Multiple Ethnicities Study Explores A Connection Between Type 2 Diabetes and Urinary Bladder Cancer

David Bogumil, PhD, from the Department of Population and Public Health Sciences, at the Keck School of Medicine of University of Southern California, details his findings on a connection between type 2 diabetes and urinary bladder cancer for African American, European American, Japanese American, Latin American, and Native Hawaiian populations.

Transcript:

David Bogumil: So, my name is David Bogumil. I'm a postdoctoral researcher over at the University of Southern California and at the time of this study I was also doing a postdoc between University of Southern California and University of Hawaii, primarily working with the multi-ethnic cohort study. So that's a little bit about me, and then the institution, like I mentioned the primary one I'm at now is the university of Southern California and I'm actually in the Department of population, public health sciences. So over here we do epidemiology, health behavior, biostatistics, research, and at the moment my primary focus is understanding racial and ethnic differences in cancer and then also looking at genetic risk and absolute risk estimates within these populations.

Can you give some background for your study and what prompted you to undertake it?

David Bogumil: So, the vast majority of research that's out there on the association between diabetes and bladder cancer, it's for the most part among European ancestry individuals, and these could be like just Europeans or white Americans and this actually isn't just specific even to the Bladder cancer association, there's just not that much research out there about bladder cancer, non-european ancestry populations, too. So that's one of the motivating factors is to fill that gap in the literature or to help fill it and then I have an interest in this cancer and exposure from my doctoral work, also. Where I looked at the association between bladder cancer and type, 2 diabetes and metoregression and meta-analysis work and then, also, too, besides those factors, as I mentioned, you know. The most of the research out there on bladder cancers among European ancestry, and part of the reason for this, too, is that the incidence is highest among European ancestry, individuals in terms of like the more commonly reported on populations and So we do understand there to be racial and ethnic heterogeneity and incidence of this cancer between the populations and then we also know for a fact that the exposure varies between the population. So for this study, we're looking at the association between type 2 diabetes and bladder cancer in a multi ethnic population, you know. European not only European ancestry individuals, but also for their racial and ethnic groups, which I'll get into in a sec. But we know the prevalence of type, 2 diabetes, and also the severity of type. 2 diabetes to differ between these populations. So, given the differences in the exposure, distribution, and incidence of the cancer, this study will. The goal is to, you know, identify if these associations that we've seen previously in Europeans are consistent in non-european ancestry populations. So it's quite a few different motivating factors for this study.

Can you briefly describe how this study was conducted?

David Bogumil: For this study we used a large prospective cohort study, the multi-ethnic cohort study and it's important to understand. You know, how unique this cohort study is. In order to see how to kinda fill this gap in a literature that we identified in the multi-ethnic cohort study. This is a large prospective cohort study in Southern California and Hawaii, and it started in the early nineties. So, enrollment began in 1993, and the study contains over 200,000 individuals from a variety of racial and ethnic groups. So there's primarily 5 that were sampled for this prospective cohort study, whites or European Americans, Japanese Americans, Latinos, native Hawaiians and African Americans. So, the study was initiated primarily to look at possible causes for differences in the incidence rate of all cancers between these populations. Not only just for having that information, for understanding the biology, the cancer, but also for the public health aspect of having that information to seeing what communities might benefit from interventions, especially. So that's a little bit of the background of the data source. So, this study every few years, it is merged. The data is merged to information from state cancer registries and that's how we get information on if somebody experienced a cancer event or not and for the exposure measurement app. The beginning of the study. So back in the nineties, when these individuals enrolled, all of them received a questionnaire in the mail, and for the most part the primary part of the questionnaire was on assessing dietary factors, but it also contained information on comorbidities like type 2 diabetes status and this question is as simple as asking the participants. You know, have they read this form, and have they had to bubble in a response, yes or no to the question. Do you have? Or have you been told? You have? Type 2 diabetes from a healthcare provider? So it's like one general kind of question to assess this exposure whether or not somebody has it and then so one of the things we did in the study is, we compare different types of type 2 diabetes measurement. So, we have this kind of simple measurement from baseline, but we also assess another exposure measurement which is use of this baseline information. But also additionally including information from Medicare data and for a subset of the cohort, we do have information on billing codes from Medicare which allow us to identify if they've received billing codes for type 2 diabetes, and if they had 3 of these, you know, in this alternate type 2 diabetes definition, we classified them as having diabetes, and then last, in addition to these, there also were follow-up surveys sent to the cohort participants and this asks the same type 2 diabetes question. So, we kind of have this more, broad baseline definition of type 2 diabetes and then we have this alternate definition, where individuals could potentially be classified as having type, 2 diabetes based on a variety of data sources. So here we have our exposure, estimation or exposure measurement and outcome measurement and for this study, we're just doing a time to event analysis, where we're basically seeing, do the rates of bladder cancer differ by type. 2 diabetes status, you know, do individuals with type 2 diabetes experience higher, faster rates of bladder cancer. Yeah, bladder cancer relative to those without type 2 diabetes and in addition to this, you know, one of the main things we're looking at in the study is, does this association differ by racial ethnic group? You know? Do the associations we see in European ancestry individuals hold true for these other populations? Or is it possible that the pattern looks different in those groups?

What were the key findings of this study?

David Bogumil: I think the key findings for the study can be split into 2 categories. You can think about the first set of findings being kind of like descriptive findings. Almost well, we're not necessarily doing analysis. But we're looking at patterns in the data and just seeing what the data has to show, so similar to what's reported in the literature, we do observe European ancestry Americans to have the highest rate of bladder cancer in comparison to the other groups. So even European ancestry Americans without type 2 diabetes experience a higher rate of bladder cancer than the other racial ethnic groups we looked at who do have diabetes. So, the rate is so much higher among European Americans that even among those without type 2 diabetes, they're still experiencing a higher risk of bladder cancer than other populations who don't have type 2 diabetes. One of the other ways, to look at this, besides, just comparison of rates is to look at a measure called cumulative risk or absolute risk and basically what this measure is quantifying. What is the probability of somebody experiencing an event, at a certain age or earlier? So again, mirroring the results from our incidents analysis. We compare the incidence rate between the groups. European Americans had the highest or the earliest age to achieve 1% risk of bladder cancer in comparison to all the other groups. So those are the main descriptive type of findings and then the main part of the paper is a lot of association analyses in order to identify, is the association between type 2 diabetes and bladder cancer the same between populations. So in in our analyses we found there was no statistically significant difference in the rate of bladder cancer and diabetes status between the different racial and ethnic groups. So, then we tested to see, does association differ by population? We found that it didn't, at least in a statistical testing sense. We did find quite a variety in the hazard ratios, which is  the comparison of rates between bladder cancer between type 2 diabetes and bladder cancer between the populations. Among native Hawaiians, those with type, 2 diabetes at 1.89 times the rate of bladder cancer in comparison to those without type 2 diabetes. So, the association was strongest within this population, which hasn't been reported on before, and is a very key, interesting finding and then earlier, you also remember me mentioning that we looked at different measures of type 2 diabetes. We had this simple measure where it was just type, 2 diabetes based on self-reported baseline and then we had this alternative measure which used follow-up surveys and Medicare billing codes and when we compared the association between type 2 diabetes and bladder cancer we found that for the main effect measure, the hazard ratio, it actually didn't matter what definition we used for our association results. Finally, in the end they came out to be the same based on measure. However, we did observe a greater precision in our measure of association when we used the more detailed measure of type 2 diabetes which pulled in information on Medicare and the follow-up surveys, too. Those are the main important findings, I would say from this study.

How do you think your research will impact future studies?

David Bogumil: One of the analyses we did was looking at what portion of bladder cancer could be attributable to type 2 diabetes and in this study, you know, just as what's reported in the literature we've identified quite varied bladder cancer risk between populations and our analysis showed that type 2 diabetes likely doesn't explain the reason for this difference in incidence rates between these populations. So, you know this, research impacts future studies by letting other scientists know that if they want to understand the disparities or variation in bladder cancer incidence between populations, whether just for public health or for understanding disease mechanism that they need to keep searching. You know, type 2 diabetes doesn't seem to be an exposure that explains this difference in bladder cancer incidence between populations and then, of course, you know, this is the results just from this one study, and, as I mentioned earlier the sample size for the multi ethnic cohort study is very large. Our final analysis included over 180,000 individuals across these 5 groups. But, you know, given the unique results we found for the strong association between type 2 diabetes and bladder cancer in native Hawaiians. Future studies should seek to replicate this association. Using other data sources to make sure that what we've identified in our data isn’t a chance finding and that this is likely a characteristic of the actual native wine or Pacific Islander populations and then I would say, last one way in which this study may impact future work is highlighting the potential value of using alternative exposure measures for diabetes. These large prospective cohort studies will use a single type 2 diabetes measure from baseline when they're looking at type 2 diabetes as an exposure. But there is the potential benefit in pulling in other measures of type 2 diabetes, such as billing information or additional follow-up surveys from what we found as potential to increase the precision of your association analysis, and possibly, you know, find associations for your main analysis or sensitivity analysis that you might not otherwise be able to find.

Was there anything you found in your results that surprised you?

David Bogumil: As like, I keep mentioning that probably the most surprising thing, is the strength of the association. So, like the magnitude among native Hawaiians, you know, diabetes really seems to affect bladder cancer risk in the native Hawaiian population and we discuss this a little bit in the discussion portion of the paper to try to understand potentially why this could be the case and you know, one of the main things to consider is the severity of type 2 diabetes in the native Hawaiian population, and then also potentially access to care or ability, or treatment protocols or ability to treat the type 2 diabetes correctly in this population. You know the association between type 2 diabetes and bladder cancer may appear strongest in native Hawaiians. But it might have something to do with diabetes not being adequately treated in that population. So again, this is like, this needs further investigation. Because, if diabetes isn't being treated the best it could be in this population that not only has an effect for bladder, cancers and outcome, it could potentially affect all diabetes, associated cancers, like pancreatic cancer, for example. So, there would be quite a bit of benefit to investigating further. How well is, you know, type 2 diabetes and blood sugar control implemented in this population. So that's surprising and it really, you know, we have the data to further investigate some of this, and I hope other researchers do, too. Because there's benefit to be gained within this population, if this could be figured out.