Does pre-existing diabetes affect prostate cancer prognosis? A systematic review.

To summarize the influence of pre-existing diabetes on mortality and morbidity in men with prostate cancer. We searched MEDLINE and EMBASE from inception through 1 October 2008. Search terms were related to diabetes, cancer and prognosis. Studies were included if they reported an original data analysis of prostate cancer prognosis, compared outcomes between men with and without diabetes and were in English. Titles, abstracts and articles were reviewed independently by two authors. Conflicts were settled by consensus or third review. We abstracted data on study design, analytic methods, outcomes and quality. We summarized mortality and morbidity outcomes qualitatively and conducted a preliminary meta-analysis to quantify the risk of long-term (>3 months), overall mortality. In total, 11 articles were included in the review. Overall, one of four studies found increased prostate cancer mortality, one of two studies found increased nonprostate cancer mortality and one study found increased 30-day mortality. Data from four studies could be included in a preliminary meta-analysis for long-term, overall mortality and produced a pooled hazard ratio of 1.57 (95% CI: 1.12-2.20). Diabetes was also associated with receiving radiation therapy, complication rates, recurrence and treatment failure. Our analysis suggests that pre-existing diabetes affects the treatment and outcomes of men with prostate cancer.

Evaluation of a measurement strategy for Parkinson's disease: assessing patient health-related quality of life.

This study evaluated the feasibility and psychometric properties of self-completed and telephone interview versions of a patient health-related quality-of-life (HQL) questionnaire for Parkinson's disease that included the SF-36 Health Survey (SF-36), the Parkinson's Disease Questionnaire (PDQ-39), and the Medical Outcomes Study Sexual Function Survey. Parkinson's disease patients (n = 150) completed the questionnaire twice: once at the study site and once over the telephone in a randomized order. Ninety-four percent of enrolled patients completed the first HQL assessment and 88% completed both assessments. Cronbach's alpha exceeded 0.70 for all scales except for the self-completed PDQ-39 Social Support subscale (0.57) and the telephone interview PDQ-39 Social Support (0.60) and Cognitions (0.67) subscales and the SF-36 General Health (0.60) and Social Function (0.61) subscales. There were no statistically significant differences in mean HQL scale scores across the two modes of administration. Mean scores for 3 of the PDQ-39 subscales and the Summary Index were significantly poorer (p < 0.05) for patients at later clinical stages. Similarly, patients with dyskinesias reported significantly poorer scores for 4 of the PDQ-39 subscales and the Summary Index and patients with self-reported comorbidities reported poorer SF-36 Physical Function and General Health subscale scores than patients without dyskinesias and comorbidities, respectively. This study suggests that the self-completed and telephone interview versions of the patient HQL questionnaire are feasible and valid for future clinical trial applications.