Results: Over 30 years of follow-up, we documented 163 incident cases of HCC over 3,891,069 person years in both cohorts. Compared with non-diabetics, diabetics had a multivariable HR for HCC of 3.52 (95%CI 2.44-5.08, p<0.0001) after adjustment for age, sex, BMI, aspirin use, smoking status, and alcohol intake. The association of DM and PD98059 nmr HCC appeared similar in women and men. Compared to those without DM, the multivariable HRs for HCC were 3.09 (95%CI 1.60-5.98)

for those with diabetes for 1-4 years; 3.85 (95%CI 2.04-7.29) for 5-8 years; 3.67 (95%CI 1.81-7.42) for 9-12 years, and 3.57 (95%CI 2.07-6.15) for more than 12 years (P linear trend among diabetics=0.65). Conclusions: In this large US prospective cohort study, DM was associated with an increased risk of HCC over 30 years of follow-up. The association was independent of duration of diabetes and did not appear to be mediated by BMI. Disclosures:

Raymond T. Chung – Consulting: Abbvie; Grant/Research Support: Gilead, Mass Biologics Andrew T. Chan – Consulting: Pfizer Inc, Bayer Healthcare, Pozen Inc, Millennium Pharmaceuticals The following people have nothing to disclose: Lindsay Y. King, Hamed Khalili, Edward S. Huang Purpose: AASLD guidelines recommend biannual buy ABT-263 HCC screening for cirrhotic patients. Previous data from government sponsored health plans suggests adherence to these guidelines is suboptimal. The objective of this study was to evaluate HCC surveillance rates in a nationally Edoxaban representative cohort of commercially insured cirrhotic patients. Methods: We used the Truven Health Analytics databases from 2006-2010, using 1/1/2006 as the anchor date for evaluating outcomes given the publication of AASLD screening guidelines in 11/2005. Surveillance patterns were characterized using categorical and continuous outcomes. The categorical outcome was: 1) complete (one ultrasound every 6-month interval after 1/1/2006); 2) incomplete (≥1ultrasound); or 3) none. The continuous measure was defined

as the proportion of time “up-to-date” with surveillance (PUTDS), with the six months immediately following each ultrasound categorized as “up-to-date.” Results: During a median follow-up of 22.9 (IQR: 16.3-33.9) months among 8,916 cirrhotic patients, only 785 (8.8%) patients had complete surveillance, 4,943 (55.4%) incomplete, and 3,188 (35.8%) none. During follow-up, the mean PUDTS was 0.34 (SD: 0.29), and the median was 0.31 (IQR: 0.03-0.52). Multinomial logistic regression models identified two significant access to care factors, insurance type (p=0.03) and provider subtype (p<0.001). Patients with consumer-directed, high-deductible, capitated point-of-service, or equivalent premium income health insurance were significantly more likely to have incomplete or no surveillance (p=0.