© 1996 Nature Publishing Group http://www.nature.com/naturebiotechnology • FEATURE VACCINE DEVELOPMENT TIMES Vaccine R&D success rates and development times
The real time saving in vaccine development is not in the preclinical development itself, but in the accelerated clinical development and reduced regulatory process of dossier preparation and review.
Mark-M. Struck
Two of the most important factors in compa• give life-long protection against disease from database1' . Success rate statistics were derived nies' decisions about whether or not to very few administrations. However, because from a total of 266 projects-not all projects, develop vaccines---the time of development vaccines are given to healthy people, adverse just those for which reports could be traced and the success rate in development-are events associated with vaccinations must be in the literature, from company reports, or relatively poorly characterized. This article lower than for drugs. Although vaccines other sources. In addition, I included 14 vac• endeavors to remedy that by providing esti• comprise a specialized pharmaceutical mar• cine projects (7 launched, 3 discontinued, 4 mates of these factors from data on the ket, they can achieve substantial sales (for ongoing) from our company; these are not progress of vaccine projects between 1983 instance, the hepatitis B vaccine). listed in Pharma Projects but have been pub• and 1994. Similar data have also been used to lished elsewhere" ". predict more generally the growth rate for The analysis In this survey, vaccine candidates are con• the pharmaceutical and biopharmaceutical I have designated seven stages of develop• sidered to have failed if a project has been industries'". Combined with other consider• ment through which any vaccine project has terminated or suspended. The rare with• ations, such as return on investment, price, to progress to reach the market place: pre• drawals of vaccines from the market in the potential market volume, achievable vaccine clinical development; clinical phases I,II, and postmarketing surveillance period have been coverage, and acceptance and distribution', III; preregistration; registration; and launch. excluded from the analysis'•. such data can aid rational decision processes There is a limited preclinical phase in which It is difficult to estimate the extent to in development. The data indicate that to evaluate potential toxicological, pharma• which failed projects are underreported in although vaccine development-from pre• cokinetic, and pharmacodynamic effects'. publicly available databases'". However, there clinical studies to launch-takes an average Then follow clinical studies in phase I and is some reassurance in the fact that the inclu• of two years less than biopharmaceutical phase II', in which investigators look at sion of data on 14 extra vaccine projects development, only half as many preclinical immunogenicity, generate limited safety from our company did not significantly alter vaccine candidates make it through the data, and study adjuvant effects, dosage, and the transition probabilities and development process. The disease prevention-driven vaccination schemes. Subsequent clinical times. approach underlying vaccine development, phase III trials evaluate efficacy in field stud• and the nature of the data that its clinical tri• ies, household contact studies, or challenge Progression through the pipeline als yield, account for much of the differences study settings10• The next phase is preregis• The numbers of projects per development in the development pattern. tration, during which data is evaluated, ana• stage (Table 1) decreases along the research Alarming reports of waning immunity in lyzed, and compiled for submission to and development pipeline, as expected. adults', the emergence of new diseases, and regulatory agencies for review"·". Once regu• The first interesting aspect of the phase the resurgence of old ones'·' have put vacci• latory approval has been granted, batches of transition data (Table 2) is the similarity of nation back in the public spotlight. This has vaccines are released by prompted numerous vaccination initiatives the national control Table 1. Breakdown of vaccine projects per development worldwide, including Article 129 of the authority and market stage. European Community treaty of Maastricht. launch is initiated. Change in stage Successful Suspended Discontinued Vaccination is the most elegant and cost• Of course, vaccine Preclinical to phase I 27 3 111 effective way to prevent outbreak of a disease development does not Phase I to phase II 17 3 32 and sequelae resulting from it. The Mercer always follow this scheme Phase II to phase Ill 16 0 15 Report commissioned by the US Department exactly (for reasons dis• Phase Ill to preregistration 5 0 10 of Health and Human Services (Washington, cussed previously for Preregistration to 2 0 2 registration DC) found that vaccines offer very high eco• biopharmaceuticals1). Registration to launch 22 nomic and health returns on investment'. Multicomponent vac• 0 They often contain multiple active ingredi• cines, especially, may not ents (in contrast to single active ingredients always fit into this Table 2. Transition probabilities of vaccines compared to used in drugs or biopharmaceuticals) and scheme"·". biopharmaceuticals. The analysis here is Transition Vaccine Biopharmaceutical based on 577 vaccine Preclinical to phase I 0.57 0.57 Mark-M. Struck is head of registration and projects between 1983 Phase I to phase II 0.72 0,88 regulatory affairs at Swiss Serum and Vaccine, and 1994 using source Phase II to phase Ill 0.79 0.86 Phase 111 to registration 0.71 0.93 79, CH-3018 Bern, data from the publicly Inc. , Rehhagstrasse Registration to launch 0.96 1.00 Switzerland. available Pharma Projects
NATURE BIOTECHNOLOGY VOLUME 14 MAY 1996 591 © 1996 Nature Publishing Group http://www.nature.com/naturebiotechnology FEATURE VACCINE• DEVELOPMENT TIMES the preclinical to phase I transition probabili• development, the product has only been test• pharmaceutical groups for the phase III to ties for vaccines and biopharmaceuticals and ed in healthy volunteers up to this point. preregistration and preregistration to regis• the fact that the probability in both cases is Although surrogate markers of efficacy such trations phase are statistically significant (by much higher than for "non-bio" pharmaceu• as serum titers, animal challenge data, or the Student t-test). ticals. For both vaccines and biopharmaceu• neutralizing antibodies for vaccines are mon• The overall development (preclinical to ticals, species specificity means that there is itored in early development phases, they may launch) time for vaccines is less-by roughly little animal toxicological data that could be misleading as indicators of the real efficacy two years-than biopharmaceuticals and stop a project from developing past preclini• of the vaccine. Pharmaceuticals and biophar• "non-bio" drugs; ten years compared with cal studies. Equally, pharmacological investi• maceuticals will have been given to affected twelve. Two years less in development means gations with vaccines are rarely performed' patient subsets in phase I or II". Therefore, two years earlier into the market, a signifi• so projects are unlikely to fail on that count. the decision to put vaccines into phase III is cant advantage in cost-saving, sales genera• The other main difference is that "lead struc• based on generally less decisive data. tion, and market positioning. tures" for vaccines and biopharmaceuticals are much easier to identify at the beginning The chances of ultimate success Conclusions of the preclinical phase than are drug leads. So what are the chances that a vaccine at a Vaccine development, like the development We can know at the outset that an attenuated given stage of development will make it to the of other pharmaceuticals, is still a complex organism has a chance of being a vaccine or market? Market entrance probabilities (Table process that takes substantial resources and that a protein will correct a protein-deficien• 3) can be calculated by multiplying the vari• requires serious and enthusiastic commit• cy disease in a way that we cannot know it ous phase transition probabilities. The mar• ment over a period of roughly ten years. for small molecule drugs. Thus, proportion• ket entrance probabilities for preclinical The contexts for vaccine and biopharma• ally fewer "hopeless" vaccine and biophar• vaccine candidates is about half that for pre• ceutical development are very different. The maceutical candidates will enter the clinical biopharmaceuticals1, largely, of general clinical settings and study endpoints preclinical phase. course, because of the lower phase I to phase for vaccines are clearer from the start and In later phases transitions, there are dif• II and phase III to registration transition seem to be better controlled than for bio• ferences between vaccines and biopharma• probabilities. In other words, preclinical bio• pharmaceuticals. Researchers already know ceuticals (see Table 2), differences that are pharmaceuticals are nearly twice as likely to about the disease, the responsible pathogens, statistically significant despite the sparseness succeed than preclinical vaccine candidates. and the immune response elicited in humans of some of the data. Vaccine projects have a The strikingly high success rate for bio• during infection. Disease staging is not a 72% chance of progressing from phase I to pharmaceuticals is explained, partially at common feature of vaccine development. phase II, and a 71 % chance of going from least, by the fact that many of the biophar• There are also far fewer indications for phase III to registration. For biopharmaceu• maceuticals developed and marketed vaccine than for other pharmaceuticals. New ticals, the likelihoods of progression were between 1983 and 1991 have been substitutes indications for vaccines do not spring forth statistically significantly higher (Chi-squared or replacements for natural molecules in from elaborate clinical research, and vaccine test) at 88% and 93%, respectively. Twenty humans-less risky markets. The road ahead candidates cannot be rescued from other percent more vaccine projects than biophar• for biopharmaceuticals may be rockier. indications. Thus, vaccine development is maceutical projects are lost at these two highly focused on its clinical indication from phases. All other transition probabilities are Development time the beginning in a way that drug develop• within the same range. How long does it take to develop vaccines ment may not always be. With its target The lower probability for vaccines getting and bring them on the market? The average highly defined, vaccine development is from phase I to phase II may be explained by time for all development phases for vaccines already the rational process that pharmaceu• the fact that this first human exposure to the appear shorter than for biopharmaceuticals tical development would like to be. vaccine candidate will reveal grounds for (Table 4) with phase III to preregistration The total development time is shorter for project failure such as hyperattenuation of and preregistration to registration taking the vaccines than for biopharmaceuticals. One organisms or viruses, insufficient immuno• shortest time. The difference between the major time saving occurs in the phase III to genicity, or insufficient attenuation". development times of the vaccine and bio- preregistration transition. Herc, the most The probability that vac• pertinent difference between vaccines cines will make the transition Table 3. Market entrance probabilities for vaccines and and other pharmaceuticals is the inci• from phase II to phase III biopharmaceuticals. dence of disease with which investiga• trans1t:1on probabilities is Transition Vaccines Biopharmaceuticals' tors have to work. High incidences slightly lower than that for Preclinical to launch 0.22 0.4 and high enrollment rates make short biopharmaceuticals. One Phase I to launch 0.39 0.71 clinical trials feasible. For vaccines, explanation may be the num• Phase II to launch 0.54 0.80 incidence is pathogen-related. There• ber of variables that are intro• Phase Ill to launch 0.68 0.93 fore, in challenge studies, incidence duced into phase II vaccine Registration to launch 0.96 1.00 can be set as desired. And in field trials: The candidate is studies, investigators can select loca• administered to more patients Table 4. Duration of development phases for vaccines and tions where incidence is sufficiently using different dosing biopharmaceuticals. high to allow rapid conclusion of effi- schemes and dosages, various Stage Vaccines Biopharmaceuticals cacy trials. Furthermore, as healthy routes of administration, and Preclinical to phase I 2.4 2.3 volunteers are included in the study, different adjuvants. Phase I to phase II 2.0 1.8 enrollment in clinical studies is gener• That vaccines should fail Phase II to phase Ill 1.8 2.2 ally not a limiting factor. For new with a higher probability to go Phase Ill to pre-registration 1.4 2 chemical entities, average duration of Preregistration to registration 1.1 from phase III to registration 2 clinical phase III is roughly 3-5 Registration to launch 1.3 1.6 years". For vaccines (in this survey) it than biopharmaceuticals is Total 10.0 11.9 unsurprising. In vaccine is 1.4 years.
592 NATURE BIOTECHNOLOGY VOLUME 14 MAY 1996 © 1996 Nature Publishing Group http://www.nature.com/naturebiotechnology • FEATURE VACCINE DEVELOPMENT TIMES
The other main source of time saving in authorities, EFPIA, Avenue Louise 250, Brussels. design, and social value, pp 135-139 in Vaccine vaccine development is in the transition 15. Pharmaceutical Projects, I. Hutton (Ed.). PJB Publi• design: The subunit and adjuvant approach, Powell, cations, London. M.F. and Newman, M.J. (Eds.). Plenum Press, New from preregistration to registration. Because 16. GIOck, R. et al. 1992. lmmuno-potentiating reconsti• York. there are no large toxicological and pharma• tuted Influenza virus vlrosome vaccine delivery sys• 21 . Tacket, C.O. et al. 1993. Safety and immunogenicity tem for immunization against hepatitis A. J. Clln, of live oral cholera candidate CVD110, a DctxA Dzot cological studies in vaccine development, Invest. 90:2491-2495. Dace derivative of EITor Ogawa V. choleras, J. regulators expect less preclinical data in the 17. Levine, M.M. et al. 1988. Safety, lmmunogenlclty, Infect. Dis. 188:1536-1540. and efficacy of recombinant live oral cholera vac• 22. Pharmaceutical R&D Statistical Sourcebook 1995, final dossier. Fewer variables and "cleaner" cines, CVD 103 and CVD 103-HgR. Lancet 2:467- Mathieu, M.P. (Ed.). Parexel International Corpora• endpoints also mean that the compilation of 470. tion. dossiers for regulatory submission take less 18. Levine, M.M. et al. 1990. Comparison of enteric• 23. Gupta, R.K. and Siber, G.R. 1995.
1. Struck, M.-M. 1994. Biopharmaceutical R & D suc• cess rates and development times. Bio/Technology 12:674-677. 2. Drews, J. 1995. The impact of cost containment on IDle lays in submitting projects to the lark Report pharmaceutical research and development, in 10th Center for Medicines Research Annual Conference FDA cost compani es time, money Results are presented in a com• Lecture, June 1995. and competitive advantage prehensive report that can 3. Mercer Report. 1995. Report on the United States in the market place. Vaccine Industry, June 14. Department of Health be incorporated into an and Human Services, Washington, DC. app li cation to be sub• 4. World Health Organization. 1995. Weekly Epidemio• lark FDA Submission mitted to the FDA. logical Record 70:252-255. Quality Sequencing 5. Efstratiou, A 1995. Diphtheria in Europe: An update. PHLS Microbiol. Digest 12:233-235. La rk specializes in Minimize the risk of 6. Jenum, P.A., et al. 1995. Immunity to diphtheria in handling FDA delays on FDA northern Norway and northwestern Russia. Eur. J. Submission Clin. Microbiol. Infect. Dis. 14:794-798. Submission projects. 7. Desselberger, U. 1995. Emerging infectious diseases Quality DNA Contact Lark for PHLS Microbial. Digest 12:141-144. Sequencing pro• additional informa• 8. Bussiere, J.L., McCormick, G.C., and Green, J.D. jects and can help tion or a quote on 1995. Preclinical safety assessment Considerations in vaccine development, pp. 61-79 in Vaccine you minimize co tly your next FDA Design: The subunit and adjuvant approach, Powell, delays. Submi ssion Quality DNA M.F. and Newman, M.J. (Eds.). Plenum PreFs, New Sequencing project. York. 9. Fast, P.E., Sawyer, L.A., and Wescott, S.L. 1995. Clini• lark Guorantee cal considerations in vaccine trials with special ref• Projects completed by Lark for FDA erence to candidate HIV vaccines, pp. 97-134 in ubmission are guara nteed to be 100% Vaccine design: The subunit and adjuvant approach, Powell, M.F. and Newman, M.J. (Eds.). Plenum accurate even in areas of high GC con• Press, New York. tent or difficul t secondary structure 10. Orenstein, W.A. et al. 1984. Field evaluation of vac• wi th I 00% double stranded coverage Lark cine efficacy, World Health Organization, Geneva, Switzerland. of the seq uenced region TECH OLOGIES 11 . Davenport, L.W. 1995. Regulatory considerations in I c. vaccine design, p.81 -96 in Vaccine Design: The lark Quality Control subunit and adjuvant approach, Powell, M.F. and U A: 9545 Kary Fwy, re. 465 Newman, M.J. (Eds.). Plenum Press, New York. All work performed at Lark i done in Housron, TX 77024-9870 12. Parkman, P.O. and Hardegree, M.C. 1994. Regula• compliance with Good Laborato1y 1-800-288-3720 or 7 13-464-7488, tion and testing of vaccines, p. 889-901 in Vaccines, Fax: 713-464-7492 Plotkin, S.A. and Mortimer, E.A. (Eds.). W.8. Saun• Practices (GLPs) as published by the Email: [email protected] ders Company. FDA. Europe: Tel: +44 (0) 1273-707500. 13. Food and Drug Administration. 1995. Points to con• READER INQUIRY NO .• Fax: +44 (0) I 27.~-707499, sider for the evaluation of combination vaccines, Email: Scqucnce@lsirkO A.CiryScapc.co.uk Office of Vaccines Research and Review, Center for Biologics Evaluation and Research (CBER), FDA, © 1996. Lark Technologies, Inc. All righrs reserved. Lark is a rcgisccrcd rrademark of Rockville, MD. L1rk Technologies, Inc. 14. European Vaccine Manufacturers Association. 1995. Pharmaceutical aspacts of combined vaccines: EVM view to be considered as a basis for discussion with
NATURE BIOTECHNOLOGY VOLUME 14 MAY 1996 593