Note: CME credit is no longer available for this program.
Introduction
Theodore R. Fields, MD
Since biologic agents were introduced, rheumatologists have had to decide whether to use them alone or in combination with other disease-modifying agents. Early studies of combination therapy raised problems in interpretation because they used methotrexate failures as the patients to receive combination treatment with a biologic agent. The study discussed here by Dr. Markenson avoids these problems by using methotrexate-naïve patients, a methodology that can help rheumatologists in making real-world decisions. After reviewing the study's methodology and conclusions, Dr. Markenson discusses the clinical implications for treating RA patients and whether or not methotrexate doses should be changed after a biologic agent is added.
***************
Lecture
Joseph A. Markenson, MD
The TEMPO trial, a trial of etanercept and methotrexate with radiographic patient outcomes. This trial was presented at the last EULAR meeting in June of 2003. Why is it unique? Why is it exciting? What does this have to do with you, the practicing rheumatologist, and why is it of some importance to understand not only the biology but the actual results of the trial? I am going to try to go over these points in this lecture.
Why is it unique? In the July 2003 issue of Arthritis & Rheumatism, there was an editorial from Maarten Boers. If you remember, Maarten Boers conducted the COBRA trial, which started initial thinking about the importance of early intervention in rheumatoid arthritis. In his editorial, Maarten Boers talks about trial methodology. There are trials where we compare one drug against a combination of drugs, and we know that either the combination works better than a single drug or they are equivalent. What has never really been answered, however, is that when you test one drug against a combination, how do you know that you wouldn't have done as well with one of the components in that combination? Why do you need the combination? Does the combination add anything over its components, or does it just give you increased side effects? That question had never been answered in any of the previous trials conducted with any of the biologics in rheumatoid arthritis. It had not been done for a variety of reasons, some of which are expensive and increased number of patients needed. This trial is the first trial that actually did that, and that is why this trial is so important.
What was the objective? The objective was to evaluate radiographic and clinical responses when we treat a patient with etanercept versus etanercept plus methotrexate as compared to methotrexate alone. What this trial did is that it looked at etanercept alone, methotrexate alone, and the combination -- so you could really decide whether the combination etanercept and methotrexate was better than etanercept or methotrexate alone or could you have done equally well by just using one of the single drugs. That is important.
The multi-center trial lasted for a year. That is important because it wasn't a short trial. The treatment grids were: etanercept 25 mg subq twice a week, the way you dose etanercept, or methotrexate once a week, or etanercept plus methotrexate. 682 patients participated at 92 different sites. This is called an "intention to treat" analysis. That is important because what it means is that if you got a dose of drug, you were in the trial. Whatever result you got, side effect and/or efficacy, was counted even if you only got one dose of drug. This next slide describes the fact that the methotrexate was given at a very important dose. We have learned that 7.5 mg of methotrexate probably is somewhere near a homeopathic dose, but you really need to prescribe 15-20 mg to get efficacy. So when you are doing a trial you can't do it against a sub-optimal dose of another drug. What this trial did is match etanercept against a very high dose of methotrexate, average dose being 17 mg per week. In Europe, where they had not used folic acid before, (which as you know decreases the side effects of methotrexate), folic acid was added in a little different regimen than we use here in the United States. 5 mg b.i.w. was used instead of 1 mg a day.
End-points -- we will spend some time on methodology again. In the ERA trial, we used something called an "area under the curve." I will describe that for you in a few minutes. It is a little different than the ACR-20, -50 and -70, which we are used too. The trial also looked at radiographic outcome at one year using the modified Sharp score, which you have seen in other trials. The secondary end-points are going to be what you are used to. That is the ACR-20, ACR-50 and the ACR-70, both at 6 months and 12 months, and we will also look at radiographic change at 24 and 52 weeks.
Since this was a European trial, the DAS score (Disease Activity Score) was also used; I will describe that in a minute. To enter the trial, participants had to have an adequate response to at least 1 DMARD, other than methotrexate, 10 tender joints, at least, which meant active disease, and a sedimentation rate of greater than 28, or a high CRP, as well as morning stiffness. Participants had to be off DMARDs for at least one month. Patients were excluded from this trial if they had taken methotrexate before but had an inadequate response. I think that is important, because we didn't load the trial with people who were going to obviously fail because they couldn't take methotrexate to begin with. They could have taken methotrexate in the past, but if they discontinued methotrexate within the last six months they were not accepted in this trial. This eliminated the patients who were getting an obviously better response because they just started the drug. They were obviously excluded if they were intolerant of methotrexate, if they ever had taken a TNF inhibitor before, or if they were on more than 10 mg of prednisone.
So these were the exclusion and inclusion criteria. So let's talk a little bit about the end-points. The difference between the ACR, the area under the curve and the ACR-20. Do you remember an ACR-20 means that at any one point in time you are 20% better than at baseline, and is a composite score that includes swollen joint, tender joint, VAS that the patient does, and possibly a sed rate and CRP. ACR-20 also means that if you are anywhere between a 20 and a 49 ACR you are listed as an ACR-20. What if, then, you are a 49% response? You still would have been categorized as a 20% response. So in an effort to try to make a continuous curve on this -- to get those patients who are 20, 25, 38, 40 -- we do an ACR-N. You see that the ACR-N is a little bit different. It is the lowest improvement of swollen and tender joint as well as third highest criteria of the other variables.
So this gives you more of a continuous variable and not just one point in time. On the next slide I just illustrate what we were doing. Let's take patient 1, for example, and patient 2 and calculate ACR-Ns. The swollen joint count in patient 1 is 28%, patient 2, 42%, and the tender joint count is 25 versus 45, and you can see the differences. What is important here is that in the ACR-20 you would take the tender joint count of 25 and yes there would be an ACR-20. But if you calculated their ACR-N, they really are an ACR-25. Patient 2, if you calculated their ACR-N, is really ACR-42, but both of these are being listed as an ACR-20. So the ACR-N is thought to be a better way of identifying actually what you are. And what is more important is that you have to think of it as an integration. I will show you that later. What the ACR under the curve does is take all the ACR-Ns and - (for most of you that suffered through calculus if you may remember -- it sums up. So it's not a mean value at point in time for two weeks, four weeks and six weeks, but the area under the curve integrates the values to indicate what your response was over time. It is a continuous variable, and it is thought to be a better measure of a patient's response.
DAS is a European score. It is fairly complicated. You have to use a calculator to get it, but it is validated. It measures a sed rate, a swollen joint count, and a tender joint count. The DAS score looks at one point, whereas the ACR-20, because it represents an increment over baseline the investigator, has to go back and look at what the patient was originally. That is easy to do if you saw the patient three weeks ago, but harder to do if you are comparing to where they were a year ago. The DAS looks at the patient at that point in time only, and it is found to correlate maybe a little bit better than the ACR does with radiographic damage and also with clinically significant outcomes, such as remission. (In other words changes in DAS correlate with severe, moderate disease and remission).
If the DAS score dropped from a high score to below 2.4 this would indicate a change to a low disease activity. A clinical remission (which we almost never talk about,) occurs with DAS of less than 1.6.
It is important in looking in a trial, to be sure, that we have a group of patients that are put into the study that are equivalent in terms of demographics. The number of patients who were on methotrexate was 228, 223 on etanercept and 231 in the combination group. There was no difference in the age, female ratio, RA duration, which was about 6 years, rheumatoid factor positivity, previous amount of DMARDs, or anything else. So the groups going into this study were thoroughly well matched in terms of their demographics. The baseline clinical measurements show you that patients going into each group had about the same tender joint counts, swollen joint counts, and global assessments, and they were pretty evenly matched as well in CRPs, sed rates, DAS scores and HAQs.
How many patients completed the study at six months? 70% of patients completed the methotrexate arm, 76% the etanercept arm, and 84% completed the combination arm -- and that begins to tell you something. Look at discontinuation. We have a lack of efficacy and adverse events, but you can see that more people discontinue in the methotrexate arm or etanercept alone arm, rather than the combination because of lack of efficacy. That was statistically significant. Adverse events were equal across the board. So at least we know why people left the study.
Let's look at what happened to these patients and talk about the methodology. The primary outcome at 24 weeks was this area under the curve. Again, it was like an ACR-20, but it is an integrated ACR-20, and here are the results. The patients on methotrexate had an area under the curve 12.2, etanercept 14.7, slightly significantly better -- but look at the combination, 18.3. No question that the combination is better. Remember this is the area under the curve and there is no question that the combination is much better than methotrexate alone or etanercept alone. So what this tells you, going back to that first discussion, is that in clinical efficacy you get more of a "bang for your buck" by using etanercept plus methotrexate than etanercept or methotrexate alone.
Why is that important? We started using biologics; we replaced the methotrexate or steroids with the biologic; in other words, start the biologic and try to taper steroids and methotrexate. Whether many of us were successful or not is another story, but what this trial begins to tell us is that we were wrong. This trial begins to tell us that in clinical response, the addition of methotrexate plus etanercept is better than either alone. I'm not saying that the combination is synergistic, (although you might make an argument for that), but it is definitely better than monotherapy with either etanercept or methotrexate alone.
Let's look at some other results. Let's go to something you are familiar with, the ACR-20. We skip from 2 weeks to 24 weeks and go to 52 week data. ACR scores of 85 versus 75 versus 76. 85 is on the extreme right, showing you that the combination therapy beats the ACR-20 score of methotrexate alone. The percentage of patients achieving ACR 20's with previous regimens are 60 and 70%, not 85 % as seen in this trial. This was quite phenomenal. For ACR 50's, 69% in combination versus 43 and 48. No doubt that we achieve an ACR-50 patients will do better with the combination. The ACR-70s, no question as seen here, 43 versus 19 and 24. Remember, the ACR-70 obviously includes everybody who is an ACR-50 and a 20. But nevertheless, look at these responses not seen in previous trials. Again although the end point was 24 weeks the 52 week data is essentially the same. I showed you the 24- and the 52-week data, and you can see that there is no question that when you look at ACR-20, 50 or 70, the combination is statistically significantly better than either methotrexate alone or etanercept alone.
HAQ scores at the end of the year -- the improvement in HAQ is much better for the combination group than either the methotrexate alone or etanercept alone groups. Remember a change of 0.22 is clinically significant in HAQ, and well demonstrated. Here we have the combination which is -1.1 from baseline versus -0.8 and -0.9, methotrexate alone versus the combination. Again the combination gives you a better improvement in HAQ than either of the components of that combination alone.
This is where DAS comes in. Remember I said remission is defined as a patient achieving a DAS of less than 1.6. What it looked at is how many of those patients achieved a clinical remission: with methotrexate you had 14%; on etanercept, 18%; the combination, 37%. It is very difficult to achieve. I don't think we've seen this with any other clinical trial.
What is also interesting here is the 2.4 DAS cutoff line. Before I showed you remission. Look at the number of people that developed low activity of disease within 20 weeks of the combination and maintained it for a year. This was not seen in the methotrexate alone arm nor in the etanercept alone arm, but in the combination arm -- a result that is very important and impressive.
The next slide basically goes through all of the little clinical measurements that we do to encompass the ACR-20. What is important here is that if you look at tender joints, morning stiffness, swollen joints, global assessment, all of the components, the combination was statistically better than methotrexate alone or etanercept alone, with the exception only of morning stiffness, and at that point it shows the combination is better than etanercept alone. Every variable is statistically different than the singular components -- very, very impressive.
So what have we learned? Combination therapy was statistically better than methotrexate or etanercept monotherapy. That had never been shown before -- never shown in the trial with infliximab and methotrexate nor in any of the adalimumab trials. So this is a unique trial, which I think will set a standard for the way trials should be done in the future. Combination therapy demonstrated clinical remission rates of 37% as noted by DAS.
It was thought for a long time that we can suppress inflammation, but we don't always suppress damage. Damage more than inflammation correlates with disability. So our new standard right now is not only to control inflammation but control damage as well. For example, do you remember the patient that came into your office week after week who said "I feel better doctor"? You felt the swollen joint count and the sed rate was less. You thought everything was going very well and then the patient would show up a year or two or three later and need a hip replacement. And you say "Why? You were doing well. I controlled everything." But we have learned that you have to control damage, and damage does not correlate, unfortunately, with decreasing the indicators of inflammation such as high sed rate or swollen joints.
So the new bar that we have standardized now is that we want to control damage. The way to control damage is to get an x-ray at the start of a therapy, and I do x-rays every year on my patients. I take them down to my radiologists and ask them to compare. I want to see if any more erosions have developed or any more damage done. I am interested in damage, a new and important measurement.
So even though the TEMPO trial showed you superior clinical efficacy of the combination therapy, did it reproduce those values in terms of x-ray regression or stopping x-ray progression? Usually we take radiographs for each patient and in a trial they are read in a in a random order. There were two independent readers. (Remember in your practice to include x-rays of the feet as well as hands, and by doing so you will be able to approximate 90% of damage in any other joint in the body.) X-rays of hands and feet were obtained at baseline and one year later. What is important is that x-rays of each group beginning the study were similar in terms of Sharp scores. Here are the results on this next curve. Again, the red bar shows you the methotrexate group. What changed from baseline was 2.8 Sharp units and then you have the pluses and minuses to demonstrate the confidence intervals or range of change. Look at the etanercept -- no question that it did somewhat better than methotrexate in slowing x-ray progression. But look at the combination, no question the results were far superior to either methotrexate or etanercept alone. The results are below the baseline, and some will argue that this even represents healing. Healing is something that we still have to define and probably not correct to conclude at this moment, but the importance is to note the difference. There is no question that methotrexate plus etanercept stopped radiographic progression much more effectively than either of the components by themselves.
Joint space narrowing plus erosion scores are components of the total Sharp score Look at patients over a year course, and the joint space narrowing and erosion score follows the same trend. Why do we measure and record joint space narrowing and erosions? Because earlier in this disease, erosions appear more than joint space narrowing. Later you see more joint space narrowing, and we usually do both. In this group of patients with about 6 years of disease, both were slowed.
Now here is the more important question: Can we demonstrate that any of the patients in this study had zero progression, not a decrease but zero, no progression over the year in each group? At 52 weeks, what percentage of patients had no x-ray progression in each group? 57% in the methotrexate group, 68% in the Enbrel alone group and, look at the combination again, 80% of patients had no progression in their x-rays. Again, we see a piece of data that we had not seen with such contrast before. Added to the clinical material and it helps to validate that the combination of this TNF inhibitor, etanercept, plus methotrexate, is much better at preventing damage as well as improving clinical response than either etanercept or methotrexate when given alone. The summary on this slide basically discusses what I just went over for you.
The third part of this discussion: we know it works, we know it's fantastic, but we have heard before about whether the combinations that are toxic. So the question is, is it tolerable? Can patients take methotrexate plus etanercept and have no more side effects than either of the components alone? This is very important.
The first curve I am showing is adverse events. These are mostly nuisance events, and they are listed as those reported in over 10% of the patients. This includes things such as abdominal pain, cough, headache and injection site reaction, nausea, rash and vomiting. I can just ask you to look across the three tables and you will see no asterisk, which means there is no clinically significant difference between each.
If you are adding an immunosuppressive plus another drug which is immunomodulating, now are you getting a higher incidence of serious infection? We define a serious infection as a patient that either gets admitted to the hospital because of their infection or needs an intravenous antibiotic. That means it is serious. So let's look at the serious infections. These are the ones we worry about a lot and you will see that when we relate this to the rheumatoid population, which has a baseline higher level of serious infections (regardless of what medication the patient is on), is there a difference in the group -- 4.4 versus 4.5, versus 4.3. No difference. The addition of methotrexate to etanercept does not cause any more infection or serious infections with either methotrexate alone or etanercept alone -- very important to know. You could argue that there is not enough time in a one year trial to really count the malignancies, but the answer is that the addition of tumor necrosis factor to the methotrexate doesn't give you any more malignancies over that period of time than we have with either drug used alone.
So, in summary, etanercept alone or in combination with methotrexate was generally well tolerated. Serious infections were not increased in the combination, nor were regular infections. No tuberculosis or opportunistic infections were reported. What is interesting in this population is that they weren't required to screen for tuberculosis ahead of time, which is something I don't think any of us would do anymore with a TNF inhibitor. But nevertheless, there was no tuberculosis, and rare cases occur in post marketing use, as you well know. Adverse events were seen more often in the methotrexate group and in combination -- not a surprise -- and malignancies were comparable to that seen in previous studies.
Okay, you could ask the question about clinical efficacy, radiographic efficacy, stopping of damage and side effects. Is there a possibility that the combination could have altered the drug level of either or of the combinations and made a difference? For example, if you add methotrexate to etanercept, do you do something to the level of methotrexate in the blood that would make it lower if you just took methotrexate alone? And would it account for some of the differences? This is answered in this slide, where they look at the serum concentrations of etanercept both in the etanercept and methotrexate group, and the bottom line is there is no difference.
In summary, I will show you this slide which shows you the efficacy variables to demonstrate to you that clinically 85% of patients achieve the ACR-20 response at 52 weeks versus 43 on methotrexate alone. I showed you the DAS, where the percentage of patients achieving remission was 37% -- unbelievable. Change of HAQ score, which is important, and the total Sharp score with radiographic data that we see at one year, the combination of etanercept plus methotrexate resulted in the higher level of efficacy than either therapy alone. Etanercept monotherapy was more effective than methotrexate alone as measured by the area under the curve, but the combination beat the pants off both, and that is I think that is the important message.
Although pharmacokinetics did not interfere, there was some question whether past methotrexate use loaded the deck. In other words, if you had taken methotrexate for three years, are you slightly ahead of the game, so that when you got into the methotrexate + etanercept combination, could you have skewed the results because you had methotrexate on board? We did a sub-analysis to look at the groups that had taken a lot of methotrexate beforehand and those that didn't, and there was no difference. So it appeared that regardless of whether you had previously taken methotrexate or not, you still did better with the combination versus either of the two components alone. Lastly, etanercept was well tolerated when given alone or in combination with methotrexate.
So, in summary, this trial really shows for the first time that a combination regimen is much better than either of the components separately. If you remember our previous combination studies that were conducted years ago, with gold and cyclophosphamide, we were never able to show benefit to combination therapy over mono therapy with DMARDS. This study with these drugs definitively shows it. The indications are that the combination is better and that it is just as safe. The implications are that the combination is better at controlling radiographic damage. And probably the most important implication is that when you have a patient on a combination, you need to think twice about whether you are going to reduce your dose of methotrexate, because you may be doing the patient a disservice, much different than what we talked about before.
Thank you very much.
