Introduction
Within seconds of vascular injury, hemostasis is initiated. The first step, known
as primary hemostasis, involves platelet adhesion to the damaged vessel wall, mediated
primarily by von Willebrand factor (vWF), and platelet aggregation facilitated by
soluble fibrinogen.[1 ]
[2 ] The initial platelet plug is inherently unstable without the contribution of secondary
hemostasis, also known as the coagulation cascade. This cascade culminates in the
activation of thrombin and the conversion of soluble fibrinogen into an insoluble
polymer fibrin.[3 ] Fibrin stabilizes the growing blood clot by forming a robust yet flexible three-dimensional
scaffold that effectively stops bleeding and allows wound healing to proceed.[4 ] Finally, to maintain and restore normal vessel function, fibrinolysis is initiated.[5 ]
Fibrinogen is a hexamer assembled in hepatocytes from two copies each of three polypeptide
chains. The structure of fibrinogen, its conversion to fibrin monomers after thrombin
cleavage, and the steps allowing the formation of the three-dimensional fibrin network
have been well described elsewhere.[4 ] Once assembled and after quality control, fibrinogen is secreted into the bloodstream,
where it circulates at concentrations of 2 to 4 g/L in healthy individuals with a
half-life of 3 to 5 days.[6 ]
[7 ] In addition to its role in hemostasis, fibrinogen plays a part in a range of processes
that, when disrupted, can contribute to cardiovascular disease, obesity, infection,
tumor growth, and neurological disorders.[7 ]
[8 ]
[9 ] In this light, the quantity, localization, and structure of fibrinogen all hold
significant clinical relevance and are influenced by both environmental factors and
genetics. Indeed studies suggest that 30 to 50% of the variation in circulating fibrinogen
levels is heritable.[10 ]
[11 ] This review will explore our current understanding of congenital fibrinogen deficiencies
(CFDs) and reassess the estimated prevalence of these conditions.
Fibrinogen Genes and Transcripts
Fibrinogen production requires the expression of three closely linked genes on the
long arm of chromosome 4: FGB , FGA , and FGG , ordered from centromere to telomere. These genes are organized in a 75-kb cluster,
including proximal promoters and four enhancer elements—CNC12, PFE2, E3, and E4.[12 ]
[13 ]
[14 ] The cluster is organized as a small topologically associating domain, a self-interacting
genomic region flanked by CTCF sites, which contributes to the coordinated regulation
of the fibrinogen-encoding genes in fibrinogen-expressing cells.[15 ] Fibrinogen production occurs primarily in hepatocytes. Basal mRNA levels are maintained
through constitutive expression with production rates increasing significantly during
the acute phase response to inflammation.[16 ]
[17 ]
The major FGA transcript (encoding the Aα chain) consists of five exons, FGB (encoding the Bβ chain) comprises eight exons, and the major FGG transcript (encoding the γ or γA chain) includes ten exons ([Table 1 ]). Alternative splicing produces minor isoforms for both FGA and FGG . The FGA Aα-E isoform is a longer transcript containing an additional sixth exon and encodes
the Aα-E chain. It is found in homodimeric form (Aα-EBβγ)2 in around 1% of circulating fibrinogen in adults but is more abundant in fetal blood.[18 ]
[19 ]
[20 ] Although its functional role remains to be fully elucidated, Aα-E has been shown
to play a role in hemostasis during development in zebrafish embryos.[21 ] Interestingly, the relative abundance of the Aα-E isoform compared with Aα has been
shown to increase during infection.[22 ]
Table 1
Comparison of major and minor isoforms of fibrinogen genes (FGB , FGA , FGG )
Gene
Transcript
Exons
Length (bp)
Length (amino acids)
RefSeq ID
Relative circulating levels
FGA
Aα
5
2,209
644
NM_021871.4
Major isoform
Aα-E
6
3,654
866
NM_000508.5
1%
FGB
Bβ
8
3,641
491
NM_005141.5
Major isoform
FGG
γA or γ
10
1,565
437
NM_000509.6
Major isoform
γ'
9
2,072
453
NM_021870.3
10–15%
The FGG minor isoform, γ', consists of nine exons, with 20 C-terminal residues encoded by
exon 9 replacing the last four amino acids of γ encoded by exon 10.[23 ]
[24 ] Fibrinogen containing γ' is found in ∼10 to 15% of the total circulating fibrinogen,
mostly in heterodimeric form (AαBβγ/AαBβγ').[25 ]
[26 ] The γ′ chain lacks a platelet-binding site normally present in γ and modifies thrombin
and factor XIII activity, affecting clot structure and potentially contributing to
thrombosis.[27 ]
Quantitative Defects
Afibrinogenemia, characterized by the complete absence of fibrinogen, follows an autosomal
recessive inheritance mode and is mostly caused by homozygous variants, with compound
heterozygosity being reported less frequently.[30 ] For example, in a cohort of 74 afibrinogenemic patients, 98.6% were homozygous,
mostly for null variants (i.e., nonsense, splice site, frameshift, or large deletions).[31 ] As previously observed, within this cohort, most variants were identified in FGA , with the two most frequent being the intron 4 donor splice site variant c.510 + 1G > T
(23.6%) previously known as IVS4 + 1G > T or the large 11-kb deletion of FGA (12.2%).[32 ]
[33 ]
[34 ]
Hypofibrinogenemia is characterized by reduced fibrinogen levels in circulation: <0.5 g/L
for severe hypofibrinogenemia, between 0.5 and 0.9 g/L for moderate hypofibrinogenemia,
and between 1 and 1.5 g/L for mild hypofibrinogenemia.[35 ] If one considers the fibrinogen level as the phenotype of interest, hypofibrinogenemia
is primarily inherited in an autosomal dominant manner, with heterozygous variants
in the fibrinogen gene cluster being sufficient to cause reduced fibrinogen levels.[36 ] These variants impair transcription, splicing, hexamer assembly, or protein secretion
all resulting in lower circulating levels of a normal protein.[7 ] In contrast to dysfibrinogenemia (see below), there is no dysfunctional fibrinogen
in circulation; thus, the fibrin clot formed upon thrombin activation does not contain
dysfunctional molecules.
Rare cases of homozygosity have also been reported in hypofibrinogenemia, often resulting
in the severe form of the disease.[36 ]
[37 ] When considering clinical phenotypes (e.g., bleeding), penetrance is not complete
since individuals with heterozygous variants can remain asymptomatic.[38 ] Although many hypofibrinogenemic patients are heterozygous for variants that cause
afibrinogenemia in homozygosity, fewer null variants are identified in hypofibrinogenemia
compared with afibrinogenemia.[31 ]
[39 ] Instead, missense variants, mostly located in the C-terminal domains of the β and
γ chains, are the most common variant type, identified in 60.4% of a cohort of 44
patients.[31 ] In a subset of hypofibrinogenemia, specific variants in the C-terminus of the γ
chain escape the normal degradation pathway used to eliminate mutant or misassembled
fibrinogen, causing the accumulation of fibrinogen aggregates within the endoplasmic
reticulum of hepatocytes and a liver disease known as hepatic fibrinogen storage disease.[35 ]
Qualitative Defects
Dysfibrinogenemia, where fibrinogen is produced at normal levels but is functionally
defective, typically follows an autosomal dominant inheritance mode. In contrast to
hypofibrinogenemia, heterozygous variants result in the production, assembly, and
secretion of dysfunctional fibrinogen molecules into circulation. Even in the presence
of normal fibrinogen hexamers, upon thrombin cleavage the incorporation of mutant
fibrin molecules into the clot results in an impaired fibrin network.[40 ] Rare cases of homozygosity have been reported in dysfibrinogenemia.[41 ] The vast majority of genotyped cases (>70%) are heterozygous for one of the “hotspot”
missense variants affecting either residue FGA p.Arg35 at the Aα thrombin cleavage site crucial for the first steps of fibrin polymerization
or residue FGG p.Arg301 crucial for the D:D interaction, also essential for fibrin network formation.[31 ]
[42 ]
[43 ] While most fibrinogen variants do not demonstrate a clear correlation between genotype
and clinical phenotype, a few dysfibrinogenemia-causing variants are strongly associated
with an increased risk of thrombotic disease.[44 ] The underlying mechanisms for these particular cases are likely variant-specific
and may involve impaired interactions with thrombin, plasminogen, or tissue plasminogen
activator, resulting in abnormal fibrin clot formation and altered clot architecture,
or impaired clot lysis.[35 ]
[45 ]
[46 ]
[47 ] Hypodysfibrinogenemia, a disorder that combines both quantitative and qualitative
defects, can be inherited either as autosomal dominant or recessive, depending on
whether a single variant is sufficient to cause reduced levels of defective fibrinogen
in circulation or not. A patient's genotype may be heterozygous, compound heterozygous,
or homozygous. Variants in two different fibrinogen genes may be present, with one
variant leading to a “hypo” phenotype and the other to a “dys” phenotype. This explains
why many variants observed in hypodysfibrinogenemia patients are also present in those
with other fibrinogen disorders.[48 ]
Clinical Heterogeneity
Congenital fibrinogen disorders present a spectrum of clinical phenotypes, ranging
from asymptomatic individuals to those experiencing mild to severe bleeding and/or
thrombotic events.[49 ] Clinical heterogeneity is particularly evident within hypofibrinogenemia and dysfibrinogenemia,
while afibrinogenemia consistently leads to a more severe bleeding phenotype due to
the complete absence of fibrinogen in circulation.[39 ]
[42 ]
[50 ]
[51 ] Afibrinogenemia cases are therefore typically diagnosed following bleeding episodes,
as is the case for most symptomatic hypofibrinogenemia patients while dysfibrinogenemia
is more often detected through preoperative screening or family diagnosis.[39 ] Both quantitative and qualitative fibrinogen disorders can present with thrombosis,
which can seem paradoxical, particularly in cases of absent fibrinogen.[35 ] This can be explained by clots forming through vWF-mediated platelet aggregation,
but without fibrin resulting in loosely packed, unstable thrombi prone to embolization.[38 ]
[52 ] Additionally, without fibrinogen in circulation, its antithrombin function is lost,
leaving thrombin free to promote further platelet activation and thrombus formation.[53 ]
Diagnosis is based on the assessment of both functional and antigenic levels of fibrinogen,
alongside clinical presentation, as recommended by established guidelines.[35 ] Capturing detailed information with clinical questionnaires and the use of standardized
tools, such as the ISTH bleeding assessment tool, is important for evaluating clinical
variability and to test correlations between genetic information and the observed
phenotype.[54 ] For instance, it was shown in a multivariant analysis of dysfibrinogenemia cases
that sex, fibrinogen level, activity/antigen ratios, and the common “hotspot” missense
variants were not associated with bleeding outcomes or the risk of thrombosis, suggesting
additional reasons exist for the clinical heterogeneity observed.[42 ]
Identification of Pathogenic Variants and Modifier Alleles
Genotyping has traditionally relied on Sanger sequencing to identify pathogenic variants,
many of which are single nucleotide variants (SNVs). The visualization of more than
330 causative SNVs placed on the coding sequences of FGB , FGA , and FGG reveals important structure-function insight for several domains of the fibrinogen
molecule.[29 ]
[53 ] With such allelic heterogeneity, the number of newly identified causative variants
continues to grow.
Next-generation sequencing (NGS) approaches, such as whole-exome sequencing (WES)
and whole-genome sequencing (WGS), now enable more comprehensive detection of variants,
including those beyond the primary pathogenic variant. This is especially important
in cases where a single variant may not fully explain the phenotype, emphasizing the
need to explore other alleles.[55 ] In the general population, the fact that variants such as factor V Leiden and prothrombin
G20210A can tip an individual's hemostatic balance toward thrombosis is well established.[56 ]
[57 ] In the presence of a bleeding disorder, such as hemophilia, the presence of factor
V Leiden or prothrombin G20210A results in a milder bleeding phenotype, a discovery
that has inspired novel therapeutic approaches for treating the disease.[58 ]
[59 ] Similarly, the genetically determined ABO blood group of an individual is known
to influence vWF and consequently factor VIII plasma levels, thereby modifying the
clinical severity of von Willebrand disease.[60 ]
[61 ] While knowledge of the impact of these variants preceded NGS, this analysis now
allows the simultaneous identification of well-known genetic modifiers, alongside
the primary pathogenic variant, as well as additional variants that may further influence
the phenotype. However, the increasing detection of variants of uncertain significance
presents a significant challenge since proving the functional impact of such variants
remains difficult, especially if they are rare.
Prevalence of CFDs
The increased accessibility and use of NGS has not only made it easier to identify
additional variants beyond the primary pathogenic variant but has also led to an explosion
in the availability of genetic data.[62 ] This has been driven by large, publicly accessible (de-identified) human sequencing
datasets, pioneered by initiatives such as the 1000 Genomes Project in 2010.[63 ] Previously, the frequency of pathogenic variants was assessed through candidate
gene sequencing within small case–control cohorts.[64 ] Current access to NGS data from hundreds of thousands of individuals from diverse
populations greatly improves the accuracy of variant frequency estimates and helps
identify new disease-causing variants.[64 ]
[65 ]
[66 ]
[67 ]
[68 ] Given the rates of natural variation, it is highly probable that all genetic variants
compatible with life exist in the human population, making population-scale genetics
incredibly valuable for uncovering the impact of variants in the genome.[64 ]
Public datasets such as the UK Biobank, which includes genetic and extensive phenotypic
information from over half a million individuals, have been pivotal in identifying
genetic risk factors for conditions like obesity.[69 ]
[70 ] These datasets enable the linkage of specific genotypes to clinical traits, as demonstrated
in a recent population-scale study leveraging data from over 937,000 individuals.
The study assembled the largest cohort to date of double heterozygotes for factor
V Leiden and prothrombin G20210A variants, providing precise effect size estimates
for the risk of venous thromboembolism (VTE). It was shown that double heterozygous
genotypes may be as frequent as factor V Leiden homozygotes, with a similar increased
risk of VTE. This example underscores the power of population-scale genomics to revisit
established genetic risk factors, while delivering far more accurate risk assessments
than were previously achievable.[71 ]
While the Genome Aggregation Database (gnomAD) does not provide individual genotype
or phenotype data, its aggregated summary data provide variant allele frequencies
which have become important for understanding genetic variation.[72 ] The database is an essential resource for variant interpretation, greatly improving
the ability to distinguish between common variants and those that are rare and potentially
pathogenic.[64 ] gnomAD data also provide insight into gene constraint (intolerance to variation)
and allows a better estimate of prevalence (i.e., the percentage of a population with
a causal genotype for a genetic disorder). Regarding CFDs, the global incidence of
the recessively inherited deficiency, afibrinogenemia, was historically estimated
at 1 in a million individuals.[73 ] The prevalences of hypofibrinogenemia and dysfibrinogenemia were difficult to estimate,
largely since many individuals are asymptomatic and only a few specialized clinical
centers were reporting the data. However, as these disorders typically follow an autosomal
dominant inheritance pattern (a pathogenic variant on one of the two fibrinogen alleles
is sufficient to cause the deficiency), they are obviously more common.
Previously gnomAD data have been used in the field of thrombosis and hemostasis to
estimate disease prevalences including for CFDs.[74 ]
[75 ]
[76 ]
[77 ] Using data from gnomAD v.2.0 which included exome/genome data from ∼140,000 individuals,
Paraboschi et al suggested that genotypes that could lead to a recessively inherited
fibrinogen deficiency were ten times higher than previously reported, with significant
variation across the eight genetic ancestry groups listed.[75 ] Prevalences ranged from 1 per million in East Asians to 24.5 per million in non-Finnish
Europeans (which excludes Finns due to the distinct genetic characteristics of this
population, shaped by historical bottlenecks and isolation). The global prevalence
of genotypes that could lead to a dominantly inherited fibrinogen deficiency, such
as hypofibrinogenemia or dysfibrinogenemia, was estimated to be around 11,000 per
million individuals.
We aimed to reevaluate these estimates using gnomAD v4.1.0, released in 2024, which
includes WES and WGS data for 807,162 individuals (730,947 exomes and 76,215 genomes).
A significant increase in sample size comes from the inclusion of 416,555 exomes from
the UK Biobank, alongside a threefold increase in non-European individuals. This dataset
reveals an average of two SNVs every three bases, offering comprehensive coverage
of human variation. Using this extensive resource, we generated new global prevalence
estimates for homozygous and heterozygous genotypes that could result in a fibrinogen
disorder ([Fig. 2 ]).
Fig. 2 Filtering process for the identification of high-confidence variants in FGB , FGA , and FGG focusing on those likely to cause a fibrinogen deficiency. First, coding variants
were selected from RefSeq transcripts (+15 bp either side), including major and minor
transcripts. Variants were then filtered based on their predicted consequences, ClinVar
classifications, and in silico predictions. The cumulative allele frequency of potentially
deleterious variants was calculated for major transcripts (Aα, Bβ, γ') only. Frequencies
of individuals heterozygous or homozygous for these variants were obtained using the
Hardy-Weinberg principle.
To this end, gnomAD v4.1.0 data were downloaded, variants predicted to be likely pathogenic
were selected, and the cumulative allele frequency (cAF) for each gene was calculated
by summing the allele frequencies of all filtered variants. To ensure high-confidence
data and minimize false positives, only short variants (SNVs and indels—small insertions
and deletions) that passed all filters in both exomes and genomes or passed in one
dataset and were absent from the other were included. Coding regions (+15 bp either
side, so splice sites are included) for all major and minor isoforms ([Table 1 ]) were initially analyzed for completeness to capture variants in all expressed transcripts,
while also observing those specific to each isoform. Additional variant annotation
was performed using the Ensembl Variant Effect Predictor (VEP), incorporating supplementary
data and in silico predictions.[78 ] Several strict filtering criteria were applied. First, null variants were selected
based on predicted consequences (e.g., frameshift, stop gained, start lost). Splice
variants not already classified as null but predicted to be high impact or likely
pathogenic were then selected, defined by a SpliceAI score > 0.8 or a Combined Annotation
Dependent Depletion (CADD) score > 30.[79 ]
[80 ] For the inclusion of missense variants and other remaining variant types, additional
filters were applied by selecting those labeled as pathogenic in ClinVar, excluding
those classified as benign, and including variants meeting specific in silico criteria:
classified as “likely pathogenic” by AlphaMissense, CADD score >30, or predicted as
“deleterious” by SIFT and at the same time “probably damaging” by PolyPhen.[81 ]
[82 ]
[83 ]
[84 ]
Among the 807,162 individuals, 5,383 different variants were identified across all
fibrinogen transcripts ([Table 2 ]). Of these, the three major isoforms accounted for 4,211 variants: 1,557 in Aα,
1,641 in Bβ, and 1,013 in γA. The two minor isoforms, AαE and γ', contributed 868
and 304 unique variants, respectively. Focusing on the three major isoforms, 798 variants
were classified as potentially pathogenic after applying the described filtering criteria.
The cAF of the selected variants in the cohort is 0.007515, generated by summing the
cAF for each gene. Using the cAF and the Hardy-Weinberg equilibrium equation (p2 + 2pq + q2 = 1), the global prevalence of fibrinogen disorder genotypes was estimated. For dominantly
inherited disorders such as dysfibrinogenemia, heterozygotes (2pq) are of primary
concern, as the presence of a single variant allele in the fibrinogen genes is likely
sufficient to cause the disorder.[36 ] Accordingly 15,000 individuals per million are expected to be heterozygous for a
potentially pathogenic variant and could theoretically exhibit a disease phenotype
of, dysfibrinogenemia or moderate hypofibrinogenemia. This estimate is higher than
previously calculated.[75 ] Likewise for recessively inherited forms, such as afibrinogenemia or severe hypofibrinogenemia,
the global frequency of homozygotes (q2 ) for the filtered variants is estimated at ∼29.4 individuals per million, higher
than the figure of 24.5 per million in non-Finnish Europeans previously reported.[75 ]
Table 2
Global estimated prevalence of inherited fibrinogen disorders
Transcript
Total variants
Total likely pathogenic variants
Cumulative allele frequency[a ]
Heterozygotes[b ] (per million)
Homozygotes[b ] (per million)
FGA
1,557
303
0.001158
2,300
1.3
FGB
1,641
272
0.001188
2,400
1.4
FGG
(without Ala108Gly)
1,013
(1,012)
223
(222)
0.005169
(0.002138)
10,300
(4,270)
26.7
(4.57)
Total
(without Ala108Gly)
4,211
(4,210)
798
(797)
0.007515
(0.004484)
15,000
(8,970)
29.4
(7.27)
a The cumulative allele frequency was calculated by summing the allele frequencies
of all variants predicted to be likely pathogenic for each gene in the gnomAD v4.1.0
dataset.
b Calculated based on Hardy-Weinberg equilibrium.
This increase in the estimated disease prevalences can be attributed to several factors.
First, the number of identified genetic variants associated with disease continues
to grow. Indeed, Baxter et al showed that for 13 nominated genes, the average number
of predicted loss-of-function variants per gene increased approximately fourfold between
gnomAD v2 and v4.[68 ] Similarly, there was a twofold increase in the number of variants per gene listed
as pathogenic or likely pathogenic in ClinVar.[68 ] Many of these newly identified variants are rare, with the average allele frequencies
of v4 variants being markedly lower than those already present in v2. These variants
have the potential to significantly impact a phenotype, and although rare, the cumulative
sum contributes to an increased incidence of potentially pathogenic alleles.
Regarding the estimated frequency of homozygotes, not all homozygous individuals will
necessarily present with afibrinogenemia, the most severe form of CFD. In some cases,
individuals may carry two pathogenic alleles resulting in severe hypofibrinogenemia
or dysfibrinogenemia.[37 ]
[41 ] This is particularly relevant for the FGG p.Ala108Gly variant, which has been linked to hypofibrinogenemia in case reports,
as well as lower fibrinogen levels in several GWAS studies, with the variant predicted
to cause a 0.2- to 0.7-g/L reduction in fibrinogen levels per Gly allele.[85 ]
[86 ]
[87 ]
[88 ]
[89 ]
[90 ]
[91 ] In gnomAD v2, no homozygous individuals were identified for FGG p.Ala108Gly. In contrast, gnomAD v4 identified 12 homozygous individuals for this
variant. These homozygous cases are more likely associated with severe hypofibrinogenemia
than afibrinogenemia. This prediction is based on our study of a patient with a large
14.8 Mb deletion including the entire fibrinogen gene cluster and FGG p.Ala108Gly on the non-deleted allele, who had fibrinogen levels of 0.7 g/L, consistent
with hypofibrinogenemia despite a homozygous-like genotype for .Ala108Gly.[55 ]
Further examination of the FGG p.Ala108Gly variant underscores the role of genetic ancestry in disease prevalence,
with a frequency of 0.38% in non-Finnish Europeans compared with 0.06% in South Asians,
a more than sixfold difference. This disparity is likely the result of a founder effect,
which enriches certain variants within specific populations, as this variant is inherited
as part of a distinct haplotype.[92 ] The inclusion of 416,555 individuals from the UK Biobank, where this variant is
relatively frequent (0.39%), increases the overall estimated prevalence. For example,
when this variant is excluded from the analysis, the cAF decreases to 0.0045 ([Table 2 ]). Consequently, the frequency of heterozygotes (2pq) drops to ∼8,970 individuals
per million, and the frequency of homozygotes (q2 ) falls to 7.3 individuals per million. These findings clearly demonstrate how population-specific
variants significantly influence global prevalence estimates.
An additional factor contributing to the increased prevalence estimates is the more
recent inclusion of databases like the UK Biobank which do not exclude individuals
based on phenotype. This broader inclusion may lead to greater representation of individuals
with rare diseases, contrasting with the Exome Aggregation Consortium (ExAC), which
primarily focused on exome sequencing data from “ostensibly healthy” individuals.[93 ] In the latest gnomAD dataset, numerous homozygous cases for disease-causing variants
have been identified that were previously only observed in heterozygosity in gnomAD
v2, as exemplified by FGG p.Ala108Gly.
A more precise estimate of disease prevalence for specific congenital fibrinogen disorders
can be obtained by analyzing well-characterized causative variants. As already discussed,
more than 70% of diagnosed dysfibrinogenemia cases result from variants affecting
the FGA p.Arg35 or FGG p.Arg301 amino acids, the so-called hot-spots.[31 ] In gnomAD, variants affecting these positions include p.Arg35His, p.Arg35Cys, and
p.Arg35Ser in FGA , and p.Arg301His and p.Arg301Cys in FGG . The cAF for these five variants in gnomAD is 3.47 × 10−5 , indicating that ∼69 per million individuals are expected to have dysfibrinogenemia
due to these five variants alone. Notably, no individuals in gnomAD were homozygous
for these variants.
There are several limitations to our global prevalence estimates based on gnomAD v4
data. First, our analysis considers the presence of the genotype, regardless of whether
individuals will develop the disease. If the genotype is not fully penetrant, genetic
estimates may be higher than actual disease prevalence, as shown in studies on penetrance
and variable expressivity in monogenic diseases.[94 ] Second, our analysis includes only the coding regions and splice sites of the three
major isoforms, as these transcripts are the most abundant in circulation and, therefore,
the most clinically relevant. Variants outside the coding regions, affecting for instance
regulatory elements (e.g., the enhancers CNC12, PFE2, E3, and E4 or the fibrinogen
gene promoters) which were not analyzed here, could also contribute to fibrinogen
deficiencies, although no cases with variants in these elements have been reported
to date. Our analysis also excludes structural variants (SVs) which include the FGA 11 kb deletion and copy number variants (CNVs). While gnomAD has released SV and
CNV data for some samples, these are not yet available for the entire cohort. Indeed,
identifying SVs and CNVs through short-read sequencing, such as exome sequencing,
remains imperfect, complicating their inclusion in such analyses. It is also important
to consider that the variants included in this analysis are not entirely specific
to congenital fibrinogen disorders; the filtering process also captures potentially
pathogenic variants associated with other fibrinogen-related conditions, such as renal
amyloidosis. This disorder, characterized by amyloid deposits formed from abnormal
fibrinogen, predominantly affects the kidneys.[95 ] Consequently, the reported cAF reflects the prevalence of fibrinogen-related diseases
more broadly. Finally, as previously mentioned, since a substantial portion of the
gnomAD v4.1.0 data is contributed by the UK Biobank, our estimates may not be accurate
for the underrepresented populations in these genomic databases.
