Response to: Oba,Y. and Lone,N. (2013) Efficacy and safety of roflumilast in patients with chronic obstructive pulmonary disease: a systematic review and meta-analysis. Ther Adv Respir Dis 7: 13–24.

In the February 2013 issue of Therapeutic Advances in Respiratory Disease, we were pleased to see the article entitled ‘Efficacy and safety of roflumilast in patients with chronic obstructive pulmonary disease: a systematic review and meta-analysis’ [Oba and Lone, 2013]. This article is valuable for educating readers on the therapeutic benefits of roflumilast in reducing moderate-to-severe exacerbations in chronic obstructive pulmonary disease (COPD) patients with chronic bronchitis and a history of exacerbations; in support of this, we would like the published information to be accurate and would like to suggest the following points for consideration.

The statement that roflumilast efficacy ‘appears modest compared with other available therapies’ (pp. 13 and 22) is misleading because it implies that all drugs used to manage COPD will function comparably in all patient populations. Since COPD is a complex disease comprising many different symptoms and phenotypes [Agusti et al. 2010], not all patients respond equally to available therapies with many different mechanisms of action [Miravitlles et al. 2012; Vestbo et al. 2013]. Roflumilast is not a bronchodilator and should not be compared with other bronchodilators; rather, it is the first and only available oral phosphodiesterase-4 (PDE4) inhibitor and has demonstrated efficacy in reducing exacerbation risk specifically within patients with severe COPD associated with chronic bronchitis and a history of exacerbations [Calverley et al. 2009]. Any comparison of roflumilast efficacy outside of its target patient population is therefore inappropriate.

Similarly, comparisons of roflumilast to theophylline should be made with caution. Roflumilast selectively inhibits PDE4 [Hatzelmann and Schudt, 2001], the PDE isoenzyme present in many cell types implicated in the COPD inflammatory response [Diamant and Spina, 2011], and is recommended by the Global Initiative for Chronic Obstructive Lung Disease (GOLD) to reduce exacerbations in patients with chronic bronchitis, severe and very severe COPD, and frequent exacerbations inadequately controlled by long-acting bronchodilators [Vestbo et al. 2013]. In contrast, theophylline acts as a weak bronchodilator, a weak nonselective inhibitor of PDE isoenzymes 1–11, and has inconsistently demonstrated nonbronchodilator actions [Barnes, 2003; Vestbo et al. 2013]. However, due to its narrow therapeutic index and the availability of more effective bronchodilators, theophylline use in COPD patients is not recommended if inhaled long-acting bronchodilators are available [Vestbo et al. 2013].

The statement that ‘the risk–benefit ratio of the entire class of orally active PDE4 inhibitors appears to be in question’ (p. 21) is also deceptive. While cilomilast development was terminated due to inconsistent results and an unfavorable risk–benefit ratio [Rennard et al. 2008; USFDA, 2003], this is not the case with roflumilast. In the US Food and Drug Administration (FDA) evaluation of the efficacy and safety of roflumilast, the agency had access not only to individual publications cited by this article, but also to large data sets of >12,000 patients and detailed patient-level data to assess rare but potentially significant events that cannot be evaluated in individual studies. Based on these data, regulatory agencies for 61 countries (including the United States and the European Union) have concluded that roflumilast has a favorable risk–benefit balance for the approved indication and patient population [Forest Laboratories, Inc., 2010].

The authors are correct in their observation that serious adverse events (AEs) such as atrial fibrillation and suicidality were rare and more common with roflumilast than placebo. With respect to atrial fibrillation, close examination of patients in the roflumilast COPD safety pool who experienced this serious AE found that all events (placebo, n = 9; roflumilast, n = 24) were judged by investigators to be either unlikely related or not related to study medication [Forest Laboratories, Inc., data on file]. Furthermore, results of 24-hour Holter electrocardiogram monitoring of 55 patients (M2-124) demonstrated no between-group differences in heart rate or occurrence of arrhythmias [Forest Laboratories, Inc., 2010]. With respect to suicidality, it should be noted that a comprehensive re-examination of possible suicide-related AEs in the roflumilast COPD safety pool using the Columbia Classification Algorithm of Suicide Assessment [Posner et al. 2007] was provided to the FDA prior to marketing authorization. FDA psychiatry consultants confirmed that this re-evaluation [Rosebraugh, 2011] did not identify any possible suicide-related AEs beyond those already identified in the original FDA submission [Forest Laboratories, Inc., 2010]. The FDA ruled that roflumilast had an acceptable risk–benefit ratio for market authorization, and the product labeling states that the risks/benefits of roflumilast treatment should be carefully weighed in patients with a history of depression and/or suicidal thoughts or behavior [Forest Laboratories, Inc., 2011]. This evidence is consistent with and lends further credence to the authors’ suggestion (p. 22) that such atrial fibrillation and suicidality events as observed in the roflumilast development program may have happened merely by chance.

An additional cardiovascular (CV) safety analysis was omitted from this article. All deaths and serious nonfatal CV events in the roflumilast COPD safety pool were evaluated by an independent adjudication committee blinded to study and treatment [White et al. 2013]. This analysis found that a composite measure of major adverse CV events (including all CV deaths, nonfatal myocardial infarctions, and nonfatal strokes) was significantly lower with roflumilast versus placebo (hazard ratio 0.65; 95% confidence interval 0.45–0.93; p = 0.019), suggesting not only a lack of CV safety signal associated with roflumilast treatment, but also a potential for roflumilast-associated CV benefits for evaluation in future controlled clinical trials. In addition to ongoing pharmacovigilance activities for roflumilast, a prospective, double-blind, postmarketing study is currently underway, which includes a component to further characterize major CV effects (if any) associated with roflumilast.

Finally, two recently presented abstracts suggest a possible contradiction to the authors’ statement that ‘roflumilast has not been shown to reduce hospitalizations’ (p. 22). A budget impact model assessing the impact of adding roflumilast to existing COPD treatment suggests decreased medical costs with roflumilast, driven in large part by a reduction in inpatient hospitalizations [Sun et al. 2012]. Post hoc analysis of the pivotal trials (M2-124/125) has also demonstrated that roflumilast versus placebo significantly reduces the rate of severe exacerbations leading to hospitalization [Bateman et al. 2012]. Additional real-world healthcare utilization studies are currently being conducted to assess the impact of roflumilast in reducing hospitalization and readmission rates.

We understand that authors and editors make every effort to provide scientifically rigorous and concise reporting and that inaccuracies in such reporting may occur. At the same time, it is important to identify and correct these inaccuracies so that medical professionals can make informed prescribing decisions. Thank you for the opportunity to bring these points to your attention.