Downloaded from adc.bmj.com on February 6, 2014 - Published by group.bmj.com

Leading article

measures involve some subjectivity, and Blinding in pharmacological trials: becomes less so to reduce observer bias for objective criteria.7 However, even the devil is in the details then, the lack of participant or healthcare provider blinding can lead to other pro- 1 2 3 2 Mandy Wan, Mine Orlu-Gul, Helene Legay, Catherine Tuleu blems, such as differential attrition and cointervention bias, which can likewise influence the assessment of clinical trial While the methodological principle of science underpinning blinding is very outcomes.710 ‘blinding’ for minimising bias in randomised much pharmaceutical based, and may be controlled trials (RCT) is widely debated technical at times, but part of this paucity BLINDING WITH PLACEBOS and accepted, rarely is this depth of thinking of knowledge is undoubtedly due to a The use of placebo in RCTs would appear applied to its correct handling with respect certain lack of recognition for the as a seemingly simple experimental to design, reporting and analysis. Beyond pharmaceutical properties of a medicine. scheme. Such perception is so deeply fi the general comprehension of the de nition This article, written with clinicians in embedded that we often implicitly accept of the word, the practical aspects of estab- mind, presents a discussion on the prac- clinical trial reporting with rather loose lishing blinding in investigator-initiated tical considerations for blinding in paedi- descriptions of blinding procedures as pharmacological trials are often grossly atric pharmacological trials, with the aim 1 adequate indication for the success of underestimated, and very much an after- to facilitate paediatricians in improving blinding.23Rarely do we ask ourselves thought. The resulting inadequacy can the success and timely delivery of blinded 1 any questions concerning the placebo, nor lead to trial delayed and increased costs, RCTs in children. Furthermore, this do we think much, if at all, about the and can also potentially place the entire trial article aims to encourage detailed disclos- work involved in producing them. The at risk. ure of blinding methodology in clinical fact is that the seemingly simple concept It is worth nothing that reporting of trial reporting. does not necessarily hold true for obtain- blinding methodology is often absent or ing placebo supply. One cannot just pur- of low quality in published articles of 23 BLINDING chase matching placebo in an RCTs. Thus, in such context, when Blinding in clinical trials refers to the ‘off-the-shelf’ manner; its provision needs these fundamental details of trial design fi 23 process of withholding information about to be speci c to each trial and the chal- are inadequately reported, it becomes the assigned treatment from specificgroups lenge is in the word ‘matching’.To easier to see why it should be of no sur- of individuals. The first blinded experiment achieve its purpose, one must ensure the prise that there are, equally, very few pub- was conducted by Benjamin Franklin who placebo matches the sensory specifications lished commentaries addressing the literally blindfolded participants to shield to that of the medicine under test. The challenges of blinding that those embark- them from knowledge in their assessments ability to maintain blindness throughout ing on pharmacological trials face. So, it of the therapeutic claims made for applying the complete clinical trial must focus on would seem that little is available, at least mesmerism. Quite understandably, the use the visual aspects of the product, and in biomedical literature, to guide investi- of blindfolds is less favourable today. should also take into consideration the gators on this particular methodological Instead, identical-appearing treatments, be other human senses. The relative import- aspect. A review of RCTs published over it matching placebo or masked active com- ance of these specifications will very much fi 15 years in this journal identi ed only parator, are important tools in modern-day depend on the route of administration 176 RCTs in children involving pharma- 4 pharmacological research. and the dosage form concerned; this may ceutical interventions ; the authors sug- When performed correctly, blinding is include: shape, size, colour, texture, fi gested that dif culties in obtaining intended to minimise the occurrence of weight, taste, smell11 and not merely a adequate placebos without the collabor- conscious and unconscious bias in the ‘sugar pill’. ation of pharmaceutical companies may 4 conduct and interpretation of a trial until An appropriate source for supply would have contributed to the low number. Our all such opportunities for bias have be the original manufacturer of the medi- experience over the years has also high- passed. The biases associated with prior cine. After all, they would have manufac- fi lighted, among investigators, a de ciency knowledge of treatment assignment are tured placebo for their own marketing in the awareness of how pharmaceutical well known, and the benefits of blinding authorisation trials, and have all the ramification can constrain trial design and 37 56 have been presented elsewhere Indeed, required technical data on the manufactur- its validity. There is no denying that the it is acknowledged that the relevance of ing and analytical methods to support clin- blinding will vary according to the clinical ical trial application. However, for many 1National Institute for Health Research—Medicines for trial context. In general, blinding of parti- pharmaceutical companies, the little incen- Children Research Network–London & South East, cipants, healthcare providers and outcome tive to manufacture placebos on the com- Evelina London Children’s Hospital, Guy’s and St assessors is considered important in paratively limited quantity required by ’ Thomas NHS Foundation Trust, King s Health Partners, explanatory trials, where the primary most clinical trials is often disappointing London, UK; 2Department of Pharmaceutics, UCL focus is to determine the efficacy of an to the independent researchers.11213 School of Pharmacy, Centre for Paediatric Pharmacy 8 Research, London, UK; 3Faculté Des Sciences intervention under ideal circumstances. Even if one is fortunate enough to have Pharmaceutiques et Biologiques de Lyon, Université By contrast, in pragmatic trials, as in the agreed a supply, commercial influence over Claude Bernard, Lyon, France real-world delivery of care, blinding of protocol design, trial results and publica- Correspondence to Mandy Wan, National Institute participants and healthcare providers are tion can be somewhat disheartening.12 14 for Health Research, Medicines for Children Research sometimes considered not necessary, so as Short of compromising the study design Network—London & South East, Guy’s and St Thomas’ fi ’ ’ to render the ndings more applicable to with an unblinded trial, an alternative NHS Foundation Trust, St Thomas Hospital, St Thomas 8–10 House, 5th Floor, London SE1 7EH, UK; usual care setting. Blinding is also par- approach is to have the matching placebo [email protected] ticularly important when outcome manufactured elsewhere. While it would

656 Wan M, et al. Arch Dis Child September 2013 Vol 98 No 9 Downloaded from adc.bmj.com on February 6, 2014 - Published by group.bmj.com

Leading article be logical to use the same ingredients copying goes, certain imprints or mark- encapsulation can potentially provide the (excipients) in the placebo formulation as ings on tablets are registered trademarks, solution to the problem. A sensible those in its active counterpart (without and therefore exclusive rights are approach would be to consider the double- the active drug), such approach does not bestowed upon the trademark owner, the dummy blinding strategy. With this always guarantee a satisfactory level of drug’s original manufacturer. approach, the comparator doesn’t need to blinding. First, blinding is particularly Consequently, third-party placebo manu- match the test product. Instead, a placebo challenging where the actual drug itself facturing is not always possible. matching the test product and a second has a characteristic taste or colour, neces- placebo matching the comparator product sitating the testing and addition of new BLINDING OF ACTIVE COMPARATORS are used. Although such a study design excipients to the placebo formulation in To many, over-encapsulation may appear may be more pragmatic, it does have some order to achieve sensory equivalent.15 16 to be the ubiquitous solution to blinding drawbacks. On the one hand, there are Second, and rather more problematically, solid oral formulations, and indeed, such considerable technical difficulties, as dis- is when such addition is unsatisfactory; in a technique can be very effective. As the cussed above, with manufacturing placebo these circumstances, it may be necessary term implies, over-encapsulation is basic- for liquid formulation. As two matching to reformulate the active treatment from ally hiding a tablet or capsule inside an placebos are now needed, this would its licensed form (eg, the addition of a opaque capsule shell (often involving the require additional research expense and new excipient to mask the taste of the addition of a backfilled excipient to time. On the other hand, the double- active drug), to what is in effect manufac- prevent rattling), so that the contents are dummy design requires participants to turing of a new product. The decision to concealed, producing products that are take two study medications; in cases where adopt such blinding strategy should, of visually identical. This method of blinding the medicines concerned require multiple course, never be taken lightly. eliminates the need for matching placebo daily doses, this raises the question of the To allow for variable dose delivery as a manufacturing, and is particularly useful potential risk of non-compliance. result of age/weight/surface area-based for active-controlled trials requiring com- For infusion therapy, blinding can trans- dosing, and to facilitate ease of adminis- parator blinding. However, over- late into a somewhat cumbersome setup. tration, liquid oral dosage form is often encapsulation increases the size of the ori- Some researchers have used opaque required in the treatment of children, ginal dosage form, potentially making tubing and the covering of infusion bag making blinding liquid formulation a administration more difficult. The use of with plastic sleeve to mask its content.18 more frequent issue in paediatric setting. such blinding method in a trial involving We question the robustness of such blind- Compared with tablets and capsules, the the younger subset of the paediatric popu- ing method, where we believe that the sensory characteristics of taste and smell lation may therefore be problematic. risk of unblinding through routine hand- are more pronounced, yet, challenging to More importantly, one must remember ling of the infusion administration set is match. Even simple flavourings, such as that a medicine is made up of ingredients substantial. Moreover, when visual inspec- strawberry flavour, can vary in taste other than the active drug, and the effects tion of the infusion bag and its adminis- depending on the manufacturer, and thus of the drug are, in part, dependent on the tration set is a standard safety precaution requiring the need to perform taste assess- properties and proportions of these other measure for nursing staff when adminis- ment studies. There is also the additional components. Thus, it is the dosage form tering infusion therapy, the feasibility and physical attribute of viscosity which one in which the active drug is administered acceptability of such blinding technique must consider. Hence, some degree of for- that determines the overall therapeutic may be limited in practice to allow the mulation development work is almost efficacy and safety of the drug. As over- trial to be delivered successfully. In these always needed in developing a placebo encapsulation brings new ingredients into situations, researchers must acknowledge liquid formulation. the question, and effectively creating a this reality, accepting the unblinded status Likewise, for placebos intended for new dosage form, demonstrating equiva- of those administering the treatment. The topical use, the property of pH deserves lence between the encapsulated and ori- use of unblinded personnel, though particular attention; the differences in pH ginal products becomes a prerequisite for imperfect, does not necessarily lead to an between the two products may lead to ensuring study result integrity.917 unblinded trial. By limiting the activity of local irritation in one group only, and Alternatively, blinding of two substan- the unblinded personnel to that of treat- therefore distinct side-effects profile will tially different looking treatments may be ment administration only, it remains pos- effectively unblind the treatment arm and achieved by bespoke manufacturing both sible to blind other key trial personnel, biased assessments of endpoints may treatments from their respective active and not limited to only outcome assessors. result. pharmaceutical ingredients into either Nonetheless, researchers should anticipate Even if developing the placebo formula- capsules or tablets.9 With this method, the that maintaining blinding by such method- tion was straightforward, difficulties can visual appearance of the two treatments ology will be labour intensive, and that arise with uniquely shaped tablets or pro- can be controlled from the outset, but guarding against inappropriate unblinding ducts with specific markings,1 where the again, there remains the question of must be a high priority. complexity of matching these shapes and equivalent bioavailability, as well as the proprietary designs may require specia- operational challenges in manufacturing THE ‘INERT’ INGREDIENTS lised equipment. The challenge is not two formulations within the resource con- There is certainly a need to consider the simply a matter of cost relating to the straints imposed by finite research suitability of an excipient for use in equipment purchase, but also the fact that funding. placebo manufacturing or in comparator the installation of new equipment will Suppose we now have a blinded active- products masking with respect to its necessitate a lengthy programme of quali- controlled comparator trial of two com- safety. Indeed, the basic considerations fication and validation work. Less obvi- mercially available liquid formulations. will not be different to that of marketed ously, perhaps, is the possible issue of This clearly differs from the case with medicines, with even greater scrutiny in a infringement on trademark law. As far as tablets or capsules, whereby over- paediatric setting in view of the potential

Wan M, et al. Arch Dis Child September 2013 Vol 98 No 9 657 Downloaded from adc.bmj.com on February 6, 2014 - Published by group.bmj.com

Leading article risk of more pronounced safety implica- equivalent or more protective packaging is how much bias might ensue owing to tions. Risk assessments of excipients used used, which itself may be difficult to absence of blinding, and to explore alter- in the context of blinding still apply, and determine, a reduction in the medicine native methods of trial design to minimise on the whole, those with identified risks expiry date cannot be avoided. Although the risk of bias in the trial. More import- relevant to the target age groups con- a reduction in the expiry date from 36 to antly, they should provide clear justifica- cerned should be avoided, as their use 24 months may have little bearing for a tions for their decisions, and report any cannot be reasonably justified for the sole trial of 12 months duration, the effect of speculation on the effect of any factors purpose of blinding. such reduction on a 30 months trial can arising from the course of the trial which Moreover, the use of any excipient be very costly. may impact on the outcome measures. requires additional concern considering Clinical trials, even when most personnel the fact that excipients have proved to be REPORTING are unblinded, can remain methodologic- anything but inert,19 and these possible In light of the challenges discussed above, ally sound in terms of controlling selec- specific effects represent a potential con- it is surprising that such topics have had so tion bias through proper randomisation founding factor that may be vital to inter- little mention in most commentaries. Then and adequate allocation concealment. preting the study results. The again, despite being included in the However, we, like others, take the view appropriateness of using nebulised saline minimum set of recommendations in the that it is almost always possible to ensure as a placebo in bronchodilator studies Consolidated Standards of Reporting that the outcome assessor in a trial is assessing symptoms relief involving Trials (CONSORT) statement,22 the blinded.3 We strongly encourage this patients with respiratory problem has been reporting of the blinding status of key trial aspect of blinding to be implemented, and questioned.20 Similarly, placebos of corn persons, and of the blinding procedures, is if not done, the rationale should be oil or olive oil in cardiac trials have raised still frequently incomplete or missing.23 clearly reported. doubts.21 The question of under what cir- Most of all, when the reporting of basic cumstances it is suitable, if ever, to use a information such as details on the dosage particular excipient will depend on our form of medicine is frequently absent in DISCUSSION current knowledge, and this may change published articles,24 it does beg the ques- The concept of blinding may appear rela- with advances in scientific research. With tion as to whether readers have simply tively simple, but in reality, substantial this in mind, the disclosure of placebo assumed that blinding has been appropri- practical difficulties can arise in construct- composition in clinical trial reporting, at ately controlled for biases without actually ing apparently identical treatment. Just as least in cases where there are differences in knowing how well it worked or even if it you would always discuss with an experi- excipients between the two experimental worked at all. enced statistician when designing a trial, arms, becomes ever so important should While it would be inappropriate to for example, the need for stratification in further excipient data emerge at a later assume inadequate blinding merely on the randomisation or the population SD in date. basis of inadequate reporting, equally, the sample size determination, the same level assumption of adequate blinding when of detailed discussion is needed with PACKAGING reporting is absent or inadequate cannot respect to blinding. Determining the most There is little merit in creating a matching be justified. If most researchers supposedly appropriate blinding strategy is often placebo or masking a comparator, when comprehend the importance of establish- complex; as highlighted by the recurring for example, the active is contained in a ing blinding, then it is only expected that theme of this article, significant timing silver aluminium blister strip whereas they would also want to know whether and cost is involved. As such, blinding white plastic blister has been used to blinding was executed successfully. We strategy deserves particular attention from package the placebo. The requirement for recognised that there are methodological investigators at the earliest stages of trial matching must go beyond the actual pro- uncertainties surrounding testing the planning, preferably before seeking ducts and extends to all associated pack- success of blinding,124but nonetheless, funding. aging and labelling. On occasions, the we support greater disclosure and trans- We recommend that investigators take a need for repackaging of medicines cannot parency in blinding methodology report- balance approach to blinding strategy with be avoided and its implications must be ing in a manner that reflects consideration of a risk-based assessment reviewed at the outset. pharmaceutical considerations. There cer- of clinical, pharmaceutical, logistical and The need for repackaging may not tainly should be explicit reporting on who resource factors. Engaging with experi- always be apparent at the time when the was blinded, with additional information enced individuals, such as research phar- pharmaceutical company agreed to on the methods involved in introducing macists, and initiating dialogue on provide the clinical trial supply. Amidst and maintaining blinding.37Where feas- formulations and blinding early in the the excitement of agreeing the supply, one ible, we also strongly encourage the trial design phase can help to avoid must request for detailed information on conduct and reporting of pretrial evalu- unworkable blinding options, as well as to the packaging to assess its suitability for ation of the identity of the two compared ensure the work required for blinding is the trial. Medicines provided in its com- treatments.325 timely delivered.1 Considering the chal- mercial form are often packaged in a We recognised that blinding of all key lenges discussed, we further recommend branded manner and as such cannot be trial personnel is not always technically or funding bodies to consider making copied. Other than the additional time practically possible, and on occasion, the pharmaceutical expert review a condition and cost for depackaging and repackaging cost of blinding is such that the trial is no of public funding for pharmacological activities, the impact of these undertakings longer feasible even with generous RCTs. At the very least, editors of bio- on product integrity and stability needs to increase in funding. Do we then simply medical journals should mandate explicit be reviewed. The original manufacturer’s disregard blinding? The issue is not blinding reporting be part of articles con- stability data supports only the medicine whether blinding should be disregarded, sidered for publication. It is only through in its original packaging, and unless but investigators need to carefully assess better reporting and critical analysis that

658 Wan M, et al. Arch Dis Child September 2013 Vol 98 No 9 Downloaded from adc.bmj.com on February 6, 2014 - Published by group.bmj.com

Leading article we can build our body of knowledge on non-commercial. See: http://creativecommons.org/ 12 Christensen M, Knop FK. The unobtainable placebo: blinding methodology. licenses/by-nc/3.0/ control of independent clinical research by industry? Lancet 2012;379:30. To cite Wan M, Orlu-Gul M, Legay H, et al. Arch Dis An article of this size cannot possibly 13 Curfman GD, Morrissey S, Drazen JM. Products at Child Published Online First: [please include Day Month address the different approaches to blind risk. N Engl J Med 2010;363:1763. Year] doi:10.1136/archdischild-2013-304037 different dosage forms or the regulatory 14 Lenney W, Perry S, Price D. Clinical trials and framework governing pharmaceutical Received 9 March 2013 tribulations: the MASCOT study. Thorax – manufacturing. Rather, the aim is to intro- Revised 6 June 2013 2011;66:457 8. Accepted 3 July 2013 15 Mahalanabis D, Lahiri M, Paul D, et al. Randomized, duce researchers to some of the key Arch Dis Child 2013;98:656–659. double-blind, placebo-controlled clinical trial of the fi pharmaceutical concepts and associated doi:10.1136/archdischild-2013-304037 ef cacy of treatment with zinc or vitamin A in infants vocabularies. Blinding should, under no and young children with severe acute lower circumstances, be seen as a means to an respiratory infection. Am J Clin Nutr 2004;79:430–6. REFERENCES 16 Morganroth J, Dimarco JP, Anzueto A, et al.A end, when in fact, pharmaceutical proper- randomized trial comparing the cardiac rhythm safety ties of the medicine should drive the 1 Whitham D, Silcocks P, Whitehouse W, et al.An fl fl example of problems that arise from clinical trials of moxi oxacin vs levo oxacin in elderly patients choice of blinding method which, in part, – hospitalized with community-acquired pneumonia. and how to avoid them. Pharm J 2009;283:129 30. – will inform the design and operational 2 Haahr MT, Hróbjartsson A. Who is blinded in Chest 2005;128:3398 406. delivery of the trial. randomized clinical trials? A study of 200 trials and 17 European Medicines Agency. Guideline on the a survey of authors. Clin Trials 2006;3:360–5. requirements to the chemical and pharmaceutical 3 Hróbjartsson A, Boutron I. Blinding in randomized quality documentations concerning investigation Contributors MW—conception, literature search, clinical trials: imposed impartiality. Clin Pharmacol medicinal products in clinical trials. London: Committee drafting of the manuscript, final approval of Ther 2011;90:732–6. for Medicinal Products for Human Use (CHMP), 2006. manuscript; MOG—conception, literature search, 4 Campbell T, Surrey SA, Royle EM. A review of 18 Boutron I. Methods of blinding in reports of randomized critical revision of manuscript; HL—design, literature randomized controlled trials published in Archives of controlled trials assessing pharmacologic treatments: a search, critical revision of manuscript; CT—conception, Disease of Childhood from 1982–96. Arch Dis Child systematic review. PLoS Med 2006;3:e425. critical revision of manuscript. 1998;79:192–7. 19 What are excipients doing in medicinal products? DTB 2009;47:81–4. Competing interests None. 5 Jefferson T, Jones MA, Doshi P, et al. Neuraminidase fl 20 Khan SY, O’Driscoll BR. Is nebulized saline a placebo inhibitors for preventing and treating in uenza in – Provenance and peer review Not commissioned; healthy adults and children. Cochrane Database Syst in COPD? BMC Pulm Med 2004;30:9 84. externally peer reviewed. Rev 2012;(1):CD008965. 21 Golomb BA. Paradox of placebo effect. Nature 6 Karlowski TR, Chalmers TC, Frenkel LD, et al. 1995;375:530. Ascrobic acid for the common cold. A prophylactic 22 Schulz KF, Altman DG, Moher D; for the CONSORT and therapeutic trial. JAMA 1975;231:1038–42. Group. CONSORT 2010 Statement: updated guidelines for reporting parallel group randomised 7 Schulz KF, Grimes DA. Blinding in randomised trials: – hiding who got what. Lancet 2002;359:696–700. trials. Ann Int Med 2010;152:726 32. 8 Sackett DL. Explanatory and pragmatic clinical trials: 23 Pandit S, Shah U, Kirby DJ, et al. Inappropriate oral formulations and information in paediatric trials. Arch a primer and application to a recent asthma trial. Pol – Arch Med Wewn 2011;121:259–63. Dis Child 2010;95:754 6. 9 Martin BK, Meinert CL, Breitner JC, et al. Double 24 Hróbjartsson A, Forfang E, Haahr MT, et al. Blinded placebo design in a prevention trial for Alzheimer’s trials taken to the test: an analysis of randomized – clinical trials that report tests for the success of Open Access This is an Open Access article disease. Control Clin Trials 2002;23:93 9. – distributed in accordance with the Creative Commons 10 Zwarenstein M, Treweek S, Gagnier JJ, et al. blinding. Int J Epidemol 2007;36:654 63. Attribution Non Commercial (CC BY-NC 3.0) license, Improving the reporting of pragmatic trials: an 25 Walter SD, Awasthi S, Jeyaseelan L. Pre-trial evaluation of the potential for unblinding in drug trials: a prototype which permits others to distribute, remix, adapt, build extension of the CONSORT statement. BMJ – upon this work non-commercially, and license their 2008;337:a2390. example. Contemp Clin Trials 2005;26:59 468. derivative works on different terms, provided the 11 Joyce C. Placebos and other comparative treatments. original work is properly cited and the use is Br J Clin Pharmac 1982;13:313–18.

Wan M, et al. Arch Dis Child September 2013 Vol 98 No 9 659 Downloaded from adc.bmj.com on February 6, 2014 - Published by group.bmj.com

Blinding in pharmacological trials: the devil is in the details

Mandy Wan, Mine Orlu-Gul, Helene Legay, et al.

Arch Dis Child 2013 98: 656-659 originally published online July 29, 2013 doi: 10.1136/archdischild-2013-304037

Updated information and services can be found at: http://adc.bmj.com/content/98/9/656.full.html

These include: References This article cites 23 articles, 8 of which can be accessed free at: http://adc.bmj.com/content/98/9/656.full.html#ref-list-1 Open Access This is an Open Access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 3.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/3.0/ Email alerting Receive free email alerts when new articles cite this article. Sign up in service the box at the top right corner of the online article.

Topic Articles on similar topics can be found in the following collections Collections Open access (66 articles) Child health (2262 articles) Clinical trials (epidemiology) (349 articles)

Notes

To request permissions go to: http://group.bmj.com/group/rights-licensing/permissions

To order reprints go to: http://journals.bmj.com/cgi/reprintform

To subscribe to BMJ go to: http://group.bmj.com/subscribe/