Communicating Comprehensive Safety Data Gained from Clinical Trials to the Scientific Community: Opportunities and Difficulties from an Example with Moxifloxacin P.M. Tulkens, 1 P. Arvis, 2 F. Kruesmann, 3 1 Université catholique de Louvain, Louvain Drug Research Institute, Brussels, Belgium 2 Bayer Santé SAS, Loos, France 3 Bayer Pharma AG, Wuppertal, Germany 2 Nov 2012 12th ISoP, Cancun, Mexico 1
The problem (in general terms) Comprehensive safety data assembled from clinical trials (phase I trough 4) and from pharmacovigilance are communicated to Regulatory Authorities These rarely appear in detail in publicly available literature (that often focuses mainly on efficacy) Yet, even if rare, the corresponding adverse effects are included in the labeling and, as such, must be taken into account by clinicians This creates disconnection between labeling and daily clinical perception uneasiness amongst clinicians (who may feel they are shown only the tips of potentially important safety issues). 2 Nov 2012 12th ISoP, Cancun, Mexico 2
The specific situation of moxifloxacin Moxifloxacin (MXF) is approved in up to 123 countries for major indications (e.g., communityacquired pneumonia [CAP], acute exacerbations of chronic bronchitis [CB], pelvic inflammatory disease [PID], skin and skin structure infections [SSSI] and complicated intraabdominal infections [ciai]) 140 million prescriptions have been issued for MXF worldwide MXF is included as an effective alternative in many guidelines Beyond known class effects of fluroquinolones, moxifloxacin has been suspected to cause cardiac toxicity (known 6-10 msec QTc prolongation) hepatotoxicity (based on rare reports and signals from PSURs). In 2008, EMA imposed a labelling change: due to safety concerns (hepatic, cardiac [in women and elderly patients], and intestinal problems), moxifloxacin should only be used when other antibiotics cannot be used or have stopped working' 2 Nov 2012 12th ISoP, Cancun, Mexico 3
The approach Objective: examine and compare the safety profile of MXF vs that of the comparators (COMP; all selected as reference therapies), providing unbiased information for comparable clinical situations Method: in-depth analysis of the manufacturer s clinical trial database for all actively controlled Phase II IV clinical trials (except one exploratory study) all approved routes of administration and all main indications including patients at risk (hepatic, renal cardiac, age, diabetes, low BMI, ) recording all treatment emergent adverse events (s), adverse drug reactions (s), serious adverse events (s), serious adverse drug reactions (s), premature discontinuations due to s, premature discontinuations due to s, s with fatal outcome, and s with fatal outcome. coding according to the Medical Dictionary for Regulatory Activities (MedDRA) detection of rare events using Standard MedDRA Queries (SMQs) and customized Bayer Medra queries) descriptive statistics (crude rates), with calculation of relative risk estimates (95% confidence intervals [Mantel Haenszel analysis stratified by study with constant continuity correction term of 0.1). 2 Nov 2012 12th ISoP, Cancun, Mexico 4
Results (1) Population: MXF: 14 981 vs. COMP: 15 023) Double blind: 75% IV and IV/PO (sequential): 29% no meaningful difference between MXF and COMP for age, sex, BMI, race, indications, and pre-existing risk factors (renal or hepatic impairment, diabetes mellitus, cardiac disorders, low BMI). distribution mirroring the different main indications (with corresponding risk factors) 2 Nov 2012 12th ISoP, Cancun, Mexico 5
Results (2) global comparisons 2 Nov 2012 12th ISoP, Cancun, Mexico 6
Results (3) - global comparisons, and were mainly "gastrointestinal disorders" and "changes observed during investigations" such as asymptomatic QT prolongation). Incidence rates of hepatic disorders, tendon disorders, surrogates of QT prolongation, serious cutaneous reactions and Clostridium difficile-associated diarrhoea were similar with moxifloxacin and comparators. 2 Nov 2012 12th ISoP, Cancun, Mexico 7
Results (4): patients at risk PO sequential IV age (> 65 y) n = 2551 vs. 2403 n = 1373 vs. 1334 n = 170 vs. 191 1050 / 1021 929 / 900 83 / 81 440 / 448 348 / 307 27 / 31 207 / 184 298 / 290 32 / 24 16 / 18 49 / 30 4 / 6 116 / 109 131 / 104 10 / 10 78 / 74 62 / 42 4 / 6 29 / 32 100 / 98 13 / 10 death. 3 / 1 2 / 3 0 / 1 diabetes n = 777 vs. 717 n = 926 vs. 917 n = 80 vs. 72 355-310 587 / 565 42-35 158-126 196 / 174 13-14 78-56 198 / 182 16-11 11-3 22 / 11 2-2 34-26 78 / 64 6-6 22-14 38 / 20 1-4 10-6 46 / 23 9-4 death. 2 / 2 2 Nov 2012 12th ISoP, Cancun, Mexico 8
Results (5): patients at risk PO sequential IV renal impairment n = 1283 vs. 1229 n = 889 vs. 863 n = 203 vs. 218 1283-1229 572-549 102-92 259-229 196-181 31-32 94-80 202-180 26-22 9-9 30-23 2-1 49-53 75-78 11-7 27-33 28-25 2-3 12-14 58-67 10-7 death. 0-3 3-3 hepatic impairment n = 146 vs. 163 n = 183 vs. 196 n = 46 vs. 46 69-70 183-196 23-18 37-32 43-43 7-6 5-7 60-53 7-7 1-1 10-7 6-7 24-24 1-1 6-3 11-7 2-4 14-24 2-0 death. 0-1 1-2 2 Nov 2012 12th ISoP, Cancun, Mexico 9
Results (5): patients at risk PO sequential IV cardiac disorders n = 1476 vs. 1404 n = 1476 vs. 1136 n = 106 vs. 104 707-655 804-804 63-57 340-297 315-293 16-25 132-110 251-246 23-11 14-8 43-35 3-2 70-64 119-96 7-3 43-45 59-43 1-1 11-25 69-75 11-8 death. 0-2 3-4 0-1 BMI < 18 n = 318 vs. 365 n = 116 vs. 115 n = 45 vs. 53 113-171 89-83 17-10 70-96 26-27 5-3 11-28 36-30 3-3 0-5 5-4 14-27 10-11 12-20 6-9 3-5 15-15 death. 2 Nov 2012 12th ISoP, Cancun, Mexico 10
Results (6): drug comparisons A. oral therapy 1. moxifloxacin vs -lactams risk factor: age > 65 y (n= 909 vs 788) diabetes (n = 282 vs 217) renal impairment (n = 347vs 380) hepatic impairment (n = 47 vs 53) cardiac disorders (n = 526 vs 444) BMI < 18 (n = 70 vs 76) 71-50 3-0 death 2. moxifloxacin vs macrolides risk factor: age > 65 y (n = 1252 vs 942) diabetes (n = 329 vs 255) renal impairment (n = 484 vs 427) hepatic impairment (n = 44 vs 64) cardiac disorders (n = 794 vs 623) BMI < 18 (n = 110 vs 114) death 2 Nov 2012 12th ISoP, Cancun, Mexico 11
Results (7): drug comparisons B. sequential therapy 1. moxifloxacin vs -lactam alone risk factor: age > 65 y (n= 440 vs 422) diabetes (n = 562 vs 506) renal impairment (n = 329 vs 324) hepatic impairment (n = 89 vs 73) cardiac disorders (n = 438 vs 406) BMI < 18 (n = 40 vs 36) 332-276 112-75 71-50 13-3 52-28 24-6 death 2. moxifloxacin vs -lactam alone or combined with a macrolide risk factor: age > 65 y (n = 223 vs 235) diabetes (n = 69 vs 99) renal impairment (n = 168 vs 161) hepatic impairment (n = 37 vs 42) cardiac disorders (n = 175 vs 168) BMI < 18 (n = 25 vs 25) 3-0 death 2 Nov 2012 12th ISoP, Cancun, Mexico 12
C. intravenous therapy 1. moxifloxacin vs -lactam Results (8): drug comparisons risk factor: age > 65 y (n= 92 vs 90) diabetes (n = 46 vs 33) renal impairment (n = 91 vs 85) hepatic impairment (n = 31 vs 35) cardiac disorders (n = 70 vs 61) BMI < 18 (n = 10 vs 6) 52-35 2-0 2-0 6-0 4-0 2-0 death 2. moxifloxacin vs another fluroquinolone risk factor: age > 65 y (n = 60 vs 74) diabetes (n = 27 vs 30) renal impairment (n = 77 vs 86) hepatic impairment (n = 7 vs 5) cardiac disorders (n = 32 vs 38) BMI < 18 (n = 26 vs 37) death 2 Nov 2012 12th ISoP, Cancun, Mexico 13
Conclusions (for moxifloxacin) The overall safety profile of moxifloxacin was found similar to that of comparators from clinical trials More specifically, and with regard to recent questions: Hepatic events reactions were very low and not superior in a statistically significant manner to comparators even if considering patients with hepatic disorders While QTc prolongation were observed, no increase clinical adverse effects were seen even in patients with prexisting cardiac disorders vs. the comparator(s) Specific toxicities (tendonitis, e.g.) remained exceedingly rare with no difference between moxifloxacin and the fluroquinolone comparator Skin events were extremely are and less frequent than with -lactams Full details are available from Tulkens et al. Moxifloxacin safety: an analysis of 14 years of clinical data. Drugs R D. 2012 Jun 1;12(2):71-100 (open access). 2 Nov 2012 12th ISoP, Cancun, Mexico 14
Pros and Cons of this approach Pros Unbiased (randomized) comparison of treatments with similar indications and target populations (all clinically-valid comparators) Estimation of the true incidence of relatively rare effects (equal balance of patients for known and unknown factors) Detailed assessment of the detected side-effects and documented causality Cons Populations analyzed potentially not representing the true final populations in which the drug is used Patients with known contraindications excluded by study design Number of patients too low to detect very rare effects Labor intensive process that can only be undertaken late in drug development and commercialization This approach may be useful for providing clinicians and regulators with a global analysis of actual risk factors for comparable drugs in comparable indications 2 Nov 2012 12th ISoP, Cancun, Mexico 15