Daratumumab is effective in the relapsed or refractory systemic light‐chain amyloidosis but associated with high infection burden in a frail real‐life population

Systemic light-chain amyloidosis (AL) is characterised by the deposition of immunoglobulin light chains (FLC) in multiple organs, where they lead to progressive organ failure (Merlini & Stone, 2006). In most cases, the underlying disease is a clonal plasma cell disorder and AL patients have consequently benefited from the vast therapeutic advances in multiple myeloma (MM). The use of proteasome inhibitors (PI) has increased the likelihood of deep haematological responses with combination regimens yielding more than 90% overall haematological response rates (ORR) including 50% complete responses (CR) (Gavriatopoulou et al., 2018b). Nonetheless, responses are short-lived and patients invariably relapse. More intensive approaches, on the other hand, often fail (Palladini & Merlini, 2011) and prognosis remains dire (Kumar et al., 2011). Thus, it is imperative to find a rapidly effective treatment, with a low-toxicity profile. Daratumumab has shown promising results in MM (Lonial et al., 2016; Boyle et al., 2019); data in AL are scarce, but promising (Kaufman et al., 2017). When dealing with AL, both trial and non-trial data are required as real-life patients are often more frail and less fit than their trial counterparts (Gavriatopoulou et al., 2018a). Given the multiplicity of end-organ involvement, the incidence of disability, hospital admission, and mortality, is high regardless of the patient's age. So far, the number of organs involved, and ECOG (Eastern Cooperative Oncology Group) performance status have been associated with a poor outcome but indices such as the Charlson comorbidity index (CCI) or the Cumulative Illness Rating Scale (CIRS) have yet to be used in AL despite offering a better measure of frailty in MM (Palumbo et al., 2015). This led us to retrospectively analyse all histologically proven AL patients treated with daratumumab in five Hauts-de-France centres, from November 2016 to March 2018. Fifteen patients received daratumumab (16 mg per kilogram) as an intravenous infusion. Daratumumab was administered weekly during the first two 28-day cycles, every other week during cycles three through six, and monthly from cycle seven until disease progression or unacceptable toxicity, as previously published (Lokhorst et al., 2015). All patients were given herpes zoster prophylaxis with valaciclovir. Data were collected after agreement from the French Data Protection Authority (Commission nationale de l'informatique et des libertés, CNIL) (registration n°2197989) and haematological and organ responses were defined according to guidelines (Palladini et al., 2012). Overall, the population was young with a median age of 60 years [interquartile range (IQR), 47–67] The median time from diagnosis was 21 months (IQR, 10–39) and patients had received a median of two (range, 1–5) prior therapies including a PI in 93% of cases. The median number of organ involvements at diagnosis was three (range, 1–6), including cardiac involvement in 67% (n = 10) patients, all Mayo-Clinic stage III. Overall, the population was frail with a median ECOG score of 2 (range 0–3), a CIRS score ≥6 in all patients (median 14; range, 8–27), and a Charlson score ≥2 in 87% patients (median 4; range 1–8; Table 1. The median time between last treatment received and daratumumab initiation was 1·7 months (IQR, 1–9). Reasons for initiating daratumumab treatment were refractory disease in three patients (20%), relapse in five (33%) and insufficient haematological response in seven (47%). All patients but one completed at least one cycle, and were therefore evaluable for response. The ORR was 86% (n = 12), including six CR (43%), two very good partial responses (VGPR, 14%), and four partial responses (PR, 29%; Fig 1A). When obtained, first haematological response was achieved after one cycle and best haematological response after three cycles (range 1–15). Eleven patients (73%) were evaluable for organ response. Global organ response was seen in 45%, (n = 5) after a median of four cycles (range, 3–9; Fig 1B). The CIRS was significantly lower among responders (mean ± SEM 14 ± 1·3 vs. 22·5 ± 4·5, P = 0·04). The median follow-up was 7·9 months (range, 1·5–17·2) and 87% were still alive at that time point. Causes of death include infection (n = 1) and disease progression (n = 1). Patients received a median of 12 cycles (range, 0·25–19). Infusion-related reactions were observed in five patients (33%), all grade I–II. All reactions occurred during or after the first infusions, and rapidly resolved with symptomatic treatment. No heart failure was seen after infusions, but one patient had transient onset arrhythmia. Despite standard prophylaxis, infections were the most common adverse event in our population, and 30% were grade III or more. Nine patients (60%) experienced infectious events including pneumonia (n = 7), upper respiratory tract infections (n = 8), and bacteraemia (n = 2; Fig 1C). As expected, several patients experienced progressive hypogammaglobulinaemia that might add to the risk of infections, Fig 1D. The median gammaglobulin level before treatment was 4·8 g/l (range, 1·6–9·7) vs. 3·3 g/l (range, 1·2–5·4) afterwards. In conclusion, these results are encouraging with daratumumab yielding high and deep response rates leading to frequent organ response. As reported in MM, responses are seen early, which could suggest early discontinuation for non-responders. Overall, daratumumab appears safe, including in the two haemodialysis patients, but infections were seen at twice the rate of what has been described in MM (Boyle et al., 2019) and led to significant morbidity and mortality. This may be related to the frailty of the population. In AL, it is well established that an ECOG score ≥2 negatively impacts survival, especially in elderly patients and in the autologous stem cell transplant setting (Sachchithanantham et al., 2015). For myeloma patients, a Charlson score ≥2 impairs overall survival and has been validated (Palumbo et al., 2014). The CIRS score is also used in cancer patients, where the cut-off to predict mortality varies between 3 and 4 (Balducci & Beghe, 2000). In this series, the CIRS was significantly lower among responders. As a single agent, daratumumab allows deep fast haematological responses in AL patients. Infectious complications were nonetheless noted, which should prompt close monitoring and rapid action. Optimal treatment duration has yet to be defined in future prospective clinical trials that should emphasise prophylactic measures for infection and may be guided by frailty scores. ZVDW, EMB and TF designed and conducted the study. BC, LP, PLH, IL, MS, MV, HD, LT, CH, SM, JBB, SB, PC, AW, MN, CB and JBG collected data and managed patients. ZVDW and EMB analysed the data and wrote the paper. All authors reviewed the abstract. EMB discloses lecture fees from Janssen, Abbvie and Celgene and travel fees from Amgen and Celgene, but none were in relationship to this paper. TF is on the Board of Directors or on an advisory committee for Janssen, Celgene, Takeda, Sanofi, Karyopharm, PharmaMar and Oncopeptide and has been affiliated with the Speakers Bureau for Janssen, Takeda and Celgene.