The improvement of MAO-B inhibitors in Parkinson’s disease is clinically irrelevant: a review

Clinical relevance of MAO-B inhibitors 2 ARC Publishing Introduction The monoamine oxidase-B inhibitors (MAO-B) have a prominent place in the treatment of Parkinson's disease (PD), especially in the early stages of the disease, both for their contribution to the delay of Levodopa (LD) treatment initiation, and for their much talked neuroprotective role [1-3]. However, questions have been raised for their effectiveness as antiparkinsonian medication [4]. Although several studies demonstrate statistically significant effects of MAO-B inhibitors [5-17], whether they have clinical significance remains to be seen, as a clinical trial may identify a small but statistically significant change in an outcome measure that may have little or no relevance to whether a patient actually feels improved. Thus, concern arises from the question: what is the magnitude of improvement that can be recognized as such by the clinician and/or, most importantly, by the patient as such? In the current study, we used the Minimal Clinically Important Difference (MCID) as a measurable size for evaluation of an outcome as clinically significant and applied it to the results of some of the largest studies with the most widely used MAO-B inhibitors, Rasagiline, Selegiline and Safinamide. Finally, we compared the results of the MAO-B inhibitors with those of Dopamine Agonists (DA) and LD. Studies which calculate the Minimal Clinically Important Difference (MCID) The most widely accepted method of assessment of the clinical signs and symptoms of PD patients is the Unified Parkinson's Disease Rating Scale (UPDRS) [18-20]. More specifically, UPDRS Part III (Motor Examination Part) is used to evaluate the antiparkinsonian effect of the intervention on the patients' motor abilities, which are, after all, the most important aspect of their everyday lives affected by the disease. Alternatively, in LD-treated, advanced disease patients, OFF-time reduction during 24h can be used as a measure [21, 22]. According to Jaeschke et al. [23] MCID is defined as the minimum change that can be recognized and appreciated by the clinician, while it is also important for the patient. The methods of MCID calculation are the anchorand distribution-based approaches [24, 25]. The anchor-based approach requires an external standard that is simultaneously independent, interpretable and has a clinical relevance and correlation with the object being surveyed [21]. The most widely accepted anchor for PD patients for MCID determination is the Clinician-rated Global Impression of Improvement (CGI-I) scale (1=very much improved; 2=much improved; 3=minimally improved; 4=unchanged; 5=minimally worsened; 6=much worsened; 7=very much worsened) [21, 22, 26-28]. Determining the MCID done either by calculating mean change on the UPDRS among patients with score 4 (unchanged) and 3 (minimally improved) on the CGI-I scale or by Receiver Operating Characteristic (ROC) curve analysis were developed, where cutoff values for UPDRS and OFF-time changes best distinguished minimally improvement patients from those with no change [22, 29], estimated as the point on the ROC curve closest to (0,1), calculated as the minimum value of the square root of (1 sensitivity)2 + (1 specificity)2. Distribution-based approach is based on the distribution of the measure in the study population, but is mainly used to calculate the effect size [26, 28]. The challenges arising from determining the MCID are not few. First, these methods can give different MCID, even on the same sample, while, conversely, the same method can give different MCIDs in different samples [22, 29]. Furthermore, Hauser et al. [21, 22] pointed out the importance of the placebo effect in the MCID, while highlighting that the MCID varies from drug to drug, as studies using drugs with larger efficacy tend to calculate larger MCIDs. Schrag et al. [27] used two double-blind clinical trials comparing DA Ropinirole with LD and Bromocryptine, respectively. The study concluded on a Clinically Important Difference of -8.0 points of mean change in the UPDRS Total score and -5.0 in the UPDRS Part III. However, neither those trials were placebo-controlled, nor the drugs' doses were stable over time. Shulman et al. [28] performed a cross-sectional observational study of patients and concluded to an MCID of -4.1 to -4.5 for UPDRS Total and -2.3 to -2.7 for UPDRS Part III. Nevertheless, whether these results can be applied on clinical trials remains to be seen. Rascol et al. [30] based on a double-blind, randomized Rasagiline study calculated an MCID of -1.52 for the UPDRS Part III. Nonetheless, these results were extracted by evaluation of both Rasagiline and Placebo groups. Horvath et al. [26] estimated the MCID for UPDRS Part III up of -3.25 based on within-patients score method and -3.5 based on ROC analysis. The study included all disease severity degrees, all kinds of medications but no placebo. The studies by Hauser et al. [21, 22] are the most relevant for our discussion, since their data are derived from clinical trials that include a placebo control group, include patients in early PD as well, evaluate an intervention that is mildly efficacious, utilize also subject-rated impression of change and, finally, used data from two double-blind, randomized Rasagiline studies. According to Hauser's anchor-based analysis (using both mean change within patients and ROC curve analysis) the lower limit for the MCID is -2.0 for UPDRS Part III, -3.8 for UPDRS Total and -1.0 hours (h) for reduction in OFF-time [21]. Moreover, Hauser et al. [22] using data from two randomized, double-blind, placebo controlled-trials comparing DA Pramipexole (PPX), either extended (ER) or immediate (IR) release in either advanced or early PD patients suggest that data derived using patient-rated self-impression of change as in Patient-rated Global Impression of Improvement (PGI-I; 1=very much better;