No Association Between Chronic use of Ranitidine, Compared With Omeprazole or Famotidine, and Gastrointestinal Malignancies

Yeseong D. Kim; Jiasheng Wang; Fahmi Shibli; Kamrine E. Poels; Stephen J. Ganocy; Ronnie Fass


Aliment Pharmacol Ther. 2021;54(5):606-615. 

In This Article


In this population-based study using a deidentified, aggregated EHR database, we investigated whether ranitidine confers an increased risk of developing gastrointestinal malignancies compared to patients on either famotidine or omeprazole. The study demonstrated that ranitidine does not confer an increased risk of developing gastrointestinal malignancies compared to famotidine or omeprazole. Our results also indicate that the overall incidence of gastrointestinal cancers in patients taking ranitidine were low (<0.5%) over a 10-year period, as well as in the famotidine and omeprazole groups.

To our knowledge, this is the first study to evaluate the association between ranitidine and gastrointestinal malignancies in humans. By utilising an aggregated national database, we mitigated any selection bias which are often present in local or institutional studies, conferring high generalisability to our study. Furthermore, we also further controlled for the effect of confounding factors commonly encountered in cohort studies by adjusting for the three most common risk factors for each malignancy studied, as well as demographic differences amongst each drug cohort. Another important aspect of our study are the chronological boundaries we imposed on subject queries; by incorporating a 1-year lag and a ten-year maximum follow up, we maximised the association between drug initiation and cancer development, ergo correcting for any immortal time bias effects that may be present.

As a known carcinogen, no in vivo studies directly studying the effects of NDMA in humans exist, although several prior studies have investigated its oncogenic properties as a contaminant in foods, water, and medications.[24–26] Pottegård et al. performed a nationwide cohort study investigating the association between valsartan, another medication found to contain NDMA, and various cancers of the GI tract, lungs, genitourinary tract, etc.[27] The authors concluded that there was no statistically significant increased odds for developing any malignancy in patients consuming valsartan compared to patients not taking the drug, supporting the rationale of subpathogenic levels of NDMA found in valsartan. Conversely, Zeng et al. showed that urinary excretion of NDMA increased 400-fold from 110 to 47 600 ng 24 hours after oral ingestion of a single 150 mg dose of ranitidine.[28] Given the fact that the levels of NDMA found in various ranitidine products ranged from 0 to 860 ng/tablet, the significantly increased levels found in urine samples imply that NDMA levels either increase via metabolism in vivo, or increase spontaneously during storage after production.[29]

The discrepancy between known pathogenic levels of NDMA and the levels found in sample products likely account for the lower ORs calculated. Prior studies have shown that NDMA starts exhibiting carcinogenicity in murine models at levels of 10-micrograms/kg/day.[13] Given that the maximum levels of NDMA detected in any ranitidine sample was 870 ng, or 0.87 μg, a daily or even twice-a-day dosing of ranitidine would deliver a subcarcinogenic dose. Accounting for the study in which 24 hours urinary excretion of NDMA was found to be 47 600 ng, or 47.6 μg, a carcinogenic 10 μg/kg/day dose for an average 70 kg person would amount to 700 μg/day of NDMA, far above the 47.6 μg/day found in urine samples even after metabolic and temporal enhancement of NDMA levels.

In addition to subcarcinogenic doses of NDMA, another explanation for the nonelevated odds seen in ranitidine is that only 7 of 14 ranitidine products tested by the FDA yielded above-regulation levels of NDMA, with many formulations containing well below the recommended maximum daily levels.[30] Levels of NDMA detected in ranitidine from initial tests conducted by the FDA were found to be comparable to levels found in smoked or grilled meats, and subjecting ranitidine samples to pH levels simulating the normal gastric environment did not reveal increases in NDMA levels.[9]

The magnitude of reduction in the odds of patients on ranitidine developing GI malignancies compared to famotidine (0.47 and 0.48 in first and second regression models, respectively, averaged) and omeprazole (0.67 and 0.90 in first and second regression models, respectively, overall) were greater than expected. One potential explanation is that the ranitidine cohort consistently displayed lower prevalence of common risk factors for gastrointestinal malignancies (Table 2). With the exception of BO, ranitidine cohorts displayed on average 3.2% lower rates of all risk factors assessed, with some risk factors such as cirrhosis and IBD displaying up to 50% lower prevalence rates. Similar numbers were seen when ranitidine cohorts were compared to omeprazole cohorts, with an average of 2.6% lower prevalence rates of common risk factors, with the exception of tobacco and alcohol use. The relatively smaller difference in prevalence rates of cancer risk factors in ranitidine cohorts compared to omeprazole cohorts may explain the more equivocal ORs encountered in said comparison. It may be that when adjusted for every common risk factors for each cancer, the adjusted ORs for the ranitidine cohorts may increase. It also may be that the three drug cohorts had unmeasured confounding factors which may have artificially elevated cancer incidence in the famotidine and omeprazole cohorts, such as differences in diet, environmental exposures, presence of co-morbidities, and socio-economic status.

The limitations of our study primarily involve the nature of the Explorys database system. While Explorys offers broad population-level patient information that confers high generalisability, several patient data points such as mortality rates, and medication compliance were not available at the population level. In addition, over-the-counter (OTC) use of medications were unable to be accounted for, as this behavioural factor is not formalised into the Explorys system. It may be that such variations in medication usage rates could alter the degree of exposure to NDMA. Although this may theoretically limit the study's time-to-event data, the study's intrinsic design adjusted for this by including only patients with an active anti-reflux prescription at the time of initial cancer diagnosis. This trend can be seen in Figure 2, which shows a decreasing cancer incidence in cohorts more temporarily removed from the index event. This phenomenon is most likely due to that there are less patients on continuous anti-reflux therapy for prolonged time periods, rather than an absolute reduction in cancer incidence. Furthermore, while we corrected for the 3 most common risk factors for each cancer and demographic factors such as age, gender, and race to achieve a more accurate OR for each of the drugs studied, not all risk factors for each disease were accounted for. Propensity or stratification score matching was unable to be performed given Explorys returns population-level data.

In summary, our study demonstrates that the incidences of GI malignancies is low in patients who were chronically taking ranitidine, famotidine, or omeprazole, and that the odds of developing cancers while taking ranitidine are not elevated. In order to further evaluate the potential risks imparted by ranitidine, future prospective studies should be considered.