Results

Prevalence of G+/P+ and carriers and clinical characteristics

In our cohort, 398 index cases were first identified and diagnosed with ATTRv. Following this, 1243 relatives underwent cascade genetic screening, of whom 674 tested negative and 569 (45.8%) were found to carry a pathogenic TTR variant. Of these, 461 were asymptomatic (G+/P−) at the time of genetic testing (98 [21.3%] siblings and 363 [78.7%] offspring of index cases). A total of 137/569 (24.1%) variant carriers were second-degree relatives. Of note, 108 individuals were found to be already affected at the time of genetic testing (G+/P+) (71 [65.7%] siblings and 37 [34.3%] offspring of index cases) (Figure 1). Baseline clinical characteristics according to the type of referral and genetic screening are summarized in Table 1 with a further stratification related to prescription of disease-modifying drugs (DMDs) in Supplementary data online, Table S2: of note, carriers and G+/P+ patients increased over time from before 2012 to 2019, as well as the number of individuals receiving DMDs, likely reflecting improved surveillance and diagnostic awareness.

Figure 1: Flowchart of patients enrolled in the study. The flowchart described the derivation of the study cohort, from the identification of the first index patients, to cascade genetic testing and final identification of genotype-positive/phenotype-positive (G+/P+), carriers (genotype-positive/phenotype-negative [G+/P−]) and converted carriers (at follow-up)

 | Indexn = 398 | Cascade screening | P-value Overall comparisons | P-valueG+/P− vs. G+/P+
G+/P−n = 461 | G+/P+n = 108
Age at time of genetic test (years), median [IQR] | 70 [62–76] | 47 [38–56] | 57 [49–73] | <.001 | <.001
Men, n (%) | 306 (76.9) | 210 (45.6) | 57 (52.8) | <.001 | .001
Ethnicity, n (%) |  |  |  |  | 
Caucasian | 395 (99.2) | 458 (99.4) | 108 (100) |  | 
Afro-Caribbean | 3 (0.8) | 2 (0.4) | 0 | .561 | .889
Other | 0 | 1 (0.2) | 0 |  | 
Family relationship |  |  |  |  | 
Sibling | - | 98 (21.3) | 71 (65.7) | <.001
Offspring | - | 363 (78.7) | 37 (34.3) | 
Variants, n (%) |  |  |  | .001 | .034
Ile68Leu | 126 (31.7) | 150 (32.5) | 18 (16.7) |  | 
Glu89Gln | 36 (9.1) | 45 (9.8) | 28 (25.9) |  | 
Val122Ile | 32 (8.0) | 39 (8.5) | 3 (2.8) |  | 
Phe64Leu | 62 (15.6) | 73 (15.8) | 14 (13.0) |  | 
Val30Met | 71 (17.8) | 84 (18.2) | 25 (23.1) |  | 
Other | 71 (17.8) | 70 (15.2) | 20 (18.5) |  | 
Matrilineal inheritance, n (%) (n = 518) | - | 148 (38.5) | 18 (34.6) | .584 | 
Phenotype, n (%) (n = 957) |  |  |  | <.001 | 
ATTRv-CA | 145 (37.4) | - | 26 (24.3) |  | 
ATTRv-PN | 67 (17.3) | - | 39 (36.4) |  | 
Mixed | 175 (45.1) | - | 42 (39.3) |  | 
Negative | - | 461 (100.0) | - |  | 
NYHA class, n (%) |  |  |  | <.001 | 
I | 126 (31.7) | - | 65 (60.2) |  | 
II | 187 (47.0) | - | 26 (24.1) |  | 
III−IV | 59 (14.8) | - | 6 (5.6) |  | 
PND score, n (%) (n = 447) | 347 |  | 100 | .001 | 
0 | 115 (33.1) | - | 23 (23.0) |  | 
1 | 107 (30.8) | - | 47 (47.0) |  | 
2 | 80 (23.1) | - | 23 (23.0) |  | 
3a | 25 (7.2) | - | 5 (5.0) |  | 
3b | 8 (2.6) | - | 0 |  | 
4 | 12 (3.4) | - | 2 (2.0) |  | 
eGFR (mL/min/1.73 m2), median [IQR] (n = 350) | 71 [56–91] | 94 [89–100] | 85 [67–94] | .002 | <.001
NT-proBNP (pg/mL), median [IQR] (n = 292) | 1337 [432–3141] | 27 [23–88] | 442 [141–1122] | .001 | <.001
Atrial fibrillation, n (%) | 102 (25.6) | 7 (1.5) | 6 (5.3) | <.001 | .011
Pseudo-infarction pattern, n (%) | 51 (12.8) | 4 (0.9) | 9 (8.3) | <.001 | <.001
Low voltage, n (%) | 33 (8.3) | 6 (1.3) | 7 (6.5) | <.001 | .005
IV conduction block, n (%) | 30 (7.5) | 2 (0.4) | 5 (4.6) | <.001 | .030
Type I AV block, n (%) | 28 (7.0) | 1 (0.2) | 3 (2.8) | .009 | .023
Maximum LVWT (mm), median [IQR] (n = 830) | 16 [14–19] | 10 [9–11] | 13 [11–19] | <.001 | .001

Median age at diagnosis was highest among index cases, compared to their G+/P+ relatives. Male prevalence was highest among index patients, followed by G + P+ and carriers (Table 1).

At diagnosis a mixed phenotype was observed in 175 (45.1%) of index cases, higher than G+/P+ where it was present in 42 (39.3%) of G+/P+ individuals (P < .001). Neurological symptoms were more prevalent in affected relatives, as seen in PND scores. Renal function, as assessed by estimated glomerular filtration rate (eGFR), was significantly lower in index patients (median 71 mL/min/1.73 m2) compared to carriers (94 mL/min/1.73 m2) and G+/P+ individuals (85 mL/min/1.73 m2) (P = .002).

Overall, atrial fibrillation was the most frequent abnormality at ECG, followed by pseudo-infarction pattern, low voltages, intraventricular conduction blocks and type I atrioventricular blocks: these had a significantly higher prevalence among index patients, followed by G+/P+ and carriers.

Characteristics of carriers who developed an overt disease

Among the 461 carriers (G+/P−), over a median follow-up time of 5.3 [1.7–9.8] years, 77 (16.7%) patients, including 33 (42.9%) siblings and 44 (57.1%) offspring of index cases, developed a TTR-related form of disease. The rate of conversion varied significantly by genetic variant and age group (Figure 2, Table 2 and Supplementary data online, Figure S2): among carriers, the highest conversion rate was seen in Glu89Gln (42.2%, 95% CI]28.8–56.9), followed by Phe64Leu (24.7%, 95% CI 16.1–35.8), Val30Met (13.1%, 95% CI 7.4–22.1), Ile68Leu (7.3%, 95% CI 4.1–12.8), and Val122Ile (5.1%, 95% CI 1.3–18.3). Other variants accounted for 22.9% (95% CI 14.5–34.1). A significantly higher risk of conversion of Glu89Gln was also confirmed at Cox regression analysis after adjusting for gender and age (see Supplementary data online, Table S3). The median age of conversion for all carriers was 59.6 [IQR 50.0–69.2] years, with conversion occurring earlier in carriers of Glu89Gln (median age 49.1 [IQR 47.7–52.1] years) compared to other variants.

Figure 2: Incidence of conversion of overt clinical phenotype among carriers at follow-up. Cumulative incidence curves with 95% confidence intervals (shaded areas) and numbers at risk are shown, highlighting the significantly higher conversion rates for Glu89Gln and Phe64Leu variants. Numbers in brackets represent cumulative converted cases within the time frame presented. Follow-up time started at genetic testing. Pairwise comparison (vs Val30Met): Ile68Leu: P = .641, Glu89Gln: P = .001, Phe64Leu: P = .008; Val122Ile: P = .404, Other: P = .623

 | Overall | Ile68Leu | Glu89Gln | Val122Ile | Phe64Leu | Val30Met | Other | P-value
Age of index cases at baseline (years), median [IQR] | N = 398 | N = 126 | N = 36 | N = 32 | N = 62 | N = 71 | N = 71 | 
70 [62–76] | 75 [69–79]* | 54 [50–60]* | 74 [70–78] | 69 [64–74]* | 70 [65–73] | 61 [47–71] | <.001
Age of G+/P+ at baseline (years), median [IQR] | N = 108 | N = 18 | N = 28 | N = 3 | N = 14 | N = 25 | N = 20 | 
57 [49–73] | 73 [56–78]* | 54 [49–55]* | 76 [69–82]* | 75 [65–77] | 60 [42–67]* | 48 [41–54]* | <.001
Age at baseline of carrier’s that do not convert to overt phenotype (years), median [IQR] | N = 384 | N = 139 | N = 26 | N = 37 | N = 55 | N = 73 | N = 54 | 
47 [38–55] | 50 [43–58]* | 34 [28–40]* | 47 [40–52] | 46 [37–52] | 45 [38–52] | 39 [32–49]* | <.001
Percentage of carriers that convert to overt phenotype | Overall77/461 (16.7) | Ile68Leu11/150 (7.3) | Glu89Gln19/45 (42.2) | Val122IIe2/39 (5.1) | Phe64Leu18/73 (24.7) | Val30Met11/84 (13.1) | Other16/70 (22.9) | 
Age of carriers at baseline who convert (years), median [IQR] | 54.5 [39.0–65.8] | 63.2 [53.7–68.3]* | 38.7 [36.0–48.4]* | 57, 78 | 57.5 [53.0–67.0] | 63.0 [52.5–68.1] | 38.8 [25.4–58.3]* | <.001
Age at conversion (years), median [IQR] | 59.6 [50.0–69.2] | 66.5 [58.8–71.7] | 49.1 [47.7–52.1] | 68, 82 | 62.8 [55.8–74.2] | 66.8 [61.5–72.3] | 42.8 [31.3–51.6] | <.001
Siblings, n = 33 | 67.9 [59.2–72.4] | 72.5 [70.1–73.7] | 52.2 [51.3–58.2] | 57, 78 | 71.9 [63.1–74.8] | 71.9 [65.2–74.8] | 60.8 [57.7–70.3] | <.001
Offspring, n = 44 | 52.1 [46.3–60.4] | 59.3 [57.7–62.2] | 49.3 [47.0–51.1] | - | 59.4 [54.0–63.5] | 60.4 [55.6–66.3] | 38.8 [31.5–44.8] | <.001
Time to conversion | 5.3 [1.7–9.8] | 3.2 [1.3–4.8] | 9.8 [5.6–15.0] | 11.0, 4.1 | 5.2 [1.3–7.8] | 3.4 [1.3–8.8] | 4.3 [2.2–14.4] | <.001
Diagnosed >10 years before the index relative age at symptoms onset, n (%) | 11/62 | 5/11 | 3/16 | 0/2 | 2/15 | 1/10 | 0/9 | .098
Diagnosed >10 years before the index relative age at onset and G + P+ family member diagnosis, n (%) | 17/65 | 6/11 | 1/16 | 0/2 | 3/15 | 3/10 | 4/12 | .065
Men, n (%) | 40 (51.9) | 8 (72.7) | 11 (57.9) | 1 (50.0) | 8 (44.4) | 8 (72.7) | 4 (25.0) | .104
Matrilinear inheritance, n (%) | 24/52 | 3/10 | 6/14 | - | 4/10 | 1/4 | 10/14 | .233
Phenotype at conversion, n (%) |  |  |  |  |  |  |  | .029
ATTRv-CA | 29 (37.7) | 7 (63.6) | 7 (36.8) | 2 (100) | 2 (11.1) | 2 (18.2) | 7 (43.8) | 
ATTRv-PN | 46 (59.7) | 3 (27.3) | 10 (52.6) | 0 | 16 (88.9) | 9 (81.2) | 8 (50.0) | 
Mixed | 4 (5.2) | 1 (0.9) | 2 (10.5) | 0 | 0 | 0 | 1 (6.2) | 
First test suggesting conversion, n (%) in 75 patients | N = 75 | N = 10 | N = 18 | N = 2 | N = 18 | N = 11 | N = 16 | 
Echocardiography | 24 (32.0) | 7 (70.0) | 8 (44.4) | 1 (50.0) | 4 (22.2) | 0 | 4 (25.0) | .142
Cardiac magnetic resonance | 5 (6.7) | 2 (20.0) | 1 (5.6) | 0 | 1 (5.5) | 0 | 1 (6.3) | 
Neurologic testing | 46 (61.3) | 2 (20.0) | 10 (55.5) | 0 | 16 (88.8) | 7 (63.4) | 11 (68.8) | 
Test confirming diagnosisa |  |  |  |  |  |  |  | 
Scintigraphy, n (%) | 28 (37.3) | 7 (70.0) | 9 (50.0) | 2 (100) | 2 (11.1) | 2 (18.2) | 6 (37.5) | 
Neurologic testing, n (%) | 42 (56.0) | 2 (20.0) | 8 (44.4) | 0 | 15 (83.3) | 7 (63.4) | 10 (62.5) | 
Biopsies, n (%) | 32 | - | 8 | 1 | 7 | 6 | 10 | 
Fat pad, n (%) | 21 (65.6) | - | 8 (100.0) | 1 (100.0) | 3 (42.9) | 2 (33.3) | 7 (70.0) | .057
Skin or other, n (%) | 11 (34.4) | - | 0 | 0 | 4 (57.1) | 4 (66.6) | 3 (30.0) | 
Heart, n (%) | 1 (3.1) | - | 1 (12.5) | - | - | - | - | 
Neurologic involvement, n/N |  |  |  |  |  |  |  | 
Small fibres | 22/34 | 3/5 | 6/8 | 0 | 6/9 | 3/5 | 4/6 | .802
Large fibres | 37/69 | 2/7 | 12/18 | 0 | 9/18 | 5/10 | 9/14 | .267
Neuroautonomic | 5/69 | 0 | 1/18 | 0 | 4/18 | 0 | 0 | .106
Motor symptoms | 17/69 | 1/7 | 6/18 | 0 | 4/18 | 3/10 | 3/14 | .892
Sensitivity symptoms | 25/69 | 3/7 | 6/18 | 0 | 6/18 | 5/10 | 5/14 | .704
Gastro-intestinal symptoms | 9/69 | 0 | 3/18 | 0 | 2/18 | 0 | 4/14 | .297

Remarkably, among carriers who converted to overt disease, the proportion of male (51.9%) was lower than that observed in index patients and closely resembled that of G+/P+ individuals diagnosed at baseline (Table 2).

When intra-familial age at onset was analysed to determine how many carriers converted beyond the 10-year window, a total of 11/62 carriers developed signs and symptoms before the 10 years from proband’s age (see Supplementary data online, Figure S1A). These included: Ile68Leu (N = 5/11–45.5%), Glu89Gln (N = 3/16–18.8%), Phe64Leu (N = 2/15–13.3%), and Val30Met (N = 1/10–10.0%). Similar results were obtained when carriers were analysed according to age at diagnosis by family pedigree, with 17 patients developing ATTRv disease prior to 10 years of PADO (see Supplementary data online, Figure S1B). We also determined the prevalence of converted carriers only in relatives within the 10 years of the PADO and results were clinically similar: Glu89Gln (40.1%, 95% CI 27.5–55.6), followed by Phe64Leu (21.4%, 95% CI 13.3–32.5), Val30Met (9.8%, 95% CI 5.0–18.6), Ile68Leu (3.4%, 95% CI 1.4–8.1), and Val122Ile (5.1%, 95% CI 1.3–18.3). Other variants accounted for 18.2% (95% CI 10.6–29.4).

Moreover, clinical conversion was frequently represented by evidence of neurological involvement, particularly small fibre neuropathy (22/34) and sensitivity symptoms (25/69). Accordingly, the test confirming clinical conversion was predominantly neurologic (61.3%). This may reflect both the tendency of patients to overlook or tolerate early neurological symptoms and the high prevalence of mixed phenotypes associated with the TTR variants observed in our cohort.

Long-term survival and outcomes

After disease diagnosis, patients were followed for a median of 3.8 [IQR 1.6–7.1] years (index 3.4 [1.6–5.8] years; G+/P+ 4.2 [1.3–8.5] years; G+/P− 4.6 [2.3–8.6] years). Overall, in 344 of 583 (59.1%) eligible patients (including 398 index cases and 185 [108 at G+/P+ baseline and 77 converters] affected relatives) a DMD (either stabilizer or silencer) was started (Index: n = 218, G+/P+: n = 126; median time on treatment: 2.9 [1.0–4.2] years and 3.1 [1.0–4.4] years for index and G+/P+, respectively; P = .106). In particular, 221/583 (37.9%) received tafamidis (51/221 [23.1%] tafamidis 20 mg), 88/583 (15.1%) patisiran, 21/583 (3.6%) diflunisal, and 14/583 (2.4%) inotersen.

During follow-up, 258/967 (26.7%) patients died. Index patients showed a worse survival compared to G + P+, even after adjustment for age, gender, and type of mutation (Figure 3 and Supplementary data online, Figure S3), with a median survival time of 6.0 (95% CI 5.0–6.9) years. Cumulative incidence of overall mortality according to last available phenotype were 10.5 per 100 patientyear (95% CI 9.0–12.3) vs 5.8 per 100 patientyear (95% CI 4.5–7.9) vs 0.5 per 100 patient*year (95% CI 0.3–1.0) for index vs G+/P+ vs G+/P−, respectively (P < .001).

Figure 3: Unadjusted survival analysis of patients enrolled in the study from the last available clinical phenotype. Kaplan–Meier curves show improved survival for G+/P+ patients identified through cascade screening compared to index cases, with follow-up time starting at the last available phenotype determination

Of note, when stratified by treatment, DMDs consistently improved survival both among index and G+/P+ patients (Figure 4): while cumulative incidence of all-cause mortality was reduced by >50% in the treated cohort (with index patients referred to DMDs significantly outliving G+/P+ not receiving therapy), G+/P+ patients receiving DMDs also showed a more favorable outcome at follow-up over the entire cohort. Supplementary data online, Figure S4 presents survival analysis according to phenotype (index vs G+/P+ and converted carriers) receiving DMDs.

Figure 4: Survival analysis by phenotype and disease-modifying drugs (DMDs). Kaplan–Meier curves show significantly improved survival among patients receiving DMDs, particularly in G+/P+ individuals highlighting the impact of early detection and treatment initiation (all P < .01). Follow-up time was defined from the start of DMD therapy for both index and G+/P+ patients (including converted carriers)

At time-dependent Cox multivariable regression analysis (Table 3), after adjustment for age, genotype, and disease-modifying therapy, G+/P+ patients diagnosed earlier via screening showed a net survival benefit with a reduced mortality risk (hazard ratio [HR] 0.433, 95% CI 0.238–0.788, P = .006). Of note, disease-modifying therapy was associated with reduced mortality risk (HR 0.109, 95% CI 0.005–0.173, P < .001), but treatment effect changed over analysis time, indicating a gradual attenuation of the initial protective effect during follow-up (HR 1.112, 95% CI 1.024–1.251, P = .014). Furthermore, Val30Met and a pure cardiac phenotype were associated with better survival compared to those with other variants and mixed phenotype.

Variable | Hazard ratio | 95% CI | P-value
Age at genetic diagnosis (per 1-year increase) | 1.049 | 1.034–1.065 | <.001
G+/P+ vs Index | 0.433 | 0.238–0.788 | .006
Genotype (Val30Met Reference) |  |  | 
Ile68Leu | 1.421 | 1.004–2.819 | .031
Glu89Gln | 1.710 | 1.103–3.012 | .038
Val122Ile | 1.984 | 1.141–3.452 | .015
Phe64Leu | 1.468 | 0.996–2.618 | .051
Other | 1.728 | 1.088–2.997 | .022
Phenotype |  |  | 
Pure cardiac vs mixed | 0.737 | 0.380–0.912 | .043
Pure neuropathic vs mixed | 1.049 | 0.676–1.626 | .828
Disease modifier | 0.109 | 0.005–0.173 | <.001
Time-varying covariate disease modifier | 1.112 | 1.024–1.251 | .014