The diagnosis of recurrent ipsilateral deep vein thrombosis (DVT) is challenging, because persistent intravascular abnormalities after previous DVT often hinder a diagnosis by compression ultrasonography. Magnetic resonance direct thrombus imaging (MRDTI), a technique without intravenous contrast and with a 10-minute acquisition time, has been shown to accurately distinguish acute recurrent DVT from chronic thrombotic remains. We have evaluated the safety of MRDTI as the sole test for excluding recurrent ipsilateral DVT. The Theia Study was a prospective, international, multicenter, diagnostic management study involving patients with clinically suspected acute recurrent ipsilateral DVT. Treatment of the patients was managed according to the result of the MRDTI, performed within 24 hours of study inclusion. The primary outcome was the 3-month incidence of venous thromboembolism (VTE) after a MRDTI negative for DVT. The secondary outcome was the interobserver agreement on the MRDTI readings. An independent committee adjudicated all end points. Three hundred five patients were included. The baseline prevalence of recurrent DVT was 38%; superficial thrombophlebitis was diagnosed in 4.6%. The primary outcome occurred in 2 of 119 (1.7%; 95% confidence interval [CI], 0.20-5.9) patients with MRDTI negative for DVT and thrombophlebitis, who were not treated with any anticoagulant during follow-up; neither of these recurrences was fatal. The incidence of recurrent VTE in all patients with MRDTI negative for DVT was 1.1% (95% CI, 0.13%-3.8%). The agreement between initial local and post hoc central reading of the MRDTI images was excellent (κ statistic, 0.91). The incidence of VTE recurrence after negative MRDTI was low, and MRDTI proved to be a feasible and reproducible diagnostic test. This trial was registered at www.clinicaltrials.gov as #NCT02262052.

Early identification of ultra-risk diffuse large B-cell lymphoma (DLBCL) patients is needed to aid stratification to innovative treatment. Previous studies suggested high baseline total metabolic tumor volume (TMTV) negatively impacts survival of DLBCL patients. We analyzed the prognostic impact of TMTV and prognostic indices in DLBCL patients, aged 60 to 80 years, from the phase 3 REMARC study that randomized responding patients to R-CHOP (rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone) into maintenance lenalidomide or placebo. TMTV was computed on baseline positron emission tomography/computed tomography using the 41% maximum standardized uptake value method; the optimal TMTV cutoff for progression-free (PFS) and overall survival (OS) was determined and confirmed by a training validation method. There were 301 out of 650 evaluable patients, including 192 patients classified as germinal center B-cell-like (GCB)/non-GCB and MYC/BCL2 expressor. Median baseline TMTV was 238 cm3; optimal TMTV cutoff was 220 cm3. Patients with high vs low TMTV showed worse/higher Eastern Cooperative Oncology Group performance status (ECOG PS) ≥2, stage III or IV disease, >1 extranodal site, elevated lactate dehydrogenase, International Prognostic Index (IPI) 3-5, and age-adjusted IPI 2-3. High vs low TMTV significantly impacted PFS and OS, independent of maintenance treatment. Although the GCB/non-GCB profile and MYC expression did not correlate with TMTV/survival, BCL2 >70% impacted PFS and could be stratified by TMTV. Multivariate analysis identified baseline TMTV and ECOG PS as independently associated with PFS and OS. Even in responding patients, after R-CHOP, high baseline TMTV was a strong prognosticator of inferior PFS and OS. Moreover, TMTV combined with ECOG PS may identify an ultra-risk DLBCL population. This trial was registered at www.clinicaltrials.gov as #NCT01122472.

Complement C5 inhibition is the standard of care (SoC) for patients with paroxysmal nocturnal hemoglobinuria (PNH) with significant clinical symptoms. Constant and complete suppression of the terminal complement pathway and the high serum concentration of C5 pose challenges to drug development that result in IV-only treatment options. Crovalimab, a sequential monoclonal antibody recycling technology antibody was engineered for extended self-administered subcutaneous dosing of small volumes in diseases amenable for C5 inhibition. A 3-part open-label adaptive phase 1/2 trial was conducted to assess safety, pharmacokinetics, pharmacodynamics, and exploratory efficacy in healthy volunteers (part 1), as well as in complement blockade-naive (part 2) and C5 inhibitor-treated (part 3) PNH patients. Twenty-nine patients were included in part 2 (n = 10) and part 3 (n = 19). Crovalimab concentrations exceeded the prespecified 100-µg/mL level and resulted in complete and sustained terminal complement pathway inhibition in treatment-naive and C5 inhibitor-pretreated PNH patients. Hemolytic activity and free C5 levels were suppressed below clinically relevant thresholds (liposome assay <10 U/mL and <50 ng/mL, respectively). Safety was consistent with the known profile of C5 inhibition. As expected, formation of drug-target-drug complexes was observed in all 19 patients switching to crovalimab, manifesting as transient mild or moderate vasculitic skin reactions in 2 of 19 participants. Both events resolved under continued treatment with crovalimab. Subcutaneous crovalimab (680 mg; 4 mL), administered once every 4 weeks, provides complete and sustained terminal complement pathway inhibition in patients with PNH, warranting further clinical development (ClinicalTrials.gov identifier, NCT03157635).

Daratumumab, a monoclonal CD38 antibody, is approved in the treatment of myeloma, but its efficacy and safety in light-chain (AL) amyloidosis has not been formally studied. This prospective phase 2 trial of daratumumab monotherapy for the treatment of AL amyloidosis was designed to determine the safety, tolerability, and hematologic and clinical response. Daratumumab 16 mg/kg was administered by IV infusion once weekly for weeks 1 to 8, every 2 weeks for weeks 9 to 24, and every 4 weeks thereafter until progression or unacceptable toxicity, for up to 24 months. Twenty-two patients with previously treated AL amyloidosis were enrolled. The majority of the patients had received high-dose melphalan and stem cell transplantation and/or treatment with a proteasome inhibitor. The median time between prior therapy and trial enrollment was 9 months (range, 1-180 months). No grade 3-4 infusion-related reactions occurred. The most common grade ≥3 adverse events included respiratory infections (n = 4; 18%) and atrial fibrillation (n = 4, 18%). Hematologic complete and very-good-partial response occurred in 86% of patients. The median time to first and best hematologic response was 4 weeks and 3 months, respectively. Renal response occurred in 10 of 15 patients (67%) with renal involvement and cardiac response occurred in 7 of 14 patients (50%) with cardiac involvement. In summary, daratumumab is well tolerated in patients with relapsed AL amyloidosis and leads to rapid and deep hematologic responses and organ responses. This trial was registered at www.clinicaltrials.gov as #NCT02841033.

Although thrombin is a key enzyme in the coagulation cascade and is required for both normal hemostasis and pathologic thrombogenesis, it also participates in its own negative feedback via activation of protein C, which downregulates thrombin generation by enzymatically inactivating factors Va and VIIIa. Our group and others have previously shown that thrombin's procoagulant and anticoagulant activities can be effectively disassociated to varying extents through site-directed mutagenesis. The thrombin mutant W215A/E217A (WE thrombin) has been one of the best characterized constructs with selective activity toward protein C. Although animal studies have demonstrated that WE thrombin acts as an anticoagulant through activated protein C (APC) generation, the observed limited systemic anticoagulation does not fully explain the antithrombotic potency of this or other thrombin mutants. AB002 (E-WE thrombin) is an investigational protein C activator thrombin analog in phase 2 clinical development (clinicaltrials.gov NCT03963895). Here, we demonstrate that this molecule is a potent enzyme that is able to rapidly interrupt arterial-type thrombus propagation at exceedingly low doses (<2 µg/kg, IV), yet without substantial systemic anticoagulation in baboons. We demonstrate that AB002 produces APC on platelet aggregates and competitively inhibits thrombin-activatable fibrinolysis inhibitor (carboxypeptidase B2) activation in vitro, which may contribute to the observed in vivo efficacy. We also describe its safety and activity in a phase 1 first-in-human clinical trial. Together, these results support further clinical evaluation of AB002 as a potentially safe and effective new approach for treating or preventing acute thrombotic and thromboembolic conditions. This trial was registered at www.clinicaltrials.gov as #NCT03453060.

A minority of patients with follicular lymphoma (FL) undergo histological transformation (HT). This retrospective analysis of 549 patients from the phase 3 GALLIUM study (NCT01332968) assessed the relationship between maximum standardized uptake value (SUVmax) at baseline on positron emission tomography (PET) and HT risk. Previously untreated patients with high tumor burden grade 1-3a FL received obinutuzumab- or rituximab-based chemotherapy induction. The relationship between baseline SUVmax (bSUVmax) and HT risk was assessed using cutoff values for SUVmax >10 and >20. Overall, 15 of 549 (2.7%) patients with baseline PET scans experienced biopsy-confirmed HT (median follow-up, 59 months). More than 65% of patients had bSUVmax > 10, with 3.3% of these experiencing HT. Only 1 of 74 (1.4%) patients with bSUVmax > 20 underwent HT. Median bSUVmax in patients with HT vs without HT was 12.4 (range, 8.1-28.0) vs 11.8 (range, 3.1-64.4), respectively; median bSUVrange (the difference between bSUVmax of the most and least 18F-fluorodeoxyglucose-avid lymphoma sites) was 8.0 (range, 1.1-23.9) vs 7.1 (range, 0.0-59.8), respectively. There was no temporal relationship between bSUVmax and HT. Neither bSUVmax nor bSUVrange predicted HT in GALLIUM, suggesting that there may be little benefit in rebiopsy of lesions to exclude HT based on SUVmax alone before initiating therapy in patients with high tumor burden FL.

The phase 3 ECHELON-1 study demonstrated that brentuximab vedotin (A) with doxorubicin, vinblastine, and dacarbazine (AVD; A+AVD) exhibited superior modified progression-free survival (PFS) vs doxorubicin, bleomycin, vinblastine, and dacarbazine (ABVD) for frontline treatment of patients with stage III/IV classical Hodgkin lymphoma (cHL). Maturing positron emission tomography (PET)-adapted trial data highlight potential limitations of PET-adapted approaches, including toxicities with dose intensification and higher-than-expected relapse rates in PET scan after cycle 2 (PET2)-negative (PET2-) patients. We present an update of the ECHELON-1 study, including an exploratory analysis of 3-year PFS per investigator. A total of 1334 patients with stage III or IV cHL were randomized 1:1 to receive 6 cycles of A+AVD (n = 664) or ABVD (n = 670). Interim PET2 was required. At median follow-up of 37 months, 3-year PFS rates were 83.1% with A+AVD and 76.0% with ABVD; 3-year PFS rates in PET2- patients aged <60 years were 87.2% vs 81.0%, respectively. A beneficial trend in PET2+ patients aged <60 years on A+AVD was also observed, with a 3-year PFS rate of 69.2% vs 54.7% with ABVD. The benefit of A+AVD in the intent-to-treat population appeared independent of disease stage and prognostic risk factors. Upon continued follow-up, 78% of patients with peripheral neuropathy on A+AVD had either complete resolution or improvement compared with 83% on ABVD. These data highlight that A+AVD provides a durable efficacy benefit compared with ABVD for frontline stage III/IV cHL, consistent across key subgroups regardless of patient status at PET2, without need for treatment intensification or bleomycin exposure. This trial was registered at www.clinicaltrials.gov as #NCT01712490 (EudraCT no. 2011-005450-60).

Relapse remains the most common cause of treatment failure for patients with acute myeloid leukemia (AML) who undergo allogeneic stem cell transplantation (alloSCT), and carries a grave prognosis. Multiple studies have identified the presence of measurable residual disease (MRD) assessed by flow cytometry before alloSCT as a strong predictor of relapse, but it is not clear how these findings apply to patients who test positive in molecular MRD assays, which have far greater sensitivity. We analyzed pretransplant blood and bone marrow samples by reverse-transcription polymerase chain reaction in 107 patients with NPM1-mutant AML enrolled in the UK National Cancer Research Institute AML17 study. After a median follow-up of 4.9 years, patients with negative, low (<200 copies per 105ABL in the peripheral blood and <1000 copies in the bone marrow aspirate), and high levels of MRD had an estimated 2-year overall survival (2y-OS) of 83%, 63%, and 13%, respectively (P < .0001). Focusing on patients with low-level MRD before alloSCT, those with FLT3 internal tandem duplications(ITDs) had significantly poorer outcome (hazard ratio [HR], 6.14; P = .01). Combining these variables was highly prognostic, dividing patients into 2 groups with 2y-OS of 17% and 82% (HR, 13.2; P < .0001). T-depletion was associated with significantly reduced survival both in the entire cohort (2y-OS, 56% vs 96%; HR, 3.24; P = .0005) and in MRD-positive patients (2y-OS, 34% vs 100%; HR, 3.78; P = .003), but there was no significant effect of either conditioning regimen or donor source on outcome. Registered at ISRCTN (http://www.isrctn.com/ISRCTN55675535).

Acute myeloid leukemia (AML) is a highly heterogeneous disease both in terms of genetic background and response to chemotherapy. Although molecular aberrations are routinely used to stratify AML patients into prognostic subgroups when receiving standard chemotherapy, the predictive value of the genetic background and co-occurring mutations remains to be assessed when using newly approved antileukemic drugs. In the present study, we retrospectively addressed the question of the predictive value of molecular events on the benefit of the addition of gemtuzumab ozogamicin (GO) to standard front-line chemotherapy. Using the more recent European LeukemiaNet (ELN) 2017 risk classification, we confirmed that the benefit of GO was restricted to the favorable (hazard ratio [HR], 0.54, 95% confidence interval [CI], 0.30-0.98) and intermediate (HR, 0.57; 95% CI, 0.33-1.00) risk categories, whereas it did not influence the outcome of patients within the adverse risk subgroup (HR, 0.93; 95% CI, 0.61-1.43). Interestingly, the benefit of GO was significant for patients with activating signaling mutations (HR, 0.43; 95% CI, 0.28-0.65), which correlated with higher CD33 expression levels. These results suggest that molecular aberrations could be critical for future differentially tailored treatments based on integrated genetic profiles that are able to predict the benefit of GO on outcome.

Therapeutic targeting of Bruton tyrosine kinase (BTK) has dramatically improved survival outcomes for patients with chronic lymphocytic leukemia (CLL)/small lymphocytic lymphoma (SLL). Acalabrutinib is an oral, highly selective BTK inhibitor that allows for twice-daily dosing due to its selectivity. In this phase 1b/2 study, 134 patients with relapsed/refractory CLL or SLL (median age, 66 years [range, 42-85 years]; median prior therapies, 2 [range, 1-13]) received acalabrutinib 100 mg twice daily for a median of 41 months (range, 0.2-58 months). Median trough BTK occupancy at steady state was 97%. Most adverse events (AEs) were mild or moderate, and were most commonly diarrhea (52%) and headache (51%). Grade ≥3 AEs (occurring in ≥5% of patients) were neutropenia (14%), pneumonia (11%), hypertension (7%), anemia (7%), and diarrhea (5%). Atrial fibrillation and major bleeding AEs (all grades) occurred in 7% and 5% of patients, respectively. Most patients (56%) remain on treatment; the primary reasons for discontinuation were progressive disease (21%) and AEs (11%). The overall response rate, including partial response with lymphocytosis, with acalabrutinib was 94%; responses were similar regardless of genomic features (presence of del(11)(q22.3), del(17)(p13.1), complex karyotype, or immunoglobulin variable region heavy chain mutation status). Median duration of response and progression-free survival (PFS) have not been reached; the estimated 45-month PFS was 62% (95% confidence interval, 51% to 71%). BTK mutation was detected in 6 of 9 patients (67%) at relapse. This updated and expanded study confirms the efficacy, durability of response, and long-term safety of acalabrutinib, justifying its further investigation in previously untreated and treated patients with CLL/SLL. This trial was registered at www.clinicaltrials.gov as #NCT02029443.

Ivosidenib (AG-120) is an oral, targeted agent that suppresses production of the oncometabolite 2-hydroxyglutarate via inhibition of the mutant isocitrate dehydrogenase 1 (IDH1; mIDH1) enzyme. From a phase 1 study of 258 patients with IDH1-mutant hematologic malignancies, we report results for 34 patients with newly diagnosed acute myeloid leukemia (AML) ineligible for standard therapy who received 500 mg ivosidenib daily. Median age was 76.5 years, 26 patients (76%) had secondary AML, and 16 (47%) had received ≥1 hypomethylating agent for an antecedent hematologic disorder. The most common all-grade adverse events were diarrhea (n = 18; 53%), fatigue (n = 16; 47%), nausea (n = 13; 38%), and decreased appetite (n = 12; 35%). Differentiation syndrome was reported in 6 patients (18%) (grade ≥3 in 3 [9%]) and did not require treatment discontinuation. Complete remission (CR) plus CR with partial hematologic recovery (CRh) rate was 42.4% (95% confidence interval [CI], 25.5% to 60.8%); CR 30.3% (95% CI, 15.6% to 48.7%). Median durations of CR+CRh and CR were not reached, with 95% CI lower bounds of 4.6 and 4.2 months, respectively; 61.5% and 77.8% of patients remained in remission at 1 year. With median follow-up of 23.5 months (range, 0.6-40.9 months), median overall survival was 12.6 months (95% CI, 4.5-25.7). Of 21 transfusion-dependent patients (63.6%) at baseline, 9 (42.9%) became transfusion independent. IDH1 mutation clearance was seen in 9/14 patients achieving CR+CRh (5/10 CR; 4/4 CRh). Ivosidenib monotherapy was well-tolerated and induced durable remissions and transfusion independence in patients with newly diagnosed AML. This trial was registered at www.clinicaltrials.gov as #NCT02074839.

This open-label, single-arm, prospective cohort trial is the first phase 3 safety study to describe outcomes in children treated with dabigatran etexilate for secondary venous thromboembolism (VTE) prevention. Eligible children aged 12 to <18 years (age stratum 1), 2 to <12 years (stratum 2), and >3 months to <2 years (stratum 3) had an objectively confirmed diagnosis of VTE treated with standard of care (SOC) for ≥3 months, or had completed dabigatran or SOC treatment in the DIVERSITY trial (NCT01895777) and had an unresolved clinical thrombosis risk factor requiring further anticoagulation. Children received dabigatran for up to 12 months, or less if the identified VTE clinical risk factor resolved. Primary end points included VTE recurrence, bleeding events, and mortality at 6 and 12 months. Overall, 203 children received dabigatran, with median exposure being 36.3 weeks (range, 0-57 weeks); 171 of 203 (84.2%) and 32 of 203 (15.8%) took capsules and pellets, respectively. Overall, 2 of 203 children (1.0%) experienced on-treatment VTE recurrence, and 3 of 203 (1.5%) experienced major bleeding events, with 2 (1.0%) reporting clinically relevant nonmajor bleeding events, and 37 (18.2%) minor bleeding events. There were no on-treatment deaths. On-treatment postthrombotic syndrome was reported for 2 of 162 children (1.2%) who had deep vein thrombosis or central-line thrombosis as their most recent VTE. Pharmacokinetic/pharmacodynamic relationships of dabigatran were similar to those in adult VTE patients. In summary, dabigatran showed a favorable safety profile for secondary VTE prevention in children aged from >3 months to <18 years with persistent VTE risk factor(s). This trial was registered at www.clinicaltrials.gov as #NCT02197416.

Sickle cell disease (SCD) is a genetic hemoglobinopathy leading to 2 major clinical manifestations: severe chronic hemolytic anemia and iterative vaso-occlusive crises. SCD is also accompanied by profound muscle microvascular remodeling. The beneficial effects of endurance training on microvasculature are widely known. The aim of this study was to evaluate the effects of an endurance training program on microvasculature of skeletal muscle in SCD patients. A biopsy of the vastus lateralis muscle and submaximal incremental exercise was performed before and after the training period. Of the 40 randomized SCD patients, complete data sets from 32 patients were obtained. The training group (n = 15) followed a personalized moderate-intensity endurance training program, while the nontraining (n = 17) group maintained a normal lifestyle. Training consisted of three 40-minute cycle ergometer exercise sessions per week for 8 weeks. Histological analysis highlighted microvascular benefits in the training SCD patients compared with nontraining patients, including increases in capillary density (P = .003), number of capillaries around a fiber (P = .015), and functional exchange surface (P < .0001). Conversely, no significant between-group difference was found in the morphology of capillaries. Indexes of physical ability also improved in the training patients. The moderate-intensity endurance exercise training program improved the muscle capillary network and partly reversed the microvascular defects commonly observed in skeletal muscle of SCD patients. This trial was registered at www.clinicaltrials.gov as #NCT02571088.

Clinical- and biomarker-based tools may identify a lower-risk acute graft-versus-host disease (GVHD) population amenable to novel, reduced-intensity treatments. Previous data suggest sirolimus may rival standard of care prednisone. We conducted a National Heart, Lung, and Blood Institute/National Cancer Institute-funded Blood and Marrow Transplant Clinical Trials Network multicenter, open-label, randomized phase 2 trial to estimate the difference in day 28 complete response (CR)/partial response (PR) rates for sirolimus vs prednisone as initial treatment of patients with standard risk (SR) acute GVHD as defined by the Minnesota (MN) GVHD Risk Score and Ann Arbor (AA1/2) biomarker status. A total of 127 MN-SR patients were randomized (1:1), and 122 were AA1/2 (sirolimus, n = 58; prednisone, n = 64). Others were AA3 (n = 4), or AA status missing (n = 1). The day 28 CR/PR rates were similar for sirolimus 64.8% (90% confidence interval [CI], 54.1%-75.5%) vs 73% (90% CI, 63.8%-82.2%) for prednisone. The day 28 rate of CR/PR with prednisone ≤0.25 mg/kg/day was significantly higher for sirolimus than prednisone (66.7% vs 31.7%; P < .001). No differences were detected in steroid-refractory acute GVHD, disease-free survival, relapse, nonrelapse mortality, or overall survival. Sirolimus was associated with reduced steroid exposure and hyperglycemia, reduced grade 2 to 3 infections, improvement in immune suppression discontinuation and patient-reported quality of life, and increased risk for thrombotic microangiopathy. For patients with clinical- and biomarker-based SR acute GVHD, sirolimus demonstrates similar overall initial treatment efficacy as prednisone. In addition, sirolimus therapy spares steroid exposure and allied toxicity, does not compromise long-term survival outcomes, and is associated with improved patient-reported quality of life. This trial was registered at www.clinicaltrials.gov as #NCT02806947.

Antigen-escape relapse has emerged as a major challenge for long-term disease control after CD19-directed therapies, to which dual-targeting of CD19 and CD22 has been proposed as a potential solution. From March 2016 through January 2018, we conducted a pilot study in 89 patients who had refractory/relapsed B-cell malignancies, to evaluate the efficacy and safety of sequential infusion of anti-CD19 and anti-CD22, a cocktail of 2 single-specific, third-generation chimeric antigen receptor-engineered (CAR19/22) T cells. Among the 51 patients with acute lymphoblastic leukemia, the minimal residual disease-negative response rate was 96.0% (95% confidence interval [CI], 86.3-99.5). With a median follow-up of 16.7 months (range, 1.3-33.3), the median progression-free survival (PFS) was 13.6 months (95% CI, 6.5 to not reached [NR]), and the median overall survival (OS) was 31.0 months (95% CI, 10.6-NR). Among the 38 patients with non-Hodgkin lymphoma, the overall response rate was 72.2% (95% CI, 54.8-85.8), with a complete response rate of 50.0% (95% CI, 32.9-67.1). With a median follow-up of 14.4 months (range, 0.4-27.4), the median PFS was 9.9 months (95% CI, 3.3-NR), and the median OS was 18.0 months (95% CI, 6.1-NR). Antigen-loss relapse occurred in 1 patient during follow-up. High-grade cytokine release syndrome and neurotoxicity occurred in 22.4% and 1.12% patients, respectively. In all except 1, these effects were reversible. Our results indicated that sequential infusion of CAR19/22 T cell was safe and efficacious and may have reduced the rate of antigen-escape relapse in B-cell malignancies. This trial was registered at www.chictr.org.cn as #ChiCTR-OPN-16008526.

The Platelets for Neonatal Thrombocytopenia (PlaNeT-2) trial reported an unexpected overall benefit of a prophylactic platelet transfusion threshold of 25 × 109/L compared with 50 × 109/L for major bleeding and/or mortality in preterm neonates (7% absolute-risk reduction). However, some neonates in the trial may have experienced little benefit or even harm from the 25 × 109/L threshold. We wanted to assess this heterogeneity of treatment effect in the PlaNet-2 trial, to investigate whether all preterm neonates benefit from the low threshold. We developed a multivariate logistic regression model in the PlaNet-2 data to predict baseline risk of major bleeding and/or mortality for all 653 neonates. We then ranked the neonates based on their predicted baseline risk and categorized them into 4 risk quartiles. Within these quartiles, we assessed absolute-risk difference between the 50 × 109/L- and 25 × 109/L-threshold groups. A total of 146 neonates died or developed major bleeding. The internally validated C-statistic of the model was 0.63 (95% confidence interval, 0.58-0.68). The 25 × 109/L threshold was associated with absolute-risk reduction in all risk groups, varying from 4.9% in the lowest risk group to 12.3% in the highest risk group. These results suggest that a 25 × 109/L prophylactic platelet count threshold can be adopted in all preterm neonates, irrespective of predicted baseline outcome risk. Future studies are needed to improve the predictive accuracy of the baseline risk model. This trial was registered at www.isrctn.com as #ISRCTN87736839.