Clinical background
Transthyretin amyloid cardiomyopathy (ATTR-CM) is a progressive and potentially fatal disease in which the prognosis ranges from several years of survival at early diagnosis to months in advanced disease. The treatment landscape has changed fundamentally since tafamidis was approved, and further disease-modifying treatments have followed. In this landscape it is essential to be able to classify the patient into a prognostic stage at diagnosis, both to guide the conversation about prognosis and to judge when and how quickly disease-modifying treatment should be started.
The Gillmore staging system was developed for precisely this purpose: to place the patient, using two biomarkers measured routinely at diagnosis — NT-proBNP and eGFR — quickly and reproducibly into one of three prognostic stages. In the literature the system is also called NAC staging (National Amyloidosis Centre) and is the most widely used staging system for ATTR-CM in Europe.
Applying the Gillmore staging system
The classification rests on two variables that are already in most patients' records at diagnosis: NT-proBNP in ng/L and the estimated glomerular filtration rate (eGFR) in mL/min/1.73 m². The thresholds are fixed and binary:
The derivation cohort consisted of 869 patients with confirmed ATTR-CM (553 with wild-type ATTR and 316 with hereditary variant ATTR) treated at the UK National Amyloidosis Centre [1]. The system was validated externally in a French cohort of 318 patients from the Henri Mondor Teaching Hospital in Créteil [1]. The cohorts date from the period before disease-modifying treatment was generally available, which is important to bear in mind when interpreting the absolute survival figures.
Median survival in the derivation cohort was 69.2 months for stage I, 46.7 months for stage II and 24.1 months for stage III (p < 0.0001) [1]. After adjustment for age, the hazard ratio for death was 2.05 (95% CI 1.54 to 2.72) for stage II and 3.80 (95% CI 2.73 to 5.28) for stage III, compared with stage I [1]. The hazard ratios were not appreciably affected by transthyretin genotype and were maintained in the external validation cohort [1].
Interpretation in practice
The staging translates into clinical action mainly by guiding the timing of disease-modifying treatment and the intensity of follow-up. A post-hoc analysis of ATTR-ACT and its long-term extension, comprising 350 patients with an available NAC stage, showed that tafamidis significantly reduced mortality in stage I (HR 0.43, p < 0.001) and stage II (HR 0.51, p = 0.003), with a numerical trend in stage III (HR 0.75, p = 0.298) [2]. At the end of the study, all-cause mortality was 36% versus 61% in stage I, 55% versus 74% in stage II and 69% versus 88% in stage III, for continuous tafamidis and placebo-to-tafamidis respectively [2].
| Stage | Criteria | Clinical meaning | Management |
|---|---|---|---|
| I | NT-proBNP ≤3000 and eGFR ≥45 | Early disease, the longest expected survival | Start disease-modifying treatment; the greatest documented absolute benefit from tafamidis |
| II | Only one variable abnormal | Intermediate prognosis | Start treatment; a significant benefit is documented, but the survival curves separate later than in stage I |
| III | NT-proBNP >3000 and eGFR <45 | Advanced disease, the shortest expected survival | Treatment should be considered; the numerical trend for tafamidis favours treatment, but the absolute benefit is smaller and the evidence weaker. Assess individually with regard to frailty and comorbidity |
An important message from the treatment data is that the survival curves for tafamidis versus placebo separate early in stage I but only later in stages II and III [2]. This underlines the value of early diagnosis and prompt initiation of treatment.
Validation and performance
The original external validation in the French cohort (318 patients) confirmed that the hazard ratios were maintained [1]. The system has since been tested in a contemporary cohort in the tafamidis era. Skov et al. studied 441 patients with newly diagnosed ATTR-CM at the Mayo Clinic between 2019 and 2023, of whom 85% started tafamidis and 94% had wild-type ATTR [3]. Three-year survival by the NAC system was 92.0% for stage I, 67.5% for stage II and 47.0% for stage III [3]. Compared with the modified Mayo system (which uses a high-sensitivity troponin T >65 ng/L and an NT-proBNP >3000 ng/L), the NAC system performed equivalently: the corresponding three-year survival for the Mayo stages was 92.3%, 71.6% and 43.5% [3]. Both systems therefore separated clinically meaningful prognostic groups even in a cohort in which the majority were on disease-modifying treatment.
An extended version of the NAC system with four stages has been proposed, in which stage IV is defined by an NT-proBNP >10,000 ng/L [2,3]. In the contemporary Mayo cohort, stage IV had a three-year survival of 34.4%, identifying a group with very high mortality that is not separated out within stage III in the original three-stage system [3]. This extension is not, however, part of the calculator presented here.
There is no published c-statistic or formal calibration analysis for the NAC system in the original derivation article, since the system is a threshold-based classification and not a continuous risk model. Performance has instead been evaluated by Kaplan–Meier analysis and Cox regression in all the published cohorts.
Limitations
The system is validated for ATTR-CM (both wild-type and hereditary variant) and should not be applied to light chain amyloidosis (AL amyloidosis), in which the biomarker pattern and the prognosis differ. Nor is it applicable to patients in whom NT-proBNP or eGFR cannot be interpreted meaningfully, for example in end-stage renal failure on dialysis, where the eGFR does not reflect the underlying renal function and NT-proBNP is cumulatively raised independently of the severity of the amyloid disease.
A fundamental limitation is that the derivation cohort was collected before the era of disease-modifying treatment [1,4]. The absolute survival figures from 2018 therefore reflect a natural history that no longer applies to patients starting tafamidis or other treatments at an early stage. The contemporary validation by Skov et al. shows that survival at every stage is considerably better today than in the derivation cohort [3], but that the relative ordering of the stages is preserved.
NT-proBNP is influenced by age, renal function, atrial fibrillation and untreated heart failure of other cause. A patient with ATTR-CM and concurrent chronic kidney disease of another cause may fall into stage III because of renal failure that is not amyloid-related, which exaggerates the amyloid-specific prognosis. Conversely, a patient with early ATTR-CM but concurrent atrial fibrillation may have a moderately raised NT-proBNP that pushes the value above 3000 ng/L and thereby moves the patient to stage II despite limited amyloid infiltration.
The system does not take account of functional class (NYHA), diuretic dose or frailty, all of which carry independent prognostic information [4]. The Columbia risk model, which adds NYHA class and the baseline diuretic dose to biomarker-based staging systems, has been shown to improve prognostic performance [4]. For clinical decision-making, the Gillmore staging should therefore be complemented by an assessment of functional capacity and comorbidity, and not used as the sole basis.
References
- Gillmore JD, Damy T, Fontana M, et al. A new staging system for cardiac transthyretin amyloidosis. Eur Heart J. 2018;39(30):2799-2806. PMID: 29048471
- Damy T, Wang R, Maurer MS, et al. Long-term efficacy of tafamidis in patients with transthyretin amyloid cardiomyopathy by National Amyloidosis Centre stage. Eur J Heart Fail. 2025;27(12):2998-3009. PMID: 40488446
- Skov JK, Clemmensen TS, Scott CG, et al. Utility of Prognostic Staging Systems in Transthyretin Amyloidosis Cardiomyopathy in the Era of Available Disease-Modifying Treatment. J Am Heart Assoc. 2026;15(16):e048585. PMID: 42568056
- Vaishnav J, Bampatsias D, Boursiquot BC, et al. Disease Prognosis and Progression in Transthyretin Amyloidosis: JACC: CardioOncology State-of-the-Art Review. JACC CardioOncol. 2026;4:120-136. PMID: 42017563
