study there is no mention of how many had excessive intrauterine pressure and for how long the elevated pressures were present. It may also be of value to know at what point during the surgery excessive pressures were exerted (i.e., was there excessive intrauterine pressure after the resection of the tissue?). I would ask that Dr. Vulgaropulos comment on this and give his opinion of the significance of these temporal relationships. It would seem that the cause of fluid absorption during hysteroscopic surgery is multifactorial. Certainly,

Fluid absorption during hysteroscopy 391

perforation of the fundus of the uterus and cervical tears have been shown in other case reports to lead to fluid absorption and overload, as was the case with the two nonstudy patients in this report. The cardiovascular status, age, and hormone effect of the patient may also be factors. Dr. Vulgaropulos's study has answered a number of questions about pressures, flow rates, and pumps used in operative hysteroscopy; however, I don't think we can as yet rule out excessive intrauterine pressure as having at least some role in fluid absorption in some patients. REFERENCES 1. Coldrath MH, Fuller TA, Segal S. Laser photovaporization of endometrium for the treatment of hemorrhagia. AM J OBSTET GYNECOL 1981;140:14-9. 2. Garry R. Safety of hysteroscopic surgery. Lancet 1990;336: 1013-4.

DR. VULGAROPULOS (Closing). The temporal relationship and the timing is something that we have not quite worked out in our study design. The time the procedure takes is much the same except for the quicker diagnostic procedures. Regarding the elevated pressures, we are just trying to get an idea of how much pressure can be created, and we are holding them for 30 to 60 seconds as opposed to doing the complete resection with closed outflow. After the resection is done, we are basically still at open outflows but it is just within the procedure itself that we will close down the outflow to measure and see what pressures we get for a period of time. It appears that the patient with fluid absorption had it occur over a 10-minute period.

Intrauterine pressure and fluid absorption during continuous flow hysteroscopy.

Our objectives were to document the causes of fluid absorption during continuous flow hysteroscopy and to determine under which operative conditions f...
759KB Sizes 0 Downloads 0 Views