Scope of Work

Our literature review process consisted of the following steps:

• Establish criteria for inclusion of articles in review.
• Identify sources of evidence in the scientific literature.
• Identify potential evidence with attention to controlled clinical trials using antioxidants.
• Evaluate potential evidence for methodological quality and relevance.
• Extract data from studies meeting methodological and clinical criteria.
• Synthesize the results.
• Perform further statistical analysis on selected studies.
• Perform pooled analysis where appropriate.
• Submit the results to technical experts for peer review.
• Incorporate reviewers' comments into a final report for submission to AHRQ.

Objectives

Based on a discussion with the Task Order Officer for AHRQ, the Director of NCCAM, Co-Directors of SCEPC, and project staff, we selected, as the focus for this report, the use of vitamin C, vitamin E, and coenzyme Q10 to treat and prevent cancer.

The report was guided by the following research questions:

• What kinds and numbers of study reports were available that presented research on the use of antioxidants for treating and preventing cancer?
• Were interventions used for treatment, primary or secondary prevention, or in adjunct to conventional treatment?
• Were interventions used for treatment or modification of known risk factors for cancer or pre-malignant states?
• What types of outcomes were measured for the identified condition?
• What is the methodological quality of the studies identified?
• Can statistical results from the various studies be pooled?

Literature Search Design

Identification of Literature Sources

Potential evidence for the report came from three areas: on-line library databases, the reference lists of all relevant articles, and other sources such as identified experts and the personal libraries of project staff and their associates. The reference librarian at RAND identified traditional biomedical databases as well as databases that focus on the condition of interest and alternative and complementary medicine (Table 1).

Table

Table 1. Biomedical and other databases searched.

We conducted four searches specifically on the interventions of interest. The full search strategies are displayed in Appendix B. We utilized the National Library of Medicine's controlled vocabulary thesaurus called Medical Subject Headings or “MeSH terms.” Limiting the output to human studies, we searched using the terms coenzyme Q10, vitamin E, vitamin C, and their many pharmacological synonyms (Table 2); the condition of interest (cancer); and study design or article type (randomized controlled trials, clinical controlled trials, meta-analyses, and systematic reviews). Because this report is focused on efficacy, clinical trials are preferred since they provide control groups which account for confounding factors. These searches yielded a total of 4595 titles, many of which were duplicates, because one article would appear repeatedly as each new search was added.

Table

Table 2. Additional search items for antioxidants studied.

Two reviewers (a physician and a PhD) independently evaluated deduplicated lists of 1079 titles that the on-line database searches generated as well as 258 additional titles from other sources, such as professional libraries and reference mining. The reviewers read the lists of titles and accepted articles that:

• focused on vitamin C, vitamin E, or coenzyme Q10 for treatment or prevention of cancer, or the modification of a known risk factor for cancer or improvement in a pre-malignant state;
• focused on controlled trials on humans;
• presented a meta-analysis or systematic review of the interventions and condition;
• presented historical or descriptive background information about antioxidants and their use.

Articles that either reviewer classified as meeting these criteria were accepted. Articles were rejected that both reviewers considered:

• focused on a disease state that was not the topic of interest;
• contained animal or in vitro data unless human clinical trial information or significant background information was also included.

Language was not considered a barrier to inclusion.

From this stage of the screening process, the reviewers requested a total of 1337 articles, of which we were able to obtain 1125. Selected articles were further evaluated to see if they met the inclusion criteria. Based on this evaluation, we selected 432 that went on to further screening.

Using Microsoft Access database software, we tracked requests for articles. We used Pro-Cite as a link to read the citations into the Access database as well as to manage our reference list. We also used the database to produce and store our data collection instruments. Table 3 summarizes the search strategy shown in Appendix B. The details of the screening process are discussed in the next section.

Table

Table 3. Summary of search strategy.

Evaluation of Evidence

Two physicians, each trained in the critical analysis of scientific literature, independently reviewed each article, abstracted data, and resolved disagreement by consensus. From the 432 articles accepted after the initial title screening, they accepted 36 articles for further study, based on the data collected using the screening form. These 36 articles were therefore included in the synthesis of evidence because they:

• focused on the antioxidants vitamin C, vitamin E or coenzyme Q10 and cancer;
• presented research on human subjects;
• reported the results of a clinical trial.

To be clear about our terminology: a “trial” refers to a controlled clinical trial; a “study” refers to a presentation of a specific portion of a trial's results, e.g., focused on one outcome or at a particular follow-up time; and an “article” refers to a published document. Some articles may contain more than one study, particularly if they contain results from more than one trial. Some trials, especially large ones, have many associated studies and articles.

Two articles^{11, 12}described two different trials each, so a total of 38 unique studies were referred for detailed review. Many of these studies reported on three large trials—the Alpha-Tocopherol, Beta-Carotene Cancer Prevention (ATBC) Trial, the Linxian General Population, and the Linxian Dysplasia Group Trials—which are discussed at length in the results section and displayed in the evidence table. We created a one-page data collection instrument that served as a screening form for this process. Appendix C contains a copy of this screening instrument.

Extraction of Data

Detailed information from each of the 38 studies was collected on a specialized data collection instrument (the Quality Review Form) designed for this purpose. This Quality Review Form (Appendix D) was developed in consultation with our technical experts. We included questions about the trial design; the quality of the trial; the number and characteristics of the patients; patient recruitment information; details on the intervention, such as the dose, route of administration, frequency, and duration; the types of outcome measures; and the time between intervention and outcome measurement. Two trained reviewers, working independently, extracted data in duplicate and resolved disagreements by consensus. A senior physician researcher on the project staff resolved any disagreements not resolved by consensus.

A note about equivalence of units for data extraction: dosages of vitamin E, often given as alpha-tocopherol, are reported in either milligrams or international units (IU). To interconvert these units, consider 1 milligram of alpha-tocopherol approximately equal to 1.5 IU of vitamin E.

To evaluate the quality of the trials, we collected information on the study design, appropriateness of randomization, blinding, description of withdrawals and dropouts and concealment of allocation.¹³ A score for quality was calculated for each trial using a system developed by Jadad.¹⁴ The Jadad score rates studies on a 0 to 5 scale.¹⁴ A score is based on the answer to three questions: Was the study described as randomized? Was the study described as double-blind? Was there a description of withdrawals and dropouts? One point is awarded for each “yes” answer, and no points are given for a “no” answer. An additional point is given if the randomization method described was appropriate. A point is deducted if the method is described but is not appropriate. A point is awarded if the method of blinding is appropriate and described, and one point is deducted if the blinding method is described, but inappropriate. Empirical evidence has shown that studies scoring 2 or less report exaggerated results compared with studies scoring 3 or more.¹⁵ Thus, studies with a Jadad score of 3 or more are referred to as “high” quality, and studies scoring 2 or less are referred to as “poor” quality.

The flow of articles from the point at which they entered our database, through the article ordering, screening, quality review, and statistical analysis stages is displayed in Figure 1. All articles that went on for abstraction were examined for inclusion in the data synthesis.

Figure

Figure 1. Literature Flow.

Selection of Studies for Meta-Analysis

Prior to the analysis, we entered all data on outcomes and treatments into the statistical program SAS.¹⁶ We analyzed this dataset to identify the clinically relevant outcomes that were reported most commonly and would therefore be appropriate for data synthesis. These outcomes were death, development of new tumors and progression of existing tumors, and development of adenomatous colonic polyps. The grouping of treatments and the appropriate comparison group, e.g., any combination of vitamins with vitamin E versus placebo, was based on clinical knowledge and was decided a priori. In addition to these three clinical outcomes, our review also examines intermediate outcomes. These are outcomes that are considered as precursors to such clinical outcomes as new tumor development and death, although they vary in the degree to which they are good predictors of those outcomes.

We defined the outcomes of interest as follows. “Death” from cancer was used as an outcome for any trial which gave survival results. “New tumors” includes the development of new tumors in a subject with no prior history of tumor as well as recurrence of tumor in a subject with a history of cancer or precancerous lesion. Adenomatous colonic polyps, or simply “colonic polyps” refers to new adenomatous polyps of the colon.

Several trials had multiple associated studies, so our first task was to discern what each study contained in terms of unique data for that trial. For example, two studies of the same trial might present data on deaths due to two different types of cancer, and therefore each contributed unique data to the analysis. Other studies contained duplicate data. Some studies did not contain sufficient data for a statistical analysis. The two primary causes of insufficient data were that only survival curves were presented graphically for death or new tumors rather than the number of outcomes at a specific follow-up time, or an outcome, e.g., number of deaths, was presented for all patients combined rather than separately by treatment group. After determining which studies could contribute to the analysis, we extracted data into the spreadsheet program Microsoft Excel¹⁷ and performed statistical and meta-analytic methods in the statistical package Stata.¹⁸

All three outcomes were dichotomous and we used a risk ratio to summarize each individual study, so we discuss this statistical approach jointly for all outcomes. For death and new tumors, the studies were too heterogeneous to pool meta-analytically. We did pool the colonic polyp risk ratios, and we discuss this meta-analytic approach below.