Acta Obstet Gynecol Scand 2001; 80: 563567 Printed in Denmark All rights reserved

Copyright C Acta Obstet Gynecol Scand 2001

Acta Obstetricia et Gynecologica Scandinavica

ISSN 0001-6349

ORIGINAL ARTICLE

Manual vacuum aspiration, a safe and effective alternative in early pregnancy termination
JOHAN HEMLIN1 AND BO MOLLER2

From the 1Departments of Obstetrics and Gynecology, 1Central Hospital, Vsters and 2Mlarsjukhuset, Eskilstuna, Sweden

Acta Obstet Gynecol Scand 2001; 80: 563567. C Acta Obstet Gynecol Scand 2001 Background and objective. Vacuum aspiration has become standard surgical procedure for safe early pregnancy termination. Most of these operations are performed in the operating theater using suction curettage and an electric vacuum pump. MVA (manual vacuum aspiration) is an alternative that is well suited for use as a clinical procedure, which could have advantages both for the patient and the health care system. In order to compare conventional VA and MVA in a Swedish setting, a randomized study was undertaken. Methods. Two hundred women requesting abortion in early pregnancy, gestational age less than 56 days, and choosing surgical termination, were randomized to VA or MVA. Main study outcome was frequency of complete abortion but also other variables were recorded. Results. There were no signicant differences between the two groups regarding background characteristics. Altogether 91 MVA and 88 VA operations were performed. There was no signicant difference in frequency of complete abortion; two patients in each group subsequently needed re-curettage because of incomplete evacuation. No case of ongoing pregnancy occurred. Two patients in each group received treatment for endometritis. No other complications were recorded. Conclusion. This study indicates that MVA is effective in emptying the uterine cavity, on par with the standard vacuum aspiration. The rate of complications with MVA was on the same low level as with conventional VA. Key words: manual vacuum aspiration; termination of early pregnancy; vacuum aspiration Submitted 19 October, 2000 Accepted 2 January, 2001

Where available, medical termination of early pregnancy has lately become the method of choice for many women. In surgical termination, vacuum aspiration (VA) is the standard method for safe early pregnancy termination. While most operations are performed using a plastic or metal tube with an opening close to the blunt tip connected to an electrical vacuum pump, there are alternatives. A system with special hand-held syringes is termed manual vacuum aspiration (MVA) and has been promoted so far mostly in resource poor areas of the world. The efcacy of vacuum aspiration for terminating early pregnancy has been widely shown with a
Abbreviations: VA: vacuum aspiration; MVA: manual vacuum aspiration.

frequency of complete abortion between 96 and 99.5 per cent, in most studies 98 per cent or more. The efcacy of MVA has been shown with similar results (1, 2). In Sweden most surgical abortions are performed in the operating theater using suction curettage and an electric vacuum pump. The use of the operating theater and staff implies a considerable consumption of health care resources and costs. Doing the operation as a clinical procedure in the hospital could mean considerable reductions in economical as well as psychological costs, saving the patient time on the waiting list and the embarrassment and greater exposure coming with visits to the main operating theater. As the MVA is well suited for use in the clinic it might be an alternative to VA for those wishing
C Acta Obstet Gynecol Scand 80 (2001)

564

J. Hemlin and B. Moller Since the Swedish health insurance allows the patient to stay home from work because of illness for up to one week without a doctors certicate, the patients had a free choice of staying home for up to one week or returning to work the next day. The main study outcome was efcacy in terms of complete abortion. Failure of the procedure was dened as either ongoing pregnancy or need for recurettage. Prophylactic pre- or postoperative antibiotics were not given. Anti-Rh-immunoglobulin was administered to all Rh-negative women. Patients were instructed to contact the hospital in case of postoperative problems with pain, fever or discharge. Patients with problems in the postoperative period were taken care of by any doctor of the department on duty according to standard procedures. Women were asked to keep records of bleeding and other symptoms in the postoperative phase and asked to attend a postoperative outpatient visit for follow-up, free of charge, six to eight weeks after the operation. One doctor (JH) performed all MVAs, standard VAs were performed by about ten different doctors. All follow-up visits were to the investigator. Per- and postoperative complications were recorded. Infection was dened as problems in the postoperative phase making the patient return for examination and be put on antibiotic treatment, regardless of laboratory results or elevation of temperature. The period of postoperative bleeding was taken as the number of days of use of a pad. Signicance concerning data at interval level was tested by use of Students t-test and on data at nominal level by use of Fishers exact probability. The level for testing signicance was set at 5%. The approval of the Research Ethics Committee of rebro Regional Hospital was obtained according to Swedish regulations.
Results

surgical termination of early pregnancy. A prerequisite for widespread use of MVA must be documented safety on par with VA. In order to compare the two methods in a Swedish setting we set up a randomized study.
Material and methods

After a pilot MVA-study, a protocol was set up. Women seeking early abortion were informed of the options of medical and surgical terminations. Gestational age was determined by LMP and by vaginal ultrasound at rst visit. If there was a discrepancy between these two we followed ultrasound. Women choosing surgical abortion were given the opportunity to participate as long as gestational age at the time of procedure would be shorter than 56 days and as long as the fetus was viable and the uterine anatomy normal. No other exclusions were made. After informed consent patients were randomized using numbered sealed envelopes, which were used in numerical order. According to calculation of statistical power, based on the assumption that the rate of incomplete abortion is about four percent and in order to prove a difference in efcacy between standard VA and MVA, one hundred patients were needed in each group. Between September 1997 and December 1999 a total of two hundred women agreed to participate. All operations, both MVA and VA, were performed in the operating theater. Women had a free choice of general anesthesia or paracervical block. VA was done with an 8 mm plastic VA instrument after dilatation to Hegar 9. Nulliparas allocated to VA were pre-treated with prostaglandin E2 suppositories (Gemeprost 1mg, CervagemA) vaginally for three hours. The MVA syringe has a size of 60 ml and a selflocking mechanism of the plunge. Vacuum is created by pulling the plunger and suction is controlled by use of a valve. A Karman curette is tted, starting with no. 6 as that was expected to make dilatation unnecessary. No women planned for operation with MVA had prostaglandin pretreatment of the cervix. If indicated, a change to curette 8 or conversion to conventional VA was done. If the uterine contents could be completely evacuated without having to disconnect the cannula from the syringe to empty it, i.e. that the 60 ml capacity would sufce, the procedure would be optimally swift and quiet. All patients, both those having general anesthesia and paracervical block, were observed postoperatively for a couple of hours before returning home. They were instructed to avoid bathing, intercourse and the use of tampon for a few weeks.
C Acta Obstet Gynecol Scand 80 (2001)

The study design allocated women into two groups with 100 patients in each. There were no differences between the two groups regarding age or reproductive characteristics (Table I). Gestational

Table I. Characteristics of all patients randomized for inclusion in study Patient characteristics Age, mean (range) years 1-gravid Gravida 2 Gestational age, mean (range) days
1) 2)

MVA(N99) 29.1 (1544) 34 65 49.7 (3856)

VA (N98) 28.6 (1546) 32 66 50.4 (3856)

p-value
NS1 NS2 NS1

Students t-test. Fishers exact probability.

Vacuum aspiration versus manual vacuum aspiration

Table II. Reasons for allocated patients not having surgery MVA (N9) Missing protocols Spontaneous abortion Decided not to abort Operation delayed beyond 56 days Changed to medical abortion Decided not to participate in study 1 1 2 3 1 1 VA (N12) 2 4 4 2

565

Table III. Postoperative complications to MVA and VA MVA (91) Ongoing pregnancy Re-curettage Infection
2)

VA (88) 0 2 2

p-value
NS2 NS2 NS2

0 2 2

Fishers exact probability.

age is at the date of operation based upon ndings at vaginal ultrasound at the primary visit. Altogether 179 operations were performed, 91 using MVA and 88 using VA. Three protocols were missing and 18 patients were allocated but not operated. Reasons for this are given in Table II. Women abstaining from surgery did not differ regarding age or reproductive characteristics from those undergoing abortion. In the MVA group dilatation was done in 17 instances. In three cases a change to curette size 8 was made. In 70 cases the initial 60 cc of vacuum sufced, in 21 the syringe had to be emptied during the procedure, either because aspirated volume exceeded 60 ml, or because of accidental suction of air so that vacuum was lost. In two cases conversion to conventional VA was done, one because of repeated lling of the syringe with blood before evacuation of the uterus was completed, the other because of repeated suction of air so that proper emptying was not achieved. Forty-one percent (37/ 91) of the MVAs were performed in paracervical block, in none of these conversion to general anesthesia was needed. In each group three cases of peroperative bleeding exceeding 100 ml were recorded but no
Table IV. Data on postoperative bleeding (women attending postoperative visit) Women attending postoperative follow-up Days of bleeding directly postoperative (range) Postoperative bleeding more than normal period Days until rst period First postoperative period heavier than normal
1) 2)

bleeding exceeded 300 ml. No case of perforation was recorded. No other operative complications were seen. At the end of the procedure IUD was inserted in 28 of MVA-operations and 31 of VAs. In the MVA-group 56 of 91 came to the followup visit, and 54 of 88 VAs. Patient les of those that did not come for follow-up were checked at a minimum of three months after surgery to see if they had been treated for infection or other complication. Data on complications are given in Table III. There were no differences in efcacy. In each group two patients needed re-curettage (NS). All re-curettage operations were performed in the operating theater on an outpatient basis on the grounds of bleeding and ultrasound ndings indicating remaining products of conception. No-one became anemic. No case of ongoing pregnancy was recorded. Two patients in each group received treatment for postoperative endometritis. No patient needed hospitalization. Data on bleeding pattern in the postoperative phase are given in Table IV. These data are calculated on information collected from the 56 and 54, respectively who attended the postoperative visit. The mean number of days of bleeding directly postoperative was 8.9 and 9.1 in the MVA and VA groups respectively (NS). In each group three patients reported bleeding in the direct postoperative phase exceeding normal periodical bleeding. The interval to rst menstruation was 27.5 and 29.0 respectively (NS). In each group 13 patients reported that the rst menstrual bleeding exceeded that of normal menstruation. The women who did not attend the postoperative visit did not differ from those who did regarding age, reproductive characteristics, gestational age at operation, operative characteristics or postoperative complications.
Discussion

In this prospective randomized study of early surgical termination of pregnancy we compared conventional VA with an electric pump and MVA, where a Karman catheter is connected to a spe-

56 8.9 (042) 3 27.5 (2156) 13

54 9.1 (138) 3 29.0 (2264) 13

NS1 NS2 NS1 NS2

Students t-test. Fishers exact probability.

C Acta Obstet Gynecol Scand 80 (2001)

566

J. Hemlin and B. Moller their decision upon abortion early, in most cases as soon as they know that they are pregnant (8). It has also been shown that a short waiting period is in their interest (9). Performing the abortion as a clinical procedure would save costs. The MVA set costs US$ 18 and can be reused several times except for the cannulae that are single-use and cost US$ 2.5. Clinical time for the operating is US$ 120, to be compared to US$ 500 for operating theater time and postoperative surveillance. If one tenth of terminations in Sweden were done as a MVA ofce procedure, 3000US$ 380 would be saved, equivalent to 10 million SEK. By Swedish law, termination of pregnancy must take place in a health care facility licensed to do it. In practice, the vast majority of abortions in Sweden are performed in public hospitals. The clinical procedure thus implicates that it is done in a gynecological outpatient department. Providing patients and ofce staff have a positive attitude towards the change we believe that MVA also has psychological advantages. As a clinical procedure the general operating ward is avoided, time in hospital shorter. The MVA-instrument is well suited for use in gynecological outpatient services. The operation would then be done in paracervical block after suitable pre-medication. The MVA-instrument has not before been tested in a Swedish setting. This study has shown that it is effective in emptying the uterine cavity, on par with the standard vacuum aspiration. We found the instrument easy to use, well functioning and practical. With the patient awake there is the advantage of silence, no electrical pump, no metal sounds. In the operating theater environment with the patient under general anesthesia however MVA does not seem to provide any advantage.
Acknowledgments
Many thanks to Professor Emeritus Marc Bygdeman, Karolinska Institute, for valuable discussions and advice. Thanks also to Stefan Srensen, Department of Research, Central Hospital, Vsters, for assistance with statistical analysis.

cially designed self-locking syringe with a valve. The aim of the study was to compare efcacy in a setting where all other parameters were as equal as possible and both VAs and MVAs were performed in the operating theater. The two methods were equally efcient. Although altogether only 61% (110/179) of the patients who had surgery came to follow-up-visit, in our study setting it is most unlikely that ongoing pregnancy or need for re-curettage would not come to the knowledge of the investigators. There was no difference in revisits because of endometritis or recurettage and these rates were in concordance with those found in other studies (3, 4). Nor was there any difference in any of the studied parameters including the number of postoperative days with bloody discharge, which also was similar to that found in other studies (5). The results of conventional VA were good, with high efcacy in spite of the fact that they were performed by about ten different doctors. One probable explanation for this is that the procedure is so common that all gynecologists get good experience of it. After the introduction of effective aborticants, many women have opted for medical termination of early pregnancy. The advantages are discretion and contacts with only a limited number of health personnel. Disadvantages are pain and sometimes worrying bleeding over a few days or even weeks. Since the expulsion of the pregnancy sometimes occurs after the patient has left the hospital there is also a need for follow-up to ensure that the treatment was effective. In large studies of medical abortion and vacuum aspiration the frequency of failure does not seem to differ (1, 5, 6). In Sweden at present equally many women choose surgical as medical termination of pregnancy. The advantages of surgical termination are that procedure is short, bleeding after the procedure generally small, the period of pain shorter, less intense and more easily relieved and follow-up is not warranted for medical reasons. Disadvantages are waiting time for operation, the alien atmosphere of the operation ward and the risk of encountering people who might recognize you and guess the mission about which discretion is desired. Another disadvantage is the risk of surgical complications. Having found that the test method (MVA) is as good but not better than the standard method (VA), is there a cause to introduce it? International experience shows that it can be made to work well (2, 7). The utilization of the operation ward consumes health care resources and the costs are considerable. It has been shown that most women make
C Acta Obstet Gynecol Scand 80 (2001)

References
1. Freedman MA, Jillson DA, Cofn RR, Novick LF. Comparison of complication rates in rst trimester abortions performed by physicians assistants and physicians. Am J Public Health 1986; 76 (5): 5504. 2. Westfall JM, Sophocles A, Burggraf H, Ellis S. Manual vacuum aspiration for rst-trimester abortion. Arch Fam Med 1998 Nov-Dec; 7 (6): 55962. 3. Fried G, stlund E, Ullberg C, Bygdeman M. Somatic complications and contraceptive techniques following legal abortion. Acta Obstet Gynecol Scand 1989; 68: 51521. 4. Heisterberg L, Kringelbach M. Early complications after induced rst-trimester abortion. Acta Obstet Gynecol Scand 1987; 66: 2014.

Vacuum aspiration versus manual vacuum aspiration

5. Henshaw RC, Naji SA, Russell IT, Templeton AA. A comparison of medical abortion (using mifepristone and gemeprost) with surgical vacuum aspiration: efcacy and early medical sequelae. Hum Reprod 1994; 9 (11): 216772. 6. Ulmann A, Silvestre L, Chemana L, Rezvani Y, Renault M, Aguillaume CJ et al. Medical termination of early pregnancy with mifepristone (RU 486) followed by a prostaglandin analogue. Acta Obstet Gynecol Scand 1992; 71: 27883. 7. Meyer JH Jr. Early ofce termination of pregnancy by soft cannula vacuum aspiration. Am J Obstet Gynecol 1983 Sep 15; 147 (2): 2027. 8. Holmgren K. Time of decision to undergo an abortion. Gynecol Obstet Invest 1988; 26: 28995.

567

9. Holmgren K. Womens evaluation of three early abortion methods. Acta Obstet Gynecol Scand 1992 Dec; 71 (8): 61623.

Address for correspondence: Johan Hemlin Department of Obstetrics and Gynecology Centrallasarettet 721 89 Vasters Sweden

C Acta Obstet Gynecol Scand 80 (2001)