Application of GRADE in hematological malignancy: watchful waiting vs. immediate systemic treatment in indolent non-Hodgkins lymphoma

Tags: Poster
Kho M, Choong K, Friedberg J, Weeks J, Schunemann H

Background: Indolent, disseminated non-Hodgkin s lymphoma (NHL) is largely incurable with current therapies. Two fundamentally different approaches to initial management of asymptomatic patients have been advocated-- watchful waiting (WW) and immediate systemic treatment (IST). The National Comprehensive Cancer Network (NCCN) guidelines recommend WW, and, while this recommendation was based on the literature, it was not based on a systematic review (SR), and the methods for guideline development were not transparent. This may be one factor that may be responsible for varying treatment approaches across member institutions. GRADE (Grades of Recommendation, Assessment, Development and Evaluation) is a new method of grading scientific evidence and formulating recommendations [1].

Objectives: The purpose of this presentation is twofold: 1) Present results from a SR of WW versus IST in indolent NHL. 2) Describe our experience applying the GRADE approach to treatment recommendations, using the results of the SR in (1).

Methods: 1) SR Randomized controlled trials (RCTs) of WW versus ISR were identified in MEDLINE (1966 to March 2004 week 2) and EMBASE (1980 to 2004 week 11) using the Cochrane Hematological Malignancies Group highly sensitive search strategy, and the Cochrane highly sensitive search strategy for clinical trials. Reference lists from relevant studies, and contacts with content experts were also used to identify other pertinent studies. Study identification and data abstraction was completed in duplicate (KC, MEK). Data were analysed using RevMan v. 4.2. Data were pooled using odds ratios (OR) and a random effects model. 2) Using available outcome measures, the GRADE approach will be used by a group not previously familiar with the approach to develop treatment recommendations for patients and clinicians.

Results: Three RCTs were identified, representing 85 centers and four countries. WW was compared with chlorambucil, prednimustine, or ProMACE-MOPP flexitherapy followed by total nodal irradiation. No significant differences in mortality at 4 years (n=528, OR=0.85, 95%CI= 0.56 to 1.29) or at 8 years (n= 439, OR= 0.74, 95%CI= 0.5 to 1.09) were noted between patient groups. Of 261 total patients initially assigned to WW, 157 (63%) crossed over to systemic therapy. No serious methodological design flaws were noted in the largest trial, outcomes were consistent across trials, and studies demonstrated indirect comparisons. None of the studies reported outcome measurements related to quality of life or symptomatic improvement. Treatment recommendations will be presented, using the GRADE approach, reflecting input from scientists, clinicians, and patients.

Acknowledgement: We would like to thank Andy Oxman for his guidance in use of the GRADE system.

Reference [1] The GRADE Working Group. Grading evidence and recommendations. BMJ 2004 (in press).