MP69-10 GLYCATED HEMOGLOBIN (HBA1C) LEVELS ARE INDEPENDENTLY ASSOCIATED WITH UNFAVOURABLE PROSTATE CANCER AND DISEASE RECURRENCE AFTER RADICAL PROSTATECTOMY

Abstract

INTRODUCTION AND OBJECTIVES: Previous studies have shown that diabetic patients diagnosed with prostate cancer (PCa) and treated with radical prostatectomy (RP) may be at higher risk of harbouring unfavourable PCa at final pathology. We hypothesized that glycated haemoglobin (HbA1c) levels, regardless of a previous diagnosis of diabetes, may be associated with adverse PCa features and higher risk of recurrence after RP. METHODS: The study included 1,249 patients treated with RP for clinically localized PCa between 2008 and 2011 at a single tertiary care center. Pre-operative HbA1c was measured in all patients the day before surgery. Uniand multivariable logistic regression models tested the association between HbA1c and pathological stage, defined as extracapsular extension (ECE), seminal vesicle invasion (SVI) and lymph node involvement (LNI). Univariable and multivariable Cox regression analyses evaluated the role of HbA1c in predicting biochemical recurrence (BCR), defined as PSA 0.2 ng/ml and rising. Covariates consisted of total PSA at diagnosis, biopsy Gleason sum and clinical stage. HbA1c was tested as both continuous and categorical variable, according to the WHO classification (namely 6.5%) RESULTS: Mean and median pre-operative HbA1c levels were 5.73 and 5.60%, respectively (range 4.0-10.8%). Overall, HbA1c was 6.5% in 643 (51.5%), 500 (40%) and 106 (8.5%) patients, respectively. According to patient medical history, 124 (9.9%) had a diagnosis of diabetes. Mean and median PSA were 12.3 and 6.8 ng/mL, respectively. Clinical stage was T1 in 54.7%, T2 in 30.4% and T3 in 14.9% of the patients. Biopsy Gleason sum was 6 in 58.2%, 7 in 30.1% and 8-10 in 11.7% of the patients. Median follow up was 34 months (range 1-118). At multivariable logistic regression analyses, after adjusting for PSA, clinical stage and biopsy Gleason sum, HbA1c was an independent predictor of SVI (OR 1.43; p1⁄40.003) and LNI (OR 1.38, p1⁄40.01). At multivariable Cox regression analyses, HbA1c categorized according to the WHO definition was independently associated with the risk of BCR (p1⁄40.02), after accounting for confounders. Patients with HbA1c >6.5% had a 2.01 higher risk of BCR as compared to those with HbA1c <5.7%, after accounting for all confounders. CONCLUSIONS: Increased levels of HbA1c are independently positively associated with more adverse pathological features and higher BCR rates. These results should be taken into account when diabetic patients are considered for curative treatments of PCa.