Pharmacokinetics and safety of subcutaneous rituximab in follicular lymphoma (SABRINA): stage 1 analysis of a randomised phase 3 study

Pharmacokinetics and safety of subcutaneous rituximab in follicular lymphoma (SABRINA): stage 1 analysis of a randomised phase 3 study Andrew Davies, Francesco Merli, Biljana Mihaljevic, Noppadol Siritanaratkul, Phillippe Solal-Céligny, Martin Barrett, Claude Berge, Beate Bittner, Axel Boehnke, Christine McIntyre, David MacDonald Summary Background Intravenous rituximab is a mainstay of treatment for follicular lymphoma. A subcutaneous formulation that achieves equivalent rituximab serum concentrations might improve convenience and save health-care resources without sacrificing clinical activity. We aimed to assess pharmacokinetic non-inferiority of 3 week cycles of fixed-dose subcutaneous rituximab versus standard intravenous rituximab. Methods In our two-stage, randomised, open-label, phase 3 trial, we enrolled patients with previously untreated grade 1 3a, CD20-positive follicular lymphoma at 67 centres in 23 countries. In stage 1, we randomly allocated patients 1:1 with the Pocock and Simon algorithm to intravenous rituximab (375 mg/m²) or fixed-dose subcutaneous rituximab (1400 mg), stratified by induction chemotherapy regimen (cyclophosphamide, doxorubicin, vincristine, prednisone or cyclophosphamide, vincristine, prednisone), Follicular Lymphoma International Prognostic Index score, and region. After randomisation, patients received one induction dose of intravenous rituximab in cycle 1 and then allocated treatment for cycles 2 8. Patients with a complete or partial response following induction therapy continued intravenous or subcutaneous rituximab as maintenance every 8 weeks. The primary endpoint was the ratio of observed rituximab serum trough concentrations (C trough ) between groups at cycle 7 (before cycle 8 dosing) of induction treatment in a per-protocol population. Patients were analysed as treated for safety endpoints. Stage 2 follow-up is ongoing and is fully accrued. This study is registered with ClinicalTrials.gov, number NCT01200758. Findings Between Feb 4, 2010, and Oct 21, 2011, we enrolled 127 patients. Pharmacokinetic data were available for 48 (75%) of 64 patients randomly allocated intravenous rituximab and 54 (86%) of 63 patients randomly allocated subcutaneous rituximab. Geometric mean C trough was 83 13 μg/ml in the intravenous group and 134 58 μg/ml in the subcutaneous group (ratio 1 62, 90% CI 1 36 1 94), showing non-inferiority of subcutaneous rituximab. 57 (88%) of 65 patients in the intravenous rituximab safety population had adverse events (30 [46%] grade 3), as did 57 (92%) of 62 patients in the subcutaneous rituximab safety population (29 [47%] grade 3). The most common grade 3 or worse adverse event in both groups was neutropenia (14 [22%] patients in the intravenous group and 16 [26%] patients in the subcutaneous group). Adverse events related to administration were mostly grade 1 2 and occurred in 21 (32%) patients in the intravenous group and 31 (50%) patients in the subcutaneous group. Interpretation Stage 1 data show that the pharmacokinetic profile of subcutaneous rituximab was non-inferior to intravenous rituximab and was not associated with new safety concerns. Stage 2 will provide data for efficacy and safety of the subcutaneous administration. Funding F Hoffmann-La Roche. Introduction The chimeric anti-cd20 monoclonal antibody rituximab (F Hoffmann-La Roche, Basel, Switzerland) is a mainstay of therapy for B-cell malignancies, including follicular lymphoma. Rituximab increases time-to-treatment failure and overall survival in the first-line treatment of follicular lymphoma when used with chemotherapy. 1,2 Maintenance rituximab after first-line induction therapy has further improv ed progression-free survival in primary follicular lymphoma 3,4 and when treating subsequent remission. 5 Standard rituximab administration involves intravenous infusions lasting 1 5 6 h, which is incon venient to patients and a burden on health-care resources. Subcutaneous delivery could simplify administration and improve convenience, and has been assessed for rituximab, veltuzumab, trastuzumab, and alemtuzumab in various cancer settings. 6 9 Subcutaneous delivery could also reduce the incidence of severe administrationrelated reactions and costs. 10,11 Establishment of the subcutaneous route with the existing intravenous rituximab formulation was hindered by injection volumes exceeding those that are normally tolerated. Therefore, a subcutaneous rituximab formu lation was developed for delivery at a fixed dose in 5 6 min. 7,12 Subcutaneous rituximab is concentrated at 120 mg/ml compared with the intravenous formulation of 10 mg/ml, and is coformulated with recombinant human Lancet Oncol 2014; 15: 343 52 Published Online February 10, 2014 http://dx.doi.org/10.1016/ S1470-2045(14)70005-1 See Comment page 254 Cancer Research UK Centre, University of Southampton Faculty of Medicine, Southampton, UK (A Davies BM); Hematology Unit, Arcispedale Santa Maria Nuova IRCCS, Reggio Emilia, Italy (F Merli MD); Clinic of Hematology, CCS, and Faculty of Medicine, University of Belgrade, Serbia (B Mihaljevic MD); Division of Hematology, Department of Medicine, Siriraj Hospital, Bangkok, Thailand (N Siritanaratkul MD); Institut de Cancérologie de l Ouest, Nantes, France (P Solal-Céligny MD); Clinical Development (M Barrett PhD) and Clinical Pharmacology (C McIntyre PhD), Roche Products, Welwyn Garden City, UK; Clinical Development, F Hoffmann-La Roche, Basel, Switzerland (C Berge MSc, B Bittner PhD, A Boehnke MD); and Queen Elizabeth II Health Sciences Centre, Halifax, Nova Scotia, Canada (D MacDonald MD) Correspondence to: Dr Andrew Davies, Cancer Research UK Centre, Cancer Sciences Unit, Somers Cancer Research Building, Southampton General Hospital (Mail point 824), Southampton S016 6YD, UK A.Davies@southampton.ac.uk www.thelancet.com/oncology Vol 15 March 2014 343

See Online for appendix hyaluronidase (rhuph20). rhuph20 transiently degrades interstitial hyaluronan at the injection site, increasing the volume that can be administered and facilitating drug entry into the circulation. 13 To confirm efficacy and safety of subcutaneous rituximab, a pharmacokinetic-based clinical bridging approach was used to establish non-inferiority in pharmacokinetic (and similarity in clinical) endpoints for intravenous and subcutaneous administration of rituximab. Rituximab serum trough concentration (C trough ) and area under the concentration time curves (AUC) correlate with clinical efficacy 14,15 whereas maximum concentration (C max ) does not. 15 Serum C trough levels after subcutaneous administration that are at least as high as those achieved after intravenous rituximab will provide at least the same degree of target-site saturation, which is expected to achieve the same degree of efficacy. By showing pharmacokinetic non-inferiority according to the established intravenous rituximab dose and dosing interval (pharmacokinetic bridging) and excluding reduction in rituximab s antilymphoma activity in one non-hodgkin lymphoma indication (clinical bridging using follicular lymphoma as an example disease), the results can be applied to other non-hodgkin lymphoma indications. A similar approach with subcutaneous trastuzumab led to European Commission approval in HER2-positive breast cancer. 16 19 Although cytotoxic drugs are generally dose-adjusted to body surface area or weight, for antibodies such as rituximab with wide therapeutic windows, 20,21 fixed dosing could offer advantages, including avoidance of dose errors and reduced costs from production and storage of one unit dose. 22,23 A simulation study of fixeddose intravenous monoclonal antibodies, including rituximab, showed that interparticipant exposure variability was not increased significantly relative to dosing based on body size. 24 The two-stage SparkThera (NCT00930514) study 25 of fixed-dose subcutaneous rituximab (1400 mg) for untreated or relapsed follicular lymphoma suggested maintenance of serum C trough in the range of those achieved with standard intravenous maintenance dosing (375 mg/m²). The study then showed non-inferiority of this dose in terms of C trough by entering observed data into a population pharmacokinetic model to generate individually simulated serum concentration profiles, and calculated C trough and AUC following subcutaneous administration over a range of doses. Safety profiles of both formulations did not differ, apart from an increase in administration-related reactions (eg, injection-site erythema) with subcutaneous rituximab; however, most events were local, mild, reversible, and in line with the expected safety profile for subcutaneous administration. The subcutaneous rituximab 1400 mg dose identified in SparkThera was selected for further development in SABRINA. This phase 3 study was designed to assess pharmacokinetic non-inferiority of fixed-dose subcutaneous rituximab (1400 mg) versus intravenous rituximab (375 mg/m²) given every 3 weeks and to investigate if the subcutaneous route of administration would impair rituximab s antilymphoma activity. 12 SABRINA was done in the first-line induction immunochemotherapy and maintenance settings in follicular lymphoma. We report stage 1 of the study, which assessed pharmacokinetics, safety, and exploratory response rates. Stage 2 will provide additional data on safety and efficacy. Methods Study design and participants In the two-stage, phase 3, randomised, controlled, openlabel SABRINA study, we enrolled adults (aged 18 years) with previously untreated, histologically confirmed CD20-positive grade 1, 2, or 3a follicular lymphoma at 67 centres in 23 countries (appendix). Eligible patients had Eastern Cooperative Oncology Group performance statuses of 0 2, bidimensionally measured disease (by CT or MRI), life expectancy of at least 6 months, adequate haematological function for at least 28 days, and one or more of the following symptoms: bulky disease (tumour mass 7 cm in greatest dimension), B symptoms (fever, night sweats, and weight loss), increased serum lactate dehydrogenase or β2-microglobulin con centrations (more than the upper limit of normal), involvement of at least three nodal sites (diameter >3 cm), symptomatic spleen enlargement, compressive syndrome, or pleural or peritoneal effusion. We excluded patients with the presence or history of central nervous system disease, transformation to high-grade non-follicular lymphoma, or malignancies other than follicular lymphoma. The inclusion and exclusion criteria for this study are in line with phase 3 studies of intravenous rituximab in firstline follicular lymphoma patients, such as the PRIMA study. 4 SABRINA was done in accordance with the Declaration of Helsinki. The study protocol was approved by local ethics committees. Patients provided written informed consent. Randomisation and masking We randomly allocated participants 1:1 with the Pocock and Simon dynamic randomisation algorithm 26 to receive intravenous rituximab or subcutaneous rituximab. Randomisation was done centrally for all participating centres. Patients in the intravenous rituximab group received induction therapy with eight cycles of intravenous rituximab (375 mg/m²) plus cyclophosphamide, doxorubicin, vincristine, prednisone (CHOP) or cyclophosphamide, vincristine, prednisone (CVP) chemotherapy. Patients in the subcutaneous rituximab group received induction therapy of one cycle of intravenous rituximab (to allow improved management of potential administration-related reactions through control of infusion rate) plus CHOP or CVP chemotherapy followed by subcutaneous rituximab 344 www.thelancet.com/oncology Vol 15 March 2014

(1400 mg fixed dose) plus chemotherapy for cycles 2 8 (appendix). After induction, patients with a complete response, unconfirmed complete response, or partial response continued intravenous or subcutaneous rituximab every 8 weeks for up to 2 years. Patients who did not respond were treated according to local practice. We stratified enrolment by block randomisation with an interactive voice recognition system and numbers were assigned chronologically to patients within each centre as they were enrolled into the study. The chemotherapeutic regimen was chosen by the treating physician, and dynamic stratification took account of this choice when patients were assigned to treatment groups. Stratification factors were Follicular Lymphoma International Prognostic Index (FLIPI) score, chemo therapy regimen, and region. Although the trial was open-label, the study management team was masked to randomised data until analysis of the primary endpoint. Procedures CHOP consisted of cyclophosphamide 750 mg/m², doxorubicin 50 mg/m², and vincristine 1 4 mg/m² (maximum 2 mg) and was given intravenously on day 1; intravenous or oral prednisone or prednisolone (100 mg per day) was administered consecutively for 5 days starting on the day of rituximab administration before rituximab dosing. CVP consisted of cyclophosphamide 750 mg/m² and vincristine 1 4 mg/m² (maximum 2 mg) and was given intravenously on day 1, with intravenous or oral prednisone or prednisolone (40 mg/m² per day) administered consecutively for 5 days starting on the day of rituximab administration before rituximab dosing. Chemotherapy cycles lasted for 21 days. Patients in the intravenous group received 375 mg/m² on day 0, 1, or 2 of cycle 1 and day 1 of cycles 2 8. Patients in the subcutaneous group received intravenous rituximab for the first cycle, then a 1400 mg fixed dose of subcutaneous rituximab on day 0 of cycle 2 and day 1 of cycles 3 8 (appendix). Subcutaneous rituximab was administered by the study investigator or nurse. No dose reductions were permitted for intravenous or subcutaneous rituximab. Dose modifications to chemotherapy because of toxicity (delay, reduction, or discontinuation) were permitted and recorded. All clinically relevant non-haematological toxicities must have resolved to grade 1 before patients could start a new treatment cycle. If chemotherapy cycles were delayed, rituximab administration was also delayed. During induction, up to eight cycles of CHOP or CVP chemotherapy were administered, but all patients received eight cycles of rituximab. The treatment schema is illustrated in the appendix. Outcomes The primary endpoint for stage 1 of the trial was rituximab C trough at induction cycle 7 (ie, before dose cycle 8). The recommended regimen for rituximab induction is eight cycles irrespective of chemotherapy (eg, six cycles of CHOP plus eight cycles of rituximab). We selected cycle 7 as the latest possible timepoint for the C trough determination in the induction clinical setting because target-specific elimination was expected to be nominal at this point and variability would be minimised. Secondary endpoints included rituximab serum concentration AUC at cycles 2 and 7, predicted pharmacokinetic parameters, peripheral blood B-cell depletion, overall response (patients who achieved a complete response, unconfirmed complete response, or partial response) at completion of induction treatment, safety (including administration-related reactions, adverse events occurring within 24 h of treatment regarded as related to rituximab), and immunogenicity (development of human antichimeric antibodies and human antihuman antibodies). During induction, we collected blood samples for pharmacokinetic analysis up to 2 h before dosing at each cycle, at the end of infusion for cycle 1 (and cycles 2, 5, and 7 for the intravenous treatment group), 24 h after dosing and days 3, 7, and 15 of cycles 2 and 7 (intensive sampling for AUC analysis), and at day 29 of cycle 8. We graded adverse events according to National Cancer 64 randomly allocated to intravenous chemotherapy 64 received allocated intervention 6 discontinued intervention 2 adverse events 2 lack of efficacy 1 physician decision 1 lost to follow-up 58 receiving ongoing maintenance treatment 150 assessed for eligibility 23 excluded 6 active or history of HBV or HCV infection 6 required other treatment 2 grade 3b follicular lymphoma 2 transformation to high-grade lymphoma secondary to follicular lymphoma 2 histologically confirmed CD20-positive follicular NHL* 1 presence or history of CNS disease 1 abnormal laboratory values 1 major surgery <28 days before study entry 2 withdrew consent 63 randomly allocated to subcutaneous chemotherapy 63 received allocated intervention 7 discontinued intervention 1 adverse event 1 death 2 lack of efficacy 2 physician decisions 1 patient decision 56 receiving ongoing maintenance treatment Figure 1: Trial profile HBV=hepatitis B virus. HCV=hepatitis C virus. NHL=non-Hodgkin lymphoma. CNS=central nervous system. *Grade 1, 2, or 3a according to WHO classification system. Serum creatinine >2 mg/dl (197 μmol/l), total bilirubin >1 5 times the upper limit of normal, or aspartate aminotransferase or alanine aminotransferase >2 5 times the upper limit of normal (or >5 times the upper limit of normal in the presence of liver involvement). One patient randomly allocated to the subcutaneous rituximab group discontinued treatment shortly after the first intravenous rituximab infusion and was analysed as part of the intravenous rituximab group in the safety population. www.thelancet.com/oncology Vol 15 March 2014 345

Institute Common Terminology Criteria for Adverse Events, version 4.0. We collected samples for immunogenicity analysis at each visit immediately before rituximab administration and every 12 weeks for 96 weeks after the last administration, and analysed samples by ELISA and electrochemiluminescence assays. We assessed tumour response at induction cycle 4, at end of induction or treatment discontinuation, and at intervals (clinical assessments were completed every 3 months and could include a CT scan) during maintenance for 96 weeks or until disease progression, with the 1999 International Working Group response criteria for non-hodgkin lymphoma. 27 We regarded the Intravenous chemotherapy (n=64) Subcutaneous chemotherapy (n=63) Demographics Age, years 57 (35 85) 54 (28 85) Patients aged >70 years 8 (13%) 8 (13%) Sex Male 33 (52%) 26 (41%) Female 31 (48%) 37 (59%) Body surface area, m²* 1 82 (1 34 2 30) 1 74 (1 37 2 32) Ethnicity White 48 (84%) 45 (83%) Asian 4 (7%) 4 (7%) American Indian or Alaska native 1 (2%) 1 (2%) Other 4 (7%) 4 (7%) Not reported 7 (11%) 9 (14%) Disease and treatment characteristics Time from diagnosis, months 1 4 (0 2 279 9) 1 7 (0 2 72 0) LDH >upper limit of normal 15 (24%) 18 (30%) Ann Arbor stage at entry I 1 (2%) 1 (2%) II 8 (13%) 3 (5%) III 18 (28%) 20 (32%) IV 37 (58%) 39 (62%) FLIPI risk group Low (0 1) 13 (20%) 13 (21%) Intermediate (2) 25 (39%) 25 (40%) High ( 3) 26 (41%) 25 (40%) Follicular lymphoma grade Grade 1 17 (27%) 22 (35%) Grade 2 35 (55%) 24 (38%) Grade 3a 12 (19%) 17 (27%) Tumour load (mm²) 5306 (940 57 080) 4851 (276 52 047) First-line chemotherapy CHOP 40 (63%) 40 (63%) CVP 24 (38%) 23 (37%) Data are median (range) or n (%). LDH=lactate dehydrogenase. FLIPI=Follicular Lymphoma International Prognostic Index. CHOP=cyclophosphamide, doxorubicin, vincristine, and prednisone. CVP=cyclophosphamide, vincristine, and prednisone. *Data for 63 patients in the intravenous group and 63 patients in the subcutaneous group. Data for 65 patients in the intravenous group and 62 patients in the subcutaneous group. Grade 1 is 0 5 centroblasts per high-power field, grade 2 is 6 15 centroblasts per high-power field, and grade 3 is 15 centroblasts per high-power field. Product of all indicator lesions. Table 1: Baseline demographics and disease characteristics 1999 response criteria as appropriate because not all centres in this study had access to a PET scanner. Responses were therefore based on the 1999 criteria with CT assessments, although optional PET scans were permitted and any deviations recorded separately. In addition, the unconfirmed complete response category was removed in the 2007 criteria (therefore allowing only complete or partial responses) and use of the 1999 criteria also allowed comparison to historical study data. Study investigators assessed responses. Radiological data were independently reviewed, followed by an internal clinical data review. Statistical analysis The aim of this stage 1 primary pharmacokinetic analysis was to show pharmacokinetic non-inferiority of subcutaneous treatment compared with intravenous treatment at induction cycle 7 (before cycle 8 dosing), defined as the lower limit of the two-sided 90% CI of the geometric mean C trough ratio of at least 0 8. 50 patients per group were needed to provide 80% power, assuming a 0 56 coefficient of variation with a geometric mean C trough for subcutaneous treatment of 5% higher than geometric mean C trough for intravenous treatment. We assumed a 5% difference in C trough would not lead to a different safety or efficacy profile and would ensure that patients with higher adjusted doses were not at risk of underdosing after the switch from dosing adjusted on the basis of bodyweight to flat dosing. Assuming 80% of patients would have valid C trough measurements before cycle 8, about 125 patients were required in stage 1. We did our analysis in the per-protocol population of patients with pharmacokinetic samples collected at cycle 7 (within 24 h of day 21) to ensure a consistent comparison between treatment groups. Samples from patients that were obtained outside this 48 h time window were not included because any chemotherapy treatment delays were in 7 day intervals and therefore inclusion of these samples could have resulted in skewed C trough values. We used standard non-compartmental pharmacokinetic methods to analyse C trough and AUC. Details on pharmacokinetic futility analysis are provided in the appendix. We did statistical analyses with SAS software (version 8.2) for all patients with available pharmacokinetic assessments (patients analysed as treated). Secondary efficacy endpoints were analysed for the intention-to-treat population (all randomly allocated patients). The safety population included all patients who received at least one dose of intravenous or subcutaneous rituximab (patients analysed as treated). We did exploratory subgroup analyses that were restricted to the patients included in stage 1, and were thus subject to high variance. Therefore, we did no comparative statistical analyses and firm conclusions should not be drawn on these exploratory subgroups. This study is registered with ClinicalTrials.gov, number NCT01200758. 346 www.thelancet.com/oncology Vol 15 March 2014

C trough, μg/ml (cycle 7) AUC, μg per day/ml (cycle 7) Intravenous rituximab plus chemotherapy Patients Geometric mean Subcutaneous rituximab plus chemotherapy Patients Geometric mean Geometric mean ratio (90% CI)* Intravenous rituximab plus chemotherapy Patients Geometric mean Subcutaneous rituximab plus chemotherapy Patients Geometric mean Geometric mean ratio (90% CI)* Overall 48 83 13 54 134 58 1 62 (1 36 1 94) 58 2734 55 3779 1 38 (1 24 1 53) Low BSA 13 68 52 23 158 83 2 29 (1 49 3 52) 16 2607 23 4429 1 66 (1 33 2 07) Medium BSA 21 87 52 14 114 27 1 31 (1 00 1 72) 25 2775 15 3240 1 17 (0 97 1 41) High BSA 13 91 02 17 123 07 1 41 (1 14 1 75) 16 2785 17 3492 1 32 (1 15 1 51) C trough =trough concentration. AUC=area under the concentration time curve. BSA=body surface area. *Subcutaneous/intravenous; based on logarithmic scale, adjusted for tumour load at baseline. BSA data were missing for one patient in the intravenous group. For BSA, patients were grouped into low ( 33rd percentile), medium (between 33rd [1 7 m²] and 66th [1 9 m²] percentile), and high ( 66th percentile). Table 2: Exploratory analysis of pharmacokinetic parameters, by study group and patient population Role of the funding source SABRINA was funded and sponsored by F Hoffmann-La Roche. The sponsor designed the study, provided study drugs, contributed to protocol development, regulatory and ethics approval, and safety monitoring, and did the data collection and analysis. All authors had full access to study data and contributed to data interpretation, and the decision to submit for publication. Results We enrolled 127 patients between Feb 4, 2010, and Oct 21, 2011, with data cutoff of June 12, 2012. In stage 1, 64 patients were randomly allocated to intravenous rituximab and 63 were randomly allocated subcutaneous rituximab (figure 1, table 1). One (2%) of 63 patients in the subcutaneous group discontinued treatment after the first intravenous infusion and was analysed as part of the intravenous group for safety analyses. In both groups, 63% of patients received CHOP and about 37% received CVP chemotherapy. Most rituximab dose interruptions were due to infusion-related adverse events and occurred in cycle 1. 18 (28%) of 65 patients in the intravenous group and 19 (31%) of 62 patients in the subcutaneous group had modifications in their first infusion (intravenous for both groups). In later cycles, only one (2%) of 56 patients in the subcutaneous group had their injection delayed (during cycle 9), compared with dose modifications for nine patients in the intravenous group (four of 64 in cycle 2, three of 62 in cycle 3, one of 59 in cycle 7, and one of 54 in cycle 9). Overall exposure to chemotherapy was balanced in both treatment groups (data not shown). The median injection time for subcutaneous rituximab was 6 1 min (IQR 6 0 7 0 min). Assessable C trough data at cycle 7 (predose sample at cycle 8) were available for 48 (75%) of 64 patients in the intravenous rituximab group and 54 (86%) of 63 patients in the subcutaneous rituximab group. Of the remaining 25 patients, nine patients in the intravenous group and five patients in the subcutaneous group had missing C trough data at cycle 7, and seven samples in the intravenous group and four samples in the Rituximab concentration (μg/ml) 300 250 200 150 100 50 0 Intravenous Subcutaneous Cycle 1 Cycle 2 Cycle 3 Cycle 4 Cycle 5 Cycle 6 Cycle 7 Figure 2: Rituximab serum trough concentrations by induction treatment cycle Box plots show median (IQR), with vertical lines showing upper and lower adjacent values, and dots show outliers. Cycle 7 is restricted to the pharmacokinetic population, all other cycles are study population. subcutaneous group were taken outside the predefined time window. The geometric mean for rituximab C trough was higher in the subcutaneous group than in the intravenous group at the end of induction cycle 7 (table 2). The lower limit for the two-sided 90% CI of the subcutaneous to intravenous geometric mean ratio exceeded the prespecified margin of 0 8 (table 2), showing C trough non-inferiority of subcutaneous to intravenous rituximab. The coefficients of variation for C trough at cycle 7 were 43 2% in the subcutaneous group and 36 7% in the intravenous group, showing similar variability. AUC at cycle 7 was also higher in the subcutaneous rituximab group than in the intravenous group (table 2). Median rituximab serum C trough levels were higher in the subcutaneous rituximab group than in the intravenous group at each induction treatment cycle (figure 2). Median time to maximum concentration www.thelancet.com/oncology Vol 15 March 2014 347

Intravenous chemotherapy* Subcutaneous chemotherapy* Overall Any-grade AE 57/65 (88%) 57/62 (92%) Any-grade treatment-related AE 30/65 (46%) 45/62 (73%) Administration-related AE 21/65 (32%) 31/62 (50%) Grade 3 AE 30/65 (46%) 29/62 (47%) SAE 14/65 (22%) 14/62 (23%) Patients with low BSA Any-grade AE 14/16 (88%) 24/26 (92%) Grade 3 AE 8/16 (50%) 15/26 (58%) SAE 3/16 (19%) 5/26 (19%) Patients with medium BSA Any-grade AE 23/27 (85%) 14/15 (93%) Grade 3 AE 15/27 (56%) 7/15 (47%) SAE 9/27 (33%) 7/15 (47%) Patients with high BSA Any-grade AE 19/21 (90%) 19/21 (90%) Grade 3 AE 7/21 (33%) 7/21 (33%) SAE 2/21 (10%) 2/21 (10%) Male patients Any-grade AE 28/33 (85%) 25/26 (96%) Grade 3 AE 14/33 (42%) 8/26 (31%) SAE 4/33 (12%) 4/26 (15%) Female patients Any-grade AE 29/32 (91%) 32/36 (89%) Grade 3 AE 16/32 (50%) 21/36 (58%) SAE 10/32 (31%) 10/36 (28%) Data are n/n (%). AE=adverse event. SAE=serious adverse event. BSA=body surface area. *One patient randomly allocated to the subcutaneous rituximab group discontinued treatment shortly after the first intravenous rituximab infusion but was analysed as part of the intravenous rituximab group in the safety population. Includes AEs at cycle 1 intravenous treatment. Patients were grouped into low ( 33rd percentile), medium (between 33rd [1 7 m²] and 66th [1 9 m²] percentile), and high ( 66th percentile). Table 3: Adverse events in patients with one or more event was 48 1 h (IQR 45 9 141 4) for subcutaneous rituximab at cycle 7 and 3 0 h (2 3 4 0 h) for intravenous rituximab at cycle 7. Predicted pharmacokinetic parameter findings are provided in the appendix. Median baseline CD19-positive lymphocyte count was 0 12 10⁹ cells per L in the subcutaneous group and 0 05 10⁹ cells per L in the intravenous group. B cells were depleted before cycle 2, with median counts of 0 cells per L in both treatment groups before cycle 2 dosing (range 0 00 8 23 10⁹ cells per L in the subcutaneous group and 0 00 0 34 10⁹ cells per L in the intravenous group). B-cell depletion was maintained throughout treatment (data not shown). After a median follow-up of 8 74 months (IQR 7 1 10 8) in the intravenous rituximab group and 8 84 months (7 1 10 6) in the subcutaneous group, most patients had one or more adverse events (table 3). The most common adverse events were neutropenia, nausea, and constipation (table 4). The proportion of patients with grade 3 or worse adverse events or serious adverse events did not differ between groups (table 3). Neutropenia was the most common adverse event of grade 3 or worse (table 4), and febrile neutropenia was the most frequent serious adverse event. Although patient numbers were small in an exploratory subgroup safety analysis, the data do not suggest that patients with the lowest body surface area and highest exposure after subcutaneous rituximab administration had an increased incidence of adverse events compared with the corresponding population in the intravenous group (table 3). We noted no association between sex of participants and rates of adverse events (table 3). Three (5%) of 65 patients who received intravenous rituximab discontinued treatment during induction because of an adverse event (one each of vomiting, pneumonia, and liver enzyme abnormalities), of which two were regarded as related to treatment. One (2%) of 62 patients in the subcutaneous rituximab group discontinued because of dysphonia, which was regarded as unrelated to treatment, and one patient died after cycle 5 after a myocardial infarction, which was also regarded as unrelated to treatment. No deaths regarded as related to treatment occurred. 21 (32%) of 65 patients given intravenous rituximab had an administration-related reaction compared with 31 (50%) of 62 patients given subcutaneous rituximab; 46 (98%) of 47 individual administration-related reactions in the intravenous group and 69 (95%) of 73 individual reactions in the subcutaneous group were grade 1 or 2. Administration-related reactions of any grade occurring in at least 5% of patients in either group were erythema (two [3%] of 65 patients in the intravenous group and five [8%] of 62 patients in the subcutaneous group), pruritus (two [3%] and four [6%]), chills (four [6%] and two [3%]), injection-site erythema (none and six [10%]), and vomiting (four [6%] and two [3%]). One (2%) of 65 patients had a grade 3 administration-related reaction (vomiting) in the intravenous group, and three (5%) of 62 patients had a grade 3 administration-related reaction in the subcutaneous group (injection-site rash after first subcutaneous administration, dry mouth after first sub cutaneous administration, decreased urine output and tumour lysis syndrome after initial intravenous infusion); none of these events led to treatment discontinuation. No grade 4 administration-related reactions occurred. Two (3%) of 65 patients in the intravenous group and two (3%) of 62 patients in the subcutaneous group had a positive human antichimeric antibodies result after baseline; this finding did not affect the adverse event profile of rituximab in these patients. Testing for human antihuman antibodies was carried out during each cycle of induction therapy and periodically during maintenance therapy. In the intravenous group, the lowest proportion of 348 www.thelancet.com/oncology Vol 15 March 2014

Intravenous chemotherapy (n=65)* Subcutaneous chemotherapy (n=62)* Overall Grade 1 Grade 2 Grade 3 Grade 4 Grade 5 Overall Grade 1 Grade 2 Grade 3 Grade 4 Grade 5 Gastrointestinal disorders Nausea 15 (23%) 10 (15%) 5 (8%) 0 0 0 18 (29%) 10 (16%) 8 (13%) 0 0 0 Constipation 17 (26%) 16 (25%) 1 (2%) 0 0 0 14 (23%) 8 (13%) 6 (10%) 0 0 0 Vomiting 13 (20%) 6 (9%) 5 (8%) 2 (3%) 0 0 12 (19%) 5 (8%) 7 (11%) 0 0 0 Diarrhoea 11 (17%) 7 (11%) 3 (5%) 1 (2%) 0 0 10 (16%) 6 (10%) 4 (6%) 0 0 0 Abdominal pain 7 (11%) 6 (9%) 1 (2%) 0 0 0 10 (16%) 7 (11%) 3 (5%) 0 0 0 Dyspepsia 0 0 0 0 0 0 8 (13%) 3 (5%) 5 (8%) 0 0 0 General disorders and administration-site disorders Asthenia 10 (15%) 8 (12%) 2 (3%) 0 0 0 14 (23%) 9 (15%) 4 (6%) 1 (2%) 0 0 Pyrexia 11 (17%) 7 (11%) 4 (6%) 0 0 0 8 (13%) 3 (5%) 5 (8%) 0 0 0 Mucosal inflammation 10 (15%) 4 (6%) 6 (9%) 0 0 0 3 (5%) 1 (2%) 1 (2%) 1 (2%) 0 0 Injection-site erythema 0 0 0 0 0 0 9 (15%) 3 (5%) 5 (8%) 0 0 0 Blood and lymphatic system disorders Neutropenia 23 (35%) 3 (5%) 6 (9%) 10 (15%) 4 (6%) 0 22 (35%) 0 6 (10%) 5 (8%) 11 (18%) 0 Anaemia 8 (12%) 7 (11%) 1 (2%) 0 0 0 9 (15%) 3 (5%) 3 (5%) 0 0 0 Leukopenia 4 (6%) 1 (2%) 2 (3%) 0 1 (2%) 0 6 (10%) 1 (2%) 0 4 (6%) 1 (2%) 0 Febrile neutropenia 2 (3%) 0 0 0 2 (3%) 0 6 (10%) 0 0 5 (8%) 1 (2%) 0 Nervous system disorders Paraesthesia 7 (11%) 6 (9%) 1 (2%) 0 0 0 12 (19%) 10 (16%) 2 (3%) 0 0 0 Headache 5 (8%) 3 (5%) 2 (3%) 0 0 0 10 (16%) 7 (11%) 3 (5%) 0 0 0 Peripheral neuropathy 7 (11%) 2 (3%) 5 (8%) 0 0 0 1 (2%) 0 0 0 0 0 Infections and infestations Pneumonia 1 (2%) 0 0 0 1 (2%) 0 2 (3%) 0 0 2 (3%) 0 0 Sepsis 0 0 0 0 0 0 2 (3%) 0 0 1 (2%) 1 (2%) 0 Skin and subcutaneous tissue disorders Alopecia 7 (11%) 1 (2%) 6 (9%) 0 0 0 12 (19%) 7 (11%) 5 (8%) 0 0 0 Erythema 2 (3%) 2 (3%) 0 0 0 0 10 (16%) 7 (11%) 3 (5%) 0 0 0 Rash 2 (3%) 1 (2%) 1 (2%) 0 0 0 7 (11%) 4 (6%) 3 (5%) 0 0 0 Musculoskeletal and connective tissue disorders Myalgia 1 (2%) 1 (2%) 0 0 0 0 7 (11%) 4 (6%) 3 (5%) 0 0 0 Respiratory, thoracic, and mediastinal disorders Cough 7 (11%) 5 (8%) 2 (3%) 0 0 0 9 (15%) 7 (11%) 2 (3%) 0 0 0 Dyspnoea 5 (8%) 2 (3%) 3 (5%) 0 0 0 9 (15%) 6 (10%) 3 (5%) 0 0 0 Psychiatric disorders Insomnia 3 (5%) 2 (3%) 1 (2%) 0 0 0 9 (15%) 7 (11%) 2 (3%) 0 0 0 Ear and labyrinth disorders Vertigo 1 (2%) 1 (2%) 0 0 0 0 3 (5%) 1 (2%) 0 2 (3%) 0 0 Data are n (%). AE=adverse event. *One patient randomised to the subcutaneous rituximab group discontinued treatment shortly after the first intravenous rituximab infusion but was analysed as part of the intravenous rituximab group in the safety population. Includes AEs at cycle 1 intravenous treatment. Table 4: Most common AEs (in >10% of patients overall) and most common grade 3 AEs (in >2% of patients) patients who tested positive was 9% (five of 58 patients during cycle 5) and the highest was 17% (nine of 54 patients during cycle 7 and eight of 47 patients during cycle 8). For patients in the subcutaneous group, the lowest proportion of patients who tested positive was 4% (two of 47 patients during cycle 8) and the highest was 8% (five of 60 patients during cycle 6). However, none of these patients tested positive for neutralising anti-rhuph20 antibody. Table 5 shows overall response according to investigator assessment and after independent review. Exploratory analyses suggested response by body surface area, chemotherapy, and sex subgroups (table 6) did not differ from the intention-to-treat population, but the number of patients in each subgroup was small. Discussion Our phase 3 SABRINA study is designed to assess pharmacokinetics, safety, and efficacy in follicular lymphoma induction therapy followed by maintenance treatment (panel). In stage 1 of the study, we showed non-inferiority of a subcutaneous formulation of rituximab in terms of C trough compared with intravenous www.thelancet.com/oncology Vol 15 March 2014 349

Intravenous chemotherapy (n=64) Investigator assessment Overall response (CR, CRu, PR) Intravenous chemotherapy Independent review Subcutaneous chemotherapy Subcutaneous chemotherapy (n=63) Investigator assessment Complete response (CR or CRu) Intravenous chemotherapy Subcutaneous chemotherapy Overall 54/64 (84%) 57/63 (90%) 19/64 (30%) 29/63 (46%) Low BSA* 15/16 (94%) 22/26 (85%) 5/16 (31%) 14/26 (54%) Medium BSA* 20/26 (77%) 15/16 (94%) 7/26 (27%) 8/16 (50%) High BSA* 18/21 (86%) 20/21 (95%) 7/21 (33%) 7/21 (33%) Male 27/33 (82%) 25/26 (96%) 7/33 (21%) 11/26 (42%) Female 27/31 (87%) 32/37 (86%) 12/31 (39%) 18/37 (49%) CHOP 34/40 (85%) 37/40 (93%) 13/40 (33%) 17/40 (43%) CVP 20/24 (83%) 20/23 (87%) 6/24 (25%) 12/23 (52%) Data are n/n (%). CR=complete response. CRu=unconfirmed complete response. PR=partial response. BSA=body surface area. CHOP=cyclophosphamide, doxorubicin, vincristine, and prednisone. CVP=cyclophosphamide, vincristine, and prednisone. *Patients were grouped into low ( 33rd percentile), medium (between 33rd [1 7 m²] and 66th [1 9 m²] percentile), and high ( 66th percentile); one patient in each treatment group had a missing BSA reading. Table 6: Exploratory analysis of investigator-ass essed tumour responses at end of induction Independent review Overall response 54 (84%) 56 (88%) 57 (90%) 54 (86%) CR or CRu 19 (30%) 12 (19%) 29 (46%) 17 (27%) PR 35 (55%) 44 (69%) 28 (44%) 37 (59%) Stable disease 3 (5%) 1 (2%) 2 (3%) 4 (6%) Progressive disease 1 (2%) 0 0 2 (3%) Missing or invalid* 6 (9%) 7 (11%) 4 (6%) 3 (5%) Data are n (%). CR=complete response. CRu=unconfirmed complete response. PR=partial response. *Patients with non-assessed, invalid, or missing response assessments were classified as non-responders; a response was classified as invalid if the response assessment was >56 days after the last rituximab intake, after the first rituximab intake of the maintenance phase, or after the start of new antilymphoma treatment. Table 5: Tumour response at end of induction rituximab. Results from body surface area and sex subgroup analyses did not suggest any notably different results from the main analysis. However, these data should be interpreted cautiously because of the small subgroup sizes. In addition to C trough non-inferiority (the primary endpoint), the secondary endpoint of AUC exposure over time after subcutaneous rituximab was at least as high as after intravenous rituximab. These data support the favourable pharmacokinetic profile of subcutaneous rituximab and are consistent with the SparkThera study, 25 in which non-inferiority of fixed-dose subcutaneous rituximab 1400 mg given every 2 months or 3 months was established for follicular lymphoma maintenance. Because subcutaneous rituximab 1400 mg has a different pharmacokinetic profile and reduced administration time compared with intravenous rituximab, use of the subcutaneous route could improve tolerability (ie, reduce the risk of reactions related to infusion rate) and increase patient convenience and resource use. Such Panel: Research in context Systematic review The standard route of administration for the anti-cd20 monoclonal antibody rituximab is intravenous infusion. Subcutaneous rituximab administration could substantially reduce the time a patient spends at the hospital and eliminate the burden on hospitals associated with intravenous administration (eg, nursing time for intravenous dosing and rental of day beds). The phase 1 SparkThera trial 25 established that a fixed 1400 mg dose of subcutaneous rituximab provided non-inferior C trough levels relative to standard intravenous dosing during maintenance therapy for follicular lymphoma, with a comparable safety profile. The SABRINA study was designed to investigate the use of this dose and administration route in patients with previously untreated follicular lymphoma during induction and maintenance. Subcutaneous administration of rituximab had not been investigated previously in a randomised phase 3 study. Interpretation Stage 1 results from this study show that subcutaneous rituximab at a fixed 1400 mg dose provided non-inferior pharmacokinetics compared with standard intravenous dosing. Safety profiles for intravenous and subcutaneous dosing were similar apart from an increase in administration-related reactions in the subcutaneous group because of local, mild, and reversible injection-site reactions such as erythema, which is the expected change of the administration-related reaction profile when the subcutaneous route of administration is used. Notably, stage 1 response suggests that switching to subcutaneous administration does not seem to affect the antilymphoma activity of rituximab. These findings suggest that subcutaneous fixed-dose rituximab administration is feasible in this population of patients. Stage 2 of the study will provide confirmatory safety and efficacy data from a large additional cohort of patients. use could therefore potentially lead to improved accessibility to therapies based on rituximab, especially in areas with few facilities for intravenous administration. Investigator-assessed overall response at the end of induction was similar in both treatment groups and was consistent with findings of the independent review. Complete response at the end of induction was numerically higher in the subcutaneous group than the intravenous group; however, the study was not designed to show superiority in terms of activity. Although body surface area subgroups were small, all achieved similar response, supporting the activity of fixed-dose (1400 mg) subcutaneous rituximab. Non-inferiority of the C trough ratio was noted even in the high body surface area subgroup. Overall, stage 1 data suggest that fixed-dose subcutaneous rituximab has a favourable benefit to risk profile, which is much the same as that of intravenous 350 www.thelancet.com/oncology Vol 15 March 2014

rituximab. Longer follow-up is needed to assess the durability of the clinical tumour response at end of induction using time-to-event clinical data. We did not note any new clinically relevant safety signals and the incidence of any-grade adverse events, grade 3 or worse adverse events, and serious adverse events seemed similar between formulations. Although administration-related reactions were more frequent in the subcutaneous group than the intravenous group, they were mostly grade 1 2 local reactions, including mild-to-moderate erythema, rash, and pruritus. Subcutaneous dosing was also associated with an increase in mild injection-site skin reactions for alemtuzumab. 6 The change in administration route was therefore expected to contribute to an increased incidence of administration-related reactions (eg, local injection-site reactions) in the subcutaneous rituximab group. However, few reactions were grade 3 and none were grade 4 or 5. Exploratory subgroup analyses suggested that incidence of grade 3 or worse adverse events and serious adverse events for patients with low, medium, and high body surface area were not strikingly different. This similarity is important because exposure to rituximab is greater for patients with low body surface area than for patients with medium or high body surface area. After stage 1, patients will continue to receive maintenance rituximab subcutaneously or intravenously for up to 2 years. The second stage of this trial is fully accrued and will provide safety and efficacy data for an additional 280 patients (data are expected by the end of 2014). Clinician and nurse preference for intravenous versus subcutaneous administration of rituximab will also be assessed and reported after completion of stage 2. Further data on subcutaneous rituximab in CD20-positive haema tological malignancies is anticipated, including a randomised phase 3 trial (PrefMab; NCT01724021) investigating patient preference for sub cutaneous or intravenous chemo therapy for the first-line treatment of diffuse large B-cell lymphoma or grade 1, 2, or 3a follicular lymphoma. Contributors All authors designed the study, did the literature search, interpreted the data, wrote or revised the report, and approved the final version. AD, DM, FM, BM, NS, and PS-C recruited patients and collected data. MB, CB, BB, AB, and CM analysed data. Conflicts of interest AD has received research funding, travel grants, and honoraria from F Hoffmann-La Roche and has attended advisory boards. FM has received honoraria from F Hoffmann-La Roche, Celgene, and Teva, and has attended advisory boards. BM has no conflicts of interest to declare. NS has received research funding from F Hoffmann-La Roche. PS-C has received research funding, honoraria, and acted as a consultant for F Hoffmann-La Roche. MB and BB are employees of F Hoffmann-La Roche and have stock options. CB, AB, and CM are employees of F Hoffmann-La Roche. DM received research funding from Roche Canada, honoraria from F Hoffmann-La Roche, Celgene, Lundbeck, and Amgen, and acted as a consultant for F Hoffmann-La Roche. Acknowledgments The SABRINA trial was sponsored by F Hoffmann-La Roche. Support for medical writing assistance (Cheryl Wright; KnowledgePoint360, Macclesfield, Cheshire, UK) and copyediting (Melissa Hernandez-Warren; CodonMedical, San Bruno, CA, USA) was provided by F Hoffmann-La Roche. References 1 Hiddemann W, Kneba M, Dreyling M, et al. Frontline therapy with rituximab added to the combination of cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP) significantly improves the outcome for patients with advanced-stage follicular lymphoma compared with therapy with CHOP alone: results of a prospective randomized study of the German Low-Grade Lymphoma Study Group. Blood 2005; 106: 3725 32. 2 Marcus R, Imrie K, Solal-Celigny P, et al. Phase III study of R-CVP compared with cyclophosphamide, vincristine, and prednisone alone in patients with previously untreated advanced follicular lymphoma. J Clin Oncol 2008; 26: 4579 86. 3 Hochster H, Weller E, Gascoyne RD, et al. Maintenance rituximab after cyclophosphamide, vincristine, and prednisone prolongs progression-free survival in advanced indolent lymphoma: results of the randomized phase III ECOG 1496 Study. J Clin Oncol 2009; 27: 1607 14. 4 Salles G, Seymour JF, Offner F, et al. Rituximab maintenance for 2 years in patients with high tumour burden follicular lymphoma responding to chemotherapy (PRIMA): a phase 3, randomised controlled trial. Lancet 2011; 377: 42 51. 5 Van Oers MH, Van Glabbeke M, Giurgea L, et al. Rituximab maintenance treatment of relapsed/resistant follicular non- Hodgkin s lymphoma: long-term outcome of the EORTC 20981 phase III randomized intergroup study. J Clin Oncol 2010; 28: 2853 58. 6 Stilgenbauer S, Zenz T, Winkler D, et al. Subcutaneous alemtuzumab in fludarabine-refractory chronic lymphocytic leukemia: clinical results and prognostic marker analyses from the CLL2H study of the German Chronic Lymphocytic Leukemia Study Group. J Clin Oncol 2009; 27: 3994 4001. 7 Shpilberg O, Jackisch C. Subcutaneous administration of rituximab (MabThera) and trastuzumab (Herceptin) using hyaluronidase. Br J Cancer 2013; 109: 1556 61. 8 Goldenberg DM, Morschhauser F, Wegener WA. Veltuzumab (humanized anti-cd20 monoclonal antibody): characterization, current clinical results, and future prospects. Leuk Lymphoma 2010; 51: 747 55. 9 Pivot X, Gligorov J, Müller V, et al. Preference for subcutaneous or intravenous administration of trastuzumab in patients with HER2- positive early breast cancer (PrefHer): an open-label randomised study. Lancet Oncol 2013; 14: 962 70. 10 Lundin J, Kimby E, Björkholm M, et al. Phase II trial of subcutaneous anti-cd52 monoclonal antibody alemtuzumab (Campath-1H) as first-line treatment for patients with B-cell chronic lymphocytic leukemia (B-CLL). Blood 2002; 100: 768 73. 11 De Cock E, Kritikou P, Tao S, et al. Time savings with rituximab subcutaneous (SC) injection vs rituximab intravenous (IV) infusion: final analysis from a time-and-motion study in 8 countries. Blood 2013; 122: abstr 1742. 12 Salar A, Bouabdallah R, McIntyre C, Sayyed P, Bittner B. A two-stage phase Ib study to investigate the pharmacokinetics, safety and tolerability of subcutaneous rituximab in patients with follicular lymphoma as part of maintenance treatment. Blood 2010; 116: abstr 2858. 13 Frost GI. Recombinant human hyaluronidase (rhuph20): an enabling platform for subcutaneous drug and fluid administration. Expert Opin Drug Deliv 2007; 4: 427 40. 14 Yin A, Li J, Hurst D, Visich J. Population pharmacokinetics (PK) and association of PK and clinical outcomes of rituximab in patients with non-hodgkin s lymphoma. Proc Am Soc Clin Oncol 2010; 28 (suppl): e13108. 15 Tobinai K, Igarashi T, Itoh K, et al, on behalf of the IDEC-C2B8 Japan Study Group. Japanese multicenter phase II and pharmacokinetic study of rituximab in relapsed or refractory patients with aggressive B-cell lymphoma. Ann Oncol 2004; 15: 821 30. www.thelancet.com/oncology Vol 15 March 2014 351

16 Bittner B, Richter WF, Hourcade-Potelleret F, et al. Development of a subcutaneous formulation for trastuzumab nonclinical and clinical bridging approach to the approved intravenous dosing regimen. Drug Res 2012; 62: 401 09. 17 Bittner B, Richter WF, Hourcade-Potelleret F, et al. Development of a subcutaneous formulation for trastuzumab nonclinical and clinical bridging approach to the approved intravenous dosing regimen. Erratum. Drug Res 2013; 63: 602. 18 Wynne C, Harvey V, Schwabe C, Waaka D, McIntyre C, Bittner N. Comparison of subcutaneous and intravenous administration of trastuzumab: a phase I/Ib trial in healthy male volunteers and patients with HER2-positive breast cancer. J Clin Pharmacol 2013; 53: 192 201. 19 Ismael G, Hegg R, Muehlbauer S, et al. Subcutaneous versus intravenous administration of (neo)adjuvant trastuzumab in patients with HER2-positive, clinical stage I III breast cancer (HannaH study): a phase 3, open-label, multicentre, randomised trial. Lancet Oncol 2012; 13: 869 78. 20 Keating M, O Brien S. High-dose rituximab therapy in chronic lymphocytic leukemia. Semin Oncol 2000; 27 (suppl 12): 86 90. 21 O Brien SM, Kantarjian H, Thomas DA, et al. Rituximab doseescalation trial in chronic lymphocytic leukemia. J Clin Oncol 2001; 19: 2165 70. 22 Egorin MJ. Horseshoes, hand grenades, and body-surface areabased dosing: aiming for a target. J Clin Oncol 2003; 21: 182 83. 23 Mathijssen RH, de Jong FA, Loos WJ, van der Bol JM, Verweij J, Sparreboom A. Flat-fixed dosing versus body surface area based dosing of anticancer drugs in adults: does it make a difference? Oncologist 2007; 12: 913 23. 24 Wang DD, Zhang S, Zhao H, Men AY, Parivar K. Fixed dosing versus body size-based dosing of monoclonal antibodies in adult clinical trials. J Clin Pharmacol 2009; 49: 1012 24. 25 Salar A, Avivi I, Larouche J-F, et al. Final results of the BP22333 study demonstrate non-inferior pharmacokinetics (PK) and safety of subcutaneous (SC) administration of rituximab compared with intravenous (IV) administration as maintenance therapy in patients with follicular lymphoma (FL). Blood 2012; 120: abstr 1641. 26 Pocock SJ, Simon R. Sequential treatment assignment with balancing for prognostic factors in the controlled clinical trial. Biometrics 1975; 31: 103 15. 27 Cheson BD, Horning SJ, Coiffier B, et al. Report of an international workshop to standardize response criteria for non-hodgkin s lymphomas. NCI sponsored International Working Group. J Clin Oncol 1999; 17: 1244. 352 www.thelancet.com/oncology Vol 15 March 2014