This transcript has been edited for clarity. For more episodes, download the Medscape app or subscribe to the podcast on Apple Podcasts, Spotify, or your preferred podcast provider.
Maureen McMahon, MD, MS: Hello, I am Dr Maureen McMahon. Welcome to the Medscape InDiscussion podcast series on systemic lupus erythematosus (SLE). Today, we'll discuss the epidemiology of SLE with our guest, Dr Peter Izmirly. Dr Izmirly is a professor in the Department of Medicine at NYU Grossman School of Medicine and co-director of the NYU Lupus Clinic. Thank you for joining us today.
Peter M. Izmirly, MD: Thank you very much for inviting me.
McMahon: When I think about epidemiology and lupus, I think of all the great work that you and your team have been doing. I know when I first started in lupus and was a young faculty member, and I had to write up a talk on lupus for the medical students.
I remember the data on prevalence and incidence of lupus were varied. There were lots of different studies which had widely different numbers. I appreciate the work that you've been doing with the CDC registries. I'm curious how that got started and also a little curious to know the process. How did you gather that data?
Izmirly: Part of the reason we started this effort was exactly what you were saying. The data available on the epidemiology of lupus before the early 2000s, when these registries started getting up and running, were kind of all over the place. And it's partially because there's no gold standard for the diagnosis of lupus. For example, in cancer, you have biopsies which are reportable in a lot of states. In infectious disease, you have a culture or a blood test. For diabetes, you have an A1c. We all know that it's sometimes very difficult to diagnose lupus.
So, between that and the different methodologies of the prior studies, the different demographics they were targeting, different case finding sources, different states with different populations, and how they're doing it (self-report) led to a variety of results. And so, somebody in government became interested in it and the money was appropriated, I believe, to the CDC. Every site, except the Indian Health Service, which was slightly different, was a collaboration between a city or state Department of Health and an academic medical center.
The first two sites were in Michigan and Georgia, and they partnered with the University of Michigan and Emily Somers and Joseph McCune. In Georgia, they partnered with Emory with Sam Lim and Cristina Drenkard. Their catchment areas had very large White and Black African American demographics.
They were going to be able to get credible estimates on those two important demographics. But at some point, which was very fortuitous for me, they realized that they weren't able to capture data on two of the faster growing populations in America: Hispanics and Asian population.
They put out a [request for application] for two more sites. I was very fortunate to get funded at NYU, and we partnered with our city Department of Health, which is a separate entity from the New York State Department of Health. And then Maria Dall'Era at UCSF was funded. We were all tasked with finding the incidence and prevalence of lupus.
There was an IHS site, the Indian Health Service, which is a branch of the federal government. They had a separate mechanism that was sort of like an internal funding. We were very fortunate to come in the second round of sites, because we used a harmonized data dictionary that was already developed and fleshed out by the Michigan and Georgia teams.
We used a harmonized data dictionary that had information on how you define arthritis, rash, and all the variables. At the time we did this, which was before 2012 when the Systemic Lupus Collaborating Clinics (SLICC) classification criteria came around, we used the 1997 revised [American College of Rheumatology] (ACR) classification criteria. But the SLICC criteria were in development, and they allowed us to look at the variables that they were discussing.
We all collected most of those classification criteria variables. We went to hospitals in our catchment area, and all the rheumatologists in our catchment area, and we also did some nephrology and dermatology offices. It was a little harder to do that for some of the sites, especially like New York City in dermatology; I mean there's like 400 dermatology offices. We also got population data sets, such as the state discharge database, the U S renal data system, and vital records, which enabled us to look at death certificates; so we had access to that data. We also approached pathology for renal biopsies at various sites.
We trained the medical abstractors very thoroughly, and fortunately for New York City, we had all MDs who were looking to land residency spots. We didn't have to teach them the language of what serositis and whatnot was. We trained them, and then they went to the various case finding sources with the data dictionary in a computerized form and went through every chart page by page and abstracted labs, arthritis information, whether they had a kidney biopsy what it showed, proteinuria, and all the various classification criteria. In the end, we just did a simple numerator denominator.
McMahon: What a tremendous amount of work that must have been.
Izmirly: Tremendous. And the manpower was incredible.
McMahon: It took years to do all that. It was a very systematic and thorough approach, and I think it speaks to the quality of the data.
Izmirly: We did quality assurance and quality control, so abstractors continued to be monitored to make sure that they weren't slacking or losing their skill set over time.
McMahon: Why don't we go over some of the things that you found. I'm especially interested in the prevalence data. I believe you took all the different sites and combined them together in one meta-analysis to get a combination of data.
Izmirly: In total, we had 5417 identified cases of lupus who met the ACR criteria from the various sites. The overall prevalence of lupus in the meta-analysis was 72.8 per 100, 000 person years, with a confidence interval that ranged from 65.3 to 81.0. And women, as we already are well aware of, were nine times higher than males. The prevalence rate for females was 128.7 compared to 14.6 per males. Overall, the highest rates were seen in African American Black females and then Hispanic females.
The American Indian, Alaska Native study, because of a difference with their methodology, weren't part of the meta-analysis. But given their methodology was similar, we included their data as a comparison, and their female rate was the highest at, 270.6 per 100,000.
McMahon: Wow, had there been previous data on that in that population?
Izmirly: Very little, which is again why they funded the Indian Health Service to get that data.
Their methodology was a little different because of the way you have to capture the demographics. It did provide very good, credible data on that demographic, which hadn't been studied very much at all.
McMahon: And then I believe you went and looked at how that would translate into the 2018 census.
Izmirly: We applied the various rates for both males and females to the corresponding census demographics for 2018, which was the most currently available at the time that we published this paper. And we came up with an estimate of 204,295 individuals who had ACR classification criteria diagnosis of lupus with a range of 160,902-261,725. That number hovers around what, in America, we classify as a rare disease. A rare disease affects under 200,000 patients.
McMahon: We're so fortunate currently to have so many clinical trials in lupus, but those numbers, I think, really give us an idea of how many patients are out there when we're thinking about trials and recruiting for trials.
Izmirly: We provided data using the SLICC definition for lupus in our original paper, and the rate was higher. That provided higher rates across the board for lupus. And then we did the recent EULAR/ACR classification criteria in a subsequent paper, and that was in between the ACR rates and the SLICC rates. I know there's a lot of emotion around the difference between using classification criteria for diagnosis and some rheumatologists with a gold standard diagnosis.
We don't always see eye to eye on lupus. I mean, even in the derivation of that EULAR/ACR criteria, there's a paper with a table of mock cases where lupus experts look at cases and say whether they have lupus or not. There are patients where everyone agrees, and everyone says the person doesn't have lupus. But there were a handful of cases in the middle where people are like, yes, this person definitely has lupus, and then the corresponding amount of people said this person absolutely doesn't have lupus.
So, classification criteria can help resolve some of these discrepancies. In that paper, we went through all various permutations of patients who met EULAR criteria only, who met ACR criteria only, and who met SLICC criteria only. And I find it hard these days to come up with any scenario now where somebody reasonably thinks a patient has lupus that doesn't meet at least one of those. The only thing that jumps to my head is a low titer antinuclear antibody (ANA), like, of 1-40, and then two strong things, like DNA and arthritis. If that ANA was 1-80, you'd be like, all right, they meet your EULAR criteria, your ACR criteria; but because it's 1-40, it doesn't. I mean that, if somebody wants to label that lupus, I'm not going to roll over and scream.
McMahon: I think that's a good point. I think these classification criteria are important for studies like these, epidemiologic studies, or other studies where you're trying to have standard criteria. But I think when it comes to the actual patient sitting in front of you, it's kind of your clinical judgment. Having these different tools can help us figure out what to do with our patients on an individual basis, I think.
Izmirly: I totally agree.
McMahon: We've talked about the prevalence in the population. I believe you also looked at the overall incidence of new cases of lupus in the population.
Izmirly: We wrote two papers. The prevalence is the number of all cases in the population in a given portion of time. And then we looked at the incidence, which is the number of new cases in a given amount of time. We pulled all the data.
The pooled incidence rate, using the same methodology in the same studies, was 5.1 per 100, 000 person years, and again was higher in females than males: 8.7 vs 1.2. The highest was in African American Black females, followed by Asian Pacific Islanders, and then Hispanic, and then White females.
I think one of the things that both these studies did is that because of the pooled numbers, we were able to provide rates for males in both studies, which I think was a major advantage because a lot of the epi studies largely focused on females given the ratio. And again, we did this estimate of how many people would be newly diagnosed with lupus in 2018. We estimated this by applying the various rates to the various corresponding proportions of the census data of 14,216 people would be newly diagnosed with lupus in 2018.
McMahon: I think that's interesting data. I appreciate your point about the incidence and the prevalence data that you were able to obtain in males. Out of curiosity, do you have any information about the age of those patients?
Izmirly: We're working on that paper right now; we have these quarterly calls with the CDC cohorts, and one of the investigators made a comment that we don't have the meta-analysis or any very good data about the age breakdown. So, each site did provide in their paper the incidence and some even the prevalence based on ages, but for reasons that are long and complicated and too boring for your audience, we did not do this for the original papers.
McMahon: I also thought that there was a lot of interesting data that came out of the various individual studies, looking at causes of death in lupus patients. I know that there was a study done by the group looking at the California Lupus Epidemiology Study (CLUES). Could you talk a little bit about what they found about the causes of death in lupus patients?
Izmirly: They did two studies, one that showed that the rate of death was very high in the Hispanic and Asian demographics, which hadn't been shown before. And then they subsequently, in a follow-up paper, they explored the causes of death. This was out of the CLUES study from the California site.
What they showed, which I think many of us in the lupus world are already semi knowledgeable about, is that cardiovascular disease was the leading cause of death across all racial and ethnic groups, followed by underlying rheumatic disease and then third heme-onc disorders.
The underlying cause of death was 3.63 times higher amongst lupus cases than the general population, and cardiovascular deaths for those with lupus were 4-6 times higher for the Asian and Hispanic demographics with lupus, respectively, compared to the general population. And that sort of was corroborated, at least in the Hispanic population, by a paper that our team was working on at the same time, coincidentally. And that paper was recently published.
McMahon: Why don't you talk a little bit about that?
Izmirly: We had all this cardiovascular data in our surveillance program that we collected. Given the surveillance nature of this program, we didn't have information on traditional cardiovascular disease factors like smoking and diabetes and high blood pressure. That's not really related to the diagnosis of lupus. Unfortunately, we did not have that data available. Instead of a standardized mortality ratio, we did a standardized cardiovascular ratio.
We compared a ratio for how many people in various demographics had lupus and had cardiovascular disease. And then the control would be a National Health and Nutrition Examination Survey (NHANES), either national or New York City, who also has cardiovascular disease. We saw, like what my friends in California showed, the Hispanic demographic has a high rate of cardiovascular disease compared to the population as a whole.
And in fact, it had the highest rate of cardiovascular disease compared to the population overall. If you compared it to the New York City NHANES, it was 5.5. And if you compared it to national NHANES, the rate was 4.
McMahon: Something that we've been recognizing over the last several years is how much of an impact cardiovascular disease can have in our population.
Izmirly: We looked at various breakdowns. Young males had much higher rates of [cardiovascular disease]. If you had lupus and you're a male age 20-49, irrespective of your race or ethnicity, your risk of having a cardiovascular event defined as a heart attack or a stroke was 11.3 times higher than if you didn't have lupus. And we showed that, again, young Hispanics had a 10.9 times rate, if you had lupus, compared to the same demographic in the absence of lupus.
McMahon: I think that we've had some of that data in young women, but I feel like this is one of the first studies I remember seeing that also describes that risk in young men.
Izmirly: That's one of the luxuries of some of the epi work that myself and my colleagues at the CDC registry had — it's the large population. I mean, part of the [request for application] required your catchment area had to have a million people at bare minimum. And then you captured all that to the best of your abilities.
I'm sure we probably missed some, none of this stuff is perfect, but when we captured as many lupus patients as we could, you got a relatively substantive population of males with lupus. And so, it gives you some power to look at this stuff. It's not perfect
McMahon: But it's really important to have a better understanding. I also thought that some interesting data looking at the overall prevalence of different disease manifestations in lupus patients. I recall there was one study from the California group. Can you speak a little bit about that?
Izmirly: They looked at their California Lupus Surveillance Program registry and broke down when the severe manifestations developed and looked at the different demographics and their risk factors.
Some of the interesting findings are that African Americans, Asian Pacific Islanders, and Hispanic patients, respectively, had the highest risk of developing lupus nephritis, and that Hispanic patients and Asian Pacific Islanders also had a higher risk of developing antiphospholipid syndrome (APS).
McMahon: I think that's interesting, and I don't remember seeing previous data that gave us an idea of severe findings, such as renal involvement or APS in the Asian population.
Izmirly: In the various registries, they did publish the criteria. For example, in New York City, and I know Georgia and Michigan and California also did this in the original papers, they did look at the various ACR criteria. In every registry, all of the minorities, African Americans, Hispanics, and Asian Pacific Islanders, had the highest rates of lupus nephritis.
The rates approached almost 50% in some of the studies. But in this paper, they showed the time to the development of lupus and the risks of those populations developing it.
McMahon: As we think about wrapping up our discussion, I'm curious to know if your involvement in this work has changed the way you practice or the way you think about your patients or the way you see lupus in general in the United States?
Izmilry: A few things, even using some of the higher estimates that you might have generated from SLICC or the EULAR/ACR criteria, we know there's not a lot of lupus patients out there. And when you have 8000 CAR T studies trying to recruit some of these same sick patients, it makes you wonder how practical that is. I mean I'm not dissuading it, but there's only so many patients out there for these various studies.
And then, we don't want patients to suffer, and we don't want them to have delayed diagnosis. And I think making the front-line people, which are not us really, they're the ER, urgent care, primary care physicians.
When you see a young African American woman who has joint pains, or you do a urinalysis for a UTI and there's like four plus protein in it, you should start thinking “maybe this patient has lupus.” There was a decent amount of publicity around the publications when they came out. Maybe the message to get across to the front-line people is that young women are at certainly a higher risk for developing lupus, especially amongst certain demographics. And you should think lupus when you have unexplained kidney disease or arthritis or some rash that isn't some fungal something or other.
McMahon: I think, along those lines, if you have a young, Black patient who comes into your office with joint pains, remember to check a protein or check for proteinuria and see if there could be proteinuria there and just kind of remember some of these manifestations that we do see more commonly in patients.
Izmirly: I've learned a lot. I've learned a lot about epidemiology. I'm actually not a card-carrying epidemiologist like some of my colleagues are. I've learned a lot about the nuances of epidemiology and how difficult it is to study some of these things, especially in a disease like lupus.
I mean, maybe if we had some blood test, that was end-all-be-all, then this could be done much quicker, right?
McMahon: Right? But instead you had to get your boots to the ground and go out there and do all that work.
Izmirly: Yes, we all had our boots. There's a lot of people working on it.
McMahon: When I think about the effort you described, it's really tremendous.
Well, thank you again for talking to us about your efforts and everything that you and your colleagues learned. Today we've talked to Dr Izmirly about the epidemiology of lupus. I think we all got insights into the incidence and prevalence of lupus in the United States. I think it's important to keep in mind that especially our young patients can be at higher risk of complications. Not only complications that we think of traditionally being from lupus, such as increased renal disease, but also cardiovascular disease in those younger patients.
Please take a moment to download the Medscape app to listen and subscribe to this podcast series on SLE. This is Dr Maureen McMahon for the Medscape InDiscussion podcast.
Resources
Medscape © 2025 WebMD, LLC
Any views expressed above are the author's own and do not necessarily reflect the views of WebMD or Medscape.
Cite this: Lupus by the Numbers: What the Data Reveals - Medscape - Mar 20, 2025.