- © 2009 by American Society of Clinical Oncology
Randomized Comparison of the Stanford V Regimen and ABVD in the Treatment of Advanced Hodgkin's Lymphoma: United Kingdom National Cancer Research Institute Lymphoma Group Study ISRCTN 64141244
- Peter J. Hoskin,
- Lisa Lowry,
- Alan Horwich,
- Andrew Jack,
- Ben Mead,
- Barry W. Hancock,
- Paul Smith,
- Wendi Qian,
- Philippa Patrick,
- Bilyana Popova,
- Andrew Pettitt,
- David Cunningham,
- Ruth Pettengell,
- John Sweetenham,
- David Linch and
- Peter W.M. Johnson
- From the Mount Vernon Cancer Centre, Department of Clinical Oncology; Lymphoma Trials Office, Cancer Research United Kingdom and University College London Cancer Trials Centre; Department of Radiotherapy, Institute of Cancer Research and Royal Marsden Hospital; Medical Research Council Cancer Trials Unit; St George's Hospital Medical School; Royal Marsden Hospital; and University College London Cancer Institute, London; St James' Institute of Oncology, Haematological Malignancy Diagnostic Service, Leeds; Cancer Research United Kingdom Clinical Centre, University of Southampton, Southampton; Academic Unit of Clinical Oncology, Weston Park Hospital, Sheffield; and University of Liverpool, Liverpool, United Kingdom; and Cleveland Clinic Taussig Cancer Institute, OH.
- Corresponding author: Peter J. Hoskin, MD, Mount Vernon Cancer Centre, Mount Vernon Hospital, Rickmansworth Rd, Northwood, Middlesex, HA6 2RN, United Kingdom; e-mail: peterhoskin{at}nhs.net.
-
Presented at the Annual Meeting of the American Society of Hematology, San Francisco, CA, December, 6-9, 2008; and in part at the 9th International Conference on Malignant Lymphoma, Lugano, Switzerland, June, 7-11, 2005.
Abstract
Purpose This multicenter, prospective, randomized controlled trial compared the efficacy and toxicity of two chemotherapy regimens in advanced Hodgkin's lymphoma (HL): the weekly alternating Stanford V and the standard, twice-weekly regimen of doxorubicin, bleomycin, vinblastine, and dacarbazine (ABVD).
Patients and Methods Patients had stage IIB, III, or IV disease or had stages I to IIA disease with bulky disease or other adverse features. Radiotherapy was administered in both arms to sites of previous bulk (> 5 cm) and to splenic deposits, although this was omitted in the latter part of the trial for patients achieving complete remission (CR) in the ABVD arm. A total of 520 patients were randomly assigned and were assessed for the primary outcome measure of progression-free survival (PFS). Five hundred patients received protocol treatment, and radiotherapy was administered to 73% in the Stanford V arm and to 53% in the ABVD arm.
Results The overall response rates after completion of all treatment were 91% for Stanford V and 92% for ABVD. During a median follow-up of 4.3 years, there was no evidence of a difference in projected 5-year PFS and overall survival (OS) rates (76% and 90%, respectively, for ABVD; 74% and 92%, respectively, for Stanford V). More pulmonary toxicity was reported for ABVD, whereas other toxicities were more frequent with Stanford V.
Conclusion In a large, randomized trial, the efficacies of Stanford V and ABVD were comparable when given in combination with appropriate radiotherapy.
INTRODUCTION
The regimen of doxorubicin, bleomycin, vinblastine, and dacarbazine (ABVD) is considered the standard treatment for patients with advanced-stage Hodgkin's lymphoma (HL).1–3 The optimal regimen should combine a high response rate and failure-free survival (FFS) with low levels of toxicity and particularly should avoid late effects, such as infertility, second malignancies, and cardiovascular disease. The 3-year FFS after ABVD is approximately 75%.4 This compares favorably to mechlorethamine, vincristine, procarbazine, and prednisone (MOPP)5 and to more complex multidrug regimens.3,4 Several novel, brief-duration regimens for the treatment of advanced HL have been developed to reduce cumulative doses of several drugs implicated in causing long-term toxic effects and to increase the dose-intensity by reduction in treatment intervals and total duration, as retrospective analyses suggest a relationship between treatment outcome and dose-intensity.6
At the time of development of this trial, preliminary results from several of these regimens had been reported, including PACE BOM (prednisolone, doxorubicin, cyclophosphamide, etoposide, bleomycin, vincristine, methotrexate),7 VAPEC-B (vincristine, doxorubicin, prednisolone, etoposide, cyclophosphamide, bleomycin),8 and Stanford V (SV).9 High response rates were reported for all of these regimens in single-institution, phase II studies, but the studies reported variable durability of response. Results from the SV regimen were particularly encouraging. This regimen comprises brief-duration combination chemotherapy, which includes relatively low cumulative doses of alkylating agents, doxorubicin, and bleomycin followed by integrated, planned radiotherapy to sites of initial bulk disease.
Mature results from the original, single-center, prospective trial with the SV regimen were reported in 2002.10 All 142 patients completed the 12-week chemotherapy course, and 91% received consolidative radiotherapy. No treatment-related deaths were observed, and a single second malignancy had occurred. The estimated 5-year overall survival (OS) was 96%, and the probability of freedom from progression (FFP) was 89%. The use of SV in a multicenter setting of five centers in the Eastern Cooperative Oncology Group (ECOG) was reported in 2000.11 Forty-seven patients were treated, of whom 87% received radiotherapy. The estimated FFP at 5 years was 85%, and the OS was 96%.
Since this trial started, a report has been published of a randomized, multicenter trial that compared ABVD versus SV and versus methotrexate, etoposide, and lomustine (MEC).12 The 5-year FFS was significantly lower for patients treated with SV (54%) compared with ABVD (78%) and MEC (81%). Radiotherapy was not mandatory and was given to 66% and 62% of patients in the SV and ABVD arms, respectively. Longer-term follow-up was reported recently in abstract form, and the 8-year OS rates were 88%, 84%, and 77% for ABVD, MEC, and SV arms, respectively (P = .337).13
Here, we report the results of a prospective, randomized, multicenter trial that compared ABVD and SV in advanced-stage HL. An initial, randomized, phase II trial demonstrated SV toxicity and response rates comparable to ABVD and led to an expanded, phase III trial; the results from both phases are presented here.
PATIENTS AND METHODS
Eligibility
Consenting patients who were age 18 to 60 years with advanced HL and no prior therapy were included. Staging included a full history and physical examination; computed tomography (CT) scans of chest, abdomen, and pelvis; and bone marrow aspirate and trephine biopsy. Pregnant or lactating women and patients with known HIV infection, pre-existing cardiac or pulmonary disease, or previous malignant disease (except for basal cell or squamous cell carcinoma of the skin) were excluded. Normal full blood count, hepatic function, and renal function were required, unless an abnormality was attributable to HL. Advanced HL was defined as clinical stages IB, IIB, or III to IV or as clinical stages IA to IIA with locally extensive disease (ie, bulk mediastinal disease [> 0.33 of the maximum transthoracic diameter on chest x-ray], extranodal extension, or other poor risk features as a result of which it was considered necessary to treat with full course combination chemotherapy). Ethics committee approval and regulatory approval of the protocol were obtained. Central pathologic review of the diagnostic material was undertaken.
Random Assignment and Treatment
The Cancer Research United Kingdom and University College London Cancer Trials Centre performed random assignment by telephone. Patients were randomly assigned in a 1:1 ratio by using block random assignment. Stratification for Hasenclever score (≤ 2 v > 2)14 was employed from July 2002. Patients randomly assigned to the SV regimen were prescribed 12 weeks of chemotherapy given as follows: mechlorethamine 6 mg/m2 on weeks 1, 5, and 9; doxorubicin 25 mg/m2 on weeks 1, 3, 5, 7, 9, and 11; vinblastine 6 mg/m2 on weeks 1, 3, 5, 7, 9, and 11; prednisone 40 mg/m2 on alternate days for 9 weeks followed by a taper; vincristine 1.4 mg/m2 on weeks 2, 4, 6, 8, 10, and 12; bleomycin 5,000 IU/m2 on weeks 2, 4, 6, 8, 10, and 12; and etoposide 60 mg/m2/d on two consecutive days on weeks 3, 7,and 11. Patients randomly assigned to the ABVD regimen were prescribed six to eight cycles of chemotherapy at 4-week intervals given as follows: doxorubicin 25 mg/m2, bleomycin 10,000 units/m2, vinblastine 6 mg/m2, and dacarbazine 375 mg/m2 on days 1 and 15. After four cycles of ABVD, an assessment of response was made. If complete remission (CR) was achieved after four cycles, two additional cycles were recommended; if partial remission (PR) was achieved after four cycles, two more cycles were given and were followed by additional assessment. If there had been no additional response, no additional chemotherapy was given. For patients with a continued response between cycles 4 and 6, two more cycles of ABVD were recommended, up to a total of eight cycles.
Granulocyte colony-stimulating factor (GCSF) was used to maintain treatment intensity if more than one dose reduction was needed for hematologic toxicity or if treatment had to be delayed because of inadequate blood counts. Patients were considered eligible for involved field radiotherapy (IFRT) if they had achieved a CR or PR to induction chemotherapy and had one or more of the following: initial bulky mediastinal disease; initial nodal mass ≥ 5 cm in diameter; and initial splenic disease, defined as macroscopic nodules on CT scan, which was not surgically removed.
The indication for radiotherapy was assessed prospectively at each center in combined consultation with a radiation oncologist. The recommended total dose to initial bulky disease was 34 to 36 Gy delivered in 1.5- to 2-Gy fractions. It was recommended that sites of extralymphatic extension, which were included in the initial bulky sites of disease, should be treated to at least 15 Gy (for lungs and pleura) or 20 Gy (for other extralymphatic sites). Radiotherapy quality assurance was performed independently.15
Initially, the protocol indications for radiotherapy in both treatment groups were identical. During the course of the phase III trial, evidence emerged from a randomized study that suggested that radiotherapy could be omitted safely for patients who achieved CR with combination chemotherapy.16 Patients randomly assigned to ABVD after January 2004 were only offered IFRT if they had not achieved CR after chemotherapy.
Assessment of Response
Tumor response was assessed at the completion of chemotherapy and again at the end of IFRT, if given. Response was defined according to the Cotswolds criteria.17
Statistical Design and Methods
The primary outcome measure was progression-free survival (PFS); secondary outcome measures were OS and toxicity. The target sample size was originally 700 patients, but this was reviewed at the independent data monitoring committee meeting in March 2007, with the knowledge that the estimated 5-year PFS for both arms combined was 75%. To detect a 10% absolute improvement in 5-year PFS from 75% with ABVD to 85% with SV, with a two-sided significance level of .05, a total of 97 or 130 events (ie, progression or death) were required to have a power of 80% or 90%, respectively. The target sample size was updated to 580 patients. However, because of persistent problems with the supply of mechlorethamine, the trial was closed early. Formal analysis was planned when the total number of events was greater than 97. At the time of reporting, 116 events had been recorded, which provided a power of at least 85%.
PFS was calculated from the date of random assignment to the date of progression, relapse, or death as a result of any cause; at the time of the analysis, survivors without disease progression were censored at the date last known to be alive. OS was calculated from the date of random assignment to the date of death as a result of any cause. The log-rank test stratified by phase II versus phase III was applied to compare the Kaplan-Meier curves for PFS and OS. The χ2 test for interaction or trend was implemented to assess in an exploratory manner the differences of the relative benefits of SV in different subgroups of patients. The tumor responses after chemotherapy and at the end of treatment were compared by using the Mann-Whitney test.
All analyses were done on an intention-to-treat basis except for the analyses of response and toxicity, which were restricted to patients who received at least one cycle per week of allocated chemotherapy treatment. All P values were two sided (Fig 1).
Role of the Funding Source
Cancer Research United Kingdom, as funder of this trial, reviewed and approved its design, and the execution was overseen by an independent data monitoring committee. The data analysis, preparation of the manuscript, and interpretation were performed independently from the funders.
RESULTS
Patients
Between March 1998 and October 2006, a total of 520 patients (SV, n = 259; ABVD, n = 261) were randomly assigned from 74 centers in the United Kingdom; 150 of these patients were treated in the phase II study. Patient characteristics at random assignment were balanced between the two arms (Table 1).
Treatment Received
A total of 248 patients were treated with the SV regimen (243 patients [95%] completed 12 weeks; two patients completed 11 weeks; and three patients completed 9 weeks). Six patients randomly assigned to the SV arm did not receive treatment; four patients received ABVD because of failure of mechlorethamine supply, and two patients were withdrawn as ineligible. Data were missing for the remaining five patients. In the ABVD group, 252 patients received ABVD chemotherapy (six cycles, n = 179; six to eight cycles, n = 59; and less than six cycles, n = 14). One patient was withdrawn because of a change in diagnosis, and this patient had no trial treatment; treatment details are missing for eight patients.
After chemotherapy, IFRT was given as part of primary treatment to 308 patients (63%) overall (SV arm, n = 178 [73%];ABVD arm, n = 130 [53%]). In the phase II study, there was no significant difference in rates of radiotherapy administration (79% v 73%) when the indications for IFRT were identical in the two arms.
Adverse Effects
Immediate toxicities were comparable between the two arms. The incidence of reported nonpulmonary, grade 3 to 4 toxicity was greater with SV (19%) than with ABVD (8%). Five patients (2%) in the SV arm had grade 4 toxicity reported (cytopenias, n = 3; peripheral neuropathy, n = 1; and hyponatraemia, n = 1). Two patients (1%) in the ABVD arm had grade 4 toxicity reported (neutropenia, n = 1; mood alteration, n = 1). Pulmonary toxicity was reported in four patients (2%) in the SV arm; one of these patients discontinued bleomycin as a result. Pulmonary toxicity was reported in 23 patients (10%) in the ABVD arm; 20 of these patients discontinued bleomycin. There was one death as a result of treatment-related toxicity, which occurred in the ABVD arm.
A total of 185 patients (75%) who received ABVD also received GCSF at some point during chemotherapy, compared with 97 patients (41%) who received SV. There were eight second malignancies in the trial cohort. After ABVD, occurrences of diffuse large B-cell lymphoma (DLBCL), rectal carcinoma, lung cancer, and colonic adenocarcinoma were diagnosed 8 months, 15 months, 5 years, and 7 years, respectively, after random assignment. Within the SV arm, diagnoses were made of follicular non-Hodgkin's lymphoma and pancreatic carcinoma at 8 and 9 months after random assignment, respectively, and two occurrences of DLBCL were diagnosed at 4 and 8 months after random assignment.
Tumor Response
The overall response rates (ie, CR, complete remission unconfirmed (CRu), and PR) at completion of all treatment were 91% for SV and 92% for ABVD. The CR/CRu rate was higher in the ABVD group at the end of chemotherapy (55% v 36% on SV; P < .0001), and a significant difference persisted at the end of all treatment (67% v 57%; P = .036; Table 2).
Survival
At the time of analysis, after a median follow-up of 52 months (which was similar between the two groups), 79% of those randomly assigned to ABVD and 77% of those allocated to SV were alive without disease progression. Eighteen patients had died in the SV arm: 15 as a result of HL, one as a result of a road traffic accident, one as a result of suicide in CR, and one as a result of DLBCL. Twenty-four patients had died in the ABVD arm: 15 as a result of HL, one as a result of non-Hodgkin's lymphoma (revised initial diagnosis; did not receive protocol treatment), one as a result of treatment toxicity, two as a result of complications of subsequent high-dose therapy, and five as a result of other causes.
Estimated 5-year PFS did not significantly differ at 76% for ABVD and 74% for SV (Fig 2). There was no significant difference in outcomes between the phase II and phase III trials. The 5-year OS rates of 90% with ABVD and 92% with SV were not significantly different (Fig 3). Exploratory subgroup analyses showed no evidence that any subgroup of patients (according to Hasenclever score, stage, mediastinal bulk, or radiotherapy received) benefited from one treatment compared with the other (Fig 4).
DISCUSSION
This large, randomized trial has shown no significant difference in PFS or OS between SV and ABVD for the initial therapy of advanced HL. The trial was designed to detect an improvement in outcome after SV, as suggested by a single-institution, single-arm study.9,10 There are several interpretations of the negative result. The first, most likely explanation is that there is truly no difference in outcomes for the two regimens when compared in a randomized fashion. The PFS and OS rates in the ABVD arm were fully in keeping with previous reports,3–5 but the results for SV were not as good as those reported in the original North American series. The 5-year PFS and OS rates were 89% and 96%, unlike the 74% and 92% PFS and OS rates in this trial. Extension from single-center to multicenter recruitment may have resulted in differences in patient selection, although there was no evidence that compliance with the chemotherapy protocol was diminished. The second possibility is a false-negative result. This trial was closed early because of persistent drug supply problems; therefore, it did not reach the planned recruitment target to achieve 90% power to detect a 10% improvement in PFS in the SV arm compared with the ABVD arm. However, the risk of a false negative is low, as longer follow-up had allowed more events to occur, and the study already had more than 85% power. Finally, it is possible that the lower use of radiotherapy in this trial than in the original series could have affected outcomes. Although the specified indications for radiotherapy were the same in the SV arm of this trial as in the original regimen,9 the uptake was considerably lower at 73% versus 91%. Although the protocol outlined indications for radiotherapy and recommended doses, the decision on whether to treat was left up to teams in the individual centers. It is, however, impossible to directly compare the patients from the two trials in terms of radiotherapy indications, and it remains unlikely that increased radiotherapy use in the SV arm would have led to significant improvement in PFS and OS compared with the ABVD arm.
The finding of no difference in PFS and OS between the two arms is at variance with results from the other randomized trial involving (modified) SV and ABVD, published in 2005,12 in which the 5-year PFS was significantly worse with SV. The indications for radiotherapy in this Italian trial were different than those in this trial, and the aim was to determine whether radiotherapy could be safely reduced. A maximum of two sites of disease were treated, and clinicians were authorized to omit radiotherapy in patients with unequivocal CR, regardless of initial bulk. Only 66% of patients in the SV arm and 62% of patients in the ABVD arm received radiotherapy, compared with 73% and 53%, respectively, in this trial. It is possible that differences in the application of radiotherapy account for some of the differences in outcome for patients treated with SV in the two trials.
After the randomized study, which suggested that patients with a CR after multiagent chemotherapy did not benefit from consolidation radiotherapy,16 many centers have moved to omit this step. Nonrandomized data has recently been presented that suggest a benefit from radiotherapy persists, irrespective of the degree of response.18 The omission of radiotherapy from the SV regimen is not recommended for patients with initial sites of bulk disease or splenic deposits. In this study, rates of radiotherapy administration in the SV arm were lower than with the original Stanford series but higher than with the Italian trial. For these reasons, we conducted exploratory analyses that compared PFS in those who received radiotherapy with that of patients who did not (data not shown). The results suggest some detriment from the omission of radiotherapy in the SV arm but not the ABVD arm. Thus, SV should be used only when in combined modality treatment with planned radiotherapy.
In this trial, the overall response rates did not differ between the two arms. The CR and CRu rates were higher in the ABVD arm, which likely reflects the treatment of patients in the SV arm for only 12 weeks, rather than for 24 to 32 weeks as with ABVD. This persisted at the final disease assessment after completion of all therapy, as 67% and 57% of patients attained CR/CRu on ABVD and SV, respectively (P = .036).
There was more pulmonary toxicity reported in the ABVD arm than in the SV arm. ABVD patients were also more likely to require GCSF during therapy. More nonpulmonary grades 3 to 4 toxicities were reported with SV (19%) than with ABVD (8%). As more patients received radiotherapy in the SV arm, it is possible more late effects may become apparent on longer follow-up.
For many, ABVD still represents the standard initial treatment regimen for advanced HL; however, excellent responses and PFS have been documented with standard and escalated regimens of bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine, and prednisone (BEACOPP).19 Several groups have compared BEACOPP with ABVD in randomized trials, but neither has been established yet as definitely superior.20–22 Although BEACOPP may be associated with an improved PFS, it has greater toxicity; because of the success of salvage regimens, improvements in PFS may not translate to improved OS.
In this trial, ABVD was not bettered by the SV regimen, despite the use of radiotherapy in the majority of instances. ABVD offers the potential to avoid radiotherapy in patients who experience CR and, therefore, is likely to remain standard therapy. However, for some patients, SV will be the first-choice regimen because of the brief duration of treatment and reduced risk of acute pulmonary toxicity. The two arms were comparable in terms of PFS and OS; SV, therefore, remains a valid option.
AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST
The author(s) indicated no potential conflicts of interest.
AUTHOR CONTRIBUTIONS
Conception and design: Peter J. Hoskin, Alan Horwich, Ben Mead, Barry W. Hancock, Ruth Pettengell, John Sweetenham, David Linch, Peter W.M. Johnson
Administrative support: Philippa Patrick, Bilyana Popova
Provision of study materials or patients: Peter J. Hoskin, Alan Horwich, Andrew Jack, Ben Mead, Barry W. Hancock, Andrew Pettitt, David Cunningham, Ruth Pettengell, John Sweetenham, David Linch, Peter W.M. Johnson
Collection and assembly of data: Paul Smith, Philippa Patrick, Bilyana Popova
Data analysis and interpretation: Peter J. Hoskin, Lisa Lowry, Paul Smith, Wendi Qian, Peter W.M. Johnson
Manuscript writing: Peter J. Hoskin, Lisa Lowry, Wendi Qian, Peter W.M. Johnson
Final approval of manuscript: Peter J. Hoskin, Lisa Lowry, Alan Horwich, Andrew Jack, Ben Mead, Barry W. Hancock, Paul Smith, Wendi Qian, Philippa Patrick, Bilyana Popova, Andrew Pettitt, David Cunningham, Ruth Pettengell, John Sweetenham, David Linch, Peter W.M. Johnson
Acknowledgment
We thank all patients who participated in this trial and research staff at the clinical centers who helped to recruit patients and collect data. We thank the National Cancer Research Network. We thank the staff of the Lymphoma Trials Office, CR UK and University College London Cancer Trials Centre for trial and data management.
Appendix
Phase II and III Trial Management Groups
Peter Hoskin (chief investigator), John Sweetenham, Alan Horwich, Ruth Pettengell, Barry Hancock, David Linch, Paul Smith, Wendi Qian, Philippa Patrick, Andrew Pettitt, and Peter W.M. Johnson.
Independent Data Monitoring Committee
Howie Scarffe, MD, Rob Glynne-Jones, MD, and Robin Prescott, PhD.
Principal Investigators at Local Centers
G. Follows (Addenbrooke's), P. Cervi (Basildon Hospital), P. Kettle (Belfast City Hospital), D. Bareford (Birmingham City Hospital), D.W. Milligan (Birmingham Heartlands), C. Price (Bristol), S. Elyan (Cheltenham General Hospital), J.A. Radford (Christie Hospital, Manchester), J. Beard (Conquest Hospital), D. Gilson (Cookridge Hospital), M. Persic (Derbyshire Hospital), H. Ciepluch (Diana Princess of Wales Hospital), J. Joseph (Doncaster Royal Infirmary), P. Murray (Essex County Hospital), C. Hoyle (Glan Clywd Hospital), M. Lumley (Good Hope Hospital), J. Harrison (Hemel Hempstead Hospital), R. Patmore (Hull Royal Infirmary), J.S. Morgan (Ipswich Hospital), S. Sadullah (James Paget Hospital), C. Pocock (Kent and Canterbury Hospital), M. Lyttelton (Kettering Hospital), G. Smith (Leeds General Infirmary), M. Dyer (Leicester Royal Infirmary), Adelman (Lincoln County Hospital), I. Singer (Monklands Hospital), P. Hoskin (Mount Vernon Hospital), C. Brammer (New Cross Hospital), J. Wimperis (Norfolk and Norwich Hospital), A. Milne (North Hampshire Hospital), J. Ross (Northampton General Hospital), A. McMillan (Nottingham City Hospital), A. Brownell (Oldchurch Hospital), C. Hatton (John Radcliffe Hospital, Oxford), P. Hillmen (Pinderfields Hospital), F. Al-Refaie (Princess Alexandra Hospital), F. Clark (Queen Elizabeth Hospital, Birmingham), P. Coates (Queen Elizabeth Hospital, Kings Lynn), S. Killick (Royal Bournemouth Hospital), A. Kruger (Royal Cornwall Hospital), M. Joyner (Royal Devon and Exeter Hospital), S. MacKinnon (Royal Free Hospital), A. Pettitt (Royal Liverpool University Hospital), D. Cunningham (Royal Marsden Hospital), J. Simpson (Royal Sussex County Hospital), C. Knechtli (Royal United Hospital, Bath), S. Fernandes (Russells Hall), J. Cullis (Salisbury District Hospital), F. Wandroo (Sandwell General Hospital), S. Al-Ismail (Singleton Hospital), P. Johnson (Southampton General Hospital), A. Eden (Southend), A. Lister (St Bartholomew's Hospital), R. Pettengell (St George's Hospital), A. O'Calloghan (St Mary's Hospital, Portsmouth), P. Bevan (St Richard's Hospital), A. Timothy (St Thomas' Hospital), P. Revell (Staffordshire General Hospital), D. Turner (Torbay Hospital), D. Linch (UCLH), B.E. Woodcock (University Hospital Aintree), T. Maughan (Velindre), C. Irwin (Walsgrave Hospital, Coventry), A. Borg (Warwick Hospital), J. Harrison (Watford General Hospital), M. Moody (West Suffolk Hospital), B.W. Hancock (Weston Park Hospital, Sheffield), G. Satchi (Whiston Hospital), S. Shafeek (Alexandra, Kidderminster and Worcestershire Royal Hospitals), S. Narat (Worthing), L. Bond (York District Hospital).
Footnotes
-
Supported by Cancer Research United Kingdom Grant No. C2422/A2858.
-
Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
-
Clinical trial information can be found for the following: ISRCTN64141244.
- Received March 31, 2009.
- Accepted June 4, 2009.