Letters Diagnosing miscarriage

Authors’ reply to Fountain and colleagues

BMJ 2015; 351 doi: https://doi.org/10.1136/bmj.h5852 (Published 04 November 2015) Cite this as: BMJ 2015;351:h5852
  1. Tom Bourne, professor and consultant gynaecologist1,
  2. Jessica Preisler, clinical research fellow1,
  3. Ben Van Calster, professor2,
  4. Dirk Timmerman, professor2
  1. 1Queen Charlotte’s and Chelsea Hospital, Imperial College, London W12 0HS, UK
  2. 2Department of Development and Regeneration, KU Leuven, Belgium
  1. tbourne{at}imperial.ac.uk

We appreciate Fountain and colleagues’ comments.1 2 We did consider how the dates of the last menstrual period might affect a possible diagnosis of miscarriage, and we draw readers’ attention to the sub-section of the results that deals with this—the revised criteria for diagnosing miscarriage contained in table 5 and fig 2.

We did not consider menstrual cycle length as a variable. We thought that making assumptions about the timing of ovulation in these circumstances, coupled with the known variations in the ovulation-implantation interval,3 would make this difficult to integrate into reliable guidelines that would be 100% specific.

The date of a first positive pregnancy test is certainly worth consideration, although we did not have the data in our study to examine this. This concept of “minimal menstrual age” was described in a retrospective analysis,4 and as might be expected is influenced by when the pregnancy test is carried out. Unfortunately this approach has not been externally validated or assessed prospectively in a large enough cohort for its test performance, with or without other markers of miscarriage, to be known.

Finally, it is of course important that clinicians and women have realistic expectations about what can be visualised using ultrasound in early pregnancy. We refer to this point in the discussion section of the paper and cite our paper on the timing of ultrasound scans, pointing out the intrauterine pregnancy of uncertain viability rate at different gestations.5 That paper showed that it is reasonable to delay the scan to an optimal gestation of 49 days in women with no symptoms. But for women with pain or bleeding, or those at high risk, as was the case for most women in our study, ectopic pregnancies would be missed if such a policy were applied. We have also published prediction scores and models that can be used in clinic to give women information about the likely outcome of their pregnancy if a follow-up scan is needed.6

Currently, in the UK and elsewhere, neither cycle length, nor the date of the first positive pregnancy test, nor gestational age form any part of national guidance on which to base a diagnosis of miscarriage. Indeed NICE guidance specifically states: “Do not use gestational age from the last menstrual period alone to determine whether a fetal heartbeat should be visible.” It also advises clinicians to “inform women that the date of their last menstrual period may not give an accurate representation of gestational age because of variability in the menstrual cycle.”7

Therefore, although the points raised by Fountain and colleagues are of interest, in the context of current guidelines they play no role in the diagnosis of miscarriage. We would welcome any large prospective studies to investigate whether any of the features they mention can reliably improve the diagnostic test performance of the ultrasound based criteria currently used to diagnose miscarriage.


Cite this as: BMJ 2015;351:h5852


  • Competing interests: None declared.


View Abstract

Sign in

Log in through your institution