Searches were performed without language restriction in the following databases: PreMEDLINE & MEDLINE (1966 through January 2002, Week 2), CANCERLIT (1983 through October 2001), EMBASE (1980 to October 2001), Current Contents (1993 to October 2001), the Cochrane Library (Issue 4, 2001), Best Evidence (1991 to October 2001) and an unpublished theses database (UMI ProQuest®[40]). The following terms were used for MEDLINE and CANCERLIT: "lymphoma, non-Hodgkin" (Medical subject heading [MeSH], text word), "lymphoma" (text word) combined with "aged" (text word) or "older" (text word) combined with "chemo:" (text word). These terms were then combined with search terms for the following study designs: practice guidelines, systematic reviews, meta-analyses, and randomized controlled trials. The detailed search strategy has been described in Appendix I of the original guideline document.
Bibliographies of major textbooks, review articles, and primary studies were hand searched. Conference proceedings of the American Society of Hematology (1993-2001), American Society of Clinical Oncology (1993-2001), International (Lugano) Conference on Malignant Lymphoma (1996, 1999), and the European Cancer Conference (ECCO 1995, 1997, 1998, 2001) were searched. A manual review of the table of contents was performed for the following journals from 1993 to 1998: American Journal of Hematology, Annals of Oncology, Blood, British Journal of Hematology, Cancer, European Journal of Cancer, European Journal of Hematology, Journal of Clinical Oncology, and the New England Journal of Medicine. The Physician's Data Query (PDQ, National Cancer Institute, USA) clinical trials database on the Internet (http://www.cancer.gov/search/clinical_trials/) was searched for trials in progress using the terms "non-Hodgkin's lymphoma, adult" and "chemotherapy."
A separate search for studies assessing risk factors predictive of fever and neutropenia in elderly lymphoma patients was undertaken to assist the Hematology Disease Site Group (DSG) in evaluating the role of primary prophylaxis with growth factors. The following terms were searched in MEDLINE (1966 through September 2001) and CANCERLIT (1984 through September 2001): "lymphoma, non-Hodgkin" (MeSH, text word), "lymphoma" (text word) combined with "neutropenia" (text word) and "risk factor" (text word). Abstract publications were not included. Specific parameters to assess the quality of these studies were not applied.
Inclusion Criteria
Articles were selected for inclusion in this systematic review of the evidence if they were fully published reports or published abstracts of randomized controlled trials (RCTs) involving newly diagnosed patients with aggressive histology [intermediate- and high-grade, Working Formulation] lymphoma who were 60 years of age and older. The age threshold of 60 years was chosen in order to remain consistent with the findings of the International Prognostic Index (IPI).
- To assess the role of chemotherapy, RCTs must compare at least two chemotherapy regimens.
- To assess the role of colony-stimulating factors, RCTs comparing the use of granulocyte colony-stimulating factor (G-CSF) or granulocyte-macrophage colony-stimulating factor (GM-CSF) with a control group were sought. In the initial phase of this guideline, nonrandomized studies utilizing colony-stimulating factors that included at least ten patients (chosen arbitrarily) were also eligible. These trials were subsequently made ineligible in February 2001 when data from three randomized trials became available.
- Randomized studies assessing the use of monoclonal antibodies (e.g., rituximab) were eligible.
- Subgroup analyses based on age or histology were eligible.
The outcome measures of interest included at least one of the following: overall survival (OS), disease-free (DFS) or failure-free survival (FFS), time-to-treatment failure (TTF), relapse-free survival (RFS), response rate, toxicity, or quality of life measures.
Exclusion Criteria
Studies were excluded if:
- Patients included had indolent lymphoma, refractory or relapsed lymphoma, human immunodeficiency virus (HIV) related lymphoma, Hodgkin's disease, multiple myeloma, or other hematological malignancies;
- Transplantation, maintenance chemotherapy, or interferon were used as interventions; or
- Radiation therapy was used unevenly in experimental and control groups.
Studies assessing the role of chemotherapy were excluded if they incorporated growth factors as part of the primary therapy in all randomized groups. Also, letters and editorials were not considered.
Article Selection
Citations were blinded for authors, journal name, institution, and results by one author. An assessment was made by two independent observers who scored each blinded citation as: "yes" (inclusion criteria were met, no exclusion criteria were met); "no" (one or more exclusion criteria were met); or "maybe" (unclear from citation if article meets any criteria). The full-length article was retrieved if the citation scored "yes" or "maybe" by at least one observer. Inclusion and exclusion criteria were applied again to the full article if necessary. Interobserver kappa coefficients (quadratic weighted) were calculated using PCAgree© for the MEDLINE, CANCERLIT, and EMBASE databases, and an intraobserver coefficient was calculated from a random sample (random numbers table) of twenty MEDLINE citations for the citations assessing the role of chemotherapy. Acceptable kappa coefficients were 0.60 or greater. The citation lists for subsequent search updates were reviewed by one author using the same inclusion/exclusion criteria outlined previously.