Background:In recent years, the use of second‐generation TKIs has increased in both first and second‐line settings in CML. Extended follow‐up of the phase III study of first‐line nilotinib (NIL) (ENESTnd) saw an unexpected increase in cardiovascular events (CVE) in the NIL arms, particularly in patients (pts) with higher Framingham risk scores. However, ‘real‐life’ data on the risk of CVE in Nil treated pts are still limited.Aims:We aimed to determine the probabilities of all CVE and ischaemic arterial events (IAE) at any time from NIL start, in a ‘real‐life’ setting.Methods:We retrospectively analysed 193 CML pts (75 male, 49.2%) given NIL in our centre. Univariate analysis was performed by Kaplan‐Meier and multivariate analysis by Cox proportional hazard model. For both univariate and multivariate analysis, the impact of cardiovascular risk factors (CVRF) was analysed independently of age.Results:The median ages at diagnosis and start of NIL were 46 (18–89) and 51 (20–89), respectively. The overall follow‐up from diagnosis was 110 mths (12–286) and from start of NIL 37 mths (0.1–139). Baseline CVRF were known for 143/193 pts (74%). At least 1 CVRF was present in 94/143 pts (65.7%) and were more common in pts > 60 (p < 0.001). A total of 70 CVE occurred in 50/193 pts (25.9) (Figure 1a). The median duration of NIL therapy in those with CVE was 35.3 months vs 38.5 months in unaffected pts. The median times to occurrence of any CVE or IAE were 19.6 months (0.1–88) and 28.6 months (0.17–85), respectively. The 3 and 5 probabilities of any CVE were 18.1% (95%CI: 17.5–18.7) and 27.9% (95%CI: 27.2–28.6), respectively, while for IAE these were 10% (95%CI: 9.6–10.5) and 13.7% (95%CI: 13.1–14.3), respectively. The duration of NIL doses of 600–800 mg in pts experiencing IAE was 25.8 months (0.17–85), compared to 14.7 months (0.2–124) in unaffected pts (p = 0.86); for CVE, these periods were 24.1 months (0.1–88) and 12.4 months (0.2–124), respectively (p = 0.45). In univariate analysis, age ≥51, presence of any CVRF and male sex were significantly associated with a higher probability of CVE (p = 0.001, p = 0.025, p = 0.005, respectively), while only age ≥51 was significant for the occurrence of IAE (p = 0.0001). ROC analysis of age and occurrence of IAE identified 59 years as the cut‐off with highest sensitivity and specificity (AUC 0.738, p < 0.001). Using this information, the 5 year probability of any CVE was 47.5% (95%CI: 46.1–48.9) for pts ≥59 years and 17.7% (95%CI: 16.9–18.5) for pts <59 years (p = 0.000058), while for IAE this was 27.5% (95%CI: 26.3–28.8) vs 6.4% (95%CI: 5.9–6.9), respectively (p = 0.000028)(Figure 1b). In multivariate analysis, age ≥ 59 and male sex remained significantly associated with occurrence of CVE (HR = 2.2[95%CI: 1.4–4.8], p = 0.001 and HR 1.8 [95%CI: 0.9–3.2], p = 0.047, respectively); age ≥ 59 was again the only variable significantly associated with IAE (HR = 5.23[95%CI: 2.2–12.7], p = 0.0001). At last follow‐up 182 pts (94.3%) are still alive, of whom 82 (47%) remain on NIL, 2 (1%) on imatinib, 33 (18%) on dasatinib, 24 (13%) on bosutinib, 9 (5%) on ponatinib, 1 on ABL001, 31 off‐TKI (20 in TFR and 11 post‐allogeneic transplant). Eleven patients (5.7%) have died, due to CML‐progression (6), secondary malignancy (1), cardiovascular disease (1) and unknown causes after loss to follow‐up (3).Summary/Conclusion:Although requiring validation by larger studies, our data confirm the significant association of NIL with long‐term CVE and identifies age as the most critical factor for their occurrence.image
Read full abstract