Randomized clinical trials with run-in periods: frequency, characteristics and reporting

Figures & data

Table

• What is a run-in period?

◦ A run-in period is a time period after inclusion, but before randomization, used to exclude certain patients. Other pre-randomization periods exist, for example, extended screening periods and washout periods. These different pre-randomization periods may overlap in purpose, design and terminology

• What types of run-in periods exist and which patients are excluded?

◦ During the run-in period, all patients receive the same intervention, for example, active treatment, placebo treatment or no intervention. Patients are excluded due to, for example, noncompliance to treatment or data collection, non-response to treatment or response to placebo

• What are the reasons for using a run-in period?

◦ By excluding certain patients, for example, noncompliers or placebo responders, run-in period may increase a study’s power, that is, chance of detecting a potential treatment effect

• What other terms are used for a run-in period?

◦ Similar terms used are lead-in periods, single-blind placebo periods and enrichment periods

• What potential problem does run-in period cause?

◦ The use of a run-in period may affect external validity, by exclusion of patients from the clinical study population, as well as internal validity, by the risk of unblinding or exaggeration of the intention-to-treat effect estimate

• What is needed to assess the possible impact of a run-in period on trial results?

◦ We propose that the study reader would want to study the number of excluded patients, reasons for exclusion and their baseline characteristics. These aspects should, therefore, be reported in the trial publications

Figure 1 Flowchart of screening for randomized clinical trials and for run-in periods.^a

Notes: ^aWe screened PubMed publications from 2014 one by one in random order until we had obtained 25 randomized clinical trials reporting a run-in period. The PubMed query used was: “(randomized controlled trial[Publication Type]) AND (“2014/01/01”[Date - Publication]: “2014/12/31”[Date - Publication])”.

Table 1 General characteristics of randomized clinical trials with and without run-in periodsa

Characteristic	Trials reporting a run-in period (n=25)	Trials not reporting a run-in period (n=100)	P-value
Number of randomized patientsb	217 (133-502)	90 (46-354)	0.01^*
Trial design
Parallel	21 (84%)	90 (90%)	0.31
Crossover	4 (16%)	7 (7%)
Split-body	0 (0%)	3 (3%)
Number of treatment arms	2 (2-3)	2 (2-2)	0.13
Control group comparatorsc
Active	15 (60%)	63 (63%)	0.82
Placebo	15 (60%)	18 (18%)	<0.01^*
Standard therapy	2 (8%)	20 (20%)	0.24
No treatment	1 (4%)	6 (6%)	1.00
Treatment classd
Pharmacological	17 (68%)	51 (51%)	0.18
Non-pharmacological	8 (32%)	49 (49%)
Clinical specialty
Medicine	19 (76%)	44 (44%)	0.02^*
Surgery	3 (12%)	28 (28%)
Other	3 (12%)	28 (28%)
Language
English	25 (100%)	91 (91%)	0.67
Chinese	0 (0%)	6 (6%)
German	0 (0%)	1 (1%)
Russian	0 (0%)	2 (2%)
Industry status
Industry trials	16 (64%)	23 (23%)	<0.01^*
Not industry trials	9 (36%)	59 (59%)
Uncleare	0 (0%)	18 (18%)

Notes:

a For categorical data, absolute numbers are shown, as well as percentages in parentheses and P-values from Fisher’s two-sided exact test. For numerical data, medians are shown, as well as interquartile ranges in parentheses and P-values from the Mann–Whitney U test. The P-value is followed by an asterisk when P<0.05. For the trials without a run-in period, we extracted data from a random sample of 100 trials out of the total 445 trials.

b The number of randomized patients was reported for 24 (96%) of the trials with a run-in period and 98 (98%) of the trials without a run-in period.

c One trial could have more control group comparators.

d Non-pharmacological trials include trials having more treatment arms where at least one of them is non-pharmacological.

e Trials were classified as unclear if no information was available on study designers, funding, sponsorship, support or similar.

Table 2 Characteristics of randomized clinical trials with run-in periods

Trial	Subject: treatment	Industry or not	Study design	Study arms	Number of patients enrolled, excluded in run-in, randomized	Run-in type, duration	Comments on reporting of run-in period	Purpose of run-in period
Angelin et al^Citation9	Hypercholesterolemia: eprotirome	Industry	P	Two drug arms with different doses, one placebo arm	NR, NR, NR	Dietary lead-in, 4 weeks	Number of patients excluded, baseline characteristics and exclusion reasons not reported. Possibly, no exclusions during run-in period, but this is unclear	“Lead-in” to diet, washout
Bleecker et al^Citation10	Asthma: fluticasone furoate/vilanterol	Industry	P	Drug and co-drug vs drug vs placebo	730, 120, 610	Inhaled corticosteroid only, 4 weeks	Baseline characteristics and detailed exclusion reasons not reported	To ensure symptom stability. Baseline data collection. Minimize placebo response
Casabé et al^Citation11	Benign prostatic hyperplasia: tadalafil/finasteride	Industry	P	Drug and co-drug vs placebo and co-drug	NR, NR, 696	Placebo, 4 weeks	Number of excluded patients, baseline characteristics and exclusion reasons not reported	NR
Church et al^Citation12	COPD: umeclidinium and tiotropium	Industry	X	Six drug arms with different doses, one alternative drug arm, one placebo arm	163, 0, 163	NR, NR	No exclusions of patients during run-in. Reporting on exclusion reasons and baseline characteristics deemed complete	NR
De Boever et al^Citation13	Asthma: an anti-IL-13 mAb	Industry	P	Drug vs placebo	NR, NR, 198	Inhaled corticosteroid, 4 weeks	Number of excluded patients unclear (flowchart numbers do not add up). Baseline characteristics and detailed exclusion reasons not reported	Stable symptomatic disease
Diamond et al^Citation14	Endometriosis: elagolix	Industry	P	Two drug arms with different doses, one placebo arm	NR, NR, 155	Placebo, 4 weeks	Number of patients excluded, baseline characteristics and exclusion reasons not reported. Possibly, no exclusions during run-in period, but this is unclear	Minimize placebo effect
Dodick et al^Citation15	Migraine: LY2951742, an mAb to calcitonin gene-related peptide	Industry	P	Drug vs placebo	367, 149, 218	No treatment, 28–38 days	Baseline characteristics and exclusion reasons not reported	Inclusion based on symptoms over 28 days. To ensure compliance to data collection
Fitzpatrick et al^Citation16	CVD in diabetes: intensive weight loss program	Not industry	P	Experimental lifestyle treatment vs standard lifestyle treatment	5,579, 434, 5,145	No treatment, 2 weeks	Baseline characteristics and explicit exclusion reasons not reported. Exclusions during run-in given as percentage not absolute number	Complete self-monitoring
Haab et al^Citation17	Overactive bladder: netupitant, a neurokinin-1 receptor antagonist	Industry	P	Three drug arms with different doses, one placebo arm	325, 79, 246	Placebo, 2 weeks	Baseline characteristics and exclusion reasons not reported	Inclusion based on symptoms. Baseline symptoms data collection
Halmos et al^Citation18	Irritable bowel syndrome: a diet low in FODMAPs or a typical Australian diet	Not industry	X	Experimental diet vs other control diet	45, 1, 44	No treatment, 1 week	One patient withdrew during baseline/run-in period. Missing baseline characteristics, but period was used for baseline characteristic collection. However, some baseline characteristics might have been reported	Baseline data collection
Hanhineva et al^Citation19	Impaired glucose metabolism: diet with whole grain, fatty fish and bilberries	Not industry	P	Two experimental diet arms, one control diet	NR, NR, 131	No treatment, 4 days	Number of patients excluded, baseline characteristics and exclusion reasons not reported. Maybe no exclusions during run-in, but unclear	Baseline data collection
Hoare et al^Citation20	Depression in HIV/AIDS: escitalopram	Industry	P	Drug vs placebo	NR, NR, 105	Placebo, 7±3 days	Number of patients excluded, baseline characteristics and exclusion reasons not reported. Maybe no exclusions during run-in, but unclear	To exclude early placebo responders. To ensure compliance to intervention
Julius et al^Citation21	Hypertension: candesartan	Industry	P	Drug vs placebo	NR, NR, 809	No treatment, 3 weeks	Number of excluded patients, baseline characteristics and separate exclusion reasons not reported	To ensure stable clinical condition: high blood pressure. Possibly to exclude of patients with adverse events
Laurent and Boutouyrie^Citation22	Hypertension: olmesartan	Industry	P	Three drug arms with different doses	202, 69, 133	Placebo, 2 weeks	Baseline characteristics and exclusion reasons not reported	NR
Maneechotesuwan et al^Citation23	Asthma: generic SFC	Not industry	X	Drug vs other drug	NR, NR, 51	Rescue medication only, 2 weeks	Number of patients excluded, baseline characteristics and exclusion reasons not reported. Maybe no exclusions during run-in, but unclear	Washout of previous medication
Marrero et al^Citation24	Type 2 diabetes: lifestyle intervention, metformin (and troglitazone)	Not industry	P	Drug plus standard treatment vs placebo plus standard treatment vs lifestyle (vs other drug, this arm was discontinued)	4,719, 900, 3,819	Placebo, 3 weeks	Baseline characteristics and separate exclusion reasons not reported	To ensure compliance to intervention. To ensure compliance to data collection
Martinez et al^Citation25	COPD: Internet-mediated walking program	Not industry	P	Active non-drug treatment vs waitlist control	307, 68, 239	No treatment, 1 week	Baseline characteristics not reported	Baseline data collection. To ensure compliance to data collection
Mugie et al^Citation26	Functional constipation in children: prucalopride	Industry	P	Drug vs placebo	NR, NR, 215	Laxative or similar if needed, 1–2 weeks	Number of excluded patients, baseline characteristics and separate exclusion reasons not reported	Documentation of symptoms. Washout
Oluleye et al^Citation27	Heart failure: irbesartan	Industry	P	Drug vs placebo	NR, NR, 4,128	Placebo, 1–2 weeks	Number of patients excluded, baseline characteristics and separate exclusion reasons not reported. Maybe no exclusions during run-in, but unclear	Stable clinical condition
Poulsen et al^Citation28	Nutrition: new Nordic diet	Not industry	P	Experimental diet vs control diet	190, 9, 181	Active control treatment, 1 week	Baseline characteristics not reported	Standardization to same diet
Reznik et al^Citation29	Type 2 diabetes: insulin pump treatment	Industry	P	Same drug in two different administration forms	495, 164, 331	Active control treatment, 2 months	Baseline characteristics not reported	Achieve optimum injection treatment. To ensure compliance to data collection
Saneei et al^Citation30	Childhood metabolic syndrome: DASH diet	Not industry	X	Experimental diet vs standard diet	60, 0, 60	No treatment, 2 weeks	Seemingly no exclusions	Baseline data collection. Learning data collection
Siproudhis et al^Citation31	Fecal incontinence: NRL001	Industry	P	Three drug arms with different doses, one placebo arm	NR, NR, 466	No treatment, 2 weeks	Design paper. No data reported even though study enrollment is finished	NR
van Gool et al^Citation32	Urinary incontinence: bladder training, oxybutynin	Not industry	P	Drug vs other treatment vs placebo	NR, NR, 97	No treatment, 3 months	Number of patients excluded, baseline characteristics and exclusion reasons not reported. Maybe no exclusions during run-in, but unclear	To ensure symptom stability: not regained continence
Zhu et al^Citation33	Hypertension: telmisartan/amlodipine	Industry	P	Drug and co-drug vs co-drug	381, 57, 324	Active control treatment, 6 weeks	Explicit baseline characteristics not reported. States that full analysis set was similar to treated set and to run-in set of patients, but data not shown	To exclude responders to A5 monotherapy lowering the blood pressure. To ensure compliance to intervention

Abbreviations: CVD, cardiovascular disease; mAb, monoclonal antibody; NR, not reported; P, parallel; SFC, salmeterol/fluticasone combination; X, crossover.

Table 3 Characteristics of the run-in period in the 25 trials reporting such a period

Characteristic	Measurea
Patient flowb
Number of patients enrolled to run-in period	325 (190-495)
Number of patients excluded during run-in period	69 (9–149)
Percentage of patients excluded during run-in period	16% (5%–24%)
Duration of run-in period in daysc	14 (11–28)
Intervention during the run-in period
Control (placebo)	7 (28%)
Control (active)	3 (12%)
Experimental intervention	0 (0%)
Other interventiond	5 (20%)
No intervention	9 (36%)
Unclear^e	1 (4%)
Main purposes of the run-in periodf
Extended screening, symptom stability	7 (28%)
Baseline data collection	5 (20%)
Compliance to data collection	3 (12%)
Exclusion of responders to non-active treatment	2 (8%)
Unclear	8 (32%)

Notes:

a For categorical data, absolute numbers are shown, as well as percentages in parentheses. For numerical data, medians are shown, as well as interquartile ranges in parentheses.

b Patient flow numbers were reported for 13 out of 25 trials (52%).

c Run-in period duration was reported for 24 out of 25 trials (96%).

d For example, inhaled corticosteroid in an asthma trial where this treatment was not an intervention arm during the randomized phase.

f Some trials reported more than one rationale of the run-in period. Here, we present the purpose most prominently mentioned. For example, four additional trials had compliance to data collection or compliance to run-in intervention noted as a secondary purpose.

Table 4 Completeness of reporting on the run-in periodsa

Characteristic	Number (percentage)
Overall level of reporting
Complete reportingb	2 (8%)
Incomplete reportingc	23 (92%)
Reports number of patients excluded during run-in period
Yes	13 (52%)
No	12 (48%)
Reports reasons for exclusion of patients
Yes	7 (28%)
No	18 (72%)
Reports baseline characteristics for excluded patients
Yes	2 (8%)
No	22 (88%)
Uncleard	1 (4%)

Notes:

a Absolute numbers are shown, as well as percentages in parentheses. A trial run-in period was completely reported if the publication clearly described 1) the number of excluded patients, 2) reasons for exclusion and 3) baseline characteristics of the excluded patients.

b Two publications reported trials with run-in periods, but stated to have no exclusions during this phase. They were, therefore, deemed to be completely reported on all three criteria.

c One publication was a rationale and design paper reporting on a trial where patient recruitment and randomization was completed. The publication did not present information on the trial’s 2-week run-in period. We deemed reporting to be incomplete for all three criteria.

d Data were not presented, but excluded and randomized patients were stated to be similar.

Table S1 Industry status in trials with and without run-in periodsa

Type of analysis	Trials reporting a run-in period (n=25)	Trials not reporting a run-in period (n=100)	P-value
Main analysis
Industry trials	16 (64%)	23 (23%)	<0.01*
Not industry trials	9 (36%)	59 (59%)
Unclear	0 (0%)	18 (18%)
Sensitivity analysis, broader definition
Industry trials	22 (88%)	41 (41%)	<0.01*
Not industry trials	3 (12%)	38 (38%)
Unclear	0 (0%)	21 (21%)
Subgroup analysis, pharmacological trials
Industry trials	16 (94%)	18 (35%)	<0.01*
Not industry trials	1 (6%)	28 (55%)
Unclear	0 (0%)	5 (10%)

Notes:

a Absolute numbers are shown, as well as percentages in parentheses and P-values from Fisher’s two-sided exact test. The P-value is followed by an asterisk when P<0.05. For the trials without a run-in period, we extracted data from a random sample of 100 trials out of the total 445 trials. Trials were classified as unclear in the main analysis if no information was available on study designers, funding, sponsorship, support or similar. In a sensitivity analysis, we reclassified half of the trials with unclear industry status as industry trials and the other half as not industry trials. In that analysis, the difference between the two groups remained significant (P<0.01). With the broader definition, trials were classified as industry trials if they received any support, financial or otherwise, from a commercial company.

AngelinBKristensenJDErikssonMReductions in serum levels of LDL cholesterol, apolipoprotein B, triglycerides and lipoprotein(a) in hypercholesterolaemic patients treated with the liver-selective thyroid hormone receptor agonist eprotiromeJ Intern Med2015277333134224754313

 PubMed Web of Science ®Google Scholar

BleeckerERLötvallJO’ByrnePMFluticasone furoatevilanterol 100-25 mcg compared with fluticasone furoate 100 mcg in asthma: a randomized trialJ Allergy Clin Immunol Pract20142555356125213048

 PubMed Web of Science ®Google Scholar

CasabéARoehrbornCGda PozzoLFEfficacy and safety of the coadministration of tadalafil once daily with finasteride for 6 months in men with lower urinary tract symptoms and prostatic enlargement secondary to benign prostatic hyperplasiaJ Urol2014191372773324096118

 PubMed Web of Science ®Google Scholar

ChurchABeeraheeMBrooksJMehtaRShahPDose response of umeclidinium administered once or twice daily in patients with COPD: a randomised cross-over studyBMC Pulm Med2014141224393134

 PubMed Web of Science ®Google Scholar

De BoeverEHAshmanCCahnAPEfficacy and safety of an anti-IL-13 mAb in patients with severe asthma: a randomized trialJ Allergy Clin Immunol2014133498999624582316

 PubMed Web of Science ®Google Scholar

DiamondMPCarrBDmowskiWPElagolix treatment for endo-metriosis-associated pain: results from a Phase 2, randomized, double-blind, placebo-controlled studyReprod Sci201421336337123885105

 PubMed Web of Science ®Google Scholar

DodickDWGoadsbyPJSpieringsELSchererJCSweeneySPGrayzelDSSafety and efficacy of LY2951742, a monoclonal antibody to calcitonin gene-related peptide, for the prevention of migraine: a Phase 2, randomised, double-blind, placebo-controlled studyLancet Neurol201413988589225127173

 PubMed Web of Science ®Google Scholar

FitzpatrickSLJefferyRJohnsonKCBaseline predictors of missed visits in the look AHEAD studyObesity201422113114023996977

 PubMed Web of Science ®Google Scholar

HaabFBraticeviciBKrivoborodovGPalmasMZufferli RussoMPietraCEfficacy and safety of repeated dosing of netupitant, a neurokinin-1 receptor antagonist, in treating overactive bladderNeurourol Urodyn201433333534023765630

 PubMed Web of Science ®Google Scholar

HalmosEPChristophersenCTBirdARShepherdSJGibsonPRMuirJGDiets that differ in their FODMAP content alter the colonic luminal microenvironmentGut20156419310025016597

 PubMed Web of Science ®Google Scholar

HanhinevaKLankinenMAPedretANontargeted metabolite profiling discriminates diet-specific biomarkers for consumption of whole grains, fatty fish, and bilberries in a randomized controlled trialJ Nutr2015145171725527657

 PubMed Web of Science ®Google Scholar

HoareJCareyPJoskaJACarraraHSorsdahlKSteinDJEscitalopram treatment of depression in human immunodeficiency virus/acquired immunodeficiency syndrome: a randomized, double-blind, placebo-controlled studyJ Nerv Ment Dis2014202213313724469525

 PubMed Web of Science ®Google Scholar

JuliusSEganBMKacirotiNANesbittSDChenAKTROPHY InvestigatorsIn prehypertension leukocytosis is associated with body mass index but not with blood pressure or incident hypertensionJ Hypertens201432225125924275841

 PubMed Web of Science ®Google Scholar

LaurentSBoutouyriePVascular Mechanism CollaborationDose-dependent arterial destiffening and inward remodeling after olmesartan in hypertensives with metabolic syndromeHypertension201464470971625001274

 PubMed Web of Science ®Google Scholar

ManeechotesuwanKAssawabhumiJRattanasaengloetKSuthamsmaiTPipopsuthipaiboonSUdompunturakSComparison between the effects of generic and original salmeterol/fluticasone combination (SFC) treatment on airway inflammation in stable asthmatic patientsJ Med Assoc Thai201497Suppl 3S91S10024772585

 PubMedGoogle Scholar

MarreroDPanQBarrett-ConnorEImpact of diagnosis of diabetes on health-related quality of life among high risk individuals: the Diabetes Prevention Program outcomes studyQual Life Res2014231758823709097

 PubMed Web of Science ®Google Scholar

MartinezCHMoyMLNguyenHQTaking healthy steps: rationale, design and baseline characteristics of a randomized trial of a pedometer-based Internet-mediated walking program in veterans with chronic obstructive pulmonary diseaseBMC Pulm Med20141411224491137

 PubMed Web of Science ®Google Scholar

MugieSMKorczowskiBBodiPPrucalopride is no more effective than placebo for children with functional constipationGastroenterology201414761285129525239590

 PubMed Web of Science ®Google Scholar

OluleyeOWRectorTSWinSHistory of atrial fibrillation as a risk factor in patients with heart failure and preserved ejection fractionCirc Heart Fail20147696096625225241

 PubMed Web of Science ®Google Scholar

PoulsenSKDueAJordyABHealth effect of the New Nordic Diet in adults with increased waist circumference: a 6-mo randomized controlled trialAm J Clin Nutr2014991354524257725

 PubMed Web of Science ®Google Scholar

ReznikYCohenOAronsonRInsulin pump treatment compared with multiple daily injections for treatment of type 2 diabetes (OpT2mise): a randomised open-label controlled trialLancet201438499501265127224998009

 PubMed Web of Science ®Google Scholar

SaneeiPHashemipourMKelishadiREsmaillzadehAThe Dietary Approaches to Stop Hypertension (DASH) diet affects inflammation in childhood metabolic syndrome: a randomized cross-over clinical trialAnn Nutr Metab2014641202724686130

 PubMed Web of Science ®Google Scholar

SiproudhisLJonesDShingRNWalkerDScholefieldJHLibertas Study ConsortiumLibertas: rationale and study design of a multicentre, Phase II, double-blind, randomised, placebo-controlled investigation to evaluate the efficacy, safety and tolerability of locally applied NRL001 in patients with faecal incontinenceColorectal Dis201416Suppl 1596624499497

 PubMed Web of Science ®Google Scholar

van GoolJDde JongTPWinkler-SeinstraPMulti-center ran-domized controlled trial of cognitive treatment, placebo, oxybutynin, bladder training, and pelvic floor training in children with functional urinary incontinenceNeurourol Urodyn201433548248723775924

 PubMed Web of Science ®Google Scholar

ZhuDGaoPHoltbrueggeWHuangCA randomized, double-blind study to evaluate the efficacy and safety of a single-pill combination of telmisartan 80 mg/amlodipine 5 mg versus amlodipine 5 mg in hypertensive Asian patientsJ Int Med Res2014421526624391142

 PubMed Web of Science ®Google Scholar