Mortality Among Veterans With Rheumatoid Arthritis

In this podcast Dr Johnson discusses a study he and his colleagues conducted into all-cause and cause-specific mortality among veterans with rheumatoid arthritis with and without interstitial lung disease.

Tate Johnson, MD, is an assistant professor with the University of Nebraska College of Medicine.

TRANSCRIPT:

Welcome to this podcast from the Rheumatology and Arthritis Learning Network. I'm your moderator, RALN, and today I'm pleased to welcome Dr. Dr Johnson, who is an Assistant Professor in the Division of Rheumatology and Immunology at the University of Nebraska Medical Center and Veterans Administration Nebraska-Western Iowa Health Care. Dr. Johnson and colleagues recently published a study on all cause and cause-specific mortality among patients with rheumatoid arthritis that resulted in some interesting findings. Thank you for joining us today, Dr. Johnson.

Dr Johnson: Yeah, of course. Appreciate you having me. Excited to be here.

RALN: What initiated your interest in examining mortality among patients with RA?

Dr Johnson:

Sure. I think it's been pretty well established for many years that patients with rheumatoid arthritis are at higher risk of mortality. RA has been shown to shorten patients' lifespan. Obviously, treatment of RA has improved substantially over the last couple of decades. So with advancements in treatments with new disease-modifying agents, we've certainly seen improvement in RA-related outcomes as far as disease activity and functional status.

There's been some conflicting reports as mortality studies have been updated over time as to whether this so-called mortality gap between patients with and without RA has closed with improvements in RA-related care. We aimed or we are interested in looking at that in our study, in addition to looking at the trends in mortality over time in patients with RA many large studies looking at RA-related mortality tend to look at what we'd call categorical causes of mortality—for instance, cardiovascular disease, respiratory disease, cancer. So one of the things we were interested in looking at on a more granular level in this study was looking at more specific causes of death. Within cardiovascular disease comparing coronary artery disease versus heart failure, versus other causes, or different types of malignancies. So those were a couple of the items that sort of motivated this study.

RALN:

Can you give us an overview of the study in terms of the numbers of patients you looked at, the kind of cohorts, the length of time it ran?

Dr Johnson:

Yeah, of course. This study, we used national-level Veterans Affairs data, study windows from 2000 to 2018. We had basically an 18-year time period. And within the national VA data we identified about 30,000 patients who were newly diagnosed with rheumatoid arthritis. Then we matched those patients based on age, sex, and then their enrollment year into the VA to about 250,000 non-RA patients. So over that 18-year study period, we had about 2 million person-years of follow-up time is kind of what that amounts to. So it's a large study. We followed patients from the time they were diagnosed with RA and until either the end of the study or until they died in order to look at causes of death.

Then we linked our national VA data to the National Death Index, which captures causes of death amongst patients in the United States. Then we also linked to that a tool through the Healthcare Cost and Utilization Project called the Clinical Classifications Software, which allowed us to take that categorical cause of death that we oftentimes will see from the National Death Index and then tease out more specific causes of death underneath those more broad categories. Within that study then we analyzed the association of rheumatoid arthritis with these categorical and more specific causes of death. And the nice part about the Veteran's Affairs data, it's a very large data set, but it has really robust clinical data from the electronic health record.

So we were able to account for things like age, sex, race, ethnicity, body mass index, smoking status, and then the burden of comorbidities as well, which we know is certainly increased in patients with rheumatoid arthritis. So we looked at that association between RA and causes of death in an overall perspective over our time period. And then to get at trends in rheumatoid arthritis-related death, we then split up our study period into 5-year time blocks. So based on when a patient was diagnosed with RA we created 3 separate cohorts over the study period and then compared the risk of mortality between those to look at trends in RA-related deaths.

RALN:

And what were your findings?

Dr Johnson:

Amongst patients with rheumatoid arthritis we identified about 9,500 deaths. So we had 30,000 RA patients, about 9.5 thousand deaths. Unsurprisingly, the most frequent causes of death were cardiovascular disease, cancer, respiratory deaths. That's not terribly dissimilar from non-RA patients. But in our adjusted model where we are accounting for all the different risk factors for these causes of death, we found that patients with RA were at a 23% increased risk of all-cause mortality. Amongst that, that's when we looked at number of deaths in these patients, looked at incidence rates and compared between the 2 groups.

That amounted to over 2,500 excess deaths in patients with rheumatoid arthritis. So amongst those 9500 deaths that occurred, we estimated about 2500 of those would be, quote, excess of what they should be based on the non-RA population rates of death. So that's sort of the all-cause mortality estimates. We had a number of analyses, obviously was our objective in this study. So when we looked at categorical causes of death, we found about a 20% increased risk in cardiovascular disease and cancer-related death, a 45% increased risk of respiratory-related mortality, and then a nearly 60% increased risk of infection-related death.

When we then broke that down into more granular causes in the different categories, and I won't obviously go into each specific cause that we looked at, but the most overrepresented cause of death in rheumatoid arthritis was interstitial lung disease-related death. So this is obviously an area of interest in rheumatoid arthritis and certainly a burdensome condition that we're trying to learn more about. We found over 3-fold higher risk of ILD-related death in this study. Then obviously we were interested in looking at temporal trends in rheumatoid arthritis-related mortality and saw some encouraging findings there.

We saw that there's a significant improvement in all-cause mortality, and then favorable trends in cardiovascular and cancer-related mortality over time where most contemporary cohorts seemed to be at a little bit lower risk than the earlier diagnosed rheumatoid arthritis patients. But one thing we point out in this study is that there's actually, even in our most contemporary cohort, the most recently diagnosed RA patients from 2012 to 2017, there's still a statistically significant increased risk for all-cause mortality, cardiovascular disease-related mortality, respiratory, and infection-related death.

RALN:

So the good news is that the RA-related mortality is decreasing. But the bad news is, there's still that mortality gap. Do you have a hypothesis about why that gap exists, and is it in fact shrinking, and do you have an idea as to why that might be?

Dr Johnson:

That's a great question. So as you mentioned, good news, things are improving over time. And I think that improvement likely or simply that we're treating RA a lot more effectively than we were 20 years ago. So obviously our sort of armamentarium of disease-modifying agents continues to increase. And so I think our ability to control RA disease activity has improved. I think that's reflected in the improvements in mortality estimates that we're seeing here. And obviously I think the best thing we can do for any individual patient is to control their RA because clearly, uncontrolled RA is going to be probably the biggest risk factor that we see as far as early death in these patients.

But then beyond that, as far as when we see that even these really recently diagnosed RA patients are still at increased risk of death, then the question becomes, are there other things that we need to be doing to help these patients from maybe a more holistic standpoint as far as not just their RA-related disease? So I guess beyond just controlling the clinical RA disease activity is, there's sort of this population of patients with either what I like to refer to like subclinical inflammation that we can't pick up on our exam, or genetic predisposition that puts them at risk for some of these extra-articular manifestations beyond what we can evaluate at the bedside.

So that's where the interest in developing prediction models and risk stratification strategies really comes into play. And beyond some of those areas, there's also been some study, for instance, in cardiovascular disease in rheumatoid arthritis patients. Some of the preventative strategies, even through primary care providers, such as statin use, may be suboptimal in rheumatoid arthritis patients.

Or secondary prevention strategies—for instance, after an RA patient gets admitted for myocardial infarction, there's been some studies that have shown they are less likely to get what we'd consider guideline-directed management with things like beta blockers. So those may also influence some of this persistent mortality gap. So I think understanding those things best we can and ensuring we're continuing to move forward in not only sort of our ability to treat RA effectively, risk stratify patients with RA, but also doing all the things we already know are helpful to the non-RA population in a more comprehensive manner in these folks.

RALN:

The relationship of RA to ILD is pretty well-known now. But is it possible that the cardiopulmonary disease burden is increased by perhaps veterans' exposures to toxins that the general population may not have?

Dr Johnson:

That's a great question. I think obviously it's a question of particular interest, especially in the veterans as far as exposure history, and military exposures, and how that influences particularly respiratory-related outcomes long term. Obviously in this study we weren't able to specifically tease out how exposure history was associated with respiratory outcomes. I think that's something that anyone that's working in the VA is interested in doing so in the future. I think one of the things I'd point out in this study is that these were what I call internal controls. So we were looking at veterans with RA and comparing them to veterans without RA.

So presumably that control population has probably experienced a lot of the same exposures that our patients with RA did in this study. And we still see this significant, substantially increased risk of respiratory-related mortality and ILD. So there's obviously some nuances there because then you could probably ask, "Okay. Well, do exposures sort of influence RA onset?" So there's a lot of things that tease out there. But even in our study, even looking at RA versus non-RA veterans, we still see this 3-fold increased risk of ILD, which I think is significant.

RALN:

So the RA is definitely the operative issue there.

Dr Johnson:

Yeah, and I think at least from our findings, yeah, from the study I think that that's where we would certainly focus our attention.

RALN:

You mentioned infection as well, and of course some of the most effective drugs for treating RA are known to be very immunosuppressive. Is that one of the factors that has to be considered in looking at your findings?

Dr Johnson:

Yeah, absolutely. I think it's interesting. In the most recently diagnosed RA cohort in this study, infectious-related deaths are actually one of the—maybe as far as categorical causes—the most overrepresented causes of death in our study. But when we looked at trends, that risk was pretty steady over time. So I think you can look at that a couple of ways. I think you can look at that, well, even with the introduction and wider use of things like biologic DMARDs and more potent immunosuppressive therapies, we haven't necessarily seen this dramatic increase in the risk of infection. So I think in a way that's sort of good news.

But as you alluded to, I mean there is sort of an unavoidable risk with immunosuppression as far as infectious complications go that doesn't negate the importance of controlling RA disease activity. Because I think some would argue that a patient with uncontrolled RA is also at increased risk of infection and would probably do worse. So this sort of points to the fact that when you're deciding to start these agents and you're beginning patients’ immune-suppressive therapy, it is an important discussion to have with patients. And tailoring, we don't have a sort of magic crystal ball when we see patients at the bedside to say, "This is the best DMARD for you."

A lot of it is tailoring what we know about side effects, in particular the administration of agents to the lifestyle or comorbidities of a given patient. Educating patients on that, having a shared decision-making conversation, and then talking about strategies that, the ways patients can reduce their risk. Obviously, in recent years vaccination status is extremely important, and I guess it always has been. But I think it's been highlighted more now during the COVID-19 pandemic.

So all those things, I think it just sort of highlights provider vigilance in some of the health care maintenance issues that we can do to mitigate that infectious risk and minimizing things that might increase their risk, for instance, glucocorticoid use. If we can minimize glucocorticoid use, there's been some good studies that have shown even low-dose glucocorticoids over time can increase infection risk. So if there's anything we can do to mitigate that infection risk, I think that's important to us as providers.

RALN:

So you've pre-empted my next question, which is, how can this knowledge be applied in the clinic by the practicing rheumatologist and even the primary care physician? And you've just named a number of things that I think would be within that purview. What else would you suggest for those physicians to be particularly cognizant of?

Dr Johnson:

Obviously that's a great question. First and foremost for the rheumatologist, it's about controlling RA disease activity. I think we've all sort of met that patient that comes in, and we see a few swollen joints and we think, "Oh, their RA's not perfect." But the patient doesn't tend to complain all the time. It depends on the patient. But I think we've all met that patient says, "I'm fine, Doc. I don't want more medicine." And then I think at least that conversation with patients saying, "We're seeing this inflammation here. I know you're feeling okay, but here's what might happen. These are sort of the long-term outcomes that we see in patients with persistent disease activity."

So discussing with patients the value of escalating treatment in those scenarios is important. And then I had started to mention, as you were alluding to, this holistic care approach between rheumatologists, primary care providers, and other specialists, is really important. So when I think about how that looks in a clinic, you think about the main takeaways from our study as far as what patients are dying from, and it's cardiovascular disease, cancer, and respiratory disease. So in patients with a smoking history, making sure that you're at least adhering to the recommended screening for low-dose CTs as far as picking up on lung cancer in patients with respiratory complaints, having kind of a low threshold for things like pulmonary function tests and more advanced imaging to catch interstitial lung disease when it's early.

Rheumatoid arthritis patients have been shown to present from a cardiovascular standpoint and a myocardial infarction standpoint have been shown to actually present in atypical ways. And then when we think about the overall rheumatoid arthritis population being middle-aged females, we'll often also sometimes talk about middle-aged females present with atypical symptoms of heart attacks like GI upset, or maybe it's fatigue, or something to that effect. So taking those complaints seriously and having a low threshold for stress testing and risk stratifying from a cardiovascular perspective, following blood pressure and lipid guidelines really closely.

Oftentimes I think where this really becomes important is, I think patients with rheumatoid arthritis see their rheumatologist so frequently that it's not uncommon for an RA patient to consider their rheumatologist their primary care provider. So realizing that and having a gestalt about when these things need to be done and either if you're comfortable performing these screening tests on your own, great. If not, having a close collaborative relationship with primary care providers and specialists.

RALN:

And then, of course, making sure that the vaccines are kept up to date to keep that infection risk low.

Dr Johnson:

Yeah, absolutely. Absolutely.

RALN:

What other research do you have planned for this area?

Dr Johnson:

I guess as far as our research group goes here, long-term outcomes in RA particularly as it relates to ILD, and cardiovascular disease, and this concept of sort of multimorbidity, is a particular area of research interest for our group. I mean I think these findings in this study help us to frame the comparative burden of these issues and make sure we're staying focused in our steps forward. So I think broadly the areas that we're interested in is looking at, are there strategies to either predict patients that are going to develop these comorbidities that influence these outcomes? Can we risk-stratify patients that are going to have more severe versions of these comorbidities? Coming up with strategies to manage these patients is really the end goal.

And whether or not that's coming up with a better understanding of which RA-related treatments improve respiratory outcomes versus cardiovascular outcomes, I think is always sort of the Holy Grail. Or if there's simply ways to understand when a patient with RA needs to be started on non-RA related medications to mitigate their risk. Like we mentioned, things like statins and lipid-lowering therapies. It's been well established that the same cardiovascular risk calculators that we use in the non-RA population tend to underestimate risk in RA. And so if there's anything we can do to improve those strategies, I think is broadly an area of future research that our group is very interested in.

RALN:

This has been very interesting. I hope we'll have an opportunity to talk again as you continue your research.

Dr Johnson:

Absolutely. Well, thanks again for having me. Appreciate your interest.