Background: Alterations to the tumor suppressorTP53 in mantle cell lymphoma (MCL) are one of the strongest predictors of poor responses to immunochemotherapy (ICT) and early mortality. Recent BSH guidelines (Eyre et al, 2024) recommend TP53 mutation analysis is undertaken on all new MCL diagnoses, and in the United Kingdom (UK) testing is now commonplace. With adverse outcomes observed with conventional ICT in clinical trials, the optimal management for these patients (pts) remains poorly defined. There is limited real-world data to guide clinical practice, providing an imperative to review outcomes of this cohort treated within an integrated healthcare framework. This study aims to provide insights on treatment efficacy; refine prognostic understanding and guide future management. Aims: Assess outcomes of pts receiving first-line (1L) systemic therapy (tx) for TP53 aberrant MCL treated within the UK in the modern era. Methods: A multicenter, retrospective analysis of anonymised data from pts with TP53 aberrant MCL treated with 1L systemic tx at 28 centers across the UK. Patients commenced 1L tx between January 2018 and March 2025. Response to tx was defined as per Lugano classification (Cheson et al, 2014). Data was collected on baseline characteristics and subsequent lines of therapy. Primary outcome was overall survival (OS) from start of 1L tx. Secondary outcomes included tx failure free survival (FFS), defined as time to next line tx, progression of disease or death. Predictors of OS were determined using univariate Cox regression. Results: Data was collected on 120 patients: median age 65 years (range 41-82) with 70% male. TP53 alterations included 113 pts with mutations, and 7 pts with deletions. Prior to 1L tx clinical phenotype was defined as nodal in 69% and leukemic, non-nodal in 31%; LDH ratio >1.5 in 40%; ECOG performance status ≥2 in 17%; blastoid morphology in 21%; Ki67 ≥50% in 46%; and MIPI high risk in 68%. 1L tx was high intensity ICT (rituximab + high dose cytarabine based) in 60 pts (50%), intermediate intensity ICT (R-CHOP/R-Bendamustine/R-BAC/VR-CAP) in 39 pts (33%), non-ICT (BTK inhibitor (BTKi) based) in 17 pts (14%), and low intensity ICT in 4 pts (3%). Ten pts were consolidated at 1L with autologous SCT; 8 pts with allogeneic SCT. Overall response rate to 1L was 68%, complete response rate was 45%. At a median follow up of 23 months (mo), 69 pts progressed, of which, 49 pts had died. Four pts died without relapse. The median 1L FFS was 13.9 mo (95% CI 8.9-18.9), and estimated median OS was 29.8 mo (95% CI 20.8-38.8). The median FFS according to 1L tx was 22.3 mo (95% CI 6.2-38.4) for high intensity ICT; 8.9 mo (95% CI 3.8-14.0) for intermediate intensity ICT; and 14.6 mo (95% CI 4.5-24.7) for non-ICT. On univariate Cox regression, LDH ratio >1.5 was predictive of OS (HR 2.3, 95% CI 1.3-4.3; p=0.007). Age >65 years, nodal vs non-nodal, blastoid morphology, Ki67 ≥50% and MIPI high risk were not predictive Of 69 pts at relapse: 11 pts died without 2L therapy; 48 pts received BTKi, 7 pts received ICT, 3 pts had not commenced 2L. Of 48 pts receiving 2L BTKi, 41 pts progressed and 33 had died. The median FFS from start 2L BTKi was 2.5 mo (95% CI 1.5-3.5), and median OS 5.6 mo (95% CI 4.0-7.2). Of 41 pts progressing on BTKi, only 10 were subsequently infused with 3L brexucabtagene autolecel (brexu-cel) (24%). Conclusion: Relative to contemporary real-world datasets, the OS analysis highlights that current management of TP53 aberrant MCL is inadequate, and there is urgent need to evolve novel strategies that combat this uniquely aggressive biology. Relatively high initial response rates to 1L tx were generally not durable, and subsequent outcomes with 2L BTKi were poor. The low proportion of pts ultimately proceeding to 3L+ brexu-cel intimates the challenge of delivering this therapy to pts with rapidly progressive disease, and provides a rationale for earlier use in this subgroup. Baseline characteristics demonstrated frequent markers of poor prognosis, with LDH ratio >1.5 the only characteristic predictive of OS. Pts presenting with markedly elevated LDH may be most suitable for aggressive 1L management strategies.

Regulatory B (Breg) cells are potentially implicated in the pathogenesis of immune thrombocytopenia (ITP). We analysed a prospective cohort of newly diagnosed steroid naïve ITP patients enrolled in the multicentre FLIGHT trial and found that the numbers of Bregs in their peripheral blood were similar to healthy controls. In contrast, Breg numbers were significantly reduced in ITP patients treated with systemic immunosuppression (glucocorticoids or mycophenolate mofetil). We also demonstrate that glucocorticoid treatment impairs Breg interleukin-10 production via an indirect T-cell-mediated mechanism.

Appendix S1: Survey questions Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

A single 1 g/kg dose of intravenous immunoglobulin is a safe and effective treatment for immune thrombocytopenia; results of the first HaemSTAR 'Flash-Mob' retrospective study incorporating 961 patients Key Messages 1 A one off 1 g/kg infusion of intravenous immunoglobulin (IVIg) may be as effective as two consecutive 1 g/kg doses. 2 This is the largest ever study of the efficacy of IVIg for immune thrombocytopenia (ITP).3 There is poor adherence to the 2016 NHS England guidelines on IVIg dosing.

BACKGROUND:Extra-corporeal photopheresis (ECP) requires anticoagulation to prevent circuit clotting. Unfractionated heparin (UFH) is currently the only anticoagulant licensed for the ECP system in use in the United Kingdom (UK). Acid citrate dextrose-A (ACD-A) is the preferred anticoagulant for most other apheresis procedures. Anecdotal evidence suggested variability in ECP practice across the UK with some providers using off-label ACD-A.AIMS:We developed a survey together with the UK Photopheresis Society to establish current practice.MATERIALS & METHODS:This was distributed to all 17 ECP providers covering 34 UK sites.RESULTS:Significant variability in practice was demonstrated with only 36% of responding providers (5/14) using UFH exclusively and 29% (4/14) using ACD-A as standard.CONCLUSION:This survey highlights the need for a UK consensus.

The British Society of Blood and Marrow Transplantation and Cellular Therapy (BSBMTCT) Registry collects and coordinates information on all haematopoietic stem cell transplantation (HSCT) activity in the UK and the Republic of Ireland. This includes details of peripheral blood and bone marrow stem cell harvest procedures, referred to as ‘harvest’ from here on, such as those required for relateddonor allogeneic HSCT (alloHSCT). In December 2019 the emergence of the coronavirus disease 2019 (COVID19) pandemic led to major concerns about the potential impact of the virus on stem cell harvests. National and international registries underwent rapid policy adaptations to prepare for potential harvest cancellations due to donor availability, donor or patient illness and operational restrictions. The BSBMTCT therefore decided to capture details of all relateddonor harvests that did not proceed as planned over a 12month period from 1 April 2020 to 1 April 2021, capturing the first and second ‘waves’ of the COVID19 pandemic in the UK. Data on unrelateddonor harvests are overseen by Anthony Nolan and therefore not included in this report. A questionnaire (Appendix S1) was developed to answer the following research questions: (i) How many harvests did not proceed as planned? (ii) What were the reasons for this? (iii) What was the impact of COVID19 on stem cell harvests? Questions were designed and reviewed by subject matter experts at the BSBMTCT Executive Committee to ensure validity. Surveys were sent in PDF and Microsoft Word® format to transplant programme directors and data management staff at all 32 adult and paediatric transplant centres performing relateddonor alloHSCTs in the UK. Centres were asked to return data as scanned forms. To increase the number of responses, four individualised reminders were sent to nonresponding centres. No financial remuneration was provided. The data collection period spanned 12 months from 1 April 2020 to 1 April 2021 and complied with General Data Protection Requirements (GDPR) requirements. The survey achieved a response rate of 94% (30/32) for questionnaire return with all responders answering all individual questions relevant to them. Selection of multiple answer options was possible in questions exploring donations described as ‘failed due to COVID19’ in more detail. Centres reported 429 planned donorrelated stem cell harvests of which 15 (3%) did not take place. Four responding centres reported no failed donations over the study period. There were 15 failed harvests from six matched sibling, five matched other relative and four mismatched relative donors. Of the 15 failed harvests, seven (47%) were reported to be due to COVID19. This was due to positive COVID19 swab in the donor (two of seven), recipient (two of seven) or household contact (one of seven), or due to suspension of all nonurgent procedures at the responding centre (two of seven). Symptoms of COVID19 were reported in two of seven of the harvests that did not proceed due to COVID19. The remaining eight failures were due to a change in patient condition (four of eight), donor failure to mobilise (two of eight) or procedural failure due to clotting or toxicity (two of eight). None of the harvests were cancelled due to personal reasons. When considering the timing of the failed harvest in context of stem cell mobilisation, there were nine failures before starting granulocytecolony stimulating factor (GCSF; nine of 15) with four failures after starting GCSF (four of 15). Two donors were planned to undergo bone marrow harvest and therefore did not require mobilisation (two of 15). None of the four cancellations after starting GCSF were COVID19 related. Figure 1 illustrates whether transplant occurred following initial harvest failure. Of the seven harvests that did not occur as planned due to COVID19, four proceeded at a later date and resulted in transplant. Of the eight harvests that failed for reasons other than COVID19, seven resulted in transplant at a later date (six with the same donor, one with a different donor). The median (range) length of postponement was 144 (35– 291) days for COVID19related delays and 68 (2– 277) days for nonCOVID19 delays. Four harvest procedures, and subsequent transplants, were abandoned all together. Three of these were COVID19 related, with patient outcomes described as ‘disease progression’ (two of three) and ‘no further treatment – patient had COVID19’ (one of three). Stem cell donation failures are a known challenge for transplanting physicians and donation centres and are well reported in the literature.1 While reasons for harvest procedure cancellation are not always clearly reported to registries, evidence suggests that many of these are caused by Received: 14 April 2022 | Accepted: 23 May 2022

Since its inception in 1968, haematopoietic stem cell transplantation (HSCT) has been the most widely used type of cellular therapy, providing a potentially curative treatment option for many haematological and non-haematological diseases. In Europe, 4200 HSCTs were reported to the European Blood and Marrow Transplant Group in 1990, a number that rose to 48 512 in 2019. The Human Tissue Act 2004 serves as a good example of how national legislation for cell and tissue collection and processing operates. International identification systems such as ISBT-128 or the Single European Code have been used to standardise identification and labelling of cells and tissues for human application since the early 1990s. The Food and Drug Administration (FDA) recognised the need for regulatory oversight in cell, gene and tissue therapies and products. The FDA recognised the need for regulatory oversight in cell, gene and tissue therapies and products.

Hyperkalaemia following transfusion is widely reported in the literature. Our objective was to critically review recent evidence on hyperkalaemia in association with transfusion and to assess whether specific aspects of transfusion practice can affect the likelihood of developing hyperkalaemia. We searched 9 electronic databases (including MEDLINE, Embase, and Transfusion Evidence Library) using a predefined search strategy, from 2010 to April 8, 2021. Three reviewers performed dual screening, extraction, and risk of bias assessment. We used Cochrane risk of bias (ROB) 2 for assessment of RCTs, ROBINS-I for non-RCTs, and GRADE to assess the certainty of the evidence. We report 7 comparisons of interest in n = 3729 patients from 28 studies (11 RCTs, 4 prospective cohort studies, and 13 retrospective cohort studies): (1) age of blood, (2) washing, (3) filtration, (4) irradiation, (5) fluid type, (6) transfusion vs no transfusion, (7) blood volume/rate. Of the 28 studies included, 25 reported outcomes of potassium (K+) concentration, 17 the number developing hyperkalaemia, 13 mortality, 10 cardiac arrest, and 10 cardiac arrhythmia. Only 16 studies provided analysable data suitable for quantitative analysis. Evidence addressing our outcomes was of very low certainty (downgraded for incomplete outcome data, baseline imbalance, imprecision around the estimate, and small sample size). While 5 studies showed a difference in K+ concentration up to 6 hours posttransfusion for 3 comparisons (age of blood, washing, and transfusion volume/rate), and 3 studies showed a difference in the diagnosis of hyperkalaemia for 2 comparisons (age of blood, and transfusion volume/rate), the evidence was inconsistent across all included studies. There was no difference in any reported outcomes for 4 comparisons (filtration, irradiation, fluid type, or transfusion vs no transfusion). Overall, the reported evidence was too weak to support identification of groups most at risk of hyperkalaemia or to support recommendations on use of short-storage RBC. For other commonly used risk mitigations for hyperkalaemia in transfusion medicine, the (low certainty) evidence was either conflicting or not supportive.

Consensus clinical guidelines recommend high-dose glucocorticoids as first-line therapy for adult patients with immune thrombocytopenia (ITP).1 However, frequent side effects and heterogeneous responses are clinical challenges with approximately 20–30% failing to respond at tolerable doses.2 Although addition of mycophenolate or rituximab to glucocorticoid may increase response rates, quality of life was worse with mycophenolate and more adverse events were associated with rituximab.3, 4 Therefore, it would be clinically valuable to predict those patients expected to fail glucocorticoid monotherapy who may benefit from early additional treatment. Studies have suggested that CD4+ T helper cell subsets from ITP patients respond differently to glucocorticoid treatment in vitro. Cells expressing the pro-inflammatory cytokines interleukin (IL)-17 and interferon (IFN)-γ in the absence of the anti-inflammatory cytokine IL-10 are resistant to inhibition.5-7 Furthermore, we recently reported that activated CD4+ T cells from glucocorticoid-refractory ITP patients showed a relative abrogation of IL-10 with persistent IL-17 in response to in vitro glucocorticoid compared to responsive patients.8 However, this study was limited by the heterogeneity of the ITP patient cohort in terms of disease stage and treatment. Here, we aimed to prospectively validate our findings in a cohort of newly diagnosed ITP patients recruited to the laboratory sub-study of the FLIGHT trial (ClinicalTrials.gov number: NCT03156452). Following written informed consent, peripheral blood (20 ml) was collected in EDTA-containing tubes at baseline (when randomised; n = 87) and at 2 months follow-up (n = 80). Full methods of the FLIGHT trial were previously published9 – in brief, patients were randomised to receive glucocorticoid (dexamethasone or prednisolone) alone or in combination with mycophenolate. A glucocorticoid response (R; n = 34) was defined as platelet count of >30 × 109/l and at least a twofold increase from baseline with glucocorticoid monotherapy. Non-responsive (NR) patients were defined by platelet count of <30 × 109/l, or less than a twofold increase from baseline with glucocorticoid therapy either alone (n = 11) or in combination with mycophenolate mofetil (n = 2). CD4+ T cells were isolated from peripheral blood using a RosetteSepTM Human CD4+ T cell enrichment cocktail (StemCell Technologies, Vancouver, British Columbia, Canada). Cytokine expression was examined immediately ex vivo by 4-h incubation with phorbol 12-myristate 13-acetate (PMA) (20 ng/ml), ionomycin (1 µmol/l) and GolgiStopTM (BD Biosciences, Oxford, UK) (2 µmol/l). Alongside this, CD4+ T cells were activated with Human T Activator CD3/CD28 Dynabeads (ThermoFisher Scientific, Loughborough, UK) and cultured in vitro in the presence of recombinant IL-2 (Roche, Welwyn Garden City, UK) and dexamethasone, a relatively pure glucocorticoid (1 µmol/l; Sigma-Aldrich, Gillingham, UK), for 96 h under tissue culture conditions with PMA, ionomycin and GolgiStopTM added for the last 4 h of culture. Intracellular cytokine staining was measured by flow cytometry using a BD Fortessa X20 (BD Biosciences, Oxford, UK). Masked analysis was carried out using FlowJo software (version 10; Tree Star Software, Ashland, OR, USA). Statistical analysis was carried out using GraphPad Prism software (version 9; GraphPad Software, San Diego, CA, USA). Following in vitro culture with dexamethasone treatment, CD4+ T cells from baseline samples of R patients compared to NR patients showed higher proportions of IL-10+ cells, no significant difference in IL-17+ cells and a higher IL-10:IL-17 ratio (Table I). An area under the receiver operating characteristic (ROC) curve of 0·7115 [95% confidence interval (CI) 0·5301–0·8930, P = 0·0358] demonstrates significant discrimination of the IL10:IL17 ratio between R and NR patients. This is consistent with our previous findings.8 There were no significant differences in the expression of interleukin (IL)-4, IL-22, interferon (IFN)-γ, granulocyte-macrophage colony-stimulating factor (GM-CSF) or tumour necrosis factor (TNF) between R and NR patients. Clinical response CD4+ T cells examined immediately ex vivo from baseline samples of R patients compared to NR patients, also demonstrated a higher IL-10:IL-17 ratio although with no significant difference in the individual proportions of IL-10+ and IL-17+cells (Fig 1A–D). The ex vivo IL-10:IL-17 ratio shows discrimination between patient cohorts, with an area under ROC (AUROC) of 0·7889 (95% CI 0·6432–0·9346, P = 0·0038; Fig 1E). NR patients had higher proportions of IL-4+ cells than R patients with no significant differences in any of the other cytokines. CD4+ T cells examined following in vitro culture with dexamethasone from 2-month follow-up samples of R patients compared to NR patients again demonstrated a higher IL-10:IL-17 ratio (AUROC 0·7642, 95% CI 0·5830–0·9454, P = 0·0096) with higher proportions of IL-10+ cells but no significant difference in IL-17+ cells. CD4+ T cells examined immediately ex vivo from samples at 2 months follow-up of R patients compared to NR patients, also demonstrated a significant difference in the IL-10:IL-17 ratio, with lower proportions of IL-17+ cells and no significant difference in IL-10+ cells. There were no significant differences in the other cytokines studied. These results provide evidence of an adaptive immune basis in the clinical diversity of glucocorticoid responsiveness, particularly related to IL-10 and IL-17 cytokine expression. While limited by a relatively small number of patients, the prospective cohort of newly diagnosed patients with blood sampling at two time points and robust clinical response data are strengths. We confirm our previous findings where CD4+ T cells from NR patients have a reduced IL-10:IL-17 ratio following in vitro treatment with glucocorticoids. Furthermore, we have expanded on these observations and shown that immediately ex vivo CD4+ T cells also demonstrate a reduced IL-10:IL-17 ratio in NR patients. Results from 2-month follow-up samples are consistent with this finding. This suggests that CD4+ T cells from NR patients are predisposed to produce sustained low levels of IL-10 and higher levels of IL-17. A key mechanism underlying glucocorticoid efficacy is likely to be the induction of IL-10 in a range of immune cell types, including CD4+ and CD8+ T cells,10, 11 and B cells.12 Indeed, failure to upregulate IL-10 in response to glucocorticoids has been identified in glucocorticoid-resistant asthma.7 CD4+ T cells upregulating IL-10 in response to glucocorticoids often co-produce T helper (Th)-17-associated cytokines (IL-17, IL-22 and IFN-γ).13 These ‘non-pathogenic Th17 cells’ are important in limiting inflammation and autoimmunity.14 Contrastingly, Th17 cells that do not upregulate IL-10 in response to glucocorticoid have been shown to be both highly pathogenic and glucocorticoid-resistant.5, 6 It may be possible to develop the observed ex vivo T cell phenotype into a clinically applicable biomarker to help predict NR patients, which could then inform the clinical decision to initiate alternative therapy early. Our data may also be applicable to other autoimmune illnesses treated with glucocorticoids8 and highlight the potential importance of CD4+ T cells in both ITP pathogenesis and therapy. Intracellular cytokine expression of CD4+ T cells from ITP patients activated with anti-CD3/CD28 beads for 96 h in the presence of 1 × 10−6 mol/l dexamethasone (in vitro culture with Dex), or directly ex vivo (immediate ex vivo phenotype) was assessed by flow cytometry on a BD Fortessa X20. The mean proportion of positive cells and the SD are shown. Patients are divided by their clinical response into those recruited to the corticosteroid-alone arm who responded (R), or patients recruited to both the glucocorticoid-only or glucocorticoid + mycophenolate arm and had no response (NR) to treatment. Samples taken at baseline (R n = 34, NR n = 13) and 2 months follow-up (R n = 29, NR n = 10) are shown. Groups were compared by Mann–Whitney U-test and significance indicated; a P value of less than 0·05 was taken as significant. CB and RL jointly led this research. CB led the FLIGHT clinical trial, JP and IT were responsible for the clinical trial delivery (Cardiff CTU), RG is the trial statistician and has provided expert input into statistical analysis of the data. MS led the laboratory processing of samples and wrote the first draft of the manuscript. JW wrote the final version of the manuscript. PL, EW and LSB contributed to sample analysis. All authors have made valuable contributions to the research design, delivery and provided feedback on the manuscript. Ethical approval from NRES Committee South West (IRAS number 225959). EudraCT Number: 2017-001171-23. This project is partly funded by the National Institute for Health Research (NIHR) under its Research for Patient Benefit (RfPB) Programme (Grant Reference Number PB-PG-0815-20016). The views expressed are those of the author(s) and not necessarily those of the NIHR or the Department of Health and Social Care. In addition, project-specific funding for biomarker development was received through a grant from the British Medical Association. ELW, PJPL, LPS-B and RWJL also received salary support from the NIHR Biomedical Research Centre at Moorfields Eye Hospital NHS Foundation Trust and UCL Institute of Ophthalmology. RWJL and LPS-B are inventors of an IL-17-based method to identify patients likely to be resistant to glucocorticoid treatment (US Patent App. 15/106,411). The other authors have no conflicts to declare.

Infection-related morbidity and mortality are increased in older patients with diffuse large B-cell lymphoma (DLBCL) compared with population-matched controls. Key predictive factors for infection-related hospitalization during treatment with rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP) and deaths as a result of infection in older patients during and after treatment with R-CHOP remain incompletely understood. For this study, 690 consecutively treated patients age 70 years or older who received full-dose or attenuated-dose R-CHOP treatment were analyzed for risk of infection-related hospitalization and infection-related death. Median age was 77 years, and 34.4% were 80 years old or older. Median follow-up was 2.8 years (range, 0.4-8.9 years). Patient and baseline disease characteristics were assessed in addition to intended dose intensity (IDI). Of all patients, 72% were not hospitalized with infection. In 331 patients receiving an IDI ≥80%, 33% were hospitalized with ≥1 infections compared with 23.3% of 355 patients receiving an IDI of <80% (odds ratio, 1.61; 95% confidence interval, 1.15-2.25; P = .006). An increased risk of infection-related admission was independently associated with IDI >80% across the whole cohort. Primary quinolone prophylaxis independently reduced infection-related admission. A total of 51 patients died as a result of infection. The 6-month, 12-month, 2-year, and 5-year cumulative incidences of infection-related death were 3.3%, 5.0%, 7.2%, and 11.1%, respectively. Key independent factors associated with infection-related death were an International Prognostic Index (IPI) score of 3 to 5, Cumulative Illness Rating Scale for Geriatrics (CIRS-G) score ≥6, and low albumin, which enabled us to generate a predictive risk score. We defined a smaller group (15%) of patients (IPI score of 0-2, albumin >36 g/L, CIRS-G score <6) in which no cases of infection-related deaths occurred at 5 years of follow-up. Whether patients at higher risk of infection-related death could be targeted with enhanced antimicrobial prophylaxis remains unknown and will require a randomized trial.

Diffuse large B-cell lymphoma (DLBCL) and osteoporotic fracture are both more common in older patients. Exposure to R-CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisolone) is likely to increase the risk of fracture, but evidence is lacking to define fracture incidence in this group. Data on consecutive patients with DLBCL aged ≥70 years treated with 1 to 8 cycles of full or attenuated R-CHOP were retrospectively collected across 10 UK centers (2009-2019). Patients were followed up from starting R-CHOP for a minimum of 6 months and censored at 18 months; at last follow-up if <18 months; or at progression or death. Of 877 patients identified, 148 were excluded: 121 had progression or died before 6 months; 23 had follow-up <6 months. Across 729 remaining patients, the median age was 77 years, and 68% had an Eastern Cooperative Oncology Group performance status of 0 to 1. Eighty-one fractures occurred within 18 months of follow-up; 42 were symptomatic, including 30 requiring hospital attendance or admission. The cumulative fracture incidence was 6.2% (95% confidence interval [CI], 4.7-8.2) at 6 months; 9.7% (95% CI, 7.8-12.1) at 12 months; and 11.4% (95% CI, 9.3-14.0) at 18 months. Multivariate analysis identified a predisposing history (osteoporosis, osteopenia, prior fracture, and rheumatoid arthritis [RhA]), DLBCL bone involvement at baseline, and receipt of prephase steroids as independent risk factors for fracture. There is a clinically relevant fracture risk and significant associated morbidity in older patients receiving R-CHOP. Careful attention to bone health is warranted in older patients receiving R-CHOP. Randomized studies are required to better define the most effective strategies to reduce fracture risk.

Carfilzomib is a second generation irreversible proteasome inhibitor, NICE-approved in 2017 to treat myeloma at first relapse (NICE, 2017). Approval was based on data from ENDEAVOR trial which compared carfilzomib plus dexamethasone (CarDex) given twice-weekly with bortezomib plus dexamethasone (BorDex) (Dimopoulos et al., 2016). Median progression-free survival (PFS) showed superiority of the CarDex arm (18·7 vs. 9·4 months, P < 0·0001). However, carfilzomib was associated with high cardiovascular adverse events (CVAEs) with a range of severity grades (G): hypertension (G1-2: 16%; ≥G3: 9%), cardiac failure (CF; G1-2: 3%; ≥G3: 4·8%) and ischaemic heart disease (G1-2: 0·9%; ≥G3: 1·7%), (Dimopoulos et al., 2016). To further explore carfilzomib dosing and frequency, ARROW trial randomised patients to once weekly dosing (at 70 mg/m2) or twice weekly dosing (at 27 mg/m2) (Moreau et al., 2018). Median PFS was statistically higher in the once weekly arm (11·2 vs. 7·6 months; P = 0·0029). However, ≥G3 of any adverse events (AEs) (once versus twice-weekly) occurred in 68% and 62% of patients, respectively. In addition, ≥G3 hypertension and CF were (6% vs. 5%) and (3% vs. 4%), respectively (Moreau et al., 2018). Given the efficacy and toxicity outcomes reported in ENDEAVOR and ARROW and in order to improve tolerability and clinical outcomes in routine practice, carfilzomib dosing and frequency need to be optimised. The European Myeloma Network (EMN) has recently published guidance on the prevention, monitoring and management of CVAEs in carfilzomib patients, which includes a comprehensive clinical assessment to identify risks of CVAEs, and the use of dose interruptions/reductions where necessary (Bringhen et al., 2019). A better understanding of the tolerability and efficacy outcomes of this novel therapy in unselected relapsed patients in the real world is required, particularly in those with pre-existing cardiac morbidities or advanced age. We reviewed carfilzomib dosing, efficacy and toxicity outcomes in a cohort of myeloma patients treated with CarDex at first relapse. To our knowledge, there are no published data describing real-world carfilzomib outcomes in this setting. Data were retrospectively collected across seven UK centres for a period of 22 months (July 2017 to April 2019). Baseline characteristics are presented in Table 1. Outcomes evaluated included response rates, time to best response, and all grade toxicities. Survival outcomes were not examined because the follow-up period was too short. We also reviewed carfilzomib dosing schedules, treatment duration, relative dose intensity (RDI), and dose reductions/discontinuations. The study received NHS service evaluation approval at each participating site. Data were censored on 1 April 2019. The method is further described in the Data S1. Thirty patients received carfilzomib therapy for myeloma at first relapse as per NICE approval. Median age at initiation of therapy was 70 years (range 53–86). Thirteen patients (43·3%) presented with at least one pre-existing cardiac morbidity, of whom 10 had hypertension. Median follow-up period was 8·7 months (range 1·2–20·2). Detailed baseline characteristics are presented in Table 1. Twenty-nine patients (97%) received at least two cycles of therapy and one patient discontinued carfilzomib after one cycle. Responses were not evaluable in 3/30 patients – two discontinued very early due to toxicities [one after only one cycle (due to G2 dyspnoea, G2 worsening pre-existing left ventricular failure, and G3 nausea and vomiting), the second after two cycles (due to G2 ulcerated swollen leg and G2 neuropathy which was possibly an exacerbation of previous thalidomide-induced neuropathy)], and the third patient moved away from the geographical area. The latter two patients did not undergo response assessments. For those with evaluable responses (27/30), overall response rate (ORR) was 92·6% (25/27). Response categories were as follows: complete response (CR) 7/27, very good partial response (VGPR) 11/27, partial response (PR) 7/27, and minor response (MR) 2/27. Median time to best response was 75 days (range 30–210), i.e. after two to three cycles. Carfilzomib therapy is ongoing in 17 patients and was discontinued in 13 patients. Median number of cycles for those still on treatment is nine (range 3–20). Reasons for treatment discontinuation were: autologous stem cell transplant (ASCT) 5/13, disease progression 3/13, palliation due to a second primary malignancy (SPM) 1/13, poor compliance/attendance 1/13, and toxicity 3/13. Two of these patients experienced early toxicities as described above, and one experienced recurrent infections. Two of the 13 patients died – one due to SPM requiring palliation; another received four cycles of carfilzomib and completed harvesting for ASCT, but later relapsed (not deemed carfilzomib-refractory), and started a next line of therapy, then died of progressive disease. Carfilzomib dosing was variable across the cohort. Five patients received full twice-weekly dosing throughout therapy, and six patients were dose-reduced to weekly upfront (i.e. from cycle 1). The remainder of the cohort (19/30 = 63·3%) received the full dose for a variable number of cycles (range 1–12) and later underwent a dose level reduction (to 42 or 36 or 27 mg/m2) or a frequency reduction to weekly, or both. Reasons for dose reductions were infections (G2/G3), hypertension (G2/G3), dyspnoea (G1/2), chest pain (G1/G2), history of GI perforation (G3), fatigue (G2), increase in baseline creatinine by nearly 50%, and diagnosis with pulmonary embolism (PE) (G3). The median relative dose intensity (RDI) for the total cohort throughout the follow-up period was 67·4% (range: 26·6–100%). Median RDI according to responses for CR, VGPR and PR patients were 61·1%, 62·2% and 51·7%, respectively. The average dose intensity for the cohort from one cycle to the next showed a downward trend (Figure S1). This correlates with dose reductions due to toxicity. Sixty AEs were associated with CarDex therapy and are presented in Table 2. Sixteen AEs were ≥G3, of which the most frequent were: anaemia (4), infections (2), hypertension (2) and PE (2). The median nadir haemoglobin (Hb) on therapy was 104 (range 72–122). Time to Hb recovery to baseline was 19 days (range 5–60). Of the seven hypertension events, four were due to worsening pre-existing hypertension (G1-2) and three were new (one was G2 and two were G3). Depending on the severity of hypertension, management consisted of introducing an antihypertensive (e.g., amlodipine) in two patients, an antihypertensive dose increase in two patients, carfilzomib dose reduction in two patients, and an antihypertensive plus carfilzomib dose reduction in one patient. This is the first study to report real-world outcomes of carfilzomib in the UK. To our knowledge, there are no published UK data describing myeloma outcomes in this setting. Our study demonstrated that carfilzomib resulted in a high ORR. It was also reasonably well tolerated (total of 16 AEs of ≥G3). Higher grade toxities (≥G3) of (this cohort vs. ENDEAVOR trial) were comparable for thrombocytopenia (10% vs. 8·4%), anaemia (13·3% vs. 14·5%), and hypertension (6·7% vs. 9%). Our patients experienced a higher rate of ≥G3 infections compared to the trial population (6·7% vs. 2%), although sample size of our cohort is small. In addition, our study demonstrated that cardiac co-morbidities need to be carefully taken into account but should not be a reason to exclude carfilzomib as a treatment option. However, and as per recent EMN guideline, risk assessment prior to initiation of treatment is imperative, as well as comprehensive monitoring and proactive management which includes carfilzomib dose interruptions and/or reductions, prior to early discontinuation. Our study is limited by its retrospective, non-randomised nature with the inherent possibility of unmeasured confounding factors, patient selection bias, and its limited sample size. Despite this, we show that this therapeutic option is clinically robust for relapsed myeloma. FD: Takeda: Honorarium, Travel to scientific conferences, Celgene: Honorarium, Research support, Travel to scientific conferences. Novartis: Honorarium, Amgen: Honorarium. HH: Nil. PN: JW: Nil. MO: Nil. AK: Nil. .EL: Nil. GV:Nil. JK:Nil .SM:Nil. KR:Nil. Not required. FD designed the study. FD, HH, PN, JW, MO, AK, GV, JK, SM and KR collected data and managed patients in this study. FD analysed data. FD wrote the manuscript, which all authors critically reviewed and approved. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

SummaryCentral nervous system (CNS) relapse following R‐CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine, prednisolone) occurs in 2–5% of patents with diffuse large B‐cell lymphoma (DLBCL). Many patients aged ≥70 years are unsuitable for high‐dose methotrexate (HDMTX) prophylaxis and therefore often receive stand‐alone intrathecal prophylaxis. The CNS international prognostic index (CNS‐IPI) is a clinical CNS relapse risk score that has not specifically been validated in elderly patients. The value of CNS prophylaxis in patients aged ≥70 years remains uncertain. Data on 690 consecutively R‐CHOP‐treated DLBCL patients aged ≥70 years were collected across 8 UK centres (2009–2018). CNS prophylaxis was administered per physician preference. Median age was 77·2 years and median follow‐up was 2·8 years. CNS‐IPI was 1–3 in 60·1%, 4 in 23·8%, 5 in 13·0% and 6 in 3·3%. Renal and/or adrenal (R/A) involvement occurred in 8·8%. Two‐year overall CNS relapse incidence was 2·6% and according to CNS‐IPI, 1–3:0·8%, 4:3·6%, 5:3·8% and 6:21·8%. Two‐year CNS relapse incidence for R/A was 10·0%. When excluding HDMTX (n = 31) patients, there remained no change in unadjusted/adjusted CNS relapse for intrathecal prophylaxis effect according to CNS‐IPI. CNS‐IPI is valid in elderly R‐CHOP‐treated DLBCL patients, with the highest risk in those with CNS‐IPI 6 and R/A involvement. We observed no clear benefit for stand‐alone intrathecal prophylaxis but observed an independent increased risk of infection‐related admission during R‐CHOP when intrathecal prophylaxis was administered.