Minimizing blood loss during surgery is important because of the associated morbidity. In addition, intraoperative bleeding can obstruct the view of the operative field and lead to complications. In particular, dissection during vaginal hysterectomy can be hampered by blood loss, because the main blood supply is not ligated until after much of the dissection has begun. Several methods to control blood loss have been used, including hydrodissection with saline as well as with the injection of vasoconstrictors. Previous reports have demonstrated that the use of the vasoconstrictor vasopressin in other gynecologic surgery has resulted in decreased operative blood loss. Procedures that have shown reduced blood loss with the use of preoperative vasopressin include loop electrocautery of the cervix, hysteroscopic myomectomy, myomectomy, and abdominal hysterectomy (Table 1).1–5
Currently there is debate as to whether the use of vasopressin in vaginal hysterectomy is effective in this task or if the possible complications outweigh the benefits. One of the first reported uses of a vasoconstrictive agent in an attempt to decrease blood loss in vaginal hysterectomy was in 1983 by England et al.6 In this study women were randomly assigned to receive either saline or epinephrine intracervically before the start of the surgery. However, this study failed to demonstrate a significant difference in blood loss with use of epinephrine. Furthermore, it showed a significant increase in postoperative infections, particularly vaginal cuff cellulitis requiring antibiotics. As a result of this study, many gynecologists were reluctant to use vasoconstrictive agents in vaginal hysterectomy. However, more recent studies have demonstrated the use of vasopressin in vaginal hysterectomy is associated with decreased blood loss, does not increase risk of infection, and may decrease operative time.7–9
Although some surgeons believe that injecting saline into the cervix makes it easier to develop surgical planes, others believes that it may obscure dissection, prolonging the procedure and increasing the risk of complications. All of the previous studies involved injecting saline in the control group. It is possible that the decreased blood loss found in the vasopressin groups in the previous studies may simply have been the result of overcoming the potentially increased blood loss by injecting anything into the cervix. We proposed to compare the preoperative injection of vasopressin against no preoperative injection of any medication or saline. The purpose was to estimate the amount of operative blood loss as well as the effect on operative time, postoperative pain medication, and complications rates.
MATERIALS AND METHODS
The study was conducted at Columbia Campus of New York Presbyterian Hospital between January 2004 and January 2005. The protocol was approved by the institutional review board, and informed consent was obtained from all patients. Patients were randomly assigned by a computer-generated randomization scheme in blocks of four patients. They were placed in groups to receive either vasopressin injection intracervically before incision or no preoperative injection. All women were invited to participate in the study if they were aged older than 18 years undergoing a vaginal hysterectomy for any indication, with or without concomitant procedures. Patients excluded were those women with significant medical conditions, including severe liver disease, congestive heart failure, documented coronary artery disease, impaired renal function as determined by an elevated serum creatinine, asthma with a history of steroid use within the past year, and a history of recurrent migraines. No medical devices or study questionnaires were employed. Furthermore, there were no additional costs to the patient, nor did patients receive compensation for their participation. The primary outcome measures for this study were estimated blood loss while undergoing vaginal hysterectomy.
Fifty-eight women were enrolled, with 29 women assigned to each arm of the study. Those patients assigned to the vasopressin arm received an injection of dilute vasopressin solution (20 units of vasopressin in 50 mL of normal saline). A total of 20 mL of vasopressin solution (8 units of vasopressin) was injected in 5 mL increments at 2, 4, 8, and 10 o'clock circumferentially around the cervix at the cervicovaginal junction. Two senior attending surgeons supervised senior residents, and the same standard steps were used in completing the vaginal hysterectomy by both surgeons. Most women received spinal anesthesia; however, seven in each group received general anesthesia.
All patients undergoing vaginal hysterectomy were included, even if other concurrent procedures were scheduled, as long as the hysterectomy was the first procedure. Vaginal hysterectomy was performed on all patients, and important steps were timed, including time until entering the posterior cul-de-sac, removal of the uterus, and reattachment of the cardinal ligaments to the vaginal cuff. Only data from the hysterectomy portion of the surgery were included in the analysis. Concomitant procedures are included in Table 2 and are similar between the study groups.
All patients without an allergy to cephalosporins or penicillin were given 1 g of cephalexin intravenously before the original incision. There was one patient in each group who reported an allergy, and each received 80 mg of gentamycin and 600 mg of clindamycin intravenously before the original incision. All patients were given Venodyne stockings (Microtek Medical, Columbus, MS) for deep vein thrombosis prophylaxis. They were begun before the induction of anesthesia and were kept in place until ambulation.
Estimated blood loss was determined after conclusion of the hysterectomy and culdoplasty with the reattachment of the cardinal ligaments to the vaginal cuff. The study was not blinded to the operating room staff because there was no placebo injection as in previous studies. To ensure the integrity and accuracy of estimating the blood loss, an independent researcher blinded to the assignment group was asked to come into the operating room at the conclusion of the vaginal hysterectomy. The same person estimated the blood loss in every case. Estimated blood loss was determined by measuring the blood in graduated collection canisters, weighing lap pads, and estimating the amount of blood on the operative field and on surgeon gowns.
All patients were treated with the same protocol for pain medication. They were treated with intravenous morphine in the recovery room and then were given a demand-only patient-controlled anesthesia (PCA) device with morphine for the first 24 hours after the surgery. After 24 hours they were converted to oral oxycodone/acetaminophen 5/325 mg tablets. All patients were treated with this protocol and no patients required supplementation with additional medications.
Data collected included preoperative and surgical variables, such as indication for surgery, parity, race, menopausal status, history of hormone therapy use, clinical estimation of uterine size, type of anesthesia given, uterine weight, change in mean blood pressure, time to cul-de-sac, time to uterus, and time to cardinal ligament. Postoperative data incorporated determination of postoperative pain medication, both intravenous and oral, change in hematocrit, febrile episodes and length of stay.
The primary outcome was blood loss, and this was used to determine the sample size need for the study. Previous studies demonstrated a range in the difference in blood loss between the saline and vasopressin groups, from 124 mL to 292 mL. This range of difference in blood loss was significant in these studies and would be an acceptable difference in this study as well. We therefore chose 150 mL as a conservative difference in blood loss between the two groups with a standard deviation of 200 mL. Assuming an alpha of 0.05 for a two-tailed test and a power of 0.8, a sample size of 29 was required in each arm of the study. Data were analyzed using an independent t test, Fisher exact test, and Pearson χ2 as appropriate. P<.05 was considered statistically significant.
There were no significant demographic differences between the two groups with respect to patient age, weight, parity, history of prior abdominal surgery, previous use of hormone therapy, or estimated uterine size (Table 3). There was also no difference in indication for surgery (Table 4) or race (Table 5). Table 6 demonstrates the results of the intraoperative data. The mean uterine weight for the vasopressin group was 180 g compared with 133 g for the control group. However, the difference was not significant. Estimated blood loss for the vasopressin arm was significantly less than the group who did not receive any injection by an average of 100 mL. Operative time, including the time from incision until the cul-de-sac was open and total operative time of the hysterectomy, which ended at reattachment of the cardinal ligaments to the vagina cuff, was not significantly different between the two groups.
Another important element of the study is the effect of vasopressin systemically as seen by a change in blood pressure. As show in (Table 6), there was a significant difference seen in the change of mean blood pressure both at 1 minute after injection and at 5 minutes after injection. The vasopressin group showed a mean increase in blood pressure at 5 minutes of 10.3 compared with 2.5 in the control group, P=.043. There were no serious cardiovascular intraoperative complications related to the change in blood pressure or vasopressin administration.
Postoperative use of pain medication was also recorded. Significantly more PCA morphine was used by women who received vasopressin than those women in the control group. The change in postoperative hematocrit, as measured on postoperative day 1, although lower in the vasopressin group, was not significantly different from the group that did not receive an injection. Among the 58 women, three women in the vasopressin group experienced intraoperative and postoperative complications, including rectal perforation during a posterior colporrhaphy, postoperative hallucinations, and persistently elevated blood pressures in one patient each. With the short half-life of vasopressin, it is difficult to assign the latter two complications to its use. Only one woman in the noninjection group had a complication, an incidental cystotomy. There were no cases of postoperative cuff cellulitis in either group. Finally, length of stay in the hospital was not significantly different between the two groups.
Vasoconstrictors to minimize blood loss have been shown to be effective in many gynecologic procedures. Specifically, vasopressin has been shown to decrease blood loss in myomectomy, hysteroscopy, and abdominal hysterectomy.2–5 Historically, vaginal hysterectomy has been performed without any intracervical injection or with a saline injection intracervically to create a mechanical tamponade and potentially to assist in creating an easier plane of dissection. Currently, there is debate regarding the use vasoconstrictors, such as vasopressin, in an attempt to decrease blood loss during vaginal hysterectomy. The concern stems from an initial study by England et al6 that examined the use of epinephrine as a vasoconstrictive agent. This study demonstrated a significantly increased risk of postoperative cuff cellulitis in the epinephrine group in comparison with the saline group, with a relative risk of 5.5. Furthermore, this study did not show a significant decrease in estimated blood loss during the case, nor was it seen in the postoperative change in hematocrit with the use of epinephrine.
More recent studies have used vasopressin, a vasoconstrictor with a shorter half-life than epinephrine, to examine the effect on blood loss during vaginal hysterectomy. One of the first was a small study by Julian et al10 that used retrospectively matched control subjects to demonstrate a decreased blood loss with the use of locally injected vasopressin during vaginal hysterectomy when compared with saline. Another retrospective study by Potter et al9 also demonstrated that the use of vasopressin in comparison with a saline injection decreased blood loss as well as shortened operative time without an increase in morbidity. More importantly, a prospective, randomized control trial by Kammerer-Doak et al8 examined the effects of vasopressin compared with normal saline injection during vaginal hysterectomy in terms of both infection risk and blood loss. In this study no increased risk of infection was demonstrated with the use of vasopressin. Additionally, it showed a significant decrease in blood loss when vasopressin was used, both as an estimate of blood loss intraoperatively as well as reflected in a change in the postoperative hematocrit. Speer and Unger,7 in another prospective randomized study, again compared women who received local injection of vasopressin to those who received an injection of saline and demonstrated a significant decrease in estimated blood loss in women who received vasopressin. This study was not powered to examine postoperative infection rates; however, no postoperative infections were reported.
Although there now seems to be more convincing data supporting the role of vasopressin in vaginal hysterectomy, there are many gynecologists who do not use any injection at all into the cervix before the procedure, believing that the injected material affects surgical planes, making the surgery more difficult and thereby increasing operative time, blood loss, and complication rates. Even in Te Linde's Operative Gynecology,11 there was no recommendation in the description of the steps to complete a vaginal hysterectomy to include injection of saline or other medications as late at the 8th edition. Only in the 9th edition does the author comment on the use of vasoconstricting agents. It is possible that there is an increased blood loss in the placebo arm of these previous studies secondary to obscured planes from any type of hydrodissection, and vasopressin is merely correcting an iatrogenic situation. Our study examines this question by comparing no injection to an injection of vasopressin. We demonstrated that the use of vasopressin results in a significant decrease in blood loss, which is similar to previous studies.7–9
Although this study was not powered to evaluate difference in operative times, the vasopressin group did show an insignificant trend toward increasing operative time. This suggests that the use of an injection may not clarify the planes of dissection or decrease operative difficulty. Furthermore, it implies that the use of vasopressin to decrease blood loss, and subsequently create a cleaner surgical field, does not perhaps achieve this goal, or if it does, this improvement does not affect operative time. These last findings do highlight the limitations of our study. Because the surgeons cannot be blinded, it may have caused a change in operative technique due to their knowledge of whether the patient had received vasopressin. Although it is always the goal of proper surgical technique to minimize blood loss and safely but expeditiously execute the procedure, subtle changes affecting the results are possible. Furthermore the study was only powered to determine changes in estimated blood loss, not other outcome variables. This may account for a lack of a statistically significant difference, because the sample size may be too small.
Some surgeons have been reluctant to use vasopressin because of concern regarding potential adverse reactions, in particular serious cardiovascular events, such as myocardial infarction, bradycardia, angina, hypertension, arrhythmias, and anaphylaxis. There have been case reports of intraoperative myocardial infarction and severe hypotension following coronary vasospasm after the use of vasopressin in gynecologic surgery. However, these adverse events were likely after intravascular injection.12,13 In our study the patients in the vasopressin group showed a significant increase in mean blood pressure at both 1 and 5 minutes; however, there were no adverse events related to this increase.
We also demonstrated a significant increase in the use of postoperative patient-controlled morphine in the patients receiving vasopressin, although immediate recovery room morphine requirements were not significantly different between the two groups. This finding is in conflict to a previous report which suggests that mechanical compression of the tissue by distension was a component of immediate pain control. 14 If this was the case, we would have expected a decrease in morphine requirement by those in the vasopressin group immediately postoperatively. A cause of the increased PCA morphine requirement is unclear at this point, although possibly an increase in local inflammatory agents several hours postoperatively occurs as the vasoconstrictive effects of vasopressin diminish. Further research into postoperative analgesia after vaginal hysterectomy as well as the effect of vasopressin on pain will need to be completed to better answer this question.
Our report is consistent with previous studies and demonstrates that the use of vasopressin in vaginal hysterectomy results in a significant decrease in blood loss in comparison with the same procedure completed without any injection. In addition, the use of vasopressin does not significantly affect operative time, nor does its use cause an increase in infection, but it may cause an increase in postoperative pain.
1. Howells RE, Tucker H, Millinship J, Shroff JF, Dhar KK, Jones PW, et al. A comparison of the side effects of prilocaine with felypressin and lignocaine with adrenaline in large loop excision of the transformation zone of the cervix: results of a randomised trial. Br J Obstet Gynecol 2000;107:28–32.
2. Phillips DR, Nathanson HG, Milim SJ, Haselkorn JS, Khapra A, Ross PL. The effect of dilute vasopressin solution on blood loss during operative hysteroscopy: a randomized controlled trial. Obstet Gynecol 1996;88:761–6.
3. Corson SL, Brooks PG, Serden SP, Batzer FR, Gocial B. Effects of vasopressin administration during hysteroscopic surgery. J Reprod Med 1994;39:419–23.
4. Fletcher H, Frederick J, Hardie M, Simeon D. A randomized comparison of vasopressin and tourniquet as hemostatic agents during myomectomy. Obstet Gynecol 1996;87:1014–8.
5. Okin CR, Guido RS, Meyn LA, Ramanathan S. Vasopressin during abdominal hysterectomy: a randomized controlled trial. Obstet Gynecol 2001;97:867–72.
6. England GT, Randall HW, Graves WL. Impairment of tissue defenses by vasoconstrictors in vaginal hysterectomies. Obstet Gynecol 1983;61:271–4.
7. Speer P, Unger JB. Does saline really decrease blood loss at vaginal hysterectomy as much as vasopressin. J Pelvic Surg 2001;7:340–3.
8. Kammerer-Doak DN, Rogers RG, Johnson Maybach J, Traynor Mickelson M. Vasopressin as an etiologic factor for infection in gynecologic surgery: a randomized double-blind placebo-controlled trial. Am J Obstet Gynecol 2001;185:1344–7.
9. Potter T, Fletcher H, Reid M. Vasopressin as a hemostatic and dissection aid at vaginal hysterectomy. Int J Gynaecol Obstet 2004;86:65–6.
10. Julian TM, Johnson GW, Gosewhr JA, et al. Vasopressin as a chemical tourniquet during vaginal surgery. J Gynecol Surg 1993;9:161–4.
11. Thompson JD. Surgical techniques for pelvic organ prolapse. In: Rock JA, Thompson JD, editors. TeLinde's operative gynecology.
8th ed. Philadelphia (PA): Lippincott-Raven; 1997. p. 969–79.
12. Martin JD, Shenk LG. Intraoperative myocardial infarction after paracervical vasopressin infiltration. Anesth Analg 1994;79:1201–02.
13. Nezhat F, Admon D, Nezhat Ch, Dicorpo JE, Nezhat C. Life-threatening hypotension after vasopressin injection during operative laparoscopy, followed by uneventful repeat laparoscopy. Am J Assoc Gynecol Laparosc 1994;2: 83–6.
© 2009 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.
14. Miller L, Jensen MP, Stenchever MA. A double-blinded randomized comparison of lidocaine and saline for cervical anesthesia. Obstet Gynecol 1996;87:600–4.