Lab Tests Red Cell Distribution Width Test - Coefficient Variant and Standard Deviation

Using SNOMED CT
,
1

Red Cell Distribution Width Tests :- RDW-CV and RDW-SD are both a part of the CBC Panel using by local LIS. In our lab integration attempts, we realized that these were not in the SNOMED International edition. The closest I could find is 66842004 |Red cell distribution width determination (procedure)|.
I am of the view that RDW-CV and RDW-SD should be (observable entity), therefore can’t be “children” of the above concept.

From Medscape: “RDW is an RBC parameter that measures the variability of red cell volume/size (anisocytosis). An increased RDW suggests a higher degree of variability in RBC size, which is useful in classifying certain types of anemia.”
”RDW-SD (expressed in fL) is an actual measurement of the width of the RBC size distribution histogram and is measured by calculating the width (in fL) at the 20% height level of the RBC size distribution histogram. This parameter is therefore not influenced by the average RBC size (Mean Corpuscular Volume (MCV)).”
”RDW-CV (expressed in %) is calculated from standard deviation and MCV

  • RDW-CV (%) = 1 standard deviation of RBC volume/MCV × 100%

Since RDW-CV is mathematically derived from MCV, it is affected by the average RBC size (MCV).”

Now, in the LOINC Ontology, I found
525151010000109 |Distribution width of erythrocyte in blood at point in time (observable entity)| (LOINC Code ID 30385-9), with two children

  1. 613181010000104 |Distribution width of erythrocyte in blood at point in time by automated count (observable entity)| (LOINC Code ID 788-0)
  2. 559581010000109 |Distribution width of erythrocyte in cord blood at point in time (observable entity)|

However, the LOINC code ID 30385-9 has the term description (from LOINC.org) “The RDW-CV is the ratio of the variation in width to the mean width ot the RBC volume distribution curve taken at +/- 1CV.”
The LOINC code ID 788-0 ALSO has the term description The RDW-CV is the ratio of the variation in width to the mean width of the RBC volume distribution curve taken at +/- 1CV.”

For RDW-SD, there are two LOINC code ID’s, and both are deprecated.

  1. 30384-2 Deprecated Erythrocyte distribution width [Entitic volume]
    Term description: The RDW-SD is determined from the RBC volume distribution curve. It is calculated from the width of the curve at 20% of the peak count.
    Mapped to: 30385-9 which as you can see above, has the term description for RDW-CV.
  2. 21000-5 Deprecated Erythrocyte distribution width [Entitic volume] by Automated count
    Term description: The RDW-SD is determined from the RBC volume distribution curve. It is calculated from the width of the curve at 20% of the peak count.
    Mapped to: 788-0 ALSO RDW-CV.

So, my questions are:

  1. In our lab integration attempts, what SNOMED CT concepts do you recommend we map RDW-SD and RDW-CV to? (In the International edition)
  2. Is this worth sending a CRS request for the new concept of RDW-SD, and for synonym RDW-CV to be added to the SCTID 613181010000104 |Distribution width of erythrocyte in blood at point in time by automated count (observable entity)|?

Thanks for any guidance you can provide.

  • Taleya
2

Taleya,

Since the LOINC ontology is a product wholly owned by the Regenstrief Institute, any questions regarding changes to or semantic mappings to concepts within the extension should be sent to LOINC.

You can pose your question at: Contact LOINC – LOINC

1 Like
3

Hi James,

The question for us really was if this was something we should just post to the CRS? Requesting new concepts. We are using SNOMED for Laboratory data in Jamaica.

We only referenced LOINC, as we felt we had to be mindful of the collaboration. Does this make our question clearer?

4

The collaboration between SNOMED and LOINC is one in which SNOMED does a transform of the LOINC release files to SNOMED conformant Observable entity model. If there are perceived issues with the content of the LOINC concepts, these need to be addressed at the source as we do not (cannot) make changes to the LOINC ontology directly. I hope that makes sense. It sounds a bit round about, but our agreement with LOINC requires that we only make changes in the LOINC distribution files as we are not modeling directly in SNOMED.