Weakest Link in my analysis of Danish reports is nonetheless strong

Regeneron’s Science magazine article supports my analysis of the Danish reports

Regeneron’s Science magazine article supports my analysis of the Danish reports

I wrote the following content before Regeneron published their report. With this recent report, we can clearly see that the vast majority of those given passing clinical scores are not found to carry a mutation, and the majority of those who are molecular hits do not have passing clinical scores.

However, I will keep this chapter here as additional proof of the variability of the FH phenotype. In fact, the variability is so great that it is even questionable if there is such a thing as an unequivocal “FH phenotype.” Without a defined phenotype, clinical scoring systems are not sufficient to identify an FH mutation carrier.

Here’s why I previously felt it necessary to include this chapter: for my mathematical proof,[1] there was one necessary assumption which was outside of the safety of deduction. However, scientific evidence, much of it provided by the very authors, showed this link to be nonetheless very strong. What follows is my original presentation regarding FH phenotypic variability.

Weakest Link in my analysis of Danish reports is nonetheless strong

Weak Link: A very large percentage of the Ex-Top4 could have originally been above[2] the DLCN Cut-off. 

Decisive Reports: Fouchier, et al, put together a report titled, “The molecular basis of familial hypercholesterolemia in The Netherlands.”

It was an exhaustive hunt for mutations among the population. Thus, they included many more mutations than their “Top 4” most frequent mutations. Here is one of their conclusions. (The emphasis is mine.)

Although mutations in specific domains of the LDL receptor gene impair specific functions of the corresponding domain in the LDL-receptor protein, this can only partly explain the striking variable phenotypic expression of the disease. Different mutations in the same domain and even the same mutation in different patients show large differences in clinical phenotype. For example the most frequent mutation in the Netherlands, the N543H/ 2393del9 bp, exhibits a phenotype ranging from an almost normal LDL cholesterol to LDL cholesterol levels far above the 95th percentile (Umans-Eckenhausen MA, et al. submitted). With such an extent of variability in the clinical expression, it is evident that other genes and environmental factors must be involved. Examples of such genes include cholesteryl ester transfer protein (CETP) gene, lipoprotein lipase (LPL) gene and the apolipoprotein E (apoE) gene (Boer et al. 1998; Kastelein et al. 1999; lipoprotein metabolism influence lipoprotein levels, but only partly explain the striking variability in phenotypic expression. Environmental factors, such as lifestyle habits, and familial risk factors for cardiovascular disease are of great influence in determining the clinical outcome of FH, as recently demonstrated (Sijbrands et al. 2000, 2001; Umans-Eckenhausen MA, et al. submitted). ~ IBID.

And another report on The Netherlands by Sjouke, et al, specifically focused on “Homozygous autosomal dominant hypercholesterolaemia in the Netherlands: prevalence, genotype–phenotype relationship, and clinical outcome.” It shows a parallel argument where the clinically diagnosed HoADH (HoFH) will result in a patient pool inflated with HeFH.  And this is because the phenotype is much more varied than previously thought, with many of the HoADH scoring below the usual clinical cutoff: “Surprisingly, only 50% of the patients met the clinical criteria for hoADH (LDL-C .13.0 mmol/L) …”

It concluded: “… the clinical phenotype is more variable than previously assumed.”

In the discussion: “We also observed a significant phenotypical variability in patients diagnosed with molecularly defined hoADH and, in particular, the majority of patients did not fulfil the phenotypic criteria for hoADH.”

The emphasis in the passage below is mine.

Phenotypic diagnostic criteria have been used to diagnose hoADH, and LDL-C levels .13.0 mmol/L are generally accepted as a major criterion for the presence of hoADH. It is of note, however, that a minority of patients in our study met this criterion, and the range of LDL-C levels (4.4–21.5 mmol/L) in our study overlaps to a significant extent with LDL-C levels observed in heterozygous ADH patients. Interestingly, we identified a total of 69 heterozygous ADH patients [of 13 080 patients (0.53%), of whom 6 index cases] with untreated LDL-C levels . 13 mmol/L in the database of the Foundation for the Identification of Persons with Inherited Hypercholesterolaemia. Based on the clinical criteria, these patients should be considered to suffer from homozygous ADH. This clearly further shows the overlap between heterozygous and homozygous ADH. The potential misperception that a patient with LDL-C levels much lower than expected for hoADH cannot be a carrier of two pathogenic mutations has likely resulted in an underestimation of the prevalence of hoADH. It is therefore no surprise that lipid levels in our study were on average lower than generally assumed in hoADH patients. Moreover, LDL-C levels observed in our study were also significantly lower than observed in a large retrospective cohort study in hoADH patients performed by Raal and co-workers. In their study, comprising 149 hoFH patients, LDL-C levels were 16.4+ 3.9 mmol/L and the mean age of hoADH patients in this South African cohort was 26.8+14.6 years, compared with a mean age in our study of 37.4+19.2 years. The majority of patients in the South African study were molecularly diagnosed with hoADH. The large difference with our study is the fact that, in their study, sequencing of LDLR and APOB was only performed upon a clinical suspicion of hoADH. The latter will result in an inflation of the clinical phenotype associated with the molecular defect.” ~IBID.

These are exhaustive studies and were not limited to a few of the most frequent mutations. Variability includes “a majority of patients [who] did not fulfil the phenotypic criteria for hoADH,” the most severe form of FH.

The majority of the Top4 have already revealed themselves to be below clinical detection. On what basis would we assume a concentration of the Ex-Top4 above the cut-off and not the distribution of the spectrum evident thus far?

There are other points to be made. First, overturning my thesis would require an inverse statistical outcome to what we know already. The majority of the Top4 have already revealed themselves to be below clinical detection. On what basis would we assume a concentration of the Ex-Top4 above the cut-off and not the distribution of the spectrum evident thus far? [3]  It would be highly improbable. The 2nd report shows that 149 out of 174 Top4 mutation carriers were below the cutoff. (See illustration above.)  These same Top4 mutations show great variability in gene expression.

Additionally, it would defy the aim of the report for the authors to keep silent here.  Strengthening their case, why then put in the extra effort to omit this data? All their declared difficulty in genotyping would have been wrong: the clinical cutoff used would have resulted in a target-rich pool of mutation carriers. Did they miss such an opportunity for a decisive result? … Or is it just that most Ex-Top4 mutation carriers are like Top4 mutation carriers?  Out of reach due to variability in the mutation effects, with large numbers of them too mild to be detected clinically.

The results revealed in the 2nd report are not the only evidence showing that environmental factors play a large role in the FH phenotype, producing great variability. Three of the very same authors of the 1st and 2nd reports produced an Earlier Report where they showed that three[4] of these four most frequent mutations had a range of clinical phenotypes, concluding that environmental factors play a significant role. 

“Because the type of LDLR mutations were the same in carriers identified in the 3 different background populations, the increase in cholesterol levels in the patient groups was not caused by an effect of the LDLR mutations, but could be attributed to both “environmental factors,” such as dietary intake and obesity, and to other minor mutations that modulate the cholesterol phenotype in the IHD and FH populations in general.”[5]

They also held out the possibility that …

In heterozygotes identified in the general population, a different genetic makeup or environmental factors could counteract the effect of LDLR mutations by reducing synthesis or increasing breakdown rates of LDL, resulting in lower cholesterol levels. However, differences in cholesterol levels between probands identified in the general population or among patients with IHD or FH could not be explained by differences in type of LDLR mutation, because these were the same and could also not be explained by differences in the most obvious confounders ….[6]

The screen shot below contains another conclusion of variability in gene expression found in “LDL-Receptor Mutations in Europe,” by Dedoussis, et al.

"Different mutations in the same domain and even the same mutation in different patients can result in variable clinical phenotypes.

There are many sources to cite. Here are some others.

 “FH is a disease that shows great phenotypic variability.” .… “Due to the paucity of data on genotype phenotype correlations, clinical diagnosis will miss a large percentage of FH patients. It is currently estimated that only 15 to 20% of patients with FH are actually diagnosed. A study on 643 Danish probands could not even find a single phenotypic characteristic to predict the existence of a mutation.” [7]

 “For all other patients with FH caused by LDLR defects, environmental or other inherited factors seem to be more important than the type of mutation in determining the phenotype severity.”[8]

“Consequently, the phenotype of FH individuals is highly variable, probably also due to environmental factors and other genetic polymorphisms influencing the clinical outcome of FH.” .… “We here present a large, descriptive study of 1038 Danish FH individuals, who display a wide variety of phenotype regardless of mutation status.” .… “Conclusions:  No parameters could decipher mutation status a priori. All individuals fulfilling the FH criteria should therefore be referred in order to facilitate family tracing and genetic counseling.” [9]

There is a famous study of Canadian Chinese who suffered harmful effects of FH. Their relatives in China were tested, and those found with identical mutations nonetheless “do not have a markedly elevated concentration of LDL.” [10] The study concluded with the strong suggestion that environmental factors play a large role in phenotypic expression of the disease. How high or low would a mutation carrier score during clinical screening? “It has been shown that different environmental factors may moderate the phenotype in heterozygous FH.”[11]  We are not just talking about the variety of mutations, but the simultaneous variation of clinical scores among patients with the same mutations.

And APOB is typically milder than LDLR, and it makes up the majority of molecular results in the 2nd report. Here is another paper, saying essentially the same thing:

“Familial defective apolipoprotein B (FDB) caused by the R3500Q apolipoprotein B gene mutation may mimic FH but the clinical course, however, is often milder than that seen in patients with LDL receptor gene mutations.”[12]

It would seem highly improbable that all of the following would occur:

  1. That The Netherlands’ exhaustive demonstration of “striking” and “surprising” phenotypic variety would not generally represent that of neighboring Denmark.
  2. That among the entire Ex-Top4 spectrum there would be no phenotypic variety from one named mutation to the next but that each type would be at the highest intensity of disease.
  3. That among those with the same Ex-Top4 mutation there would be no variation of intensity of the disease but all cases would be at the highest intensity.

We would expect an equally wide distribution among the Ex-Top4 mutation carriers. It is either that or variety would have to stop, without precedent. 


[1] Click here.

[2] As shown in the screenshot above, Table 1 of the 2nd report listed the higher scoring DLCN scores in the lower rows, and the lower scoring in the higher rows. However, when I write, “Above the cutoff,” I mean the higher scoring Probable & Definite categories. When I say, “Below the cutoff” or “Below the detection point,” I mean the lower scoring categories, Unlikely & Possible.

[3] By “Ex-Top4” I mean the remainder of the spectrum besides the four most frequent, or to put it differently, those ADH mutations which were not targeted in the studies. Click here.

[4] The four most prevalent mutations used in the 1st and 2nd reports include the APOB mutation. This passage is about LDLR, and so it mentions only the top 3 most frequent LDLR. The APOB is irrelevant here. Table 1 of the 2nd report shows that the APOB mutations also have great variability, with most of them scoring below the clinical detection point.

[5] Phenotype of Heterozygotes for Low-Density Lipoprotein Receptor Mutations Identified in Different Background Populations; Anne Tybjærg-Hansen, Henrik Kjærulf Jensen, Marianne Benn, Rolf Steffensen, Gorm Jensen, Børge G. Nordestgaard 2005

[6] Ibid.

[7] Familial Hypercholesterolemia: The Lipids or the Genes? Akl C Fahed and Georges M Nemer; 2011

[8] Mechanisms of Disease: genetic causes of familial hypercholesterolemia; Anne K Soutar* and Rossi P Naoumova 2007

[9] No certain predictors for mutation status in a Danish cohort with familial hypercholesterolemia: A descriptive study Mads Nybo, Klaus Brusgaard, Annebirthe Bo Hansen; 2007

[10] Phenotypic Variation in Heterozygous Familial Hypercholesterolemia: A Comparison of Chinese Patients with the Same or Similar Mutations in the LDL Receptor Gene in China or Canada. Pimstone, et al. 1997.

[11] Ibid.

[12] The molecular genetic basis and diagnosis of familial hypercholesterolemia in Denmark. Jensen HK1. https://www.ncbi.nlm.nih.gov/pubmed/12553167