Domselaar, Manuel Van

Abstract Background A meta-analysis by our group has shown an elevated risk of lymphoma with thiopurine therapy for IBD (Kandiel 2005 Gut). It summed referral data and population (pop) data, but the risk of lymphoma in referral centers may be skewed higher due to referral bias (Ang 2006 PMID 16531538). Since a previous meta-analysis presented in 2012 (Kotlyar 2012, DDW), there have been two large population based studies published (Pasternak 2013 Am J of Epidemiol, Abbas 2013 Gastroenterology). The latter study reported on a nationwide database of patients from the Veterans Affairs (VA) Hospitals from the United States. Aims Calculate the standard incidence rate of lymphoma in pts exposed to AZA/6-MP in population cohort studies and contrast these results to referral center studies. Methods We searched MEDLINE and the Cochrane Database for: “lymphoproliferative and thiopurines”; and “azathioprine and lymphoma”. Included citations were IBD cohort studies, evaluated cancer as an outcome, and pts. received AZA and/or 6-MP. Additionally a comprehensive search of the literature and abstracts from international meetings (2005-2013) was done. In our study additional data were extracted from the Spanish collaborative registry ENEIDA. Pooled standardized incidence ratios (SIRs) and 95% confidence intervals (CIs) were estimated. CIs assumed a Poisson distribution. To examine for heterogeneity, the deviance statistic from Poisson regression models was examined. Results There were 507 citations, and fifteen citations were included. Two studies (Korelitz and Kinlen) were obtained from Kandiel 2005, and Lewis 2001 was replaced by Armstrong 2010 as data were of the GPRD database. In referral studies (n=7), the SIR = 6.47 (95% CI: 3.77-10.36). In pop studies (n=8), the SIR = 2.27 (95% CI: 1.76-2.88). Overall the SIR was = 2.61 (95% CI: 2.08-3.23). Data from referral centers did not show significant heterogeneity (p=0.052), while pop studies did show significant heterogeneity (p=0.003). Results between referral and pop centers showed a significant difference (p<0.01). Conclusion Pts with IBD who are treated with thiopurines have approximately a 3-fold increased risk of lymphoma as compared to the general pop. The SIR of referral studies were seen to be significantly higher than those of population based studies, consistent with possible referral bias. Also, the addition of two population based studies have added 58,349 person-years of follow-up to our analysis.With a total of 84 lymphomas observed, and with total pt-years being 184,085, this study shows 2,191 pt-yrs per lymphoma in the exposed group. Meta-analysis of Standardized Incidence Ratios (SIR) Disclosures: Lewis: Pfizer: Consultancy; Centocor: Consultancy; Allos: Research Funding; Shire: Consultancy; Takeda: Research Funding; Amgen: Research Funding; Millenium: Consultancy; AstraZeneca: Consultancy; GlaxoSmithKline: Consultancy; Dark Canyon Labs: Consultancy; Roche Pharma AG: Consultancy. Beaugerie:Ferring : Speakers Bureau; Schering-Plough: Speakers Bureau; Merck : Speakers Bureau. Loftus:Pfizer: Consultancy; Genentech: Research Funding; Braintree: Research Funding; Abbott: Consultancy, Research Funding; UCB: Consultancy, Research Funding; Bristol-Myers Squibb: Consultancy, Research Funding; Shire: Membership on an entity’s Board of Directors or advisory committees; Centocor: Research Funding; Takeda: Research Funding; Amgen: Research Funding. Lichtenstein:Abbott: Research Funding; BristolMyers: Consultancy; Centocor: Consultancy, Research Funding; Elan: Research Funding; Ferring: Consultancy; MilleniumResearch: Consultancy; Procter and Gamble: Consultancy, Research Funding; Prometheus: Consultancy, Research Funding; Salix: Consultancy, Research Funding; Warner-Chilcotte: Consultancy, Research Funding; Schering-Plough: Consultancy; Shire: Consultancy, Research Funding; Wyeth: Consultancy; UCB: Consultancy.

Background and Aims . EUS-FNA is an accurate and safe technique to biopsy mediastinal lymph nodes. However, there are few data pertaining to the role of EUS-FNA to biopsy central lung masses. The aim of the study was to assess the diagnostic yield and safety of EUS-FNA of indeterminate central mediastinal lung masses. Methods . Design: Retrospective review of a prospectively maintained database; noncomparative. Setting: Tertiary referral center. From 10/2004 to 12/2010, all patients with a lung mass located within proximity to the esophagus were referred for EUS-FNA. Main Outcome Measurement: EUS-FNA diagnostic accuracy and safety. Results . 73 consecutive patients were included. EUS allowed detection in 62 (85%) patients with lack of visualization prohibiting FNA in 11 patients. Among sampled lesions, one patient (1/62 = 1.6%) had a benign lung mass (hamartoma), while the remaining 61 patients (61/62 = 98.4%) had a malignant mass (primary lung cancer: 55/61 = 90%; lung metastasis: 6/61 = 10%). The sensitivity, specificity, and accuracy of EUS-FNA were 96.7%, 100%, and 96.7%, respectively. The sensitivity was 80.8% when considering nonvisualized masses. One patient developed a pneumothorax (1/62 = 1.6%). Conclusions . EUS-FNA appears to be an accurate and safe technique for tissue diagnosis of central mediastinal lung masses.