Article Text

Download PDFPDF

Research paper
A study of the neuropathy associated with transthyretin amyloidosis (ATTR) in the UK
  1. A S Carr1,
  2. A L Pelayo-Negro1,2,
  3. M RB Evans1,
  4. M Laurà1,
  5. J Blake1,3,
  6. C Stancanelli4,
  7. V Iodice4,
  8. A D Wechalekar5,
  9. C J Whelan5,
  10. J D Gillmore5,
  11. P N Hawkins5,
  12. M M Reilly1
  1. 1MRC Centre for Neuromuscular Diseases, UCL Institute of Neurology, London, UK
  2. 2Department of Neurology, University Hospital “Marqués de Valdecilla”, Santander, Spain
  3. 3Department of Clinical Neurophysiology, Norfolk and Norwich University Hospital, Norwich, UK
  4. 4Autonomic unit, National Hospital of Neurology and Neurosurgery, London, UK
  5. 5National Amyloidosis Centre, University College London Medical School, Royal Free Campus, London, UK
  1. Correspondence to Professor M M Reilly, MRC Centre for Neuromuscular Diseases, National Hospital for Neurology and Neurosurgery and UCL Institute of Neurology, Box 108, Queen Square, London WC1N 3BG, UK; m.reilly{at}


Background Hereditary transthyretin amyloidosis (ATTR) is usually characterised by a progressive peripheral and autonomic neuropathy often with associated cardiac failure and is due to dominantly inherited transthyretin mutations causing accelerated amyloid deposition. The UK population is unique in that the majority of patients have the T60A missense mutation in ATTR where tyrosine is replaced by adenine at position 60. This has been traced to a single founder mutation from north-west Ireland. The neuropathy phenotype is less well described than the cardiac manifestations in this group.

Methods We present the findings from an observational cohort study of patients with ATTR attending the National Hospital Inherited Neuropathy Clinic between 2009 and 2013. Detailed clinical neurological and electrophysiological data were collected on all patients alongside correlating autonomic and cardiac assessments. Follow-up data were available on a subset.

Results Forty-four patients with genetically confirmed ATTR were assessed; 37 were symptomatic; mean age at onset=62 years, range=38–75 years; 75.7% male. T60A was the most common mutation (17/37), followed by V30M (5/37). A severe, rapidly progressive, predominantly length dependent axonal sensorimotor neuropathy was the predominant phenotype. T60A patients were distinguished by earlier and more frequent association with carpal tunnel syndrome; a predominance of negative sensory symptoms at onset; significant vibration deficits; and a non-length dependent progression of motor deficit. Progression of the neuropathy was observed over a relatively short follow-up period (2 years) in 20 patients with evidence of clinically measurable annual change in Medical Research Council (MRC) sum score (–1.5 points per year) and Charcot Marie Tooth Neuropathy Score (CMTNS:2.7 points per year), and a congruent trend in the electrophysiological measures used.

Conclusion The description of the ATTR neuropathy phenotype, especially in the T60A patients, should aid early diagnosis as well as contribute to the understanding of its natural history.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Systemic amyloidosis is a disorder of protein misfolding caused by extracellular deposition of pathological fibrillar proteins.1 Transthyretin amyloidosis (ATTR), previously known as familial amyloid polyneuropathy (TTR-FAP), is a rare systemic disease due to the preferential peripheral nerve but also cardiac, vitreous and rarely renal amyloidosis manifesting as a disabling and ultimately lethal disease.

Wild-type transthyretin (TTR) is a plasma transport protein for thyroxine and vitamin A that is produced predominantly by the liver. TTR point mutations tend to increase its amyloid-forming potential via destabilisation of the soluble TTR homotetramer complex and precipitation of amyloidogenic mutant-TTR monomers.2 Liver transplantation has been the standard of care for the past two decades, but improved survival is genotype-dependent3 and neurological progression continues in 25%, possibly due to continued deposition of wild-type TTR.4 ,5 Promising therapeutic strategies are based on reduction of TTR production via TTR gene silencing (ongoing clinical trials) and reduction of the amyloidogenic potential of mutant-TTR by small molecule stabilisation of the soluble TTR tetramer (diflunisal and tafamadis).6–8 These new therapies highlight the need for early diagnosis.

ATTR was first described by Andrade in 19529 as an autosomal dominant peripheral and autonomic neuropathy which progresses to significant disability and death over a median of 12 years.10 This phenotype is now known to associate with the V30M (p.V50M) mutation that is endemic in Portugal, Japan and Sweden,11–13 and more than 100 other pathogenic TTR mutations have been described since then.14 ,15

The UK ATTR population is unique in that the majority of patients have the T60A (p.T80A) point mutation,16 first described in 1986 in an Irish family.17 Although there is a major focus in north-west Ireland,18 ,19 ATTR T60A has been identified widely around the world in areas of high Irish immigration.20 The phenotype is characterised by early cardiac, autonomic and gastrointestinal involvement; with symptoms and signs of peripheral neuropathy reported in approximately 50%.16 Descriptions of the neurological manifestations of T60A have been limited to single cases or small cohorts,18–22 with limited natural history data available.7 Previous descriptive studies in populations with small endemic foci with alternate TTR mutations have revealed specific clinical features. The A97S (p.A117S) genotype, common in Taiwan, can present with rapid deterioration in symptoms and albuminocytological dissociation without response to immunomodulatory therapy.23 G47A (p.G67A) ATTR, endemic in Mexico, is associated with a distinctive pseudosclerodermatous affection of the hands;24 other common mutations in this region, S52P (p.S72P) and S50A (p.S70A), tend to progress more rapidly and carry a worse cardiac prognosis. Although rare, careful clinical characterisation allows for appropriate monitoring and management, and facilitates early diagnosis which is paramount because of the recent emergence of disease-modifying therapies.

We describe the clinical and electrophysiological characteristics of ATTR in a UK-based cohort with follow-up in a subset and a particular emphasis on T60A patients with neuropathy.


Patients and clinical evaluation

This retrospective study included all TTR patients attending the neuropathy clinic in the National Hospital of Neurology and Neurosurgery, London UK (NHNN) between 2009 and 2013. Patients were identified either through primary diagnosis by the neuropathy clinic or via referral for neurology opinion from the UK National Amyloidosis Centre (NAC). Organ involvement by amyloid was defined according to the amyloid international consensus criteria, originally defined for light chain (AL) amyloidosis.25

After diagnostic evaluation, patients were reassessed at regular intervals (usually annually) or more frequently depending on clinical need. At each evaluation, patients underwent a detailed clinical and usually a electrophysiological review, including full neurological assessment by the same consultant neurologist (MMR), and nerve conduction studies (NCS) as part of their routine care.

Peripheral nerve involvement was established according to the following criteria. The presence of small fibre involvement was based on spontaneous sensory disturbance, cold-like pain, pins and needles, burning pains and/or electric shock-type pains in a length dependent pattern, with or without allodynia and hyperasthesia. These symptoms in the presence of normal motor, coordination and reflexes on examination were deemed consistent with predominant small fibre involvement.26 The presence of weakness, fasiculations, or muscle atrophy suggested involvement of large motor fibres, whereas sensory large fibre involvement was indicated by abnormal sensation or numbness, imbalance, and reduced vibratory and proprioceptive sensation. Abnormal NCS were required to support a diagnosis of large fibre neuropathy. A patient describing predominant sensory large fibre symptoms, with a positive Romberg's test, difficulty or inability to tandem walk, and minimal weakness was diagnosed with ataxic neuropathy. Greater attenuation of sensory nerve action potentials (SNAPs) than compound muscle action potentials on NCS was used as supportive evidence. Carpal tunnel syndrome (CTS) was diagnosed in the presence of a typical clinical syndrome with supportive neurophysiological criteria.27

Nerve conduction measures were available for ulnar and peroneal motor nerve fibres, and for ulnar and sural sensory nerve fibres obtained by standard methods. Normal limits were established from a large cohort of departmental controls. Thermal thresholds were used to assess small fibre neuropathy in some circumstances. The Charcot-Marie-Tooth Neuropathy Score (CMTNS, a composite score of clinical and electrophysiological data) which has been validated in Charcot Marie Tooth disease, another inherited neuropathy28 was either performed at the time of clinical assessment or retrospectively by case note review. We use this score in assessing all patients attending the Inherited Neuropathy Clinic at our institution rather than the Neuropathy Impairment Score (NIS) which has been used in the recent ATTR drug trials.6 ,7 ,29

All patients were also under regular review at the NAC and the Autonomic Unit, NHNN. Relevant information on diagnosis (ATTR mutation, tissue confirmation of ATTR in clinically affected tissue), and cardiac status was recorded. This included (New York Heart Association) NYHA performance status, N-terminal prohormone brain natriuretic peptide (NT-proBNP; abnormal >50 pg/mL) and echocardiogram: abnormalities characteristic of cardiac amyloidosis included granular speckling appearance of the myocardium, interventricular septal thickness >12 mm and unparalleled diastolic dysfunction.30 We did not include cardiac MR data as this test was not performed on all patients. Amyloid tissue deposition according to isotope imaging (serum amyloid P: SAP and 99mTc-3,3-Diphosphono-1,2-Propanodicarboxylic acid: DPD scintigraphy) were recorded when available. Routine autonomic screening tests in our institution included an active stand test, deep breathing test, mental arithmetic test and Valsalva manoeuvre. Data were collected retrospectively by case note review and stored anonymously on a computerised database in the Medical Research Council (MRC) centre for neuromuscular diseases, NHNN. Data were collected through a review of routine clinical practice and therefore is covered by UCLH trust and National audit policies.31

Histology and immunohistochemistry

Sections from formalin-fixed paraffin-embedded biopsies were stained for amyloid with Congo red and viewed under crossed polarised light. Immunohistochemical staining of the amyloid deposits was performed, as previously described.32

DNA analysis

Genomic DNA was extracted from whole blood and the coding regions of the TTR gene were amplified by PCR assay and sequenced, as described previously.33

Statistical analysis

Cohort characteristics were described using proportions, medians and range. Comparisons were made using Mann-Whitney test and Wilcoxon matched-pairs tests. Regression analysis was used to examine trend. Non-parametric tests were used in the main, but parametric tests were applied when appropriate. All reported p values are two sided. A p value <0.05 was considered to be significant. Statistical analysis was performed using Prism5 (1995–2009 by GraphPad software Inc.).



Forty-four individuals with potentially pathogenic mutations in TTR were assessed; 12 were diagnosed in the neuropathy clinic and 32 referred for neurology assessment after diagnostic work up in the NAC. Half of the patients were the proband for the disease in their family. Of those with a family history (22 cases), 18 were identified by screening of the known family mutation, 16 of whom were asymptomatic at time of genetic diagnosis. Thirty seven individuals had symptomatic amyloidosis affecting at least one system at neurological assessment. Seven individuals with a family history of ATTR were screened and found to have potentially pathogenic mutations but had no symptoms or signs attributable to ATTR, and were under routine review; median age at last examination=44 years, range=25–58 years. The following analysis excludes these asymptomatic cases.

DNA analysis

Direct sequencing of the TTR gene demonstrated heterozygous missense mutations in 36 patients; 1 patient was homozygous for V122I (p.V142I).34 The most common mutation was T60A and was found in 17 patients, all of whom had Irish ancestry. V30M was the next most frequent (5 patients: 3 English and single cases from Cyprus and St. Lucia). There were two small kindreds of three and two affected members with G47V (p.G67V) (Singapore) and E89K (p.E109K) (English) mutations. The rest were individual mutations from a variety of ethnicities: A34G (p.A54G), G54 (p.G74), H90D (p.H110D), I84S (p.I104S), I84T (p.I104T), L12P (p.L32P), S77Y (p.S97Y), S97T (p.117T) and two unrelated individuals with V122I mutations (figure 1). All mutations have been previously reported in association with ATTR and are recognised as pathogenic.14

Figure 1

ATTR mutations in symptomatic patients. ATTR, transthyretin amyloidosis.


In 28 cases, positive staining with Congo red showing green birefringence under polarised light confirmed the presence of amyloid in a range of tissues (in some patients in more than one tissue type), including endocardial (9 patients), nerve (8), rectal (6), abdominal fat (5), flexor retinaculum (3), vitreous (2), muscle (1), colon (1) and leptomeninges (1). In 27/28, the amyloid stained specifically with antibodies against TTR. In the other two patients, negative staining with antibodies against κ and λ AL amyloid alongside a confirmed pathogenic mutation in TTR and suitable phenotype was felt to be adequate for diagnosis. A higher diagnostic yield was obtained when symptomatic (nerve, endocardial, vitreous) tissues were sampled (18/28), than when rectal or abdominal fat pad biopsies were attempted (10/28). Only one symptomatic tissue biopsy failed to provide histological evidence of amyloid (sural nerve), while six biopsies of rectal and/or abdominal fat tissue were negative for amyloid in four patients (all T60A). Tissue confirmation was not obtained in nine cases (7×T60A, 1×V30M and 1×S97T). The presence of a pathogenic mutation in TTR, a family history of the disease and either a typical clinical presentation or histological confirmation in another family member was deemed sufficient to confirm the diagnosis.

Clinical presentation

Median (range) disease duration at first neurological assessment was 2 (0–9) years; clinical and electrophysiological details at this time point are shown in table 1.

Table 1

Clinical and electrophysiological characteristics of symptomatic cases

First ATTR symptoms

The first symptoms attributable to ATTR were established on clinical questioning. The median age at onset was 62 years, range=38–75 years; 75.7% male, with no difference between genotypes (table 1). Peripheral nervous system (PNS) involvement was the commonest cause of first symptoms(PNS=43.7%); with cardiovascular, autonomic (ANS), central nervous system and visual symptoms seen initially in 32.4%, 18.9%, 2.7% and 2.7%, respectively (figure 2). The proportions of cases presenting with a neuropathy were similar between groups, possibly reflecting a referral bias in this cohort. CTS preceded other symptoms in 72.7% of patients by a median of 7 years (range=−13 to +5 years, but without histological examination of all flexor retinaculae it is not possible to say whether it was attributable to ATTR. A higher proportion of T60A cases had CTS prior to other amyloid symptoms (64.3% compared to 55.5% in other mutations, 20% in V30M; p=0.24). In T60A, CTS developed on average 2 years earlier than in other mutations (−6.9 vs −4.9 years) and 6 years earlier in V30M patients. Two patients had CTS as their sole presenting symptom. Both had been genetically confirmed as carriers after opting for screening on the basis of a family history of systemic ATTR. The first, a 60-year-old English lady (S77Y), received histological confirmation of penetrance on finding ATTR on flexor retinaculum biopsy after carpal tunnel release surgery (CTR). Her only symptom was bilateral CTS refractory to conservative management. However, she had been under 2-yearly surveillance for 15 years and had been found to have presymptomatic cardiac amyloidosis on echocardiogram. The second, a 44-year-old German woman (I84S), was diagnosed as manifesting after amyloid, which stained positive for TTR on immunohistochemistry, was found in the flexor retinaculum biopsy on recurrence of bilateral CTS 9 years after successful CTR surgery.

Figure 2

Clinical manifestations of multisystem amyloidosis: (A) Proportion with CTS prodrome (B) First system affected with amyloid-related symptoms (C) Numbers of systems affected at first assessment: PNS, peripheral nervous system; ANS, autonomic nervous system; CVS, cardiovascular nervous system; CNS, central nervous system; CTS, carpal tunnel syndrome.

A cardiac presentation was more common in T60A (41.2%) than V30M (20%), with more frequent autonomic presentation in V30M (40%) than T60A (17.6%). There was greater clinical heterogeneity in the other group, reflecting genetic diversity.

Peripheral Neuropathy in T60A ATTR

T60A cases had median disease duration of 2 years (range: 6 months–9 years) at first neurological assessment. At this stage, 13/17 had peripheral neuropathy. Onset was sensory in all but one. Numbness was the most common first symptom, occurring in the feet in 7/13 and hands in 2/13. The other three individuals (3/13) reported of burning pain in the feet, with burning pains developing 3–6 months after numbness in five others. Spontaneous electric shock-like sensations were rare (1 case). 4/11 cases developed distal weakness within 1 year. A definite length dependent pattern of progression in sensory symptoms was observed in 11/13, with hand involvement within 2 years of foot symptoms and roughly coinciding with above-knee symptoms in the lower limbs. However, intrinsic hand muscle weakness was found in 3/13 when there was only mild distal weakness (MRC>4) at the ankle. Significant imbalance was present in 3/13 by 1 year. In those with hand onset, numbness developed in the hands at least 1 year before similar symptoms in the feet. There was asymmetry in 1/3, with bilateral involvement 3 months later. 3/3 developed pain which had a burning neuropathic quality within 6 months and hand weakness within 2 years.

On examination, 5/13 had a normal gait, three individuals had bilateral foot drop, three were predominantly ataxic, one antalgic and one wheelchair bound; Romberg's was positive in four cases. Distal muscle wasting was seen in 11/13 cases, present in upper and lower limbs in 6, intrinsic foot muscles in one and predominantly in distal upper limbs in the three cases with early hand symptoms. 4/13 were fully strong and one case had mild extensor hallicus weakness only. Eight individuals had mild (MRC: 4− to 4+/5) weakness in the distal upper limbs and five had mild distal lower limb weakness. Two individuals had moderate (MRC: 3/5) distal weakness in all four limbs. Median (range) MRC sum score shown in table 1.

Marked pinprick (PP) and vibration sensation (VS) deficit was common, with decreased PP to at least the knee in 8/13 and loss of VS in costal margins in 6/13, in anterior superior iliac spine in 3/13 and in the knee in 3/13. Proprioception was decreased at the ankle in 3/13.

Thermal thresholds were examined in nine individuals and were abnormal in seven, three of whom did not describe small fibre symptoms. Upper and lower limb SNAPs were absent in five cases. A length dependent, axonal, sensorimotor neuropathy was found in 8/13, 3/13 had a non-length dependent upper limb predominant neuropathy, 1/13 had normal large fibre studies and one individual had slow conduction velocities (CV): ulnar nerve CV=36 m/s, CMAP of 0.9 mV, absent f waves. Median CMTNS was 15 (0–27) in the T60A group, in keeping with a moderate neuropathy.

Peripheral neuropathy in non-T60A ATTR

A small fibre neuropathy was common in non-T60A patients. Distal pain and dysastheasia were the presenting features of neuropathy in 4/4 symptomatic V30M cases, the other had isolated orthostatic hypotension. Numbness and paraesthesia were concurrent or followed in 6–12 months with distal lower limb weakness between 1 and 2 years later, and hand symptoms between 6 months and 4 years after sensory disturbance in the feet. Proximal lower limb weakness and significant mobility problems were common by the time hands became weak (18 months–3 years). Neuropathy was less frequent (10/15) and less severe in the other mutations group; median (range) CMTNS=21 (0–33) in V30M; other=2.5 (0–21); p=0.02. Onset was distal and sensory with burning pain or ache in 6/10 and acral parasthesias in the others. Progression was slightly slower, spreading to the knees in 1–3 years with weakness developing 2 years after sensory symptoms. Negative sensory symptoms (numbness) was present in 13/37 (35.7%) of whole cohort, 10 of whom had T60A ATTR (p=0.005). Positive sensory phenomena (pain, paraesthesia) were less common in this group.

Examination findings for non-T60A cases are summarised in table 1. Notable differences between the groups include marked VS deficit in T60A, while PP loss was greatest in V30M. V30M individuals were weakest (p>0.005) and showed a length dependent pattern with mild hand weakness (>4 in intrinsic hand muscles) by the time proximal lower limb weakness was detectable. In contrast, T60A cases developed upper limb weakness earlier, with hand weakness not attributable to CTS in three patients with only mild weakness at the ankle (MRC>4).

A length dependent axonal sensory motor neuropathy was the norm with no significant differences in SNAP/CMAP amplitude or CV between groups. One V30M case also had slow CV, ulnar nerve CV=35 m/s, and CMAP=0.7 mV with absent f waves. The V122I homozygous case had a combination of neurogenic and myopathic electromyography abnormalities in proximal muscles: with a combination of large motor units and low amplitude units recruiting to a variable frequency. There was also increased soft-tissue tracer uptake on DPD scintigraphy and histological confirmation of ATTR in nerve and muscle.33

Extra-PNS involvement

Multisystem involvement tended to occur earlier in T60A (96% at 2 years, 1–9 years; median, range) from onset) than in other mutations (60% at 3 years, range 1–6 from onset), with a pattern more suggestive of sequential tissue involvement. Multisystem involvement in this V30M cohort may be related to disease duration (80% at 4 years, 1–5 from onset; figure 2C).

Relative frequencies of individual symptoms are listed in table 2. Cardiac symptoms were most common (12/17), with subclinical evidence of cardiac amyloidosis in a further four cases. No subclinical evidence of autonomic dysfunction was found on autonomic screening tests in asymptomatic T60A cases, but subclinical dysautonomia was found in 3/15 other mutations. Likewise, there was subclinical cardiac dysfunction in a further 2/3 V30M cases and ¾ other cases without cardiac symptoms. Characteristic abnormalities on echocardiogram were found in 7/15 presymptomatic individuals and elevated NT proBNP in 11/15, with cardiac uptake on DPD scintigraphy being most sensitive in 7/7 cases examined. ECG abnormalities were common but non-specific. All cardiac biomarkers supported more severe involvement in the T60A cohort. There was no correlation between measures of neuropathy severity and any of the measures previously shown to correlate with progression of amyloid cardiomyopathy: NYHA, p=0.63; BNP, p=0.44; interventricular septum thickness, p=0.45.

Table 2

Symptoms associated with TTR-amyloidosis in autonomic nervous, gastrointestinal and cardiovascular systems present on first assessment

Tc99m DPD scintigraphy was more sensitive than SAP scintigraphy for the detection of tissue amyloid; 13/13 of patients imaged had evidence of increased cardiac uptake on Tc99m DPD scan. No T60A patients had visceral uptake on SAP scintigraphy; only one patient in the other mutations category (L12P) had evidence of increased kidney uptake with associated proteinuria and renal dysfunction on blood tests.

Both patients presenting with isolated neuropathy had elevated NT proBNP and characteristic abnormalities of cardiac amyloid on echocardiogram (DPD scintigraphy not done). Both patients with isolated CTS had normal cardiac investigations except for increased cardiac uptake on DPD scintigraphy. Autonomic screening tests were normal in all four cases.

Disease progression

Clinical and electrophysiological follow-up was available on 20 patients (70% male); median duration of symptoms=4 years, range=2–11 years; median duration of clinical follow-up 1.9, range=0.7–4.8 years. Eight were T60A, three were V30M and nine were other mutations (three were G47V patients and others were individuals with V122I, E89K, GLY54, I84S, S77Y, S97T mutations). Nine patients in the follow-up group were on disease-modifying therapy; 4/8 T60A patients were on diflunisal (median duration of treatment=4, range=2–9 years); two brothers with G47V mutations had orthotopic liver transplant (4 years prior to last follow-up) and were also started on diflunisal at this time;  and a further three patients with other mutations were on diflunisal for 3–6 years. No V30M patients were on disease modifiers. Subgroups were felt to be too small to analyse for genotype-specific differences in progression of treated and untreated groups.

In the whole group, progression over the follow-up period was demonstrated by a statistically significant change in CMTNS (p=0.003), MRC score (p=0.006) and progression of PP sensory loss (p=0.02), but not progression of VS loss (p=0.77) when scored using the CMTNS sensory scales for PP and VS, respectively. This change was equivalent to a mean (SE) annual progression of 2.7 (0.8) points on CMTNS and −1.5 (0.52) points on MRC sum score which represent clinically detectable differences. The mean annual change of 0.2 (0.09) in PP sensory loss is unlikely to be clinically detectable in the individual when graded on the CMTNS PP sensory scale (range=0–4; where 0=normal, 1=decreased below or at ankle bones, 2=decreased up to the distal half of the calf, 3=decreased up to the proximal half of the calf including the knee, 4=decreased above the top of the patella). Electrophysiological studies did show deterioration overall with percentage annual change in SNAPs ranging from −24.6% (−1.4 µV/year) in the ulnar nerve to −30.0% (−2.7 µV/year) in sural nerve, and percentage annual change in CMAP's ranging from −34.4% (−2.5 mV/year) in ulnar nerve to −3% (−0.1 mV/years) in common peroneal nerve. These changes were small given the recognised inter-rater and intra-rater variability in NCS and are difficult to interpret in this setting. Standardised testing by a single examiner may provide useful electrical evidence of neuropathy progression in a controlled trial setting.


We present a detailed description of the neuropathy phenotype of ATTR in a UK cohort which is unique for its high proportion of patients with a T60A mutation. Novel observations from this work include earlier and more frequent association with CTS, a predominance of negative sensory symptoms at onset, significant VS deficit and a non-length dependent progression of motor deficit as discriminating features in the T60A group. Progression of the neuropathy was observed over a relatively short follow-up period (2 years) with evidence of clinically measurable annual change in MRC sum score and CMTNS, and small but congruent changes detectable electrophysiologically despite lack of robust standardisation of technique and examiner. This information is likely to be useful in diagnosis as well as in contributing to development of much needed disease-specific outcome measures.

The T60A neuropathy phenotype is a late onset (range=58–68 years), severe generalised sensorimotor neuropathy which may be length dependent or patchy in onset. It is best characterised by descriptions of seven affected individuals in the paper by Staunton and colleagues, 5/7 of whom presented with neuropathy, 2/5 concomitant with autonomic symptoms, all had cardiac involvement and 6/7 died within 7 years of onset (median=3.5, range=3–7 years).17 Two large studies have been published recently providing detailed descriptions of the cardiac manifestations in patients from UK, Canadian and US cohorts, but neuropathy information was limited.16 ,35 Sattianayagam and colleagues report symptoms of peripheral neuropathy in 23% at presentation and 54% at diagnosis (age range=56–66 years) in 60 T60A patients while Arruda-Olsen and colleagues state that 58% of their overall cohort of 282 patients with ATTR (of whom 58 had T60A point mutations) had parasthesias at diagnosis (mean, SD age=62, 7 years) without deeper phenotyping in either study. Our study concentrates on the neuropathy phenotype in this genotype. We report a neuropathic presentation in 43.7% and 41.2% of our complete cohort and T60A subgroup, respectively, with 73% and 76.5% having clinical evidence of neuropathy at first assessment (38–75 and 55–75 years). These differences can best be explained by neuropathy clinic referral bias, but it may be that directed clinical assessment by a neurologist is more sensitive and these figures are more representative of the true frequency of the neuropathy, especially in T60A. Given recent therapeutic developments, the issue of delayed or underdiagnosis is of increasing importance. Age-adjusted annual incidence of hereditary amyloidosis in the UK is 0.04/100 000, but discrepancy between NAC database and death certificate data as well as clustering of cases around expert centres provides epidemiological evidence of underdiagnosis.36

Clinical diagnosis of amyloid neuropathy is challenging; recognised diagnostic pitfalls in late-onset ATTR include no endoneurial amyloid in 18% of nerve biopsies, non-specific electrophysiology, and coincident diabetes mellitus (23%) or monoclonal gammopathy (7%) in one study of 15 genetically confirmed cases onset after the age of 50 years.22 The combination of cardiac and neuropathic manifestations in one disease process is rare and has been argued to be a useful indicator of amyloidosis,36 but the co-occurrence of neuropathy and heart disease is not uncommon in this age group. Our data suggest that, given the correct neuropathy phenotype, screening for amyloid–related cardiac abnormalities using echocardiogram or DPD-scintigraphy, even in the absence of cardiac symptoms, may be diagnostically useful. An elevated NT proBNP may be helpful but non-specific. Further clinical hallmarks for differentiation from idiopathic polyneuropathy in this age group were rapid progression to a severe neuropathy, early ambulatory loss and autonomic disturbances with genetic analysis of TTR recommended in the presence of these features even when amyloid deposits are lacking or when polyneuropathy-causing comorbidities are concomitant. It is, therefore, not surprising that we found a measurable annual change in MRC sum score and CMTNS score. However, the sensory scales within this score were either non-significant (PP deficit) or too small to be clinically measurable (VS deficit); furthermore, CMTNS lacks a pain score. In addition, routine NCS show a floor effect with up to 64% absent median SNAPs at baseline and lack sensitivity to annual progression although introduction of a protocol to improve consistency of serial measurements might improve this. As a number of our patients were on treatments, the longitudinal changes we observed may be less than the natural history; however, these finding would support the need for a disease-specific score in ATTR. While decent natural history data are available on V30M, there is less information on other mutations. Although 2 year follow-up data were collected on 15 T60A cases in the recent diflunisal randomised clinical trial, this information was summarised in the non-V30M cohort which included another 44 patients with 37 different mutations and genotype-specific outcome measures were not reported even in online supplementary data.7 There were no differences in baseline characteristics between V30M and non-V30M groups using the following measures: PND disease stage, NIS+7, Kumamoto score, modified body mass index or SF36. Intragroup heterogeneity may have masked any inter-genotype differences such as those highlighted here. It may be that more detailed genotype-specific deep phenotyping and natural history studies are required to establish the most reliable, sensitive and specific clinical and objective measurements applicable in this disease, and to determine whether a disease- specific severity score is needed, akin to the CMTNS in CMT.

Limitations to this study are the small numbers precluding further genotype comparison, especially in the follow-up cohort. The relatively short follow-up period and varying durations of disease-modifying therapies in 9/20 patients impacts on the interpretation of this natural history data. However, as the majority of TTR patients are likely to be on some form of treatment in the future, disease progression in this cohort is likely to provide further useful information over time.



  • Contributors ASC, MRBE and ALP collected the data. ASC analysed the data, wrote the manuscript, the revision and response to reviewers. MMR conceived the study and provided major input into the final manuscript and revision process. All other authors provide clinical care to patients in this cohort and contributed to the writing process.

  • Competing interests MMR is grateful to the Medical Research Council (MRC) and NINDS/ORD (1U54NS065712-01) for their support. This work was undertaken at University College London Hospitals/University College London, which received a proportion of funding from the Department of Health's National Institute for Health Research Biomedical Research Centres funding scheme.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement Additional clinical data are available on neurophysiological, cardiac and autonomic measurements. Please contact ASC: if this information is of interest.