Original ©2014 Dustri-Verlag Dr. K. Feistle ISSN 0946-1965 DOI 10.5414/CP202124 e-pub: ■■month ■■day, ■■year Received February 2, 2014; accepted April 25, 2014 Correspondence to Andrea Messori, ■■■Dr.? MD? HTA Unit, Area Vasta Centro Toscana, Regional Health System, Via San Salvi 12, 50100 Florence, Italy andrea.messori.it@ gmail.com Key words meta-analysis – equiva- lence – stress ulcer – omeprazole – pantopra- zole Intravenous proton pump inhibitors for stress ulcer prophylaxis in critically ill patients: determining statistical equivalence according to evidence-based methods Andrea Messori 1 , Valeria Fadda 3 , Dario Maratea 3 , Roberta Gatto 1 , Sabrina Trippoli 1 , Mauro De Rosa 3 , and Claudio Marinai 2 1 HTA Unit, 2 Department of Pharmaceutical Logistics, ESTAV Toscana Centro, Regional Health Service, Florence, and 3 The BW Project, Sifact, Società Italiana di Farmacia Clinica e Terapia, Milan, Italy Abstract. Background: Although intrave- nous proton pump inhibitors (PPIs) are con- sidered at least as effective as H2-receptors an- tagonists for stress ulcer prophylaxis (SUP) in critically ill patients, there is no data on whether there is also the proof of no difference among these agents. Methods: The clinical material was the same as that reported in previous me- ta-analyses and included all trials comparing intravenous PPIs vs. H2-receptor antagonists for SUP in critically ill patients. Our method- ology was a combination of meta-analysis and equivalence testing based on confidence inter- vals (CIs). The end-point was the rate of overt bleeding. All PPIs evaluated in the included trials were separately studied. The equivalence margins were derived from power calculation data of the original trials. Results: Our analysis involved 8 randomized trials for 851 patients. Two comparisons were made (pantoprazole vs H2-receptor antagonists and omeprazole vs H2-receptor antagonists). The following RDs were estimated: pantoprazole, RD = –1.2%, 95% CI: –3.5% to +1.2%; omeprazole, RD = –3.0%, 95% CI: –7.2% to +1.3%. The 95% CIs confidence intervals for RDs remained within the ± 6% margin. These results indicate that intravenous pantoprazole and intravenous omeprazole are equivalent, Conclusion: These two PPIs, when administered by intravenous route, are equivalent according to reasonable equivalence margins. This conclusion can be the basis to develop local acquisition tenders on these drugs. One advantage of this approach is that the feasibility of administrative decisions can directly be tested on clinical grounds and on the basis of standard evidence-based meth- ods. Introduction In interpreting randomized trials, the differentiation between no proof of differ- ence (an inconclusive result) and proof of no difference (demonstration of equiva- lence) is attracting an increasing interest in both theoretical [1, 2] and practical terms[3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16]. This differentiation can be made using trial- sequential analysis [3, 4, 5, 6, 7, 8, 9, 10, 11] or by application of statistical testings based on confidence intervals (CIs) [12, 13, 14, 15, 16]. Both of these approaches require that a margin (or delta) is pre-specified to define the minimal improvement in the end-point that is thought to be clinically relevant. In the Italian national health system (NHS), local tenders have long represented the standard tool for medicines procurement. In December 2012, however, a national regu- lation (“Decreto Balduzzi” [17]) has been is- sued to accelerate, across different regions, the patients’ access to reimbursed medicines of our National Formulary. In practice, how- ever, this regulation turned to be an obstacle for local tenderings run by the NHS; in this way, in fact, all tenderings involving “ho- mogeneous” drug classes have been man- datorily subjected to a preventive technical authorization by our national Agency for Medicines. Hence, AIFA has been assigned the responsibility of declaring which drugs classes are “equivalent” and can therefore be managed through local tenderings involving different agents of the same pharmacological class. This new regulation is a tough challenge in operational terms, but has also attracted the attention of evidence-based researchers because specific criteria are being sought for these demonstrations of equivalence. Initial International Journal of Clinical Pharmacology and Therapeutics, Vol. ■■ – No. ■■/2014 (1-5) • 202124Mess / 10. June 2014, 10:13 AM