Department of Obstetrics and Gynaecology, Women's Centre, Oxford Radcliffe Hospitals NHS Trust, Headington, Oxford OX3 9DU, UK
![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Key words: abortion/complications/surgical/termination/training
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
Materials and methods |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Terminations were performed using electrical vacuum aspiration with a plastic cannula under general anaesthesia. Neither infection screening nor prophylactic antibiotics were routinely used. Nulliparae received cervical priming using gemeprost 1 mg vaginally 23 h prior to surgery (MacKenzie et al., 1990). Women were instructed to see their general practitioner 6 weeks after surgery. As the John Radcliffe is the only hospital serving Oxford and its environs, patients with complications would either be referred back to us by their general practitioner or present as an emergency to the unit.
Women who had a complication while undergoing surgery or during the subsequent 6 weeks were identified and cross-checked against the gynaecology unit acute admission database. This database was prospectively collated and included the details of all women seen as emergencies in the unit. The complications recorded included the following: additional surgical procedures (curettage, laparoscopy or laparotomy); ongoing pregnancy; uterine perforation; pelvic sepsis requiring intravenous antibiotics; and blood transfusion.
Complete termination was defined as complete evacuation with no ongoing pregnancy or requirement for surgical curettage during the 6 weeks following termination.
The 2 test or MantelHaenszel
2 test for trend (when controlling for a third variable) was used for statistical analysis when appropriate.
![]() |
Results |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
|
|
|
One nulliparous woman whose pregnancy was terminated at 9 weeks gestation by a registrar presented complaining of abdominal pain 16 days after surgery. Ultrasound and magnetic resonance imaging suggested a pelvic collection. She underwent a laparotomy and a pelvic haematoma was diagnosed; an unrecognized uterine perforation at the time of abortion was considered responsible. There were no other cases of perforation diagnosed. The recognized perforation rate was 1/828 = 0.12%.
One patient returned with persistent bleeding 4 weeks post-termination and was found to have a higher than expected serum human chorionic gonadotrophin (HCG) concentration (1120 IU/l). Serial serum HCG declined slowly, an ectopic pregnancy was suspected, and a week later she underwent a laparoscopy with cervical dilatation and curettage. At surgery it was not clear whether her symptoms and signs were due to a resolving ectopic pregnancy or to retained products of conception. There were no other ectopic pregnancies suspected or diagnosed. The overall laparoscopy/laparotomy rate was therefore 2/828 = 0.24%.
No patients required blood transfusion.
Eight patients (0.97%) were admitted with a presumed diagnosis of post-abortion pelvic sepsis and commenced on intravenous antibiotics. Of the eight women, six had a curettage for presumed retained products during their admission for antibiotic treatment, one woman received intravenous antibiotics alone with no further surgical treatment, and the remaining patient was admitted with pelvic sepsis 13 days after curettage for presumed retained products following termination.
In all, during the 6 weeks after termination, a total of 14 (1.69%) women required overnight hospital admission for treatment and/or observation.
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Our complete termination rate of 94.6% is at the lower end of those previously reported (Nesheim, 1984; Heisterberg and Kringelbach, 1987
; MacKenzie et al., 1990
; Sykes, 1993
; Westfall et al., 1998
). A prospective trial from our unit of 1030 patients undergoing first trimester surgical termination, designed to assess the value of different prostaglandin preparations for cervical ripening, was reported in 1990 (MacKenzie et al., 1990
). We demonstrated a complete termination rate of 98.8% (readmission for curettage 1.0% and continuing pregnancy 0.2%), rates of uterine perforation and blood transfusion of 0.5% and nil respectively, and an overnight admission rate of 1.0%. Senior registrars performed 75% of these abortions. Apart from the now much higher rate of readmission for curettage (5.2%), the results were very similar to our latest report. A large retrospective review of 2879 surgical abortions performed in New Zealand in 1989 and 1990 reported a complete termination rate of 95.2% (Sykes, 1993
). A recent report of 1769 manual vacuum aspirations in the USA at gestations of 10 weeks or less demonstrated efficacy of 99.5% (Westfall et al., 1998
). A physician with great experience of the technique performed the majority of procedures. If all terminations in Oxford had been performed by consultant gynaecologists then the incomplete abortion rate would have been in the order of 23%, closer to the best results reported. However this is not common practice in UK hospitals; registrars or senior registrars without direct supervision perform a high proportion of abortions, and, as we have shown, these training grades have a lower efficacy rate. This relationship between complete termination rate and grade of surgeon supports the findings of Kaunitz et al. who demonstrated that the rate of continuing pregnancy following surgical abortion by residents (similar to the senior house officer and registrar grades) was 0.42% compared with 0.19% for attending physicians (Kaunitz et al., 1985
).
Overdiagnosis and/or overtreatment of retained products of conception is likely. A senior house officer or registrar alone would have assessed nearly all women on readmission. It may be that doctors too readily submitted women to a surgical re-evacuation when a diagnosis other than incomplete abortion, such as endometritis, was more appropriate. Alternatively, if retained products were present then conservative management, rather than curettage, may have been more appropriate. Unfortunately, as tissue was not routinely sent for pathological examination, we do not have information on the frequency with which retained products were confirmed at re-evacuation. However, a subsequent study from our unit suggests that products of conception are not identified in about 20% of cases of re-evacuation (Heath et al., 2000). In the New Zealand study (Sykes, 1993
) with similar complete abortion rates to ours, products of conception were confirmed in only 55% of cases undergoing curettage. Neither infectious screening nor prophylactic antibiotics, which have been shown to decrease the incidence of post-termination symptoms (Sawaya et al., 1996
), were used routinely in our patients. This management strategy has since been changed.
Other complications occurred with a similar frequency to previous reports. There was only one perforation (0.12%), although this was not recognized at the time of the operation. This is not unusual. Kaali and colleagues demonstrated, in a large study, that the recognized perforation rate at surgical termination was 0.24% (Kaali et al., 1989). However, in patients who then went on to have an immediate laparoscopic sterilization further undiagnosed perforations were found bringing the final perforation rate to 1.98% (Kaali et al., 1989
). Our rate of ongoing pregnancy (0.24%) was remarkably similar to that reported by Kaunitz et al. of 0.23% in their series of 33 090 first trimester suction abortions (Kaunitz et al., 1985
). They noted that when the abortion was performed by a resident then a continuing pregnancy was more likely. Both of our ongoing pregnancies had attempted abortions performed by staff in training. One per cent of patients required admission for intravenous antibiotic treatment for pelvic sepsis. We acknowledge, however, that an additional and unknown number of women would have consulted their general practitioner and been prescribed oral antibiotics for symptoms and signs suggestive of endometritis without presenting to the hospital.
Since this is not a randomized controlled study it is important to attempt the identification of confounding factors that may have affected the results. Potential biases include the assigning of patients to particular grades of surgeons for termination, for instance, cases considered more difficult may have been performed preferentially by a particular grade of surgeon. There could also have been differences in the likelihood of patients returning to the hospital for assessment depending on the grade of operating surgeon. Furthermore, on assessment, there may have been a difference in the willingness to recommend evacuation depending on the seniority of the surgeon who performed the termination. Though it is difficult to assess accurately the effect of these confounding factors each point must be considered. During the study period, two training grade surgeons worked for each consultant, and surgical terminations were performed as part of a general, rather than dedicated, list. Termination procedures were generally shared among the three surgeons. If anything, consultants may have performed cases considered more `difficult' which would tend to increase the incomplete abortion rate for their grade. The grade of surgeon is unlikely to have affected the decision of the patient whether or not to return to hospital, as women were unaware of the identity or grade of the operating gynaecologist. On returning to the hospital for assessment, the grade of surgeon who performed the termination is also unlikely to have affected subsequent management as women complaining of symptoms and signs suggestive of retained products generally had an ultrasound scan. An ultrasound finding of intrauterine material was then the deciding factor on whether to proceed to evacuation. Consequently, we think it unlikely that the above acted as significant confounding factors, although the possibility cannot be excluded.
A recent guideline from the Royal College of Obstetricians and Gynaecologists on induced abortion recommends that `conventional suction termination' be avoided at gestations of <7 weeks (Royal College of Obstetricians and Gynaecologists, 2000). While only 17 (2.1%) of our patients had termination at <7 weeks gestation, all had a complete abortion. The majority of terminations were performed between 810 weeks gestation. This was likely to be due to a combination of consultant preference and also delayed presentation of the patient.
We have shown that the grade of operating surgeon is a significant factor affecting outcome of surgical termination. However, within a grade, the number of surgeons is small, and there is of course variation in the skill level between individuals. Consequently, the poor performance of one or two practitioners could affect outcome for the group. Analysis shows, however, that apart from the senior registrar discussed above, none of the other surgeons' results deviated greatly from the average of the remaining surgeons of the same grade. There was no relationship between number of procedures performed during the year and complete termination rate. It appears that once trained, surgeons require in the region of 30 procedures per year to maintain their skill level.
In conclusion, we have demonstrated that the only significant factors affecting outcome were the grade of operating surgeon or terminations performed at later gestations of 12 to 13 weeks. The proportion of women undergoing a medical rather than surgical termination in England and Wales for unwanted first trimester pregnancy is increasing (Office for National Statistics, 1998). This proportion may increase still further if and when mifepristone becomes licensed for gestations between 9 and 12 weeks as currently these abortions are performed surgically. However, vacuum aspiration continues to be the mainstay of treatment worldwide for first trimester pregnancy termination. It is important that physicians performing surgical terminations are adequately trained.
![]() |
Notes |
---|
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Heisterberg, L. and Kringelbach, M. (1987) Early complications after induced first-trimester abortion. Acta Obstet. Gynecol. Scand., 66, 201204.[ISI][Medline]
Kaali, S.G., Szigetvari, I.A. and Bartfai, G.S. (1989) The frequency and management of uterine perforations during first-trimester abortions. Am. J. Obstet. Gynecol., 161, 406408.[ISI][Medline]
Kaunitz, A.M., Rovira, E.Z., Grimes, D.A. et al. (1985) Abortions that fail. Obstet. Gynecol., 66, 533537.[Abstract]
MacKenzie, I.Z., Jackson, C., McKinlay, E. et al. (1990) Cervical priming before first trimester aspiration abortion: a single blind study comparing gemeprost 1 mg and prostaglandin E2 10 mg pessaries. J. Obstet. Gynaecol., 10, 513517.[Medline]
Nesheim, B-I. (1984) Induced abortion by the suction method, an analysis of complication rates. Acta Obstet. Gynecol. Scand., 63, 591595.[ISI][Medline]
Office for National Statistics (1998) 1997 Abortion Statistics. England and Wales. Series AB No. 24. The Stationery Office, London.
Royal College of Obstetricians and Gynaecologists (2000) The care of women requesting induced abortion. RCOG Press, London.
Sawaya, G.F., Grady, D., Kerlikowske, K. et al. (1996) Antibiotics at the time of induced abortion: the case for universal prophylaxis based on a meta-analysis. Obstet. Gynecol., 87, 884890.
Sykes, P. (1993) Complications of termination of pregnancy: a retrospective study of admissions to Christchurch Women's Hospital 1989 and 1990. N. Z. Med. J., 106, 8385.[ISI][Medline]
Westfall, J.M., Sophocles, A., Burggraf, H. et al. (1998) Manual vacuum aspiration for first-trimester abortion. Arch. Fam. Med., 7, 559562.
Submitted on July 19, 2000; accepted on January 31, 2001.