The conservative management of first trimester miscarriages and the use of colour Doppler sonography for patient selection

Peter Schwärzler1,3, Des Holden1, Sven Nielsen2, Mats Hahlin2, Povilas Sladkevicius1 and Thomas H. Bourne1

1 Department of Obstetrics and Gynaecology, St George's Hospital Medical School, Blackshaw Road, London SW17 0QT, UK and 2 Department of Obstetrics and Gynaecology, Sahlgrenska University Hospital, Göteborg, Sweden


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
This study of patients with first trimester miscarriage evaluates whether conservative management is a feasible strategy and assesses the value of colour Doppler ultrasonography for patient selection. After confirmation of the diagnosis by transvaginal sonography all patients were offered the choice of immediate dilatation and curettage or conservative management. The presence of a gestational sac, the occurrence of spontaneous complete miscarriage within 28 days, detectable pulsatile blood flow within the placenta in the presumed region of the intervillous space and post-treatment complications were the main end-points. Out of a total of 108 women recruited, 23 (21.3%) elected to undergo immediate dilatation and curettage and 85 (78.7%) chose conservative management. The treatment groups were similar in age, gestational age, gestational sac diameter, serum concentrations of human chorionic gonadotrophin (HCG) and progesterone, and proportion of patients who had post-treatment complications (12–13%). Of patients in the conservative management group, 71 out of 85 (84%) had a spontaneous, complete abortion, while 37 out of 46 cases (80%) with detectable presumed intervillous pulsatile blood flow had a complete, spontaneous abortion within 1 week; this occurred in 23% of cases with no detectable flow. This suggests that conservative management is a successful approach for many patients with first trimester miscarriage; colour Doppler ultrasonography can be used to select the most suitable patients for this strategy, and thus reduce the need for hospital admission and surgery.

Key words: anembryonic pregnancy/colour Doppler ultrasonography/conservative management/missed abortion/spontaneous miscarriage


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Miscarriage is the most common complication of early pregnancy, and remains an important clinical problem. Knowledge of the pathophysiology of this condition is limited and therefore therapeutic regimens are based on few data. The overall early pregnancy loss rate is thought to be ~50%. At least 15% of fertilized ova are lost before implantation (Little, 1988Go) and 20–25% of pregnancies end before they can be clinically detected (Wilcox et al., 1988Go), leading to an incidence of spontaneous abortion of some 12–19% among clinically recognized pregnancies (Giacomucci et al., 1994Go). The management of miscarriage has changed little this century and conventional surgical evacuation of the uterus has been recommended when there are retained products of conception (McKee et al., 1992Go; Rosevear, 1994Go). However, the technique carries a small but real risk of morbidity and mortality (MacKenzie and Bibby, 1978Go). In recent years, the medical management of miscarriage, which can achieve complete uterine evacuation in 95% of early miscarriages, has been developed as a realistic alternative to surgical evacuation (el-Refaey et al., 1992; Henshaw et al., 1993Go). The method involves the combined use of the antiprogesterone, mifepristone, and the prostaglandin E1 analogue, misoprostol, and has been shown to result in lower costs per patient, compared with surgical treatment (Hughes et al., 1996Go).

More recently, it has been proposed that selected cases of spontaneous abortion can be managed using a `watch and wait' or conservative management strategy (Dickey, 1993Go; Haines et al., 1994Go). In a further randomized study, conservative management was used to treat miscarriages at <13 weeks gestation. Inclusion in this study depended on transvaginal ultrasonography showing an anterior–posterior (A–P) diameter of the endometrial cavity of 15–50 mm. Conservative management was successful in 79% of women, without any increase in complications compared with those who underwent dilatation and curettage (D&C) (Nielsen and Hahlin, 1995Go).

We used transvaginal ultrasonography (TVS) and biochemical parameters to evaluate the feasibility of conservative management for first trimester miscarriages. Our study differs in that inclusion was limited to pregnancies with what were thought to be either missed abortions or anembryonic pregnancies. Furthermore, we assessed the value of colour Doppler imaging (CDI) to identify those patients with failing early pregnancies who would be most suitable for conservative management.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
The aim was to include 100 consecutive patients who fulfilled the inclusion criteria. Patients were offered two alternative possibilities in order to manage their miscarriage: either a D&C to empty the uterus (D&C group), or to be rescanned again at weekly intervals up to 1 month to assess for spontaneous resolution of the miscarriage (conservative management group). If miscarriage had not occurred after 1 month, the woman was offered a D&C.

At inclusion, an ultrasound examination including CDI was performed, and blood was taken to assess the full blood count (FBC), serum concentrations of human chorionic gonadotrophin (HCG) and serum progesterone. Patients who were rhesus-negative received anti-D immunoglobulin (125 µg). Paracetamol 1 g in combination with codeine 60 mg was prescribed to alleviate any pain.

The ultrasound examination including CDI, as well as the assessment of serum indices was repeated at intervals of ~1 week, up to 28 days. The intervals were 5–7, 12–14, 21–23 and 28 days after inclusion into the study. If the patient had an empty uterus, a further examination with ultrasound including CDI and blood samples was carried out 12–14 days after inclusion. If evidence of retained products of conception was found (endometrial cavity A–P diameter of >10 mm) at the time of any of the examinations, the options of either continuing with conservative management or undergoing a D&C were given to the patient. If the scan suggested the persistent presence of retained products 28 days after inclusion, the patient was advised to have a D&C. All patients gave written informed consent, and the study protocol was approved by our local Ethics committee.

Subjects
A total of 108 symptomatic patients attending the early pregnancy assessment unit at St George's Hospital, London, UK, were recruited into the study. Pregnancies with an estimated gestation age of <13 weeks from the last menstrual period (LMP) were included. All women included in the study were symptomatic and had reported vaginal bleeding and/or abdominal pain. Inclusion was restricted to those shown by ultrasonography to have either a missed abortion or an anembryonic pregnancy.

Definitions
For the purposes of this study, a missed abortion was defined as the presence of an intrauterine gestation sac containing an embryo of >4.0 mm length, but with no heartbeat. An anembryonic pregnancy was defined as the presence of an empty gestation sac of >20 mm with no visible embryonic structures. Complete miscarriage was defined as having an endometrial cavity thickness of <10 mm and a negative urinary pregnancy test (Clearview, 50 IU/l; Unipath Ltd, Bedford, UK) or serum HCG value of <50 IU/l. The miscarriage was classified as incomplete if the endometrial cavity thickness was >10 mm and/or a gestational sac was present.

Transvaginal ultrasonography and colour Doppler measurements
An Aloka SSD 2000 scanner (Aloka Ltd, Tokyo, Japan) was used for all scans. A vaginal probe incorporating a 5 MHz transducer was used for B-mode and CDI, and for the production of flow velocity waveforms. B mode imaging was used to measure the maximum gestation sac diameter. The presence or absence of an embryo was noted and the crown–rump length measured when appropriate. The thickness of the endometrial cavity was the maximum distance across the cavity in the longitudinal plane, including any tissue or clot that might be present. The maximum diameter, side, and morphology (solid or cystic) of the corpus luteum was recorded.

For CDI, the return filter was set at 50 MHz. Uterine arteries were located above the supravaginal portion of the cervix. A pulsed Doppler range gate was placed over each artery in sequence and flow velocity waveforms generated. The resistance index (RI) and pulsatility index (PI) were measured: RI = (A – B)/A and PI = (A – B)/TAMXV; where A is the peak Doppler shifted frequency, B is the minimum Doppler shifted frequency, and TAMXV the time-average maximum velocity over the cardiac cycle. The angle-corrected peak systolic velocity (PSV cm/s) was also measured. The spiral arteries were recognized by their position at the base of the placenta. This entire area was surveyed by CDI. The power, gain and pulse repetition frequency were adjusted for maximum sensitivity for low velocity blood flow. The lowest velocity signals were gradually filtered out by increasing the pulse repetition frequency, and flow analysis carried out on the highest velocity signals. After positioning the pulsed Doppler range gate, minute adjustments were made to the angle of the probe until the visible and audible signals were considered optimal. The position was considered to be at the optimal angle for the probe at that location. The value with the highest PSV and corresponding TAMXV, PI and RI was recorded. A subjective assessment was made of the presence or absence of blood flow in the presumed intervillous space. In the presence of blood flow, Doppler measurements were taken using the same approach as described above. Finally, the corpus luteum was assessed for size, morphological appearance and the presence of blood flow, the highest velocity of which was recorded.

Outcome measures
The principal outcome measures were the presence or absence of a gestation sac in the uterine cavity, the thickness of the endometrial cavity, and the presence or absence of blood flow in the presumed intervillous space. The frequency of complications was also recorded. Pain was recorded using a linear scale from 1 (little or no pain) to 5 (intolerable pain).

Statistical analysis
The probability of complete spontaneous miscarriage during 1 week of conservative management was estimated by logistic regression analysis. The logistic model was applied in a stepwise manner taking into consideration all diagnostic measures included in Table IGo. The procedure included a new variable in each step and was stopped when no additional variable contributed significantly to the diagnostic power.


View this table:
[in this window]
[in a new window]
 
Table I. Characteristics of patients in both study groups at inclusion. Values are shown as means, with SD or range in parentheses
 

    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Of the 108 women recruited, 23 (21.3%) patients elected to undergo an immediate D&C, and 85 (78.7%) opted for conservative management. The two treatment groups were similar in age, gestational age of the pregnancy, gestational sac diameter and serum HCG and progesterone concentrations (Table IGo).

Outcome
By 4 weeks after inclusion, 71 (84%) patients in the conservative management group had undergone a spontaneous complete miscarriage and 14 (16%) had undergone D&C. A breakdown by gestational age and ultrasound characteristics of the miscarriage is shown in Table IIGo. For each gestational age category, the proportion of patients in the expectant management group having a spontaneous complete miscarriage after 1, 2, 3 and 4 weeks after inclusion is shown in Table IIIGo. The chances of the miscarriage resolving were independent of gestational age. After 7 days, 46 (54%) women had undergone a complete miscarriage. Of those women with an incomplete miscarriage, 33 (39%) opted to continue with conservative management and six (7%) for a D&C. The cumulative figures for days 7, 14, 21 and 28 are shown in Figure 1Go. The number of women who had undergone a complete miscarriage increased significantly each week up to 28 days.


View this table:
[in this window]
[in a new window]
 
Table II. Occurrence of spontaneous complete miscarriage associated with gestational age and ultrasound characteristics in the conservative management group. Figures in parentheses are percentages
 

View this table:
[in this window]
[in a new window]
 
Table III. Proportion of patients in the conservative management group having spontaneous expulsion after 1, 2, 3, and 4 weeks related to gestational age. Numbers in parentheses are those of patients in the expectant management group opting for a dilation and curettage (D&C)
 


View larger version (20K):
[in this window]
[in a new window]
 
Figure 1. Cumulative miscarriage rates in the two groups of patients.

 
Colour Doppler results
Indices of blood flow in the uterine arteries, spiral arteries, and corpus luteum were not significantly different between the D&C and conservatively managed groups. For those in the conservative management group, there were no differences in indices between those women who underwent a complete resolution of their miscarriage, and those who needed a D&C. In the conservatively managed group, the presence of blood flow in the presumed intervillous space was associated with the likelihood of complete spontaneous abortion within 7 days. Of pregnancies with positive presumed intervillous flow, 80% underwent spontaneous resolution, whereas this only occurred in 23% of those where presumed intervillous flow was absent (Table IVGo).


View this table:
[in this window]
[in a new window]
 
Table IV. Detection of pulsatile flow in the intervillous space and occurrence of complete miscarriage without intervention within 7 days. Figures in parentheses are percentages
 
Logistic regression analysis
Employing a stepwise logistic regression procedure two variables of the patients' characteristics at inclusion possessing prognostic power were identified: progesterone and HCG concentration. For each variable, an estimated coefficient in the logistic model was obtained. The coefficients obtained for the two variables and the linear combination in the logistic model are presented in Table VGo.


View this table:
[in this window]
[in a new window]
 
Table V. Logistic regression analysis (multivariate analysis) showing the prognostic power of progesterone and human chorionic gonadotrophin (HCG) to predict the occurrence of spontaneous miscarriage within 7 days
 
Complications
There were no significant differences in the frequency of complications between conservative management and D&C groups. Five (6%) patients in the conservative management group requested a D&C due to heavy bleeding and/or intolerable pain. Three patients (13%) of the D&C group and seven patients (8%) of the conservative management group suffered from one or more complications requiring hospital admission. Only one woman required a blood transfusion (D&C group). Details of the complication rates in the two groups are shown in Table VIGo.


View this table:
[in this window]
[in a new window]
 
Table VI. Complications experienced by patients in the two groups. Figures in parentheses are percentages
 

    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
The majority of women diagnosed with a first trimester miscarriage still undergo a D&C with the aim of avoiding potential complications such as bleeding, pain and infection. However, D&C is not without complications (Glenc, 1974Go) and there is increasing evidence that many women with early pregnancy failure will undergo a complete miscarriage within an acceptable period of time (Nielsen and Hahlin, 1995Go). Furthermore, this approach does not appear to affect future fertility (Blohm et al., 1997Go).

It has therefore been of interest to develop techniques that can be used to select suitable patients for a conservative management strategy. Nielsen et al. (1996) successfully used stepwise logistic regression analysis of five diagnostic variables: serum concentrations of progesterone, CA-125, {alpha}-fetoprotein and daily changes in serum HCG values and intrauterine diameter to assess the likelihood of spontaneous complete miscarriage occurring. Their results suggested that the outcome in cases of spontaneous miscarriages managed conservatively was associated with the viability of the corpus luteum and trophoblast (as indicated by concentrations of serum HCG and progesterone). However, the inclusion criteria for this study were based on an ultrasound examination that showed `an inevitable or incomplete spontaneous abortion with intrauterine tissue with an A–P diameter 15–50 mm'. Such criteria must lead to the inclusion of women with only a small amount of decidual tissue, blood clots, or residual villi in the endometrial cavity. This led to criticism of the validity of the study (Ankum and Van der Veen, 1995Go). Accordingly our study was restricted to pregnancies where ultrasound had shown the presence of either a missed abortion or anembryonic pregnancy. Even with these criteria, 54% of women underwent a complete miscarriage within 1 week. It is of interest that by 28 days, this number had increased to 84%. Women opting for a conservative management approach can, therefore, be told that prolonging the time of observation is associated with a realistic chance of avoiding surgical intervention. Those women who were followed up longest did not have an increased number of short-term complications. Long-term morbidity, such as posterior fertility, has not been assessed in our study. However, it has been excluded in a previous follow-up study, which showed no difference in mid- and long-term complication between conservative management and D&C (Blohm et al., 1997Go). It would seem that the follow-up time of 3 days in the Swedish study (Nielsen and Hahlin, 1995Go) was over-restrictive. The acceptability of such long periods of observation will depend very much on the patient population studied; however, we feel that having data that allow the clinician to inform a patient about the likely success of conservative management is likely to increase the number of women opting for this approach. Clearly not all women would find 28 days an acceptable period to wait. Two-thirds will complete the miscarriage within 14 days; perhaps a more acceptable duration for women in our local population. That 11 patients in our study opted for a D&C when a follow-up scan revealed products of conception were still present, may reflect our inability to give sufficient reassurance to these women that complete resolution of their miscarriage could still to be expected.

Our findings in relation to the vascularity of the presumed intervillous space are of both clinical and physiological relevance. We have observed a close relationship between the presence of blood flow in the presumed intervillous space and spontaneous complete abortion occurring within 7 days. Conversely, when blood flow was absent in the region, spontaneous complete miscarriage was unusual. This association was independent of gestational age. This finding can be seen against the background of previous physiological studies. Hustin and Schaaps (1987) used perfused hysterectomy specimens from first trimester pregnancies to show that in early pregnancy, trophoblast plugs occlude the openings of the uteroplacental arteries, and that the intervillous space is avascular. More recently, a study of the vascularity of first trimester pregnancies using transvaginal CDI showed an increase in blood flow within the placenta of complicated, compared with normal intrauterine pregnancies (Jauniaux et al., 1994Go). An examination of the histological features of the placenta of pregnancies of <12 weeks gestation, which had been shown to have blood flow within the intervillous space using CDI, showed extended dislocation of the trophoblast shell and a massive infiltration of the intervillous space and placental bed by maternal blood. The authors concluded from their study that premature entry of maternal blood into the intervillous space disrupts the materno–embryonic interface, and is probably the final mechanism causing abortion. Our data support these observations. We have shown prospectively that the spontaneous miscarriage of either missed abortions or anembryonic pregnancies is associated with the presence of blood flow in the presumed intervillous space. In contrast, in apparently similar pregnancies that do not resolve spontaneously, blood flow in the presumed intervillous space is usually absent. These data support the view that miscarriage, although not necessarily embryonic demise, follows the breakdown of the normal embryo–maternal interface. If blood does not enter the intervillous space, it seems that the trophoblast continues to function, and the spontaneous resolution of a failed early pregnancy does not occur.

In summary, we conclude that conservative management is a feasible strategy for many patients with spontaneous first trimester abortion. This includes both missed abortion and anembryonic pregnancies. Therefore, women can be advised that they can reasonably expect to undergo a complete miscarriage within 28 days from the time of diagnosis of pregnancy failure. Furthermore, in our study the presence of detectable flow within the presumed intervillous space was predictive of spontaneous complete miscarriage. Accordingly, CDI may be used to select those patients most suitable for conservative management.


    Acknowledgments
 
The authors would like to thank Gill Costello, midwife, and Suzanne Mathis, medical student, for their enthusiastic help during the study. We are also grateful to the Aloka Company (Japan) and Keymed Ltd (Southend-on-Sea) for the use of their ultrasound equipment. This work was supported by a grant of the Austrian FWF Foundation (J01428).


    Notes
 
3 To whom correspondence should be addressed Back


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Ankum, W.M. and Van der Veen, F. (1995) Management of first-trimester spontaneous abortion. Lancet, 345, 1179–1180.[Medline]

Blohm, F, Hahlin, M., Nielsen, S. et al. (1997) Fertility after a randomised trial of spontaneous abortion managed by surgical evacuation or expectant treatment. Lancet, 349, 349.

Dickey, R.P. (1993) Management of uncomplicated miscarriage. Patients safe with expectant management. Br. Med. J., 307, 259.

el-Refaey, H., Hinshaw, K., Henshaw, R. et al. (1992) Medical management of missed abortion and anembryonic pregnancy. Br. Med. J., 305, 1399.[ISI][Medline]

Giacomucci, E., Bulletti, C., Polli, V. et al. (1994) Immunologically mediated abortion (IMA). J. Steroid. Biochem. Mol. Biol., 49, 107–121.[ISI][Medline]

Glenc, F. (1974) Early and late complications after therapeutic abortion. Am. J. Obstet. Gynecol., 118, 34–35.[ISI][Medline]

Haines, C.J., Chung, T. and Leung, D.Y. (1994) Transvaginal sonography and the conservative management of spontaneous abortion. Gynecol. Obstet. Invest., 37, 14–17.[ISI][Medline]

Henshaw, R.C., Cooper, K., el-Refaey, H. et al. (1993) Medical management of miscarriage: non-surgical uterine evacuation of incomplete and inevitable spontaneous abortion. Br. Med. J., 306, 894–895.[ISI][Medline]

Hughes, J., Ryan, M., Hinshaw, K. et al. (1996) The costs of treating miscarriage: a comparison of medical and surgical management. Br. J. Obstet.Gynaecol., 103, 1217–1221.[ISI][Medline]

Hustin, J. and Schaaps, J.P. (1987) Echographic and anatomic studies of the maternotrophoblastic border during the first trimester of pregnancy. Am. J. Obstet. Gynecol., 157, 162–168.[ISI][Medline]

Jauniaux, E., Zaidi, J., Jurkovic, D. et al. (1994) Comparison of colour Doppler features and pathological findings in complicated early pregnancy. Hum. Reprod., 9, 2432–2437.[Abstract]

Little, A.B. (1988) There's many a slip `twixt implantation and the crib. N. Engl. J. Med., 319, 241–242.[ISI][Medline]

MacKenzie, J. and Bibby, J. (1978) Critical assessment of dilatation and curettage in 1029 women. Lancet, ii, 566–568.

McKee, M., Priest, P., Ginzler, M. et al. (1992) Can out-of-hours operating in gynaecology be reduced? Arch. Emerg. Med., 9, 290–298.[ISI][Medline]

Nielsen, S. and Hahlin, M. (1995) Expectant management of first-trimester spontaneous abortion. Lancet, 345, 84–86.[ISI][Medline]

Nielsen, S., Hahlin, M. and Oden, A. (1996) Using a logistic model to identify women with first-trimester spontaneous abortion suitable for expectant management. Br. J. Obstet. Gynaecol., 103, 1230–1235.[ISI][Medline]

Rosevear, S.K. (1994) Bleeding in early pregnancy. In James, D.K., Steer, P.J., Weiner, C.P. and Gonik, B. (eds), High Risk Pregnancy. Vol. 7. Saunders Company Ltd, London, UK, pp. 75–100.

Wilcox, A.J., Weinberg, C.R., O'Connor, J.F. et al. (1988) Incidence of early loss in pregnancy. N. Engl. J. Med., 319, 189–194.[Abstract]

Submitted on June 25, 1998; accepted on November 24, 1998.