search

obophenotype / human-phenotype-ontology

Ontology for the description of human clinical features
http://obophenotype.github.io/human-phenotype-ontology/
Other
293 stars 51 forks source link

Stillbirth/miscarriage in different subontologies #6842

Closed pdl closed 2 years ago

pdl commented 3 years ago

HPO term https://hpo.jax.org/app/browse/term/HP:0005268 Spontaneous abortion (miscarriage)

Suggested revision and reasons In order to be able to assign HPO terms for patients when they have died, we have looked at this term, but it does not sit within the Aging/Mortality tree https://hpo.jax.org/app/browse/term/HP:0011420 which does not have a term for a foetus miscarried before 20 weeks, which is when stillbirth https://hpo.jax.org/app/browse/term/HP:0003826 is defined to start.

It's not totally clear to me from the description whether this term can used on the parent or the child or both, but it looks as if it is only for the parent as it has a sub-term which is 'Recurrent spontaneous abortion' HP:0200067, which can only be applied to the parent; if so, another term might be needed instead.

Note also that I understand from clinicians that the cutoff for when the event is classed as a miscarriage varies between countries - US being 20w and UK being 24 ( https://www.nhs.uk/conditions/miscarriage/ ), so it might be preferable to have a term for pregnancy loss in general with stillbirth and miscarriage being subterms.

pnrobinson commented 3 years ago

As currently implemented in HPO, the term Spontaneous abortion HP:0005268 refers to the mother, and the term Stillbirth HP:0003826 refers to the fetus. We are currently in the process of hosting workshops on the prenatal phenotype with the intention of improving the HPO's representation of these concepts. I will add the prenatal label to that and we will collect suggestions/feedback. You would be welcome to join these discussions -- if so, please send Peter Robinson an email (the email can be found on the HPO page).

pnrobinson commented 2 years ago

@pdl sorry for the delayed response. Looking at this, I think that we need to update the definitions for recurrent spontaneous abortion (to refer to the mother). The terminology for stillbirth/miscarriage is not uniform across countries, but I agree that it would be good to work on terms to cover prenatal demise more comprehensively. Spontaneous abortion seems to partially overlap with stillbirth, and we will need to decide how to manage the concepts. If you have literature or suggestions, for instance, existing classifications, that might be of use, please post them.

pdl commented 2 years ago

Thanks @pnrobinson - I don't have anything to hand, but I know a few people who might and will ask.

pdl commented 2 years ago

Following your request I have been in contact with Luke Richardson and Laura Yates who are involved in a project called ConcePTION ( https://www.imi-conception.eu ) who have provided some very helpful pointers.

I'll start by listing below a number of terms I've come across looking into this.

As part of their data model they settled on a 22-week figure, comparing the FDA, EMA, and RCOG cutoffs: "Upon discussion with ConcePTION partners, it was determined that the definition supplied by the EMA was likely most widely applied in existing datasets, so 22 weeks gestational age was used as the cut-off.". Their initial proposals for their data model is at https://www.imi-conception.eu/wp-content/uploads/2020/04/D2.3-Report-describing-exisiting-coding-systems-schemes-and-regulatory-guidelines-of-reported-medication-exposed-pregnancies.pdf - p119 for relevant definitions - though Laura writes that that document is an older version, and a working document is "continually being revised and refined as we gain experience from the ConcePTION demo" (and will be published in due course).

Luke writes:

It is possibly worth noting that our work package in the ConcePTION project has focussed on European primary source pregnancy pharmacovigilance datasets (where exposure and outcome data are sampled from the population), and different definitions are used in the US (I believe it is 20 weeks), and also in the UK (the RCOG define the cut off as 24 weeks).

The EMA definition currently in use is at https://www.ema.europa.eu/en/documents/regulatory-procedural-guideline/guideline-exposure-medicinal-products-during-pregnancy-need-post-authorisation-data_en.pdf : The section ANNEX 4 - DEFINITIONS is relevant

Luke also pointed me to a draft open for comment at https://www.ema.europa.eu/en/documents/scientific-guideline/draft-guideline-good-pharmacovigilance-practices-product-population-specific-considerations-iii_en.pdf which has a terminology section P.III.A.2.Terminology.

In the UK, there is a 24-week threshhold in the law for registration of stillbirth - RCOG guidelines https://www.rcog.org.uk/media/030jxnk5/goodpractice4registrationstillbirth2005.pdf bear this out. The RCOG have other guidelines which may be relevant such as "Late intrauterine foetal death and stillbirth" https://www.rcog.org.uk/guidance/browse-all-guidance/green-top-guidelines/late-intrauterine-fetal-death-and-stillbirth-green-top-guideline-no-55/ and "Recurrent Miscarriage" https://www.rcog.org.uk/media/3cbgonl0/gtg_17.pdf.

Laura writes

This inconsistency across definitions underscores the importance of collecting raw (rather than interpreted) information ie. The gestational week of pregnancy loss and then classifying as a second step.

Luke and Laura's details are as follows:

Dr Jonathan Luke Richardson PhD ORCID ID: 0000-0003-3494-9559

  1. Senior Medical Information Scientist, UK Teratology Information Service, Newcastle upon Tyne Hospitals NHS Foundation Trust, Newcastle upon Tyne, UK.
  2. Associate Researcher, Population Health Sciences Institute, Newcastle University, Newcastle upon Tyne, UK.

Dr Laura M. Yates MBChB, DRCOG, MRCPCH, PhD, CCT Clinical Genetics ORCID ID:0000-0002-7963-625X

  1. Honorary Senior Lecturer KRISP, University of KwaZulu-Natal University, South Africa
  2. (Hopefully soon Locum Consultant Geneticist, IGM, Newcastle upon Tyne Hospitals NHS Foundation Trust, Newcastle upon Tyne, UK)

It strikes me that the most important thing for HPO as an ontology is to have a broad term which covers all prenatal death. "Foetal death" appears to me to be the most used term which covers this most broadly and neutrally (the phrase "foetal demise" you use above seems a bit more satisfying to me but I've not come across it elsewhere). Secondarily, picking up signal on how early that event occurs - there seems to be a consensus that there is an early and a late phase, but disagreement on where that boundary sits.

You could construct a structure which aligns with particular weeks in order for HPO to be able to represent some of the key cutoffs being used in practice:

I suspect there may be another useful boundary at about 12 weeks, but I have no evidence for this.

I would apply the following synonyms to Foetal death:

Each of the 'early' terms (in first n weeks) would want the following synonyms

The late terms (after first n weeks) would also want the following synonyms

In terms of the non-time-bound terms, I'm not sure if "Molar pregnancy" makes sense as a foetal phenotype. "Ectopic pregnancy and "Blighted ovum" might, as would "Induced termination" (with synonyms of "termination of pregnancy" "induced abortion", "elective abortion"), as child of Foetal death.

pnrobinson commented 2 years ago

I am adding a term Fetal death (Note we include UK spelling as a synonym type with the label "UK spelling", i.e., Foetal death etc) --Stillbirth --Miscarriage

The onset and death terms in the HPO are not meant to capture the precise dates of individual cases. For this we recommend the new GA4GH Phenopacket. The HPO terms are intended to provide the general age range in ways that can be used by algorithms. So I would suggest for now that we stick with the above three terms. A possible reason for extending the hierarchy in the HPO would be if the information of say 20 vs 22 vs 24 weeks is useful for the differential diagnosis (I do not think this is the case). Otherwise, this information can be better coded using a phenopacket for individual patients. I will add synonyms with appropriate modifications.

I have left "Recurrent spontaneous abortions" as a maternal phenotype.

Thank you for these suggestions! They will be available at the next HPO release next week.

PaulNSchofield commented 2 years ago

Just late to this... I do think that first trimester spontaneous pregnancy loss, second trimester spontaneous loss etc. would be useful as they tend to differentiate between placental and cardiovascular origins. this could also be articulated as first trimester fetal death etc.