Understanding the dandruff flare‐up: A cascade of measurable and perceptible changes to scalp health

Abstract

Objective

Clinical studies assessing the dandruff condition typically focus on the therapeutic resolution of the signs and symptoms of dandruff, starting with a symptomatic scalp as a baseline. The reverse, the progression of events leading to a scalp flare‐up after stopping the use of an anti‐dandruff (AD) shampoo, is poorly understood. This study explored onset time and progression of the signs and symptoms of dandruff in individuals who stop using an AD shampoo.

Methods

This randomized, double‐blind, parallel design study recruited self‐identified, scalp‐concerned adults. After a 2‐week run‐in with an AD shampoo, participants were randomized into two groups: The first continued using the AD shampoo and the second switched to a cosmetic (non‐AD) shampoo. At baseline and defined intervals over the 28‐day study period, objective measures of scalp condition were investigated: expert‐assessed flaking, scalp impedance, and biomarkers. These were combined with self‐assessments of scalp condition. A Stochastic Mixed‐Effect Random Forest (SMERF) machine‐learning algorithm was explored to model time‐varying technical measurements and daily self‐assessment responses from participants.

Results

Switching from an AD shampoo to a non‐AD shampoo resulted in increased scalp Malassezia load, degree of flaking, barrier disruption, and inflammatory and oxidative stress biomarkers compared to those maintaining AD shampoo usage. Signs of declining scalp condition (histamine, complement component C3) were observed as early as 3 days after ceasing AD shampoo usage despite no statistical increase in flaking (ASFS) until around 3 weeks of using a non‐AD shampoo. These changes were accompanied by self‐perceived itching and flaking, reported to be significantly worse in those using the non‐AD shampoo compared to the AD shampoo at the end of the study. No statistically significant increase from baseline was observed for any objective measure in the AD shampoo group. In the SMERF model, markers of inflammation, oxidative stress, and Malassezia load were predictive of self‐perceived scalp attributes.

Conclusion

For the first time, a comprehensive understanding of changes that occur as scalp condition goes from asymptomatic to dandruff‐affected has been established. For scalp‐concerned individuals, regular, sustained use of an AD treatment promotes optimal scalp health. When switching to a non‐AD shampoo, scalp health deteriorates rapidly, even before visible flakes appear.

A summary of the changes that occur as the scalp condition worsens after switching from an anti‐dandruff (AD) shampoo to a cosmetic, non‐AD shampoo. The data, represented as a percent change from those maintaining the use of an AD shampoo, underscores the importance of regular, sustained usage of an AD shampoo to promote optimal scalp health. AD, anti‐dandruff; C3, complement component C3; HODE, hydroxyoctadecadienoic acid; IL‐1a, interleukin‐1α; IL‐1Ra, interleukin‐1 receptor antagonist; MPO, myeloperoxidase; S100A12, S100 calcium‐binding protein A12.

INTRODUCTION

Dandruff is a common, chronic, relapsing skin disorder of the human scalp that is estimated to affect approximately 50% of the general adult population worldwide [1]. Despite its high prevalence, the pathophysiological understanding of dandruff onset has not been fully elucidated. While dandruff is generally classified as existing on the spectrum of seborrheic dermatitis, differing only in locality and magnitude, [1, 2, 3] others argue that the pathologies vary enough to warrant differentiation [4]. Regardless, it is well documented that various intrinsic and environmental factors, such as sebaceous secretions, skin surface fungal colonization by the commensal scalp yeast Malassezia, and the complex interactions between these factors are reported to contribute to the pathogenesis of dandruff [1, 5]. Studies have shown that effective therapies for dandruff reduce the Malassezia load from the pretreatment level and that recolonization of Malassezia post‐treatment results in reoccurrence of the condition, accompanied by inflammation and disruption of the skin–scalp barrier [6, 7, 8, 9, 10]. This leads to the hallmark signs and symptoms associated with dandruff, such as flaking, itch, and scalp dryness [3, 11, 12, 13].

Historically, clinical studies to assess the dandruff condition have focused on the therapeutic resolution of dandruff flakes, starting with a flared, symptomatic scalp at baseline [14]. In these traditional models, effective therapeutic dandruff treatments have been shown to alleviate the signs and symptoms of the condition, resulting in the normalization of the structure and function of the scalp skin [14, 15, 16]. Biomarkers of inflammation, oxidative stress, and barrier integrity are exacerbated by the dandruff condition (IL‐1Ra:IL‐1α, hydroxyoctadecadienoic acid [HODE] and human serum albumin [HSA]); however, a significant reduction in the levels of these biomarkers has been reported with the use of an anti‐dandruff (AD) shampoo [3, 17, 18, 19, 20].

While the progression from dandruff‐affected to healthier scalp with the use of an AD shampoo has been robustly studied and documented [11, 12, 13, 14, 15, 16] the reverse—the progression of events leading to a scalp flare‐up after stopping use of an AD shampoo—has not been reported. This study reversed the traditional treatment model to evaluate the temporal changes in scalp condition that occur after ceasing use of an AD shampoo among individuals with self‐perceived scalp concerns. The hypothesis was that scalp‐concerned individuals, who underwent a 2‐week period of consistent exposure to an AD shampoo to improve scalp health prior to the switch to a cosmetic (non‐AD) shampoo product, would demonstrate a progression from healthy to unhealthy in a manner that was both detectable through objective methods and perceptible by the participants.

The aim of this randomized, double‐blind, parallel design study was to explore the onset time and progression of the signs and symptoms associated with dandruff in individuals who stop using an AD shampoo. Objective measures to assess the time and extent of the return of the signs and symptoms of dandruff included:

• expert‐assessed visual flake grading;

• skin barrier integrity (via human serum albumin [HSA] levels and scalp impedance);

• biomarkers representing inflammation, oxidative stress, and barrier function;

• self‐assessment of scalp flaking, itching, dryness, and sensitivity.

In addition, a Stochastic Mixed‐Effect Random Forest (SMERF) machine‐learning algorithm was adapted and used to model the time‐varying technical measurements and daily self‐assessment questionnaire responses from participants. The aim was to identify which of the objective measures of scalp health might be drivers of subjective, participant‐reported outcomes such as scalp flaking, dryness, and itch.

METHODS

Study design

A single‐center, randomized, double‐blind, parallel design study was conducted over a period of 6 weeks at an independent test center in the United States between May and June 2022. The study consisted of a 2‐week run‐in phase followed by a 4‐week treatment phase. During the run‐in phase, all participants shampooed their hair three times a week for 2 weeks using a commercial zinc pyrithione‐based AD shampoo (Head & Shoulders Classic Clean, Procter & Gamble). The 2‐week AD run‐in was intended to provide enough time to blunt dandruff‐associated symptoms prior to treatment start [16, 21]. Furthermore, this 2‐week run‐in mimics consumer habits and practices, demonstrating how an AD shampoo might be intermittently used only when signals of dandruff and/or scalp discomfort become noticeable (i.e. periodic treatment of symptoms vs. continuous usage for prevention of flake reemergence). During the 4‐week treatment phase, half of the participants continued to use the AD shampoo, and the other half switched to a commercially available cosmetic (non‐AD) shampoo. All shampoos were packed in identical, opaque white bottles, and subjects were blinded to the identity of treatment assignments. The study design is shown in Figure 1, and the study schedule is shown in Table 1.

FIGURE 1.

Study design overview. AD, anti‐dandruff; ASFS, adherent scalp flaking score.

TABLE 1.

Study schedule and timeline of objective measures.

Study day	Evaluation visit	Fungal count	ASFS	Biomarker sampling	Scalp impedance
1	Baseline	X	X	X	X
3	2			X	X
5	3			X	X
8	4	X	X	X	X
11	5			X	X
17	6	X	X	X	X
22	7	X	X	X	X
29	EOS	X	X	X	X

Abbreviations: ASFS, adherent scalp flaking score; EOS, end of study.

At the screening visit, participants were evaluated using the inclusion/exclusion criteria, and those eligible for the study provided written informed consent, completed a pre‐study habits and practices questionnaire, and provided demographic data and medical history. Participants were instructed to wash their hair three times each week per a provided washing calendar with their assigned product and to keep a daily journal to track the number of washes and compliance. Conditioner was provided to be used ad libitum from the earlobes down if their usual haircare regimen included the use of conditioner; individuals whose hair did not fall below the earlobes were not permitted to condition their hair during the study.

At the baseline visit, eligible participants were provided with written and verbal instructions on the proper use of their assigned treatment, a shampoo calendar indicating which days to wash their hair, and general restrictions and instructions to follow during the treatment phase. Participants were instructed to use the study‐provided shampoo three times each week, as they did during the run‐in phase, and to use conditioner only if permitted, avoiding contact with the scalp. Training was provided for at‐home measurements (for those participants who had access to Wi‐Fi and able to use a simple mobile app on a test phone provided). A total of 100 participants were enrolled in the study. For the 4‐week treatment phase, participants were randomized into two groups of 50, balanced for sex and previous product use (those who in the month before study enrollment used an AD shampoo at least once a week and those who used a cosmetic treatment shampoo for the purpose of scalp condition improvement).

All participants were required to complete a daily diary and at‐home self‐perception questionnaire until the end‐of‐study visit. At baseline and on days 8 (week 1), 17, 22 (week 3), and 29 (week 4), participants were instructed to complete their final shampoo 72 h prior to the study visit and to refrain from applying styling products until after their scheduled measurement visit. For all other study visits, participants were instructed to refrain from getting their heads wet and applying styling products the night before and the morning of their scheduled visit.

Participants

Participants eligible for inclusion were healthy adults (aged 18–65 years) who used an AD shampoo or a cosmetic shampoo for the purpose of scalp condition improvement at least once per week and typically washed their hair 2–4 times per week. Participants were required to wash their hair according to the schedule (washing three times per week and adhering to 72‐h wash‐free periods); refrain from dyeing, bleaching, or perming their hair; swimming for the duration of the study including the wash‐in period; and avoid use of hair styling products or wetting their head the night before and morning of scheduled visits. There were no flaking entrance criteria. Exclusion criteria included an inability to comply with the protocol procedures and requirements.

At each study visit, participants were asked about any changes to their health and medications since their last study visit and if they had complied with all pre‐visit instructions and restrictions to determine whether they were qualified to continue in the study.

Endpoints

The key objective and subjective measures assessed at baseline and over the 4‐week treatment period are summarized in Figure 1 and Table 1, and briefly described here. All scalp measurements and sampling were performed in the same octant at the same location or directly adjacent to every subject at each time point.

Fungal count to track the growth of Malassezia

At baseline, 1 week, day 17, week 3, and week 4, a scalp swab was taken from two sites on the clipped scalp area, and quantitative polymerase chain reaction (qPCR) was used to detect Malassezia as described previously [22].

Expert‐assessed flaking via adherent scalp flaking score

The use of expert flake assessment to identify dandruff has been described previously [23]. Assessments were performed by a qualified grader. Briefly, flaking was judged on a 0–10 point scale (none to severe) for each of eight regions of the scalp, yielding a total adherent scalp flaking score (ASFS) ranging from 0 to 80.

Scalp biomarker analysis

For biomarker analysis, scalp skin tape strips were collected and analysed as reported previously [17, 18, 24, 25]. Scalp tape strips were collected from each participant at baseline, day 3, day 5, week 1, day 11, day 17, week 3, and week 4.

Assessment of scalp barrier integrity via scalp skin impedance assessment

A proprietary handheld device [26], utilizing commercially available impedance measurement technologies, was used to quantify changes to the impedance of the scalp skin over the course of the study. Subjects had two ~1.9 × 1.9 cm sites identified in the D and H Octant in an effort to minimize clip site noticeability. Hair was clipped with scissors as close to the scalp as possible at each time point prior to taking measurements. Following acclimation in an environment with controlled temperature and humidity with clipped sites exposed, 2 non‐invasive measurements were taken from within the lower, medial aspect of the clipped scalp sites.

Self‐assessment of perceived scalp changes

A daily self‐perception assessment questionnaire (SAQ) was designed to capture subject‐relevant outcomes (severity of scalp flaking, itch, dryness and sensitivity) using a categorical scale that varied from ‘none’ to ‘worst imaginable’ (0 = none to 10 = worst imaginable).

Statistical analysis

No formal sample size calculations were applied. The planned sample size was based on exploratory studies evaluating scalp biomarkers in participants with dandruff. Analyses were conducted on the ‘full analysis set’ which is as close as possible to the intention‐to‐treat ideal of including all randomized participants. Means and standard errors were calculated for the study measures assessed at each visit. Analyses were performed evaluating change from baseline.

For biomarker analyses, given the wide variability of biomarker levels between and within individuals, data were transformed to the log10 scale before analysis. Statistical analysis was carried out on the equivalent log10 ratio values (i.e. log10[evaluation visit/baseline]). The final reported biomarker results are converted back to their original scale (or percentage change from baseline values) for ease of interpretation. For all endpoints, statistical significance was defined as p < 0.10.

Exploratory endpoint: Modelling of longitudinal data

Over the intervention period, predictors (biomarker and biophysical measurements) and outcomes (SAQ responses) were repeatedly measured. To model time‐varying longitudinal data with clustered and correlated observations, SMERF, a novel machine‐learning algorithm, was employed. Thereafter, objective measures driving consumer‐relevant outcomes such as flaking, dryness, itch, and sensitivity were determined from the built models (see Supplementary materials).

By randomly splitting the data between training (70%) and testing (30%) sets, the quality of the built SMERF models was assessed by: (1) examining the correlation between model‐predicted and observed SAQ response values on the original 11‐point (0–10) scale, on which consumers were assessed; and (2) determining the percentage of model‐predicted values that were within 1 point of reported SAQ responses on this 11‐point scale (Table S3). The methodology is described in the Supporting Information.

Ethics

The study was conducted in accordance with the principles of ICH E6 Good Clinical Practices (GCP) with the exception of review or approval by an investigational review board (IRB) or independent ethics committee (IEC).

RESULTS

Malassezia fungal count

There was an initial marked decrease in the median amount of Malassezia detected in samples from the group continuing to use AD shampoo at 1 week (indicated by a higher cycle threshold (Ct) value) after which time the amount detected reached a plateau (Figure 2). As the data reflect average Malassezia load across the entire population, the observed decrease among AD users can be attributed to a subset of subjects that had higher basal levels of the scalp fungus and/or were slower to respond to the treatment, thus benefiting from continued AD shampoo usage. In contrast, the median amount of Malassezia detected in samples from the group using the non‐AD shampoo showed a marked increase rapidly from baseline through 1 week and day 17 (indicated by a lower median Ct value) before eventually plateauing, indicating successful recolonization of scalp Malassezia spp.

FIGURE 2.

Malassezia fungal count measured from scalp swabs taken over the course of the study. Total Malassezia load was quantified by qPCR using pan‐Malassezia probes. An increase in Ct value denotes a decrease in the overall scalp Malassezia population. Ct, cycle threshold.

Assessment of scalp flaking

Scalp flaking severity, as determined by the ASFS method, decreased significantly from baseline across all time points measured in the group using the AD shampoo (Figure 3), indicating a continued improvement in scalp flaking over the course of the study. Compared with the AD shampoo group, there was a smaller decrease from baseline in scalp flaking severity in the group using the non‐AD shampoo at 1 week, day 17, and 3 weeks, returning to near baseline levels by 4 weeks. The mean improvement from baseline in flaking severity in the AD shampoo group was significantly greater than that in the non‐AD shampoo group at week 3 and week 4.

FIGURE 3.

Assessment of scalp flaking by adherent scalp flaking score (ASFS) evaluated by an expert grader at scheduled time points during the study. Results not connected by the same letter are significantly different from each other (p < 0.10). Results with an asterisk (*) are significantly different from baseline (p < 0.10). ASFS, Adherent Scalp Flaking Score ASFS.

Measurement of scalp skin barrier integrity

An assessment of scalp impedance was conducted to provide insight into scalp skin hydration and overall barrier structure. A decrease in scalp surface impedance was observed across the study in both groups; however, the AD shampoo treatment did provide improved impedance measures compared to those using the non‐AD shampoo from week 1 onwards (Figure 4a).

FIGURE 4.

Measurement of scalp skin barrier integrity via assessment of (a) scalp impedance and (b) human serum albumin (HSA). HSA was measured from scalp tape strips taken at scheduled time points during the study. The levels of HSA are standardized to total soluble stratum corneum protein (pg/μg total protein) and reported as change from baseline (±SE). Results not connected by same letter are significantly different from each other (p < 0.10). Results with an asterisk (*) are significantly different from baseline (p < 0.10). HSA, human serum albumin.

In the AD shampoo group, HSA levels, a marker of skin barrier function, remained unchanged from baseline at all time points. In contrast, mean HSA levels significantly increased from baseline in the group using the non‐AD shampoo from day 11 onwards, and HSA levels were significantly higher in the non‐AD shampoo group than in the AD shampoo group from day 11 onwards (Figure 4b).

Biomarker analyses

Inflammatory mediators

For each of the inflammatory biomarkers, change in levels did not significantly increase from baseline at any time point in the group using the AD shampoo. In contrast, there was a significant increase from baseline in the levels of histamine, complement component C3 (C3), and the mean IL‐1Ra:IL‐1a ratio from day 11 onwards in the group using the non‐AD shampoo (with the exception of C3 on day 17) (Figure 5). Histamine and C3 levels were significantly higher in the non‐AD shampoo group than in the AD shampoo group from day 3 onwards (Figure 5a,b). The mean increase from baseline in IL‐1Ra:IL‐1a was significantly higher in the group using the non‐AD shampoo than in the group using the AD shampoo at days 17 and the 4 week time point (Figure 5c). Mean levels of S100 calcium binding protein A12 (S100A12), a biomarker of non‐optimal skin conditions, [27] were largely unchanged from baseline in the AD shampoo group; however, the non‐AD shampoo group demonstrated significant change from baseline increases in S100A12 at most time points from day 3 onwards (Figure 5d).

FIGURE 5.

Inflammatory biomarkers (a) Histamine, (b) Complement component 3, (c) IL‐1Ra:IL‐1a, and (d) S100A12 as measured from scalp tape strips taken at scheduled time points during the study. The levels of each biomarker are normalized to soluble stratum corneum protein (pg/μg total protein) and reported as change from baseline (±SE). Results not connected by same letter are significantly different from each other (p < 0.10). Results with an asterisk (*) are significantly different from baseline (p < 0.10). C3, complement component 3; IL‐1a, interleukin‐1α; IL‐1Ra, interleukin‐1 receptor antagonist; S100A12, S100 calcium binding protein A12.

Oxidative stress

The levels of myeloperoxidase (MPO), a marker of oxidative stress, did not increase from baseline at any time point in the group using the AD shampoo. There was a significant increase from baseline in the levels of MPO from day 11 onwards in the group using the non‐AD shampoo (Figure 6a). MPO levels were significantly higher in the non‐AD shampoo group than in the AD shampoo group at most time points from day 5 (day 5, week 1, day 17 and week 4).

FIGURE 6.

Biomarkers of oxidative stress (A) Myeloperoxidase (MPO) and (B) HODE (C9 + C13 HODE) as measured from scalp tape strips taken at scheduled time points during the study. The levels of each biomarker are normalized to soluble stratum corneum protein (pg/μg total protein) and reported as change from baseline (±SE). Results not connected by same letter are significantly different from each other (p < 0.10). Results with an asterisk (*) are significantly different from baseline (p < 0.10). HODE, hydroxyoctadecadienoic acid.

The mean levels of HODE, an oxidation product of linoleic acid and a marker of scalp oxidative stress status, [20] did not increase from baseline at any time point in the group using the AD shampoo. In the non‐AD shampoo group, the levels of HODE increased from baseline starting from day 3, reaching a plateau between day 11 and 17, and then returning towards baseline at weeks 3 and 4 (Figure 6b). HODE levels were significantly higher in the non‐AD shampoo group compared to the AD shampoo group at week 1, day 11, day 17, and week 3.

Subjective endpoints

At the end of the study, all subjective attributes were scored numerically higher (‘worse’) in the group using the non‐AD shampoo than in the group using the AD shampoo, although only flaking and itch reached statistical significance (Figure 7). Participants perceived symptoms associated with a deteriorating scalp condition before the appearance of visible flakes.

FIGURE 7.

Self‐perception scores for scalp attributes after 28 days of using an AD shampoo versus a cosmetic (non‐AD) shampoo. A categorical scale of 0–10 was used for the worst of the condition that panellists had experienced during the last 24 h, where 0 = none and 10 = worst imaginable. In each case, a lower score indicates a healthier scalp. Results with an asterisk (*) are significantly different from each other (p < 0.10). AD, anti‐dandruff; SE, standard error.

Exploratory endpoint: Modelling of longitudinal data

The objective measurements were mapped to the corresponding key subjective responses of flaking, dryness, itch, and sensitivity using the SMERF modelling. Inflammation markers drove scalp dryness while oxidative stress markers drove scalp itch and flaking; markers of microbial growth drove scalp sensitivity (Figure S1). Prediction accuracy of the SMERF models was generally good (R ² between model‐predicted and actual SAQ responses on the reported 11‐point (0–10 rating) scale was >0.70; and >80% of model‐predicted responses were within $\pm$1 point of actual SAQ responses on the reported 11‐point scale) (Table S3). While confirmation of the biological and physical mechanisms linked to consumer perception need to be validated in future studies, our findings precipitated a framework for which technical endpoint manifest as symptoms of the dandruff condition (Figure S1).

Safety

No participants withdrew from the study because of a study‐related adverse event.

DISCUSSION

For the first time, a comprehensive analysis of the changes that occur as the scalp condition goes from asymptomatic to dandruff‐affected has been described. The study confirms for scalp‐concerned individuals, without continued use of an AD shampoo, the scalp quickly regresses to display attributes of an unhealthy scalp, starting with the early sub‐clinical changes (e.g. underlying inflammation, itch, and discomfort) that eventually result in the emergence of visible flakes. Importantly, regular and frequent use of an AD shampoo for a prolonged period of time advanced these scalp‐concerned individuals from a stable, asymptomatic scalp condition to a state of optimal scalp health (summarized in Figure 8).

FIGURE 8.

Summary of the changes that occurred as the scalp condition transitions from asymptomatic to dandruff‐affected with the use of a non‐AD shampoo, whereas regular, sustained use of an AD shampoo promotes optimal scalp health. Dandruff‐causing microbes induce scalp oxidative stress and inflammation, disrupting the scalp barrier. Biomarkers of these sub‐clinical responses can serve as early indicators of deteriorating scalp condition that precede symptoms of itching, sensitivity and dryness as well as the appearance of visible flakes.

Differences were observed between the non‐AD shampoo and AD shampoo groups in all the objective scalp conditions reported. The observed differences were consistent with the hypothesis that sustained AD shampoo treatment promotes healthy scalp attributes, such as reduced fungal count, superior barrier integrity, reduced flaking, and lower levels of markers of inflammation and oxidative stress, compared to those using a non‐AD shampoo.

Scalp flaking severity decreased significantly from baseline at all time points in the group using the AD shampoo, indicating a continued improvement in scalp flaking over the course of the study. Notably, no significant increases in the levels of biomarkers for inflammation, MPO (oxidative stress), or HSA (barrier integrity) were observed at any time point in the group using the AD shampoo, further indicating that continued use resulted in sustained scalp health. HODE, another established marker of oxidative stress [20, 25, 28] levels remained relatively stagnant throughout the study in the AD shampoo group, suggesting that the 2‐week run‐in AD regimen likely provided the primary benefit that was comfortably maintained. While the data suggest AD users may experience scalp hydration benefits compared to those in the non‐AD group, impedance measures can also be impacted by changes to the stratum corneum itself, making it difficult to definitively interpret the impedance data.

Interestingly, HODE levels in the group that switched to a non‐AD shampoo increased dramatically within the first 11 days, plateaued for approximately a week, and were observed to return to the elevated level first observed upon switching to the cosmetic shampoo. Furthermore, scalp flaking severity also decreased up to the 3‐week time point in the group using the non‐AD shampoo, albeit at a lower magnitude, suggesting there is a lag time between switching from the AD shampoo at the start of the study and the time at which flakes visibly reappeared. The observed mild improvement in HODE and sustained reduction in ASFS among the non‐AD shampoo cohort may be attributed, in part, to the maintenance of a 3× per week wash schedule during the study, as it has been demonstrated that regular cleansing (even with a cosmetic, non‐AD shampoo) affords scalp health benefits [29]. Of note, after the 2‐week run‐in period, the mean adjusted ASFS was <13.0 at the start of the study for both groups. This was considerably lower than in traditional dandruff clinical studies, where the inclusion criteria typically specify a minimum ASFS of ≥24.0 [25]. At the end of the study, the mean (SE) adjusted ASFS was reduced to 6.6 (1.1) in the AD shampoo group, defined as non‐dandruff in clinical studies, [25] and significantly lower than in the non‐AD shampoo group (p < 0.001). This further supports that continued, frequent use of an AD shampoo continues to improve the overall health of the scalp.

Biomarkers indicative of declining scalp condition were elevated as early as 3 days (histamine and C3) after switching from an AD shampoo to a cosmetic, non‐AD shampoo. By day 11, several biomarkers (histamine, C3, IL‐1Ra:IL‐1a, MPO, HSA) were found to have increased significantly from baseline in the non‐AD shampoo group only, coincident with an increase in Malassezia load on the scalp reaching a peak level after 1 week. Although not measured in this study, Malassezia spp. are known to recolonize the scalp rapidly (within days) in the absence of AD active ingredients [3, 7, 30] leaving the scalp vulnerable to Malassezia‐derived irritating stimuli such as oxidized free fatty acids, a source of oxidative stress [31, 32, 33]. This likely accounts for the early detection of inflammatory mediators reported herein, including the release of histamine from mast cells [34] which subsequently induces hyperproliferation of skin cells characteristic of the dandruff condition [3]. No significant difference in flaking between the two groups was observed until approximately the 3‐week mark, further demonstrating that the irritants and discomfort associated with dandruff manifest long before seeing visible flakes.

Results of the SAQ indicated that participants perceived the changes in scalp condition that occurred when they switched to using the non‐AD shampoo. At the end of the study, subjective attributes for flaking and itch were significantly worse in the group using the non‐AD shampoo compared with those using the AD shampoo, with dryness and sensitivity directionally worse. It is possible that participants may perceive the scalp changes associated with the use of a non‐AD shampoo before the appearance of visible flakes, and this correlates with the early detection of biomarkers of deteriorating scalp condition. The results of the SMERF modelling suggested that self‐perceived scalp flaking, itch, dryness, and sensitivity were associated with specific markers of scalp oxidative stress, inflammation, and fungal load.

The biological and physical mechanisms responsible for the deterioration in scalp condition upon withdrawal of an AD shampoo have not yet been elucidated; however, the findings presented here support a possible mechanistic rationale based on the objective and subjective responses observed.

Within a week of AD shampoo cessation, Malassezia load on the scalp increases. Metabolic byproducts and mycotoxins emanating from the fungi can disrupt the stratum corneum and interact with nerve endings involved in the physical sensations of irritation and discomfort [30]. Malassezia produce lipase enzymes that can hydrolyze triglycerides into irritant fatty acids [32]. The formation and accumulation of species such as oxidized lipids lead to oxidative stress that can trigger the release of inflammatory mediators [35] and accelerate the turnover of skin cells, driving hyperproliferation and increased scalp flaking [3]. This increased cell turnover can disrupt the balance of moisture in the stratum corneum, resulting in scalp dryness due to compromised scalp barrier integrity and increasing scalp sensitivity [36, 37]. Underlying scalp inflammation has also been shown to exacerbate sebum production to overcompensate for scalp dryness, which may lead to some individuals experiencing an oily scalp [38]. Malassezia spp. generate reactive oxygen species (ROS), and the intensity of ROS formation was proportional to the cell density of the yeast [39, 40]. This Malassezia‐mediated oxidative stress can induce the recruitment of mast cells, initiating the release of histamine, shown to contribute to scalp itch [34]. The proposed mechanism and sequence of physiological events that lead to the recognition of dandruff symptoms and scalp discomfort are summarized in Figure S1.

Limitations

At baseline, even after a 2‐week run‐in with AD shampoo, there was considerable heterogeneity in the flaking scores (ASFS) for individuals at study entry, and individuals with broader underlying condition(s) were not identified or excluded from the study. Given no specific ASFS or dandruff entry criteria for this study, more frequent (3+/week) AD shampoo usage and/or a longer (e.g. 4 weeks) run‐in phase to account for skin cell turnover may be necessary to increase the likelihood of a largely asymptomatic population at treatment start. Nevertheless, the observed diversity in scalp flaking reflects the real‐world spectrum of dandruff severity in scalp‐concerned individuals and helps build our understanding of the changes in biomarkers as well as the commonly observed signs and symptoms associated with dandruff that may occur without frequent and consistent use of an AD shampoo. Individual‐level data analysis and modelling will be key to improving our knowledge of scalp health to better address the concerns of consumers along the dry scalp, dandruff, and seborrheic dermatitis continuum.

Future studies

It is anticipated that the regression study design will facilitate evaluations of new AD treatments and enable a more detailed understanding of temporal onset and changes to the scalp condition. It cannot be ruled out that seasonality and the heterogeneity in participants' scalp flaking at baseline may have factored into some of the data nuances, such as the unexpected overall decrease in scalp impedance observed across all subjects during the study, independent of treatment. Additional studies are required to understand these levers and determine which scalp health indicators between different sub‐populations of scalp sufferers (e.g. itchy vs. flake‐prone individuals) might best inform both the consumer and expert on how to achieve optimal scalp care practices. Ultimately, this work enables better definition and identification of clinical populations to facilitate the development of effective treatments for the dandruff condition.

According to the exploratory SMERF modelling approach, specific biomarker and biophysical measurements were found to be drivers of participant‐reported outcomes, with over 80% prediction accuracy. The results suggest that this machine‐learning modelling may be a useful tool with which to explore longitudinal data where predictors and outcomes are repeatedly assessed and time‐varying. Further studies are warranted to validate this novel approach.

CONCLUSIONS

This study uniquely adds to the existing literature by providing insights into the onset of poor scalp health and the cascade of changes that occur when switching from an AD shampoo to a cosmetic, non‐AD shampoo. These, often invisible, signs of poor scalp health indicate the importance of regular and sustained AD shampoo use. Furthermore, individuals who remained on AD shampoo for an extended period not only maintained but also strengthened their scalp integrity, advancing towards an optimal scalp health state.

AUTHOR CONTRIBUTIONS

MGD, TLC, RAB, LB, and MP contributed to the design and implementation of the research. RAB, MGD, and KCSL analyzed the interpreted the data. DJA and PS designed the computational framework and analyzed the data via the built model. All authors discussed the results and contributed to the final manuscript.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

DISCLOSURES

MGD, KCSL, RAB, TLC, LB, DJA, and PS: Employee of P&G. MP: Chief of Dermatology, Cleveland Clinic.

Supporting information

Data S1.

Locker KCS, Bacon RA, Caterino TL, Breyfogle L, Alperet DJ, Sarkar P, et al. Understanding the dandruff flare‐up: A cascade of measurable and perceptible changes to scalp health. Int J Cosmet Sci. 2025;47:703–717. 10.1111/ics.13067