How this conflation took place within the scientific record is eye-opening. Unfamiliar terms create room for misunderstandings that familiar terms would never allow. So let’s first use both a familiar context and familiar terms to outline the basic linguistic mechanism at work. Think of it as a kind of “fact-ectomy” — removing a detail, from an equation, by supplying a more general term, in the new text. When this hole in the text-body “heals” over time, the surface conclusion has moved to a new location, in this case from 1:500 to 1:200. The whole process however is only literary; the underlying population has not changed at all, only our perception of it. Next I will present a simple illustration of how this works. We will then follow up with the actual case.
|Year||Studies||Detail Added to or Dropped from the scientific record|
|1st report||Nobel prize winners Jane and John Doe make a first-of-its-kind estimation of the population of domesticated Horses in SomeCountry: 1,000,000.||Nothing is dropped. A discovery is added to the scientific record.|
|2nd report||Breakdown of prior studies under the category, Equidae:
· Domesticated Horses: 1,000,000
· Zebras: 500,000
· Donkeys: 250,000
Under the umbrella-term, Equidae, we have a total population of 1,750,000
|Nothing is dropped. A useful aggregation of prior research is added to the scientific record.|
|3rd report||Explicit combining of the terms “Domesticated Horses,” “Zebras,” and “Donkeys” — now all are called “Horses.”||Use of Equidae as the umbrella-term for the constituents is dropped for an overlapping use of the word, “Horse.” Although explained, ambiguous usage of “horses” is added to the scientific record, a subtraction from the clarity of the previous unambiguous usage.|
|4th report||Implicit combining of Domesticated Horses, Zebras, and Donkeys – all are implied to be “Horses.”||The explicit reference to blending the terms is dropped, and the idea of their possible combination is only available to those who pay close grammatical attention to ambiguity. Without such attention, the assumption that “horses” are, and always have been, Zebras and Donkeys will prevail. Substantial detail is subtracted from the scientific record. Of course this is ludicrous in this familiar context, because we are familiar with the terms. Outside of a familiar vocabulary, however, the problem is more difficult to see.|
|5th report||Now Horses are explicitly defined as Domesticated Horses, Zebras, and Donkeys. This is conflation – and we do mean this in the linguistic sense of the term: conflation is not a “scientific” concept.
Even though the same 1,000,000 domesticated horses that the Nobel Prize winners counted are still the same 1,000,000 domesticated horses in the most recent study, they have simply been added to 500,000 zebras and 250,000 donkeys.
|The implicit reference to the combining of separate terms is dropped. Zebras and Donkeys are now linguistically conflated with Horses.
Critical detail is subtracted from the scientific record.
Details are dropped from one publication to another, and this allows a “conclusion drift” from the specific definition of domesticated horses to a general, stretched definition of “Horses.” Due to successful conflation, recent scientists declare that the estimates of the Nobel Prize winners fail to compare with the “new prevalence” of 1,750,000 horses. Even though the constituent of domesticated horses within that 1,750,000 is still 1,0000,000.
|1973: Nobel Prize winner Goldstein, et al. https://doi.org/10.1172/JCI107332.||LDLR = FH with prevalence of 1:500 ~ 1:1,000||Nothing is dropped. Groundbreaking work is added to the scientific record.|
|2003: Rader et al
|Breakdown of recent discoveries under the general category, ADH:
· LDLR = FH 1:500
· APOB = FDB 1:1,000
· PCSK9 = FH3 1:2,500
|Nothing is dropped. Detail is aggregated to the benefit of the scientific record.|
|2011: “Although the term FH has, in the past, been used to refer specifically to LDL receptor (LDLR) defects, this document will use a broader definition to reflect discoveries of defects in the genes for apolipoprotein (Apo) B, proprotein convertase subtilisin/kexin type 9 (PCSK9),” Goldberg, et al. DOI:10.1016/j.jacl.2011.03.001||Explicit combining of the terms FH, FDB and FH3 as “FH”||Use of ADH as the umbrella-term for the constituents FH, FDB and FH3 is dropped. A degree of clarity is subtracted from the scientific record as ambiguity is added: cultural use of FH is acknowledged to refer exclusively to the LDLR, but is not here restricted to the LDLR.|
|2012: “FH is an autosomal dominantly inherited disorder caused primarily by mutations in the gene encoding the low-density lipoprotein (LDL) receptor, LDLR. Less frequent mutations in the APOB and PCSK9 genes have similar functional consequences.” Benn, et al. DOI:10.1210/jc.2012-1563||Implicit combining of LDLR, APOB, and PCSK9 mutations. They are adjacent to the name FH, and the names for FDB and FH3 are not mentioned, leaving room for the assumption that the APOB and PCSK9 carriers are FH patients.||Gone are the references to FDB and FH3. The constituents appear next to each other spatially. However, grammatically their “conflation” is not necessarily true. An easy-to-make assumption due to sequential sentences, however vulnerable a reader may be, does not make an airtight conclusion. The period crucially separates the two sentences, and the word “primarily,” although suggestive, leaves us with declaring an adverb to be our smoking gun. But we can say that the explicit reference to an expanded definition of FH is dropped from the scientific record. Newcomers to FH will accustom themselves to a new definition and most likely will not even be aware of the ambiguity, especially as this “new research” replaces the old.|
|2013: report by Nordestgaard, et al. “Heterozygous FH is caused either by heterozygous loss-of-function mutations in LDLR, heterozygous mutations in APOB that affect the LDL receptor- binding domain of apolipoprotein B, or heterozygous gain-of-function mutations in PCSK9.” doi:10.1093/eurheartj/eht2732016||Now FH is LDLR, APOB, and PCSK9. For a genetically inherited disease, in a prevalence study which compares the results to FH-as-Only-LDLR, this is conflation – and we mean linguistics. Now, with the aid of the new definition for “FH,” prevalence in a Danish population is cited as 1:200.||The implicit reference to the combining of separate terms is dropped from the scientific record.|
From 2011 to 2013, critical detail has been subtracted from the scientific record, enabling a “conclusion drift” from a specific to a general use of “FH.” Now in 2016 FH is said to be between 1:222 and 1:256 in the Regeneron report. And this is compared to the Nobel Prize winners’ estimate of 1:500, where in the new studies, after linguistic and citation integrity is restored, FH-as-LDLR is still 1:500.
Again, what if professor A did a first-of-its-kind prevalence study of domesticated horses, and an experienced and prominent professor B did the same, but included zebras as if they were “horses”? If professor B does not explain the conflation of “zebras” with “horses” and claims that his prevalence result is higher than professor A‘s, then he uses the omission of information to deceive his readers. For new students, professor B‘s research precedes and precludes awareness of the conflation. Now what if professor B uses his numbers to declare professor A’s work to be incorrect and obsolete?
“Most physicians believe that FH is rare and not often seen in practice. In fact, it is significantly more common than 1:500, the estimate made at the time that Brown and Goldstein identified the LDL receptor. Current studies suggest a prevalence of 1 in 200 to 300 people based on work in the Netherlands, Denmark, and other countries where genetic testing has played a significant role.”~ DOI: 10.1161/CIRCULATIONAHA.116.021673
The above passage was written by Anne Goldberg and Samuel Gidding and is aptly titled, “Knowing the Prevalence of Familial Hypercholesterolemia Matters.” After we are familiar with the terms, the problem is rendered conspicuous. Two sentences suffice:
- “Brown and Goldstein identified the LDL receptor” … and its prevalence was estimated to be 1:500.
- “Current studies suggest a prevalence of 1 in 200 to 300 people” … if we stretch the definition of “FH” to add FDB-APOB and FH3-PCSK9 to FH-LDLR.
Now, in June 2018, Dr Goldberg updates the Merck Manual, returning to the old definition of FH. FH is again identified solely by the presence of the LDLR mutation. The APOB and PCSK9 are listed immediately below in their own categories – while the Merck Manual keeps the quantities of the APOB and PCSK9 in the math for the LDLR, the “FH”: 1/200.
|June 2018 update of the Merck Manual||“FH” returns to its original usage and is defined again solely by presence of LDLR. The APOB and PCSK9 are listed separately again.||But the APOB and PCSK9 are still included in the math. FH as solely LDLR cannot possibly be 1/200.|