I have to admit that I don’t do these summaries of Arthritis & Rheumatism just for my readers. It’s also a way of forcing myself to at least glance through the journal every month and see what’s going on in the arthritis research world. I’m more likely to do it and remember it if I write something.

So here’s the November summary – a little late (the December issue is already available on line) and with the usual caveats (I summarize what interests me, which is mostly research in epidemiology and things related directly to the clinic). All links require login to access.

There is a very interesting article by Schneeweiss, et al dealing with risk of gastrointestinal (GI) and cardiovascular (CV) complications of NSAIDs. The interesting thing from my point of view is the method by which they did the analysis – they used something called an instrumental variable – but I think this is just too technical to go into here. The major finding though, was to confirm that NSAIDs cause GI bleeding (nothing new) and that some of them increase the risk CV disease (rofecoxib, aka Vioxx – old news – and diclofenac, aka Voltaren – new news). If you’re taking diclofenac and have a history of or risk factors for heart disease, this may be an article to bring up with your doctor.

Zink, et al looked at the use of TNF-inhibitors (the so-called ‘biologics’: etanercept aka Embrel, adalimumab aka Humira, and infliximab aka Remicaide) in a chohort of patients with rheumatoid arthritis. This is an interesting article because it looks at an issue that is well known to clinical researchers: patients in clinical trials are not the same as patients in a typical practice. Clinical trials select subjects who tend to have few other diseases/problems, and they do this because researchers are trying to show that the drug works and they don’t want a bunch of extraneous factors like other diseases getting in the way to ‘confound’ the results. This makes good sense. But it means that the patients in seen in a typical practice tend to be sicker and more complicated than those in the trials. The question then is whether the treatment works as well in these sicker patients as it does in the subjects of the trial.

Zink confirmed that 1) in their database of patients with RA only about 25% would have qualified for a TNF-inhibitor trial and that 2) the responses of patients in the database seemed to be comparable to those in a trial, though it was true that those patients who would not have qualified fared less well. This is reassuring though not major news.

Witt et al presented a randomized controlled trial of acupuncture for patients with hip osteoarthritis (OA). There have been several trials of acupuncture for OA in recent years – my impression (and it’s just an impression) is that good sham-controlled studies don’t find much effect. In this one there was no sham control group (i.e. patients were completely aware of what treatment they got – other trials use a ‘sham’ control meaning that the subjects think they’re getting acupuncture even though they’re not) and treatment was ‘individualized’ for each patient meaning that there was no one standard treatment. I haven’t looked at the results yet, but I bet the trial shows that acupuncture works. … And the verdict is …

Well, it’s a little hard to say. The main result they presented was comparison of those receiving acupuncture including those who refused to be randomized but got acupuncture anyway, and those not receiving acupuncture. In this case, acupuncture seemed better than control at 3 months. This is garbage. If people who got to choose their treatment assignment are included in an analyis then it’s not a controlled clinical trial and misses the entire point of doing this kind of study. By the way, the difference between groups disappeared at 6 months. When they just looked at the randomized patients, at 3 months there was still a difference, which again disappeared by 6 months. This is also crap and fulfills my prediction. Why was I right? Because the treatment groups were not blinded.

Blinding is a very important part of any randomized controlled trial and must be done if at all possible. If it’s not, then those receiving the treatment are more likely to believe it will be effective and report their outcome that way. In addition, if the investigators know who got which treatment at the end of the study, they’re more likely to assign a favorable outcome to those who got the treatment under study. End result? An entirely inadequate study which is subject to way too much bias to be at all useful. This study was a waste of someone’s money.

Binham et al examined risedronate (Actonel) in patients with knee OA. To make a long story short, risedronate did not improve symptoms of OA or radiographic progression of disease. (Risedronate is used to treat osteoporosis very effectively, and there were theoretic reasons to think it might treat/slow progression of OA.)

Zhang et al showed that Chinese people in Beijing appear to have a lower prevalence of chondrocalcinosis than white people in the USA. Chondrocalcinosis is an X-ray finding that is associated with the presence of calcium pyrophosphate crystals (CPPD) in the joint. This can cause something called pseudogout which – surprise, surprise – is a disease that gives joint pain and swelling that can looks just like gout. An interesting finding though not of much immediate practical value.

An interesting study by Tseng et al suggests that treating some lupus patients with short courses of steroids when they have blood tests that suggest they could have a disease flare in the near future could be of benefit. This is a preliminary study and I’m going to have to go through it in some detail, but if true it would be very nice to know that there is a reasonably effective way to treat some lupus flares based just on blood tests. We generally can’t do this now.

Another interesting lupus study by Clowse et al supports the idea of maintaining pregnant women with lupus on hydroxychlorine (HCQ – also called Plaquenil) to prevent flares. HCQ, according to their data, appears to be safe for the fetus, and appears to prevent flares in the mom during pregnancy. HCQ is one of our backbone drugs in lupus, and pregnancies in lupus are often accompanied by flares, so this is a very valuable finding.

There is a study by Finckh et al which suggests that silica and solvent exposure may be associated with developing lupus, but I’m going to withold comment on this one until I’ve reviewed it in detail. These can be very difficult studies to conduct and are subject to a lot of bias, so just reporting the results based on the abstract is not a good idea. I’ll try to get back to you on this one.

I think those are the highlights from my point of view. There are a few other moderately interesting articles that I have to read, but I think there’s plenty here for one post. I have, of course, skipped most of the basic science articles (labs and rats). I’ll get back to you if there’s more.

Coming soon: the Dec roundup!