Intended for healthcare professionals

General Practice

Incidence and outcome of bleeding before the 20th week of pregnancy: prospective study from general practice

BMJ 1997; 315 doi: https://doi.org/10.1136/bmj.315.7099.32 (Published 05 July 1997) Cite this as: BMJ 1997;315:32
  1. Christopher Everetta, general practitioner
  1. a Alton Health Centre, Alton, Hampshire GU34 2QX
  1. Correspondence to: Church Cottage, Holybourne, Alton, Hampshire GU34 4HD
  • Accepted 21 May 1997

Abstract

Objective: To estimate the miscarriage rate in a cohort of pregnant women and the final outcome of pregnancy.

Design: Two year prospective community study.

Setting: Women registered with four semirural practices at one health centre.

Subjects: 626 pregnant women from a population 21 448, 5140 of whom were women aged 15-44 years.

Main outcome measures: Vaginal bleeding and outcome of pregnancy.

Results: 76 of the 89 women with an unwanted pregnancy requested a termination. In the 550 ongoing pregnancies bleeding occurred before the 20th week in 117 (21%), and 67 (12%) ended in miscarriage. The risk of miscarriage was not significantly increased after a miscarriage in the previous pregnancy (11 (15%) women had miscarriage v 55 (12%) women who had not had miscarriage) who had previously had a live birth). Of the 117 women with bleeding, 64 were not admitted to hospital by the general practitioner; 42 of these women had an ultrasound examination at the health centre and 19 subsequently miscarried at home. In hospital 41 of 46 women who miscarried had evacuation of the uterus.

Conclusions: Bleeding occurred in one fifth of recognised pregnancies before the 20th week and over half of these miscarried. Treatment of women with miscarriage at home means current statistics on miscarriage in Britain are missing many cases.

Key messages

  • No national statistics for Britain are published on miscarriages

  • Extrapolations from this survey indicate that in 1993 there may have been 70 000-90 000 miscarriages in England and Wales

  • Bleeding in early pregnancy is followed by a live birth in about half the affected pregnancies

  • At least a quarter of all miscarriages were treated at home by general practitioners and would therefore not be recorded in any published statistics

  • Women who had had a miscarriage did not have a significantly higher chance of a second consecutive miscarriage

Introduction

No statistics about miscarriages are published in Britain except the hospital episode statistics,1 which report only on hospital admissions in England. Bleeding in early pregnancy is common. It occurs in a fifth of all pregnancies and with a miscarriage rate of about 15%2 represents appreciable morbidity in the community. Previous reports on the incidence of miscarriages (spontaneous abortions) in early pregnancy have been carried out on selected groups of women in hospital clinics.3 No published study could be found about the incidence of bleeding in pregnancy, and the only prospective community study of miscarriages was done in Hawaii.4 Retrospective studies after delivery have been shown to be unreliable because of the problem of recall.5

Subjects and methods

The cohort consisted of women with a positive pregnancy test result whose last menstrual period was between 1 January 1989 and 31 December 1990. A weekly check was made on the practice pregnancy test results book, hospital discharge letters about bleeding, and attenders at ultrasound, antenatal, and midwife clinics. Access to notes was possible because all Alton general practitioners are in one building. Data were recorded on an Amstrad PCW 8256 and analysed with Microsoft Foxpro 2.5 and Excel.

Results

During the two year study 657 pregnant women were seen by the general practitioners. Table 1 shows the outcome of their pregnancies and table 2 their previous obstetric history. The average age was 26.7 years. Forty eight women were aged 14-19, 143 were 20-24, 202 were 25-29, 172 were 30-34, 51 were 35-39, and 10 were 40-46 years old.

Table 1

Outcome of pregnancy from conception to 20th week and from 20th week onwards

View this table:
Table 2

Numbers of previous terminations, miscarriages, and live births among 626 women in Alton cohort

View this table:

At the first consultation 234 (37%) of the 626 pregnancies for which data were available at the 20th week were reported as unplanned. Of these, 89 were unwanted, and 70 women requested a termination. Of the 550 ongoing pregnancies, bleeding occurred in 117 (21%). Two confirmed pregnancies were lost without bleeding; one woman subsequently had a normal uterus on ultrasound and the other woman with a missed abortion had an evacuation. Four of eighteen pregnancies survived despite a heavy loss with clots and moderate pain. Bleeding occurred in half the six ectopic pregnancies. Ultrasound examinations were done at our health centre6 to establish fetal viability in 85 of the 117 women with bleeding and 49 women were admitted to hospital (table 3).

Table 3

Hospital admission, outcome, and treatment of 117 women who bled in early pregnancy

View this table:

Table 4 gives the gestational age at miscarriage. The risk of having a second successive miscarriage after a previous miscarriage was not significantly increased in the 74 women who had miscarried previously (15%, 95% confidence interval 8% to 25%) compared with that for women who had had other outcomes (table 5).

Table 4

Week of pregnancy at which 117 miscarriages occurred in 626 pregnancies

View this table:
Table 5

Outcome of last pregnancy and risk of miscarriage in study pregnancy

View this table:

The 20th week of pregnancy was reached by 477 women, after which one late termination was performed on a hydrocephalic fetus. Bleeding was experienced by 10 women, of whom three had bled before the 20th week. No bleeding occurred in 10 sets of twins or the four children born with severe abnormalities (one with exomphalos and three with heart problems).

Discussion

In this community cohort study the miscarriage rate was 12%, which is comparable with previous reports ranging from 11% to 16%.2 7 8 These figures do not include reports on unsuspected early pregnancy loss, which vary from 8% to 22%9 10 or the further 10% of women who do not contact any health professional after a miscarriage.11

Regan et al reported a risk of miscarriage of 19% in women who had had a previous miscarriage compared with 5% in nulliparous women or those who had had a live birth,12 13 14 but this fourfold increase has not been confirmed by other authors.7 15 Selection bias might have occurred in Regan and colleagues' group of 407 volunteers as women whose previous pregnancy had ended in a miscarriage accounted for half of the sample compared with 10.5% (1687/16 015) in a report by Naylor and Warburton16 and 13% (74/550) in this study.

The national incidence of miscarriages has never been published by the Office for National Statistics.17 The Department of Health collects general practitioner item of service claims for miscarriages (FP24/GMS2), and the number of these ranged from 46 440 to 98 640 a year (average 73 230) in England and Wales between 1991 and 1995. The figure would be larger if general practitioners had also recorded unbooked pregnancies under eight weeks' gestation. (Ian Hughes, personal communication)

The Birmingham Research Unit of the Royal College of General Practitioners has 53 general practices throughout England and Wales which report by electronic link on a population of 323 739 (D M Fleming, personal communication). In 1993 they recorded a rate of 55.8 miscarriages/10 000 women aged 15-44. This extrapolates to 60 134 miscarriages among the 10 769 000 women in England and Wales. This total excludes those women admitted directly to hospital and those in whom the diagnosis was delayed.

In my study there were 67 miscarriages and 449 registerable births representing 14.9 miscarriages/100 births. Extrapolation to the 675 000 births in England and Wales18 suggests that there were 70 000-90 000 miscarriages during 1993, assuming a 12.2% miscarriage rate. However, many problems exist with such extrapolations, and the estimate cannot be relied on. There would be an equal number of pregnancies that survived after some bleeding in early pregnancy.

Hospital admissions for bleeding in England only are reported in the hospital episode statistics and averaged 51 000 during 1989-95: the equivalent figure for England and Wales would be 54 000 (Suzanne Dunn, personal communication). Assuming that there are 70 000-90 000 miscarriages a year about 14 000-40 000 (23-40%) women are not included in the current Department of Health statistics and would have been cared for at home.

For many years detailed epidemiological data about every miscarriage treated has been sent to local Family Health Service Authorities by general practitioners in the maternity form FP24/GMS2. As an increasing number of practices are now transmitting this information electronically the system has great potential and could be developed by the Department of Health to produce useful epidemiological data.

In the absence of any official statistics many thousands of women who miscarry may be excluded from important health planning processes. It may also be important to monitor miscarriage rates if environmental influences are capable of altering miscarriage rates.19

Acknowledgments

I am grateful for the advice of Helen Smith and Nigel Baker of the Wessex Research Network, Alison Macfarlane, Professor John Bain, Michael Read, Michael Bull, and others. I also thank Sandy Hall and Hugh Bethell for their help. The computer data were analysed by Jonnie Thornton.

Funding: Scientific Board of the Royal College of General Practitioner and Wessex Research Network.

Conflict of interest: None.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.