Ranitidine and Risk of Bladder and Kidney Cancer: A Population-Based Cohort Study

The H2-receptor antagonist (H2-blocker) ranitidine is an acid-suppressing drug marketed in 1981 which has been widely used to treat gastroesophageal reflux disease and peptic ulcers. In 2019, the FDA became aware of independent laboratory testing that found N-nitrosodimethylamine, a probable human carcinogen, in ranitidine (1) and consequently, the FDA and in 2020, the European Medicines Agency requested withdrawal of all prescription and over-the-counter ranitidine drugs from the market (2). Clinical studies have confirmed that ranitidine may be a significant source of N-nitrosodimethylamine (3, 4). Yet, whether ranitidine use in routine clinical care is associated with an increased risk of cancer remains unclear (5).

With approximately 549,000 new bladder cancer cases and 200,000 bladder cancer deaths each year worldwide, bladder cancer is a serious disease occurring three times more often in men than in women (6). Besides age, the main risk factor for bladder cancer is tobacco smoking, estimated to account for approximately half of all urothelial cell carcinomas (7). Other risk factors include environmental or occupational exposure to carcinogens including N-nitrosamines (8, 9). Because of urinary excretion of N-nitrosodimethylamine ranitidine could potentially increase the risk of bladder or kidney cancer.

A cohort study from South Korea examined the risk of various cancers in 40,488 individuals initiating ranitidine between 2009 and 2011 compared with 10,122 individuals initiating the H2-blocker, famotidine, and found a HR of bladder cancer of 1.41 [95% confidence interval (CI): 0.88–2.24], based on 118 ranitidine-exposed bladder cancers during follow-up (10). Similarly, another study used data from the UK Biobank and compared cancer occurrence in 8,844 persons who reported ranitidine use with 26,959 who reported use of the proton pump inhibitor (PPI), omeprazole (11). The adjusted HR for bladder cancer, which was included as an explorative outcome, was 1.30 (95% CI: 0.69–2.46) based on only 13 ranitidine-exposed bladder cancers. Thus, these two existing studies suggested a 30%–40% increased hazard of bladder cancer but had low precision. Although previous studies in rodents have suggested that N-nitrosodimethylamine may also increase the risk of kidney cancer (12), neither the study from South Korea nor the UK Biobank study found increased hazards of kidney cancer with adjusted HRs of 0.78 (95% CI: 0.49–1.25; ref. 10) and 0.39 (95% CI: 0.19–0.82; ref. 11).

In light of the public health and clinical importance of this issue, we conducted a large nationwide population-based cohort study to examine whether use of ranitidine was followed by an increased risk of bladder or kidney cancer possibly through urinary excretion of N-nitrosodimethylamine. For comparison, we included users of other H2-blockers and users of PPIs as a second reference cohort. Medication exposure was based on redeemed prescriptions.

We conducted the study in Denmark which in 2008 had a population of approximately 5.6 million people of which 4.4 million were 18 years or older. All Danish residents are provided free tax-supported access to health care (13). Since 1968, the Civil Registration System has assigned a unique personal identifier, the Civil Registration Number, to all residents at birth or upon immigration (14). This identifier allows unambiguous data linkage at the individual level. The Civil Registration System also tracks changes in vital status and migration for the entire population.

The Danish National Prescription Registry contains complete data on prescription medications redeemed from community pharmacies and hospital-based outpatient pharmacies in Denmark since 1995 (15). From this registry, we identified three study cohorts including all first-time users of ranitidine as the exposed cohort and first-time users of other H2-blockers (i.e., cimetidine and famotidine) or PPIs as two separate comparison cohorts (see Supplementary Table S1 for codes). First-time users were defined as persons who redeemed at least two prescriptions of the same study drug category in 1996–2008 and had no prescriptions of this drug in 1995.Index date was date of the second prescription and we only included persons who were 18 years or older at index date. We excluded persons who 2 years before index date or between their first prescription and index date redeemed a prescription of one of the other study drug categories. We furthermore excluded all persons who redeemed a prescription for the combination drug ranitidine bismuth citrate or the H2-blocker nizatidine in the same period. Nizatidine oral solution has recently been suspected of high N-nitrosodimethylamine contamination as well (1). In Denmark, H2-blockers and PPIs have also been sold over-the counter but until 2001, at least 70% of both ranitidine and cimetidine was sold on prescription and during 2002–2010 more than 50% was sold on prescription (16, 17). Less than 3% of famotidine and PPIs were sold over the counter during the study period (17).

The outcomes of interest were bladder cancer and kidney cancer (see Supplementary Table S1 for codes) recorded in the Danish Cancer Registry (18). The Danish Cancer Registry has recorded incident cases of cancer on a nationwide basis since 1943 and has valid and almost complete ascertainment of cancer cases. Cancer diagnoses in the Cancer Registry are recorded according to the International Classification of Diseases, version 10 (ICD-10). We included both non-invasive and invasive bladder tumors. We excluded members of the cohorts if they had a cancer diagnosis recorded before start of follow-up, defined as any cancer except non-melanoma skin cancer.

The Danish National Patient Registry has recorded all inpatient admissions to all Danish hospitals since 1977 and hospital outpatient clinic and emergency room visits since 1995 (19). From this registry, we retrieved all previous diagnoses of gastrointestinal disease indicating H2-blocker treatment: esophagitis or gastroesophageal reflux disease, gastric or duodenal ulcers, and other upper gastrointestinal diseases recorded within 10 years before index date (see Supplementary Table S1 for codes). We also retrieved information regarding previous urogenital disease and other diagnoses potentially associated with bladder cancer risk, including diabetes, chronic obstructive pulmonary disease, ischemic heart disease, hypertension, and alcohol-related disease recorded within 10 years before index date.

From the prescription registry, we retrieved information on redeemed prescriptions within 90 days before index date of drugs potentially associated with risk of bladder cancer or kidney cancer, that is, non-steroidal anti-inflammatory drugs, and thiazolidinedione. We additionally included use of drugs associated with bleeding risk, as such drugs could alter the threshold for diagnosis of a bladder cancer, including low-dose aspirin and other platelet inhibitors, and oral anticoagulants redeemed within 90 days before index date (see Supplementary Table S1 for codes). To identify patients with hypertension, we also retrieved information on use of antihypertensive drugs within 10 years before index date.

We characterized initiators of ranitidine, other H2-blockers, and PPIs with respect to baseline variables measured at index date. We used logistic regression to compute propensity scores for exposure to ranitidine compared with other H2-blockers and PPI separately, and used these scores to compute stabilized inverse probability of treatment (sIPT) weights (20). The propensity score models included age, time between first and second prescription, and calendar time modeled as spline curves as well as marital status, underlying gastrointestinal diseases, urogenital diseases, other comorbidities, and use of the included medicine. We assessed the covariate balance after weighting using standardized mean differences.

We started follow-up at index date. We followed the patients until date of a bladder cancer or a kidney cancer diagnosis, any other cancer (except non-melanoma skin cancer), death, emigration, a prescription of nizatidine or ranitidine bismuth citrate, or end of study period (December 31, 2018), whichever occurred first. We furthermore censored persons in the comparison cohorts if they redeemed a prescription for ranitidine during follow-up. We plotted crude and sIPT-weighted cumulative incidence curves with death as a competing event. From the cumulative incidence curves, we estimated the 10-year risk and calculated crude and sIPT-weighted 10-year risk differences. We calculated the number of events and person years, and compared cancer occurrence between the cohorts via HRs estimated by Cox regression using both crude and sIPT-weighted observations. The 95% CIs were estimated by bootstrapping. We checked and accepted the proportionality assumptions using log–log plots.

In sensitivity analyses, we stratified analyses by sex and, as a subgroup-analysis, we restricted to patients ages 50 years or older. We also stratified the analysis by calendar period of inclusion (1996–1999, 2000–2003, 2004–2008). We furthermore changed start of follow-up to 1, 5, and 10 years after index date.

To examine whether the risk of cancer varied by intensity of ranitidine treatment we required, in separate analyses, patients to fill five and 10 prescriptions of a study drug, rather than two, and applied inclusion/exclusion criteria with the index date set to the fifth and tenth prescription. To focus on those with high exposure and long follow-up, we additionally, in a subgroup analysis, required at least 10 prescriptions and at least 10 years of follow-up after the 10th prescription. Finally, in the main population, we examined risk of invasive bladder cancer only while censoring on non-invasive bladder cancer.

The study was reported to the Danish Data Protection Agency through registration at Aarhus University (Aarhus, Denmark; record number KEA‐2017‐36/812).

We identified 31,393 initiators of ranitidine, 65,384 initiators of other H2-blockers and 509,849 initiators of PPIs (please see Supplementary Fig. S1 for a flowchart). The age and sex distribution was fairly similar in the three exposure groups (Table 1). However, calendar year of treatment initiation differed with 40.8% of ranitidine users initiating treatment in 1996–1999 compared with 52.5% of users of other H2-blockers, and 17.8% of PPI initiators (Table 1).

During a median follow-up of 14 years [interquartile range (IQR): 9–18], 15 years (IQR: 7–19), and 11 years (IQR: 5–16), respectively the proportion of patients who died without a bladder or kidney cancer was 22.7% among ranitidine initiators, 20.3% among initiators of other H2-blockers and 26.9% among initiators of PPIs.

During follow-up, we identified 270 bladder tumors in ranitidine initiators. The 10-year crude risk of bladder cancer was 0.50% (95% CI: 0.43–0.59) in ranitidine users, 0.41% (95% CI: 0.36–0.47) in users of other H2-blockers, and 0.55% (95% CI: 0.52–0.57) in users of PPIs. The crude HR of bladder cancer in ranitidine users was 1.33 (95% CI: 1.15–1.55) compared with users of other H2-blockers and 0.94 (95% CI: 0.83–1.07) compared with PPI users (Supplementary Fig. S2).

In the sIPT-weighted analyses, the HR of bladder cancer in ranitidine users was 1.11 (95% CI: 0.95–1.29) compared with users of other H2-blockers and 1.24 (95% CI: 1.04–1.48) compared with PPI users (Fig. 1). The 10-year sIPT-weighted risk difference was—0.01% (−0.06 to 0.04) between users of ranitidine and users of other H2-blockers and 0.15% (95% CI: 0.01–0.32) compared with users of PPIs (Fig. 1). The slightly increased risk when compared with PPIs was most pronounced in men with a 10-year sIPT-weighted risk difference of 0.20% (95% CI: −0.06 to 0.22) while it was 0.06% (95% CI: −0.11 to 0.22) among women. Among men, the sIPT-weighted HR of bladder cancer in ranitidine users was 1.15 (95% CI: 0.96–1.37) compared with users of other H2-blockers and 1.15 (95% CI: 0.95–1.38) compared with users of PPIs. Among women, the HR was 1.03 (95% CI: 0.79–1.34) and 1.32 (95% CI: 0.97–1.79), respectively (Fig. 1).

When we stratified the analyses by calendar period of treatment initiation, we observed no increased risk for those who initiated treatment in the period 1996–1999 while the risk was slightly increased for those who initiated later (Fig. 1). Compared with other H2-blockers the sIPT-weighted 10-year risk difference was 0.15% (0.00–0.31) for those who initiated treatment in 2000–2003 and 0.06% (−0.18 to 0.31) for those who initiated in 2004–2008. Compared with PPIs, the respective risk differences were 0.20% (−0.17 to 0.57) and 0.19% (−0.03 to 0.42).

Changing start of follow-up to 1, 5, and 10 years after the second prescription did not substantially change our estimates (Fig. 1).

When we restricted the analysis to patients who redeemed at least five or 10 prescriptions, the sIPT-weighted HRs compared with other H2-blockers were 0.99 (95% CI: 0.76–1.28) when follow-up started at the fifth prescription and 0.73 (95% CI: 0.53–1.01) starting at the 10th. Compared with PPI users, these HRs were 1.06 (95% CI: 0.77–1.44) and 0.84 (95% CI: 0.56–1.27; Fig. 1). Starting time of follow-up 10 years after the 10th prescription did not suggest any associations. The sIPT-weighted HR after more than 10 years of follow-up was 0.75 (95% CI: 0.44–1.28) in ranitidine users compared with users of other H2-blockers and 0.73 (95% CI: 0.46–1.15) compared with users of PPIs.

Restricting the outcome to be invasive bladder cancers only indicated no association with use of ranitidine. The sIPT-weighted HRs were 1.05 (95% CI: 0.84–1.31) when compared with other H2-blockers and 0.98 (95% CI: 0.79–1.22) compared with PPIs (Fig. 1).

During follow-up, we identified 107 kidney cancers in ranitidine initiators. The 10-year crude risk of kidney cancer was 0.19% (95% CI: 0.15–0.25) in ranitidine users, 0.21% (95% CI: 0.18–0.25) in users of other H2-blockers, and 0.24% (95% CI: 0.23–0.26) in users of PPIs. The crude HR of kidney cancer in ranitidine users was 1.01 (95% CI: 0.81–1.26) when compared with users of other H2-blockers and 0.84 (95% CI: 0.69–1.04) when compared with PPI users (Supplementary Fig. S3). In the sIPT-weighted analyses, the cumulative incidence of kidney cancer was similar between exposure groups (Fig. 2). The sIPT-weighted HR of bladder cancer in ranitidine users was 0.89 (95% CI: 0.72–1.10) when compared with users of other H2-blockers and 0.87 (95% CI: 0.67–1.13) when compared with PPI users. Our subgroup analyses were consistent with the main findings (Fig. 2).

In this population-based study reflecting routine clinical practice, we did not observe any consistent or substantial increase in risks of bladder cancer and we consistently across subanalyses observed no increased risk of kidney cancer in ranitidine users.

The cohort study from South Korea, matched on age, sex, diabetes, and cumulative treatment duration, (10) and the UK Biobank cohort study, with multivariable adjustment, (11) both suggested a 30%–40% increased hazard of bladder cancer which was not interpreted as a positive association by the authors due to lack of statistical significance. We observed a slightly lower increase and moreover, we found no dose–response association when restricting to persons who redeemed at least five and persons who redeemed at least 10 prescriptions. Furthermore, we assessed absolute risk differences as well with death as a competing risk with findings that reflected the relatively low absolute risk of bladder cancer and little difference between users of ranitidine and other H2-blockers. Also, when restricting to invasive bladder cancers, we observed no association. Our finding of no increased risk of kidney cancer following ranitidine exposure is in line with the two previous studies (10, 11).

Our findings are also in line with those from an early Danish study that examined the association between exposure to N-nitrosodimethylamine through contaminated valsartan products (21). This prior study found no evidence of an increased risk of bladder cancer, with a HR of 0.66 (95% CI: 0.15–2.89), or kidney cancer [HR, 1.00 (95% CI: 0.22–4.65] comparing those exposed to potentially N-nitrosodimethylamine contaminated valsartan with those redeeming prescriptions for non-contaminated valsartan albeit with limited statistical precision.

Use of nationwide medical registries allowed us to conduct a large cohort study with long and virtually complete follow-up. It additionally enabled us to use an active comparator design including users of both other H2-blockers and PPIs as reference groups to minimize confounding by indication as these drugs are used for similar indications as ranitidine.

We identified bladder cancer cases using the Danish Cancer Registry which is known to have a high level of completeness (22). Still, our study has some weaknesses that should be considered. We identified ranitidine exposure based on redeemed prescriptions and we do not know whether the patients actually took these drugs. Yet, restricting to patients who redeemed at least five or at least 10 prescriptions did not suggest any increased risk following ranitidine exposure. Another concern is the fact that ranitidine was sold over the counter and users of other H2-blockers or PPIs may have used ranitidine without our knowledge. We identified exposure in a time period with more than 50% of the ranitidine doses sold on prescription (17). Over the counter medication is more expensive than prescribed medication and because members of our comparison cohorts redeemed prescriptions for either other H2-blockers or PPIs, we would expect concomitant use of over-the-counter ranitidine to be low. The proportion of over the counter sale of ranitidine increased during our study period and we would therefore expect an increasing bias towards no association over time. Yet, we did not see an attenuating risk over time.

We used sIPT weighting to construct pseudo-populations with balanced distributions of the covariates included in the propensity scores thereby adjusting for potential confounding (20) but we lacked information on smoking, which is an important risk factor for bladder cancer, and other lifestyles factors such as body mass index. However, smoking and other lifestyle factors are unlikely to be strongly related to choice of type of H2-blocker; thus, the active comparator design likely averted potential confounding by these factors. In addition, we took chronic obstructive pulmonary disease and ischemic heart disease into account in our analyses, and both these conditions are associated with smoking.

In conclusion, this large nationwide study with complete ascertainment of bladder and kidney cancers provided little evidence of any substantially increased risk of bladder or kidney cancer in ranitidine exposed individuals. Thus, our findings suggest that the N-nitrosodimethylamine contamination found in ranitidine did not substantially impact bladder or kidney cancer occurrence in our population.

No disclosures were reported.

M. Nørgaard: Conceptualization, methodology, writing–original draft, project administration. I.T. Andersen: Data curation, formal analysis, methodology, writing–review and editing. U. Heide-Jørgensen: Formal analysis, supervision, methodology, writing–review and editing. R. Erichsen: Writing–review and editing. J.R. Rees: Writing–review and editing. M.R. Karagas: Conceptualization, writing–review and editing. H.T. Sørensen: Conceptualization, resources, supervision, writing–review and editing.

This work was not supported by any external funding.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.