Steven R. Feldman, MD, PhD
I love registries! With large numbers of participants, registries can be very helpful to identify rare side effects and to assess the efficacy and safety of medications in populations that may not be fully represented in clinical trials. I also love dupilumab; it was revolutionary in the management of patients with AD.
Vittrup and colleagues have created a registry of 347 participants treated with dupilumab. This does not yet have the large number of participants needed to identify new issues that wouldn't have been detected in clinical trials, but the study is informative about real-life use. The dramatic improvement in the Eczema Area and Severity Index (EASI) score is consistent with the high efficacy of dupilumab. The high rate of treatment persistence is also consistent with dupilumab being a very effective and safe treatment (because if the drug wasn't working well or was causing a severe problem, patients would probably stop the treatment). Though the study reported persistent head and neck involvement, the residual involvement may be quite minimal.
The EASI-75 and Investigator Global Assessment response rates reported in dupilumab trials underestimate the value of this drug. With a 2-year persistence rate of nearly 90%, it's clear that dupilumab is making a huge difference in the lives of patients with AD.
Fatigue is a fascinating issue in AD. We might wonder if all the inflammation in patients with AD would directly cause fatigue. Almost certainly all the itching in AD adversely affects sleep and would cause tremendous fatigue. It surprised me that most of the children in the study by Rangel and colleagues were reported as having no or mild fatigue; severe fatigue was very uncommon. It leaves me wondering whether the assessments of fatigue fully capture what's happening. Also, since the fatigue score was reported by the parents, I (as the parent of a child with AD) am wondering whether the parents were projecting, with the score more reflective of the parents' fatigue than with that of the child; alternatively, perhaps the child's hyperactivity leaves parents thinking there is no fatigue when there actually is (and possibly even causing the perceived hyperactivity).
The lack of a control group without AD is another major limitation in our ability to interpret the study findings. Is fatigue more common or less common in children with AD than in children without AD? I cannot tell from these findings. Does fatigue warrant, as the authors suggest, more attention in clinical practice? I don't know. If we are already treating our patients based on patients' global impressions of how they are doing — combined, of course, with our observations of their objective disease severity — I'm not sure how asking about fatigue would change anything, even if future studies were to definitively show that AD is associated with fatigue.
I hate new drugs (well, maybe not hate, but I worry about unknown long-term risks). Clinical trials that help a drug get approved can tell us a lot about a drug's efficacy, but these studies are generally limited in what they tell us about a drug's safety. Clinical trials are generally not powered enough (not enough participants and not followed for long enough) to be informative about rare risks. I love long-term studies of new drugs in large numbers of people because those studies can be very reassuring about the risks of medications. Studying nearly 10,000 patients for 5 years is quite reassuring, confirming my impression that dupilumab has a remarkable, excellent safety profile (Owji et al). Blocking interleukin 4 and interleukin 13 seems to be very specific to AD. Finding no association to cancer is what I would have expected; being able to share this information with patients is likely to be reassuring to them.
Oh, lord help me, another study that claims we should change our disease management because they've identified an increased risk for something. When you compare 70,000 patients with 270,000 controls, you have huge power to detect statistically significant associations of no clinical consequence. Let's assume for the moment that the detected association the authors found between AD and juvenile idiopathic arthritis (JIA) is real. The odds ratio is 2; the odds ratio for smoking causing cancer is on the order of 100.
In this study, over 99% of individuals in both AD and control groups did not have JIA. The difference between rates of JIA in patients with AD compared with controls was 0.3%! The authors conclude "it is important to inquire actively about symptoms not directly linked to the patients' skin disease"; based on the findings of this study, I would conclude that we don't need to worry about JIA in patients with AD even if there is a (marginally) higher prevalence of JIA in this group.