NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

MacAllister R, Clayton T, Knight R, et al. REmote preconditioning for Protection Against Ischaemia–Reperfusion in renal transplantation (REPAIR): a multicentre, multinational, double-blind, factorial designed randomised controlled trial. Southampton (UK): NIHR Journals Library; 2015 May. (Efficacy and Mechanism Evaluation, No. 2.3.)

Cover of REmote preconditioning for Protection Against Ischaemia–Reperfusion in renal transplantation (REPAIR): a multicentre, multinational, double-blind, factorial designed randomised controlled trial

REmote preconditioning for Protection Against Ischaemia–Reperfusion in renal transplantation (REPAIR): a multicentre, multinational, double-blind, factorial designed randomised controlled trial.

Show details

Chapter 2Methods

Study design

The REPAIR trial was a multicentre, double-blind, European-based randomised controlled trial assessing the impact of RIPC on kidney function following renal transplantation. It used a 2 × 2 factorial design in which the recipients and their donors were randomised to RIPC or a sham procedure both 24 hours before surgery (late RIPC) and immediately pre surgery (early RIPC). Therefore, there were four arms in total: (1) a sham procedure both 24 hours before surgery and immediately pre surgery, (2) early RIPC and a sham procedure 24 hours before surgery, (3) late RIPC and a sham procedure immediately pre surgery and (4) late RIPC and early RIPC. Both donor and recipient were randomised to the same intervention group.

Aim

The REPAIR trial investigated whether RIPC improved kidney function and other clinical outcomes following renal transplantation. RIPC is a simple, non-invasive and virtually cost-free intervention, and any improvement in graft function might ultimately lead to prolonged allograft life in these patients, with resultant economic and quality of life benefits. The findings might also have implications for the use of RIPC in other clinical ischaemic syndromes.

Participants

The study intended to recruit 400 pairs of transplant recipients and their living donors from centres in the UK, the Netherlands, Belgium and France. Patients undergoing living-donor transplantation aged ≥ 18 years from 13 tertiary care hospitals in the UK, the Netherlands, Belgium and France were invited to take part in the study.

Final inclusion and exclusion criteria

Inclusion criteria

  1. Patients undergoing living-donor transplantation.
  2. Patients aged ≥ 18 years.

Exclusion criteria

  1. Patients on KATP channel-opening or -blocking drugs.
  2. Patients on ciclosporin.
  3. Patients with a known iodine sensitivity (who cannot undergo iohexol clearance studies).
  4. Patients with ABO incompatibility.
  5. Any patient requiring human leucocyte antigen (HLA) antibody removal therapy.

There were two major protocol amendments that affected the inclusion and exclusion criteria. In December 2009, the decision was made to exclude patients who require HLA antibody removal therapy as they required a different immunosuppression regime that may have an effect on preconditioning. In September 2010, the decision was made to include patients who had had a previous transplant, who were originally excluded from the trial. It had been considered that RIPC may not be as effective in patients who have undergone a previous transplant; however, there was subsequently a change in opinions and evidence.

A full list of protocol amendments can be found in Appendix 1.

Recruitment

The 13 centres from which patients were recruited were Guy’s and St Thomas’ NHS Foundation Trust, London, UK; Leiden University Medical Centre, Leiden, the Netherlands; North Bristol NHS Trust, Bristol, UK; Cambridge University Hospitals NHS Foundation Trust, Cambridge, UK; Royal Free London NHS Foundation Trust, London, UK; Queen Elizabeth Hospital, Birmingham, UK; Vrije Universiteit (VU) University Medical Centre, Amsterdam, the Netherlands; St George’s Hospital, London, UK; Royal London Hospital, London, UK; Western Infirmary, Glasgow, UK; Centre Hospitalier Universitaire (CHU) Erasme, Brussels, Belgium; CHU de Liège, Belgium; and Centre Hospitalier Régional Universitaire (CHRU) de Lille, France. The recruitment of patients was initiated in the outpatient setting. Both the recipient and the donor were given the information sheet prior to giving consent.

Randomisation, concealment and blinding

Patients were allocated at random in a 1 : 1 : 1 : 1 ratio to the control condition (sham RIPC), early RIPC alone (immediately pre surgery), late RIPC alone (24 hours pre surgery) and dual RIPC (RIPC 24 hours before surgery and immediately pre surgery). This was performed using a web-based service provided by Sealed Envelope (Sealed Envelope Ltd, London, UK) through the Clinical Trials Unit (CTU) at the London School of Hygiene and Tropical Medicine. The method of randomisation was random permuted blocks of size four and eight stratified by recruiting centre. During the study only the unblinded statistician supporting the Data Monitoring Committee (DMC) had access to the randomisation codes.

The enrolment and preconditioning procedures were performed by an unblinded research nurse who was not involved in sample collection or data analysis. All other research personnel at each study site, including those responsible for assessing outcomes, remained blinded to the allocation of patients to either real or sham RIPC. The patients and donors were also blinded to the allocation of their randomised intervention by the use of sham procedures in those not allocated to RIPC. A limited number of staff at the CTU were unblinded to the allocations, in order to enter data onto the electronic case report forms (eCRFs), a web-based data capture system provided by Sealed Envelope. However, no member of staff at the CTU had involvement or influence in any outcome measures.

Treatment group allocation

The trial intervention was a physiological procedure and was performed on both the donor and recipient at two time points before transplantation (24 hours before surgery and immediately before surgery). At the time point immediately before surgery, the active or sham RIPC sequences were initiated before induction of anaesthesia and were completed in advance of the initiation of surgery. There were no other interventions. The interventions consisted of different combinations of the active RIPC intervention and sham RIPC intervention as described below. The active RIPC procedure consisted of four 5-minute inflations of a blood pressure cuff on the upper arm to 40 mmHg above systolic blood pressure separated by 5-minute periods when the cuff was deflated. The sham RIPC procedure consisted of four 5-minute inflations of a blood pressure cuff on the upper arm to a pressure that would not impede blood flow (40 mmHg) separated by 5-minute periods when the cuff was deflated. All patients and donors underwent either the sham or the active RIPC procedure at both time points and so were randomised to one of the following groups:

  • control group: the control group underwent sham RIPC both 24 hours before surgery and immediately before surgery
  • early RIPC group: the early RIPC group underwent sham RIPC 24 hours before surgery and active RIPC immediately before surgery
  • late RIPC group: the late RIPC group underwent active RIPC 24 hours before surgery and sham RIPC immediately before surgery
  • dual RIPC: the dual RIPC group underwent active RIPC both 24 hours before surgery and immediately before surgery.

Postintervention treatment regimens

Patients followed the same immunosuppressive protocol, which was agreed by all participating centres. Patients received methylprednisolone and/or prednisolone according to local practice and anti-CD25 antibody [basiliximab (Simulect®, Novartis)] according to the manufacturer’s recommendations (20 mg intravenously pre transplant followed by 20 mg intravenously on day 4). Patients received mycophenolate or azathioprine according to local practice. Mycophenolate was administered as mycophenolate mofetil (CellCept®; Roche), starting at a dose of at least 1 g/day, or as mycophenolate sodium (Myfortic®; Novartis), starting at a dose of at least 720 mg/day. Azathioprine was administered at a starting dose of 2 mg/kg. Patients received tacrolimus with a target concentration according to local practice. Antimicrobial and antithrombotic prophylaxis was administered in accordance with local practice. It was anticipated that patients would receive prophylaxis against Pneumocystis carinii pneumonia, oral Candida albicans and cytomegalovirus (donor positive, recipient negative). There were no alterations to routine treatment and no changes from routine practice for anaesthesia.

Data collection and management

Data management

Randomisation and completion of the interventions were performed by the unblinded research staff. The intervention data were then faxed to the CTU, where the data were entered onto the eCRFs and then stored in locked cabinets. Following this, all subsequent follow-up visits and data collection were completed by blinded research staff. A paper case report from (CRF) was provided to assist with data collection but the source data were considered to be those on the eCRF. A series of logic and range checks were built into the system to reduce the possibility of erroneous data being entered. The system also contained a log that detailed all notable events associated with the trial (including inserts, updates and deletions) and this provided a clear and complete audit trail throughout the trial. The data management processes were conducted following the principles of good clinical practice (GCP) (see www.ich.org; accessed 12 March 2015) and the Data Protection Act 1998.48

Monitoring and site visits

The first site visit was a prerecruitment visit for training (trial interventions and procedures) and ensuring that all relevant documentation was in place before the start of recruitment. After the first patient had been recruited in each site, the senior data manager provided training on the eCRF. This training was either carried out as part of a visit to the site or performed remotely using the standard operating procedure document. A further site visit occurred at each centre after five patients had been recruited. At this visit the CRF, consent forms, source data, sample storage and the site file were monitored and the data verified. Following the visit, a report was sent to the principal investigator (PI) and any other relevant research staff involved in the trial, which included recommendations for changes if any issues were raised at the visit. Regular monitoring of the data was also performed by the trial statistician and senior data manager. Further visits could be arranged at the different sites if any problems arose or if the statistical monitoring highlighted any concerns.

Baseline assessment

All recipients had a medical history taken and a clinical examination as part of usual care. The following information was recorded at baseline: weight, height, gender, ethnicity, systolic blood pressure, creatinine, urea and albumin levels, comorbidities, date of birth, date when started dialysis and type of dialysis at time of admission. All donors filled out either a UK Transplant or a Eurotransplant organ donor form, depending on location. The following information was collected from this form at baseline: weight, height, gender, ethnicity, systolic blood pressure, creatinine, urea and albumin levels, age and glomerular filtration rate (GFR).

Follow-up

Donors were followed up immediately after RIPC and at day 1 (the day of the transplant), day 2 and day 3. Recipients underwent follow-up immediately after RIPC, perioperatively and at day 1, day 2, day 3, day 5, month 3 and month 12. The 3-month and 1-year follow-ups were carried out primarily at clinic visits. If a patient was not able to attend, information was obtained directly from the patient or from the patient’s general practitioner. In addition, recipients are being followed up annually to 5 years using clinic visits, telephone calls and renal registry data to assess eGFR. Recipients at the Leiden University Medical Centre were also followed up at 6 months post transplant for renal fibrosis assessment.

Safety assessments

This was not a trial of an investigational medicinal product; therefore, by definition, all untoward occurrences were adverse events rather than adverse reactions. Safety assessments were collected from the time of randomisation to the completion of follow-up for recipients and from the time of intervention to discharge for donors. There was a list of expected adverse events (both serious and non-serious) for which information was not collected. A detailed listing of individual adverse events was supplied as part of the reports to the DMC.

Outcome measures

Primary outcome

The primary end point of the study was GFR at 12 months after transplantation as measured by iohexol clearance.

Secondary outcomes

The secondary outcomes were:

  1. Time for serum creatinine to fall by 50%. For each recipient the time for serum creatinine to fall by 50% from the value in recovery was derived. Creatinine was measured up to 24 hours post surgery.
  2. eGFR 3 months after transplantation, derived from the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation.49
  3. eGFR 12 months after transplantation, derived from the CKD-EPI equation.
  4. Plasma interleukin 6 (IL-6), interleukin 1 beta (IL-1β), interferon gamma (IFN-γ) and tumour necrosis factor alpha (TNF-α) before and 1–3 days after surgery for donors and 1–5 days after surgery for recipients.
  5. RIPC-induced protein expression changes in renal tissue [analysis in biopsy material; protein kinase C (epsilon isoform; activated/membrane-bound fraction), manganese superoxide dismutase (MnSOD), COX-2, iNOS, HSPs 27/72, reperfusion injury salvage kinases [PI3K–protein kinase B (Akt) and mitogen/extracellular signal-regulated kinases (MEK1/2)–ERK)].
  6. Renal graft cortical tubulointerstitial fibrosis at 6 months (digital analysis of Sirius red staining in biopsy material).
  7. Incidence of delayed graft function (either the need for dialysis in the first 7 days after transplantation or serum creatinine levels increase, remain unchanged or decrease by < 10% per day in 3 consecutive days in the first week after transplantation).
  8. Incidence of acute rejection during the first 12 months after transplantation. Rejection was defined as biopsy-proven rejection, clinical acute rejection or steroid-resistant rejection. Biopsy-proven rejection was defined as any rejection grade according to the Banff criteria,50 based on histopathological appearance of a needle core biopsy of the transplant kidney. Clinical acute rejection was defined as any biopsy-proven or biochemical rejection that is treated with pulsed methylprednisolone. Steroid-resistant rejection was defined as a rejection episode that did not respond to a 3-day course of pulsed methylprednisolone and which required antithymocyte globulin (Thymoglobuline®; Sanofi-Aventis) or muromonab-CD3 therapy.
  9. Serum creatinine levels and eGFR 2–5 years after transplantation.
  10. Patient survival (at 12 months and 2–5 years) and graft survival (at 3 months, 12 months and 2–5 years). These were measured from the date of transplant and were reported for all deaths and graft failures because of rejection and from all causes.

Laboratory techniques

Glomerular filtration rate assessment at 12 months

Omnipaque 240 (5 ml; GE Healthcare Ltd) was administered intravenously and blood samples were taken at 5, 120, 180 and 240 minutes after dosing. Heparinised blood (5 ml) was centrifuged at 400 g for 10 minutes and a 2-ml sample of supernatant was aspirated and collected in a labelled Eppendorf tube. Heparinised plasma was stored at –70°C/–80°C until analysis, with GFR calculated using the iohexol clearance rate between 120 and 240 minutes. Results were corrected for body surface area.

Enzyme-linked immunosorbent assay assessments

Enzyme-linked immunosorbent assays (ELISAs) were used to measure IL-6, IL-1β, IFN-γ and TNF-α in plasma (donors and recipients) and urine (recipients) before (plasma) and 1–2 days after surgery. Urine and heparinised blood (5 ml) were centrifuged at 400 g for 10 minutes and the supernatant was aspirated and collected in labelled Eppendorf tubes.

Immunoblotting

Vascular tissue was homogenised in buffer containing peptidase inhibitors, electrophoresed on a sodium dodecyl sulphate-polyacrylamide gel and transferred to a nitrocellulose membrane. Antibodies (Calbiochem, Invitrogen and Dako) were used to probe membranes for proteins activated by early and late RIPC. Analysis is ongoing.

Immunohistochemistry

Cross-sections were obtained from formalin-fixed, paraffin-embedded renal biopsy tissues. Analysis is ongoing but sections will be prepared for immunohistochemistry by dewaxing and rehydrating using xylene and alcohol. Antibodies for use in immunohistochemistry will be determined from the results of the immunoblotting analysis.

Kidney graft fibrosis

Cortical tubulointerstitial collagen deposition was assessed by Sirius red staining of tissue slices, using digital analysis software. Analysis is in progress and will be performed at baseline and at 6 months following transplantation; graft fibrosis at 6 months will be expressed relative to baseline graft fibrosis.

Sample size

Primary analyses

There were two main analyses to reflect the factorial design of the trial, using mean GFR in the first year after transplantation. These were: (1) the two arms receiving early RIPC compared with the two arms not receiving early RIPC and (2) the two arms receiving late RIPC compared with the two arms not receiving late RIPC.

Mean GFR in the first year after transplantation was estimated to be 47.3 ml/minute/1.73 m2 in those not receiving RIPC, with a standard deviation (SD) of 13.9 (data from the Cambridge transplantation programme). The calculations in Table 1 were based on either early or late RIPC increasing GFR by 10% (or 4.73 ml/minute/1.73 m2). A trial of 80 patients in each of the four arms (160 for each comparison group, 320 in total) gives 80% power (with a 5% type 1 error) to detect this difference in GFR at 12 months for either comparison allowing for a 15% dropout rate.

TABLE 1

TABLE 1

Sample size calculations for the REPAIR trial

This sample size provided reasonable power for the primary end point while retaining useful power for secondary analyses. The trial would provide > 80% power if the difference was > 4.73 ml/minute/1.73 m2 (as might be expected if the effects of early and late RIPC combine multiplicatively, i.e. a 21% increase in GFR compared with no RIPC), the SD was lower than anticipated or the dropout rate was < 15%.

Further, some allowance was made for the possibility of a moderate interaction between early and late RIPC whereby the impact might be to lessen the anticipated effect when comparing: (1) the arms receiving late RIPC and the arms not receiving late RIPC and (2) the arms receiving early RIPC and the arms not receiving early RIPC. To allow for this possibility the aim was to recruit 100 pairs of patients in each of the four arms (400 in total).

Statistical analysis

The primary analysis was conducted on an intention-to-treat (ITT) basis with all patients and donors, when information was available, considered in the groups to which they were randomised. A per-protocol (PP) analysis was undertaken including those who received the randomised intervention as specified [i.e. excluding those pairs in which the intervention was not undertaken or in which the intervention was incomplete (whether RIPC or sham)]. All p-values are two-sided.

Primary outcome

Primary analysis

The primary analysis included the comparison of mean GFR at 1 year after transplantation (1) between the two arms receiving early RIPC and the two arms not receiving early RIPC and (2) between the two arms receiving late RIPC and the two arms not receiving late RIPC.

The model used to complete the primary analysis was a two-way analysis of covariance (ANCOVA). Specifically, this was a regression model adjusted for the donor’s baseline eGFR with indicator variables for the two treatment schedules (early and late RIPC). The distribution of GFR was examined to assess whether transformations were necessary to adhere to the assumptions of the ANCOVA model.

Secondary analysis

A large interaction between early and late RIPC on mean GFR was not expected and the trial was not powered to detect small interactions. However, any interaction was formally assessed by inclusion of an interaction term between the two treatment types in the ANCOVA model. Irrespective of the result of this interaction test a secondary analysis was conducted combining all RIPC arms together compared with the control arm receiving no RIPC, that is, to address the question of whether giving early and/or late RIPC confers a benefit compared with no RIPC. In addition, for patients in whom an iohexol measurement was not available at 12 months, eGFR at 12 months was used when available.

Secondary outcomes

Estimated glomerular filtration rate 3 months and 12 months after transplantation

The ANCOVA models described for the primary analysis were also used for the analysis of these secondary outcomes.

Time for serum creatinine to fall by 50% from the value in recovery

A Cox proportional hazards regression model for time to event (fall in serum creatinine of 50%) was used with indicator variables for the two treatment types. The interaction between early and late RIPC was assessed by inclusion of an interaction term in the Cox regression model, along with indicator variables for early and late RIPC. As with the primary outcome, a secondary analysis combined all RIPC arms compared with the arm receiving no RIPC.

Plasma interleukin 6, interleukin 1 beta, interferon gamma and tumour necrosis factor alpha before and up to 3 days after surgery for donors and 5 days after surgery for recipients

The ANCOVA model described for the primary analysis was used for the analysis of these secondary end points. Values on the second day after surgery were compared between treatment groups with adjustment for baseline values of the outcome measure. As values of these outcomes showed positive skew, data were transformed into natural log before analysis. Results for IL-1β and INF-γ used bias-corrected and accelerated bootstrapped confidence intervals (CIs) based on 2000 replications, as even after transformation it was apparent that the parametric assumptions of the linear model were violated. The use of bootstrapping means that there is no p-value for these analyses, but inference of statistical significance at p < 0.05 can be made from whether the 95% CI for treatment effects cross zero.

Incidence of delayed graft function

The proportion of participants with delayed graft function during the first 7 days was analysed using a logistic regression model with indicator variables for the two treatment types. Because of the small number of events it was not possible to test formally for an interaction between treatments. As with the primary outcome, a secondary analysis combined all RIPC arms compared with the arm receiving no RIPC.

Incidence of acute rejection during the first 12 months after transplantation

A Cox proportional hazards regression model for time to event (acute rejection) was used with indicator variables for the two treatment types. The interaction between early and late RIPC was assessed by inclusion of an interaction term in the Cox regression model, along with indicator variables for early and late RIPC. As with the primary outcome, a secondary analysis combined all RIPC arms compared with the arm receiving no RIPC. Participants were censored on the date at which the outcome occurred, if they died, if they were lost to follow-up or at 12 months after transplantation.

Three-month and 12-month graft survival

Graft survival was analysed using a time-to-event framework. The log-rank test was used to evaluate difference in graft loss between: (1) the two arms receiving late RIPC and the the two arms not receiving late RIPC and (2) the two arms receiving early RIPC and the two arms not receiving early RIPC. Participants were censored on the date at which the outcome occurred, if they died, if they were lost to follow-up or at 12 months after transplantation

Twelve-month survival

Patient survival was analysed using a time-to-event framework, as described for evaluation of the incidence of graft loss.

Missing data

An analysis was undertaken by imputing the eGFR for those patients who did not have a GFR determined by iohexol clearance measured at 12 months. All other analyses were conducted on a complete case basis, excluding any participants who had missing data for the outcome or predictors of interest in that particular model.

Subgroup analyses

The main subgroup analysis was to assess whether any effect of RIPC differed according to a patient’s underlying risk of low GFR at 12 months. This was assessed using a linear regression model to predict GFR at 12 months from baseline explanatory variables. Potential predictors of GFR at 12 months were identified from the published literature.5154 For each potential predictor, a linear regression model was used to assess the independent association between the value at baseline and GFR measured by iohexol clearance at 12 months. Those factors that showed evidence of association with GFR were entered into a multiple linear regression model, starting with factors with the strongest association. The final model included all factors that were associated with GFR (p < 0.10) together with terms for early and late RIPC. From this model, a risk score was calculated for each patient as the predicted GFR at 12 months excluding the estimated effects of early and late RIPC (i.e. the prediction for that participant if they had received sham early and late RIPC). An interaction term was fitted between risk and treatment to establish if there was a difference in benefit according to the underlying risk. Patients were categorised into quartiles of baseline risk and the data were presented as the mean (SD) GFR by treatment group in each baseline risk group.

Two other specific subgroup analyses were conducted:

  1. to compare patients for whom this was their first transplant with patients who had had a previous transplant
  2. to compare patients with a ‘000’ mismatch (i.e. those with no mismatches at any of the major HLA loci) type with patients without a ‘000’ mismatch type.

These two subgroup analyses were chosen because they represent changes to the inclusion criteria of the study after it had begun. The mean difference in the treatment groups between each subgroup and 95% CIs were calculated and subgroups compared using an interaction test.

Ethical considerations

Ethical approval for the study in the UK was given by the Joint University College London (UCL)/University College London Hospital (UCLH) Committees on the Ethics of Human Research in June 2009 (reference number 09/H0715/48). Outside the UK, local research ethics committee approvals were gained at all recruiting sites and all hospitals that were involved in trial follow-up. The trial was registered with the International Standard Randomised Controlled Trial Register (reference number ISRCTN30083294). The trial had two committees overseeing its conduct: the Trial Steering Committee (TSC) and the Project Management Group (PMG). In addition, there was an independent DMC to ensure the safety of patients in the trial and to review operational issues such as recruitment. The DMC was the only group to review interim analyses broken down by treatment group during recruitment and follow-up of patients in the trial. The DMC performed interim safety analyses annually. The interim reports contained details of patient recruitment, demographic and baseline characteristics, transplant and intervention details, primary safety end points, the primary efficacy end point and other end points identified by the DMC, including adverse and serious adverse events. At the 18-month DMC meeting, a review of the assumptions on which the sample size calculations were based was carried out as requested by the Efficacy and Mechanism Evaluation (EME) programme board.

The TSC had overall responsibility for the scientific integrity and quality of the trial. This involved ensuring that the trial was conducted to the standards set out in the guidelines for GCP and that the protocol was adhered to as far as possible and having responsibility for overall patient safety as well as considering new relevant information arising throughout the duration of the trial. The TSC also had responsibility for considering any recommendations made by the DMC. The TSC met annually throughout the trial to monitor the progress and quality of the trial, to review the recruitment rate and consider protocol amendments. The PMG was responsible for the day-to-day running of the trial, meeting fortnightly during the setting up of the trial and the early stages of recruitment and then approximately monthly for the remainder of the trial.

Patient and public involvement

There were two consumer representatives on the TSC who were actively involved in all TSC activities. The consumers were recruited from the Royal Free Hospital, London, where there is a very active support network for patients who have undergone kidney transplantation. In particular, they were key in developing both the participant information sheet and consent forms. Although the consumers did not always attend the meetings, they were always willing to contribute and comment on any REPAIR trial literature. Their contribution was extremely valuable, in particular in relation to raising the profile of the trial within the kidney transplant community. Both the Royal Free Hospital Kidney Patients Association (RFHKPA) and the National Kidney Federation published articles on the REPAIR trial in their newsletters, which helped to raise further awareness of the REPAIR trial within the kidney transplant community.

Copyright © Queen’s Printer and Controller of HMSO 2015. This work was produced by MacAllister et al. under the terms of a commissioning contract issued by the Secretary of State for Health. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.

Included under terms of UK Non-commercial Government License.

Bookshelf ID: NBK294379

Views

  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (747K)

Other titles in this collection

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...