Guidance for Industry
Acne Vulgaris: Developing
Drugs for Treatment
This
draft guidance, when finalized, will represent the Food and Drug
Administration’s (FDA’s) current thinking on this topic. It
does not create or confer any rights for or on any person and
does not operate to bind FDA or the public. You can use an
alternative approach if the approach satisfies the requirements
of the applicable statutes and regulations. If you want to
discuss an alternative approach, contact the FDA staff
responsible for implementing this guidance. If you cannot
identify the appropriate FDA staff, call the appropriate number
listed on the title page of this guidance.
This document is intended to provide guidance
to the pharmaceutical industry on the development of drug products
for the treatment of acne vulgaris other than nodulocystic acne.
The information presented will help applicants plan clinical
studies, design clinical protocols, implement and appropriately
monitor the conduct of clinical trials, collect relevant data for
analysis, and perform appropriate types of analyses of study data.
This guidance does not address systemic
retinoid therapies, which may not have appropriate risk-benefit
profiles for non-nodulocystic acne therapy. Development programs
for these treatments should be discussed with the review division
before initiation.
The recommendations in this guidance are
based on careful assessment of important issues raised in the
review of clinical trials for acne vulgaris. Applicants are
encouraged to discuss development plans with the review division
before embarking on studies to ensure that the clinical trial
design and analysis plan meet defined objectives. The FDA’s
guidance for industry Format and Content of the Clinical and
Statistical Sections of an Application
and ICH E9
contain additional information.
FDA’s guidance documents, including this
guidance, do not establish legally enforceable responsibilities.
Instead, guidances describe the Agency’s current thinking on a
topic and should be viewed only as recommendations, unless
specific regulatory or statutory requirements are cited. The use
of the word should in Agency guidances means that something
is suggested or recommended, but not required.
Acne vulgaris is a chronic disease of
sebaceous follicles that is multifactorial in etiology and varies
in severity as evidenced by lesion type, size, numbers, scarring,
and post-inflammatory pigmentary changes. The severity of acne
vulgaris (i.e., amount of inflammation and number of lesions) can
wax and wane in a given patient. A variety of drug products,
topical and systemic, are currently available to treat acne. Acne
occurs more frequently on the face, but can also occur on
nonfacial skin (e.g., back, shoulders, chest).
There are two major types of acne lesions:
noninflammatory and inflammatory. Although most drug products for
acne are intended for the broad indication of acne vulgaris, some
products have been developed that only target one of these two
specific subsets of acne vulgaris lesions.
Noninflammatory lesions of acne are the open
(blackheads) or closed (whiteheads) comedones. Closed comedones
may be more difficult to detect visually and may require
stretching of the skin to aid in visualization. These lesions,
especially closed comedones, may be precursors to the larger
inflammatory lesions and therefore are of clinical importance.
Inflammatory lesions are divided into
papules, pustules, and nodules/nodulocystic lesions, depending on
the severity and location of the inflammation within the dermis.
The papules and pustules have surrounding halos of erythema
allowing for their characterization as inflammatory. Nodules are
typically erythematous and often tender and/or painful.
Additionally, they are deep-seated in the skin (i.e., centered in
the dermis or subcutis). Nodules have been defined as being
greater than 5 mm in diameter. The borders of these lesions may
be difficult to determine because of the associated erythema/inflammation.
There is no single uniform, standardized, and
reproducible grading system for severity of acne. Acne severity
is dependent on the numbers and types of lesions present and the
extent of involvement (i.e., the body areas affected). (See
Section III.A.2., Baseline Lesion Counts.)
A number of acne grading systems have been
described, some with inherent difficulties with regard to use.
Such difficulties include inadequately defined grades of severity
or excessively small differences between grades to be objectively
evaluated. An Investigator’s Global Assessment (IGA) scale that
may be useful for assessing overall acne severity is described in
Section IV.A.1.
Acne vulgaris is primarily a disease
beginning with and extending beyond puberty, but can persist past
the third decade of life. Although acne vulgaris affects both
genders, severity may be greater in male patients. Acne vulgaris
occurs in all races and across the United States.
The Agency recommends that phase 2 clinical
studies provide sufficient information to optimize dose and
duration of treatment chosen for phase 3 evaluation (with adequate
consideration given to both safety and efficacy before
end-of-phase 2 discussion). In general, a minimum treatment
duration of 12 weeks is needed to demonstrate efficacy. We
recommend considering a post-treatment follow-up period to
evaluate recurrences following treatment discontinuation.
It is important that all drug products for
acne be evaluated for safety and efficacy in the treatment of
facial acne. Applicants are encouraged to demonstrate the safety
and efficacy of the investigational drug in at least two adequate
and well-controlled studies. We recommend that these trials be
randomized, blinded, multicenter trials with suitable comparator
arms, which usually include a vehicle or placebo control.
Additional assessments for safety and efficacy in the treatment of
nonfacial acne can be included for topical products and should be
included for all systemic acne medications.
Patients are often recruited for study entry
at their worst severity and usually improve during the course of
therapy, whether the therapy is active or placebo (vehicle in the
case of topical drugs). A demonstration of superiority against a
placebo arm is generally needed for clinical studies.
We recommend that applicants enroll a
population that is representative of the age, race, gender, and
geographic location of acne patients in the United States. (See
also ICH guidance for industry E5 Ethnic Factors in the
Acceptability of Foreign Clinical Data and guidance for
industry Collection of Race and Ethnicity Data in Clinical
Trials.)
Baseline acne
severity is a key element of the enrollment criteria. Baseline
lesion counts are expected to have a major influence on efficacy
outcomes in acne trials. Since change in lesion counts from
baseline may not always translate into a successful IGA outcome,
it is recommended that power calculations be based on both
endpoints (see Section IV., Study Design). As the range of
baseline lesion counts is expected to affect the success rates of
the outcomes, and may vary from drug to drug, applicants are
encouraged to investigate the optimal range of baseline lesion
counts to demonstrate success before initiation of phase 3
studies.
The baseline score of the IGA severity scale
should be consistent with the baseline lesion counts. However, no
numerical range of lesions for categorizing the IGA is
recommended. This is because the IGA is the investigator’s or
physician’s overall general assessment of the patient’s condition
and is considered to be more subjective than the purely numerical
lesion count. It also takes into account the quality, as well as
the quantity, of the acne lesions.
If a drug product is developed specifically
for either inflammatory lesions or noninflammatory lesions of
acne, labeling should reflect the specific type of lesion studied
with reference to lack of proven efficacy for the lesion type not
studied.
For drugs specifically intended to treat
either inflammatory or noninflammatory lesions, it is appropriate
for both lesion counts and the IGA to be assessed. Superiority
will need to be demonstrated for both the targeted lesion type and
the IGA. We recommend that the IGA allow for a clinical and
statistical evaluation of the investigator’s overall qualitative
assessment of the acne severity in each patient. However,
emphasis in the IGA regarding the lesion type not being targeted
may be modified. To show that there is no worsening of the
nontargeted lesion type, we recommend the endpoint for the
nontarget lesion count demonstrate noninferiority of the active
treatment to the vehicle at the prespecified time point. It is
important that an appropriate noninferiority margin be selected to
maintain a substantial proportion of the expected improvement from
baseline for the nontargeted lesions in the vehicle or placebo
treatment group.
At the end of phase 2 and before initiation
of phase 3 trials, we recommend that the applicant specify if a
drug product would be indicated for only inflammatory, only
noninflammatory, or both types of lesions of acne.
If the drug product under development is
expected to be used together with another marketed drug therapy
for acne, it is important that the clinical study design reflect
such co-use or adjuvant use. We recommend consultation with the
Agency before the conduct of these types of trials.
As there are specific informational needs
with regard to many treatments for nodular acne, it is recommended
that applicants seek additional guidance from the Agency regarding
treatments targeted for nodular/nodulocystic acne.
Fixed combination topical products for the
treatment of acne vulgaris are considered under 21 CFR 300.50 and
require evidence for the contribution of each active component
that is claimed to provide for additional safety or efficacy. In
study design, we recommend consideration be given to active
components that may target a specific lesion type.
Although evidence for efficacy may rely on
comparison with a reference-listed combination drug product,
comparison of only test and reference-listed combination drug
product may be insufficient. The contribution to efficacy of
different active ingredients may vary among different vehicles.
Because of the complexity of clinical trial design, we recommend
applicants seek further guidance from the Agency before the
conduct of studies for fixed combination drug products.
For topical drug products, dermal safety
studies with the final to-be-marketed drug product are
recommended. It is important for these trials to provide
information regarding cumulative irritancy (at least 30 evaluable
subjects), contact sensitization (at least 200 evaluable
subjects), phototoxic (at least 30 evaluable subjects), and
photocontact allergic potential (at least 50 evaluable subjects).
These trials are usually conducted simultaneously with phase 3
clinical trials, although preliminary dermal safety evaluations
could be conducted during development of the to-be-marketed
formulation.
Dermal provocative irritation studies may be
waived if phase 2 safety data demonstrate that the product is
irritating and the Agency determines that this information is
adequate for labeling purposes. Dermal irritation and
sensitization (allergenicity) studies can be combined as long as a
sufficient number of subjects are included for sensitization
evaluation. Phototoxicity and photosensitization studies may be
waived if there is no absorption of the drug product by UVB, UVA,
or visible light (280 to 700 nm).
Safety assessments, for short-term use, of
both oral and topical acne drug products can include routine
chemistry and hematology profiles. Other safety parameters may be
appropriate depending upon any safety signals found in preclinical
studies.
Given the natural history of acne vulgaris,
acne drug products have the potential for chronic use. Therefore,
we recommend addressing long-term safety. Applicants are referred
to the ICH E1A guidance
for assistance in determining the number of subjects exposed and
duration of treatment needed to provide an acceptable safety
database.
Pharmacokinetic studies to assess the degree
of and/or potential for systemic absorption may be needed to
fulfill the requirements of 21 CFR part 320 (Bioavailability and
Bioequivalence Requirements). Under this section, a new drug
application (NDA) must either contain an assessment of in vivo
bioavailability or sufficient information which would allow the
Agency to issue a waiver of in vivo bioavailability testing. In
general, waivers of in vivo biostudies are the exception and are
only granted in specific cases. Because of the variable dosing
nature of topical products, we recommend that studies of the in
vivo assessment of systemic exposure be done under so-called
maximal use conditions. The recommended elements of such a
maximal use study are as follows:
1)
A formulation identical to the clinically
studied/to-be-marketed formulation should be used.
2)
The study should be done in an adequate number of patients
with area of involvement and disease severity index/measuring
toward the upper end of that in the proposed indication to include
at least the face, shoulders, chest, and back.
3)
The topical dosing used should represent the maximal dosing
anticipated in both phase 3 trials and in the proposed package
insert for the following:
a)
Frequency of dosing
b)
Duration of treatment
c)
Use of highest proposed strength
d)
Extent of involved area to be treated at one time
e)
Amount applied per square centimeter
f)
Method of application/site preparation
4)
The analytical method should be properly validated for both
parent compound and metabolites.
The objective of this study is to maximize
those elements affecting dermal penetration such that systemic
absorption can be determined. We recommend, when possible, that
the resulting pharmacokinetic data be analyzed using standard
pharmacokinetic metrics (AUC, Cmax, Tmax).
It is also recommended that the study protocols incorporate
evaluations for cutaneous safety.
The Agency recommends that the study protocol
for evaluation of acne vulgaris clearly specify the objectives of
the trial, the patient population, study drug dosage and duration
of treatment, primary endpoints, and key planned analyses. In
addition, we recommend the study design support the proposed
claims by taking into consideration the following factors:
1)
Sample size: It is important that the sample size
be sufficiently large to support the overall safety and efficacy
claims. We recommend the study be powered to ensure at least 80
percent power with a 2-sided Type I error rate of 0.05. It is
important that the protocol provide details concerning sample size
calculations for each of the co-primary endpoints: changes from
baseline in inflammatory and noninflammatory lesion counts and
success according to the IGA. We also recommend the study be
adequately powered for all co-primary endpoints. For each of
these co-primary endpoints, it is important that the protocol
specify estimated treatment effect for each comparator. For
noninferiority trials, it is also important for the noninferiority
margin to be prespecified and discussed with the Agency. Unequal
treatment allocation of patients to the various treatment arms
might help keep the sample size manageable. Such unequal
treatment allocation can be particularly useful for evaluation of
combination drug products (see Section III, Drug Development
Plan).
2)
Randomization and stratification: Randomization is
intended to allocate patients to treatment groups to reduce bias
and to ensure that the statistical procedures can be appropriately
applied. As baseline lesion counts are expected to have a
significant effect on outcomes, it is important to make a
considerable effort to ensure random allocation of subjects to
treatment arms to reduce bias. Although randomization is intended
to balance treatment allocation for confounding factors, there is
always a chance that randomization may fail to achieve balance,
particularly in smaller trials.
If there are known
factors that are expected to have a large influence on outcome,
stratification can be used to balance patient assignments for
these factors instead of relying solely on simple randomization.
However, we recommend that stratification be limited to the most
influential factors to avoid having a large number of strata and
consequently a small number of subjects per cell. Because
stratification implies constraints on randomization, the
statistical analysis for studies that have been stratified for
certain factors should account for these factors. Since some
degree of variation in efficacy across patients of different sites
or geographic areas is expected, we recommend randomization by
study site to balance the treatment arms in acne trials.
3)
Blinding: Because efficacy assessments of acne, in
particular the IGA, have a high degree of subjectivity, it is
important that the highest degree of patient and assessor blinding
be sought to achieve credible inference. Blinding may be
compromised if there is a marked difference in the adverse event
profile between the comparators. Every effort should be taken to
preserve blinding, such as using an independent assessor (a
qualified independent clinical assessor would provide the scoring
of record for those aspects of clinical assessment to be
blinded).
Many methods exist for
assessing the severity of acne and almost all of them use an
ordinal scale for assessing global severity (Lehmann et al.
2002). The primary difficulty in developing a standardized
ordinal scale is the pleomorphic nature of acne, as is pertinent
to the mixture of lesion types, sites of involvement, the variable
characteristics of the lesions (especially the inflammatory
types), and the variability in the progression of acne lesions.
However, this inherent difficulty provides the basis for the
categorical utility of having such a global assessment (Allen
1980; Feinstein 1977; Plewig et al. 1992).
Use of lesion count assessments alone as an
endpoint may be less than reliable because of the lack of
appreciation for the variable expression of acne vulgaris with a
strictly quantitative definition (e.g., size of lesions, intensity
of inflammation, and location of lesions). Although reduction in
lesion counts may indicate improvement of acne severity, clinical
perception of a given lesion count reduction (e.g., 50 lesions
less) is different for various baseline lesion counts (e.g., 100
versus 53 lesions). In addition, precision achieved with lesion
counts can be difficult and can vary even among clinicians who are
experienced in counting lesions of acne vulgaris. Variability of
lesion counts among raters has been shown to increase as the
number of acne lesions on a patient increase (Lucky et al. 1996).
Finally, although individual lesion counts have often been
employed successfully in the investigational setting, their
practicality and value for use in the clinical setting have been
questioned (Pochi et al. 1991).
Combining the two approaches of ordinal
global assessment scale and lesion count assessments allows for a
balanced approach toward the evaluation of acne severity. The
Agency continues to evaluate new metrics and alternative methods
as they are developed for evaluating acne severity.
We recommend for clinical studies
investigating the effect of a therapy on acne severity, co-primary
endpoints that evaluate an IGA, and acne lesion counts.
The Agency
recommends that the IGA be a static evaluation of qualitative
overall acne severity. To accomplish this, the global assessment
scale should be an ordinal scale with approximately five severity
grades (reported only in integers, e.g., 0 to 4). Each grade
should be defined by a distinct and clinically relevant
morphologic description that minimizes interobserver variability.
The grades on the scale should be sufficiently defined to
appropriately and unambiguously represent each severity grade on
the scale. Photographic examples of each grade that have been
agreed upon with the Agency before their use may be provided to
investigators. It is recommended that measures to ensure blinding
of investigators as to any previous or baseline scores with each
evaluation be submitted for review by the Agency. For
consistency, it is important that the same IGA scale be used
throughout the study, including study enrollment, evaluation at
endpoint, and for assessment of relapse. The Agency recommends
that each subject’s improvement be verifiable (e.g., via
photographic records of baseline and assessment time point) by
Agency staff for auditing purposes. Table 1 is an example of an
IGA scale that may be useful.
Table 1. Sample
IGA Scale for Acne Vulgaris
Grade |
Description |
0 |
Clear skin with no inflammatory or
noninflammatory lesions |
1 |
Almost
clear; rare noninflammatory lesions with no more than one
small inflammatory lesion |
2 |
Mild severity; greater than Grade 1;
some noninflammatory lesions with no more than a few
inflammatory lesions (papules/pustules only, no nodular
lesions) |
3 |
Moderate severity; greater than Grade
2; up to many noninflammatory lesions and may have some
inflammatory lesions, but no more than one small nodular
lesion |
4* |
Severe; greater than Grade 3; up to
many noninflammatory and inflammatory lesions, but no more
than a few nodular lesions |
*
The Case Report Forms for
acne studies can allow for reporting by investigators of lesion
worsening beyond Grade 4 with treatment. It is recommended that
enrollment of acne vulgaris patients not include patients with
nodulocystic acne. Patients who worsen beyond Grade 4 are to be
described in the safety evaluation.
For assessment of efficacy, the Agency
recommends that the IGA scale be dichotomized to success or
failure using one of the following two criteria to be
selected a priori.
1)
Clear or almost clear (Grades 0 or 1) as success:
Success is defined as “Clear” (Grade 0) or “Almost clear” (Grade
1) at the prespecified primary time point. For patients whose
baseline score is Grade 2, the clinically meaningful criterion for
IGA success is achieving a score of Grade 0 at the prespecified
primary time point because of limitations inherent to an ordinal
scale.
2)
Two grade improvement as success: Success is
defined as improvement of two grades from the baseline score at a
prespecified primary time point. Since under this alternative
definition of success not all subjects with “Severe” (Grade 4)
acne will achieve the “Clear” or “Almost clear” state, if the
product under study is approved, these outcomes would provide
useful information in product labeling.
We recommend that the IGA success criterion
not selected a priori for primary evaluation be evaluated
as a secondary endpoint. A study that fails on the primary IGA
success criterion selected would not be rescued with the other
success criterion, as this would not control for Type I error.
For targeted acne therapy (i.e., treatment of
inflammatory or noninflammatory lesions of acne alone), the IGA
selected for use may modify emphasis for the lesion type not being
evaluated.
Applicants are encouraged to discuss other
alternative IGA grading scales and study designs with the Agency
before implementation.
For the acne vulgaris indication,
noninflammatory and inflammatory acne lesion counts are co-primary
endpoints along with the IGA. When counting facial acne lesions,
it is important that all lesions be counted, including those
present on the nose.
Even if the indication is limited to only one
type of lesion (i.e., either noninflammatory or inflammatory
lesions of acne), as described in Section III.A., Clinical
Considerations, we recommend obtaining lesion counts for both
types, but only declaring one as primary in the prespecified
analysis plan.
The Agency is interested in
patient-reported outcome information; however, such
information should not be used as a substitute for objective data
or as a surrogate for efficacy. For patient-reported outcome
assessments, objective measures could be helpful tools, which may
inform both the patient and clinician.
It is important that the tools for
statistical evaluation be appropriate for analysis of the efficacy
endpoints. We recommend that the statistical analysis plan
prespecify the primary efficacy variables, the study population,
the hypothesis to be tested, and the statistical methodology to be
used. It is important to consider the following points in the
statistical analysis plan:
1)
The primary efficacy analyses for an acne indication should
be:
a)
Change from baseline in the inflammatory lesion count;
b)
Change from baseline in the noninflammatory lesion count;
and
c)
The proportion of success according to the dichotomized IGA.
Secondary efficacy
variables should be those clinically relevant outcomes that
support the validity of the primary efficacy variables.
2)
For a general acne indication, we recommend the test drug
be superior to its vehicle with respect to change, both in
inflammatory and noninflammatory lesions, in addition to success
according to the IGA (see above). It is also important to provide
secondary analysis for percent change of lesion counts.
On the other hand,
for an acne indication specific to a certain lesion type (see
Section III.A.3., Targeted Acne Therapy), we recommend the test
drug be superior to its vehicle with respect to the specified
lesion type, and be noninferior to its vehicle for the other
lesion type. It is important that the noninferiority margin be
discussed and agreed upon with the Agency before study
initiation. In addition, it is important to demonstrate
superiority for success according to the IGA.
3)
Prespecification of the statistical analysis is a key
factor for obtaining consistent and convincing evidence of product
efficacy, as data-driven analyses should not be used to support
efficacy claims. We recommend the protocol have sufficient
description of the statistical analyses of the primary efficacy
endpoints so that an independent statistician could perform the
analyses in the protocol. The description should include:
specifying the hypotheses to be tested, indicating the level of
significance to be used, and whether it is 1- or 2-sided, denoting
the mathematical expression of the statistical models, and
identifying methods for controlling Type I error rates for
multiplicity or interim analyses if needed.
4)
It is important that the protocol prospectively identify
the covariates to be used in the analysis. We recommend using all
prespecified covariates that are selected. It is also important
that the number of covariates be kept to a minimum and limited to
those whose influence on the outcome is suspected to be strong,
such as stratification factors like study center.
5)
We recommend addressing in the protocol possible
center-by-treatment interaction along with planned sensitivity
analyses to ensure robustness of the efficacy results.
6)
If multiple assessments are taken (e.g., over time) it is
important that the protocol prespecify how they will be evaluated
for efficacy. If the claim is that a win occurs if any assessment
wins, an adjustment needs to be made for multiplicity, but, if a
win occurs only if all assessments win, no adjustment in
significance level is warranted. We recommend that the method for
multiplicity adjustment be planned and specified in the protocol.
This would also include assessments for any validated
patient-reported outcome endpoints.
7)
If interim analyses are planned, it is important that the
protocol prespecify early stopping rules and penalties for such
interim analyses.
8)
We recommend provisions be made for analysis of clinically
relevant secondary endpoints and subgroup efficacy analysis, along
with safety evaluations. A multiplicity adjustment could be
appropriate if the efficacy results from multiple secondary
endpoints are intended to appear in the label.
9)
It is important that the study protocol clearly define the
study population to be analyzed, and provisions be made to handle
dropouts (see Section V.A., Handling Dropouts). We recommend
efficacy evaluation be carried out for the intent-to-treat (ITT)
population, defined as all subjects randomized and dispensed study
medication. We also recommend that a supportive analysis be
carried out for the per-protocol (or completers) population and
criteria for defining the per-protocol population be specified in
the protocol.
We recommend that efficacy and safety
evaluation be carried out on all patients randomized and dispensed
study medication. Every effort should be made to follow all
enrolled subjects until the end of the study and until the
resolution of any adverse event. However, in clinical trials, it
is anticipated that a certain percentage of enrolled subjects will
drop out.
Dropouts are common in acne trials and lead
to information loss. It is unlikely that dropouts occur randomly,
and they rarely occur completely independent of the treatment
being tested, so there is always the possibility that dropouts
introduce bias. The extent of this bias is expected to be related
not only to the magnitude of the information loss due to dropout
but also to the distribution of the dropouts among the various
treatment arms. Several methods for handling dropouts have been
proposed, but none is fully adequate.
1)
The Agency’s current approach for acne vulgaris trials is
based on using the intent-to-treat analysis with imputation of the
last observation carried forward (ITT/LOCF), along with the
per-protocol (completers) analysis. Although consistency in
efficacy findings from the two analyses can increase confidence in
the efficacy results, this does not resolve the problem of
handling dropouts. The LOCF might not be the optimal approach for
handling dropouts; however, it is frequently applied because of
simplicity. If other or additional approaches for handling
dropouts are proposed, we recommend they be prespecified in the
protocol.
2)
It is important that the effect of dropouts be addressed in
all clinical trials and analyses, and analyses be carried out to
demonstrate that the study conclusions are robust with regard to
handling dropouts. An approach that can be used to check
robustness of study findings is the worst-case rule (assigning the
best possible score to all dropouts on placebo arm and the worst
score to all dropouts on the active arm and then performing an
analysis including these scores).
We recommend
that all data from clinical trials be validated and their quality
assured. We also recommend that all data be submitted in
electronic format per Agency guidance.
It is important to consider the following points during database
formulation:
1)
In addition to efficacy and safety data, it is important
that demographic and baseline data be submitted to the Agency. It
is also important that data for derived variables be provided
along with the algorithm to generate these variables.
2)
Efficacy and safety summaries can be consolidated. We
recommend data from multiple studies use the same format, so that
data from one trial can be easily merged with data from another to
allow subset analyses based on gender, age, race, and, when
appropriate, other subgroups.
3)
Electronic photographic records should be submitted to the
Agency such that they can be readily evaluated (e.g., sufficient
resolution to allow for clinical re-evaluation), clearly labeled
(e.g., with regard to subject, study number, center, and time
taken), and organized in a retrievable fashion for storage and
archiving purposes.
Allen, AM, 1980, Clinical Trials in
Dermatology: Measuring Responses to Treatment, Int. J. Dermatol,
19:1-6.
Feinstein, AR, 1977, Clinical Biostatistics:
Hard Science, Soft Data, and the Challenges of Choosing Clinical
Variables in Research, Clin Pharmacol Ther, 22:485-498.
Lehmann, HP et al., 2002, Acne Therapy: A
Methodologic Review, J Am Acad of Dermatol, 47:231-240.
Lucky, AW et al., 1996, A Multirater Validation
Study to Assess the Reliability of Acne Lesion Counting, J Am Acad
of Dermatol, 35:559-565.
Plewig, G et al., 1992, Acne and Rosacea,
Springer-Verlag, Berlin.
Pochi, PE et al., 1991, Report of the Consensus
Conference on Acne Classification, J Am Acad of Dermatol,
24:495-500.