Module 3 - Strategic case studies in practice

Keszei et al. (2012) found an association of oesophageal cancer with dietary intake of NDMA (HR for NDMA 1.15; 95% CI: 1.05 - 1.25) in a Dutch cohort of 120,852 participants. The participants were aged 55–69, recruited in 1986 and followed for 16.3 years. The association of pancreatic cancer after dietary intake of NDEA and NDMA (OR for NDEA 2.28, 95% CI: 1.71 - 3.04, and OR for NDMA 1.93, 95% CI: 1.42 - 2.61) has been studied by Zheng et al. (2018) in a US-based case control study comprising 957 cases and 938 controls from genetically unrelated family members which were recruited between 2002 and 2009. Catsburg et al. (2014) investigated the association of bladder cancer after dietary intake of NDMA (OR for N- nitrosamines 1.03, 95% CI: 0.78 - 1.36) in a case control study with 3,246 participants from the US which were recruited between 1987 and 1996. Two studies focused on colorectal cancer. A cohort study by Knekt et al. (1999) based on 9,985 Finnish participants which were followed for 24 years found a RR after dietary intake of NDMA of 2.12 (95% CI: 1.04 - 4.33). A case control study by Zhu et al. (2014) with 4,241 Canadian participants aged between 20 and 74 years found a RR after dietary intake of NDMA of 1.42 (95% CI: 1.03 - 1.96). Rectal cancer has been studied by Loh et al. (2011) who found a HR of 1.46 (95% CI: 1.16, 1.84) after dietary intake of NDMA in a UK based cohort of 23,363 participants who were recruited in 1993–1997 and followed for 11.4 years. A comparable RR of 1.61 (95% CI: 1.11 - 2.35) was found for rectal cancer after dietary intake of NDMA by Zhu et al. (2014) who employed a case control design with 4,241 Canadian participants. The main body of research based on monitoring of dietary intake of N- nitrosamines is focused on gastric cancer. A meta-analysis by Song, Wu, and Guan (2015) comprises all recent studies concerning stomach cancer after dietary intake of NDMA and found an elevated relative risk of 1.34 (95% CI: 1.02 - 1.76). Included were cohort studies (Knekt et al. (1999), Jakszyn et al. (2006), Keszei et al. (2012)), of which only Larsson, Bergkvist, and Wolk (2006) showed a significant association. The included case-control studies [De Stefani et al. (1998); La Vecchia et al. (1995); Palli, Russo, and Decarli (2001); Pobel et al. (1995)], however, showed very heterogeneous results with exceptionally high effect sizes for the moderately sized studies of Pobel et al. (1995) and De Stefani et al. (1998). A subset of studies with at least 2000 participants, omitting De Stefani et al. 1998; Palli, Russo et al 2001; Pobel et al. (1995) showed no significant heterogeneity and no significant relative risk of gastric cancer with an RR of 1.12 (95% CI: 0.97 - 1.39) after dietary intake of NDMA. The heterogeneity between included studies in the meta-analysis by Song, Wu, and Guan (2015) emphasizes the challenge of attributing cancer risk to single ingredients in nutritional epidemiology [Schoenfeld and Ioannidis (2013), Egger, et al (1998)]. Even though those studies adjusted at least for the main confounders such as age, sex and total energy intake, residual confounding cannot be ruled out. For example, the intake of other carcinogens such as polyaromatic hydrocarbons has not been taken into account. Additionally, patient specific factors such as the intake of vitamins and ethanol, or the CYP2E1 metabolism might further modify effects [Toshihiko et al. (2005); Peto et al. (1991b); Stephens et al. (1994); Trafalis et al. (2010)]. Although, these studies suggest an association of dietary intake of NDMA with some types of cancer, definite conclusion cannot be drawn at this stage and further confirmation is required, as reported associations, effect sizes and especially dose-response relations should be interpreted with great caution. Further research based on large sample sizes and better control for confounding is needed.

EMA/369136/2020

Page 47/90