Androgenetic alopecia, often called male or female pattern hair loss, is the dominant cause of hair thinning worldwide, driven by a genetic sensitivity of hair follicles to dihydrotestosterone, a potent derivative of testosterone. Under its influence, follicles on the scalp gradually miniaturize, producing finer, shorter hairs over successive growth cycles until they eventually stop producing cosmetically meaningful hair at all. Because the condition is so widespread, affecting a large share of men by middle age and a substantial number of women, the market for supplements claiming to restore hair through nutrition has flourished, often citing older, smaller studies that reported altered levels of zinc, copper, or iron in people with hair loss.

The new investigation set out to test that idea at a scale rarely attempted before. The team retrospectively collected patient data from Sichuan Provincial People’s Hospital covering the period from October 2021 to December 2024. Patients diagnosed with androgenetic alopecia formed the experimental group, while healthy individuals served as controls. Serum concentrations of copper, zinc, calcium, phosphorus, magnesium, iron, selenium, manganese, molybdenum, nickel, chromium, cobalt, and lithium were quantified using inductively coupled plasma mass spectrometry, an analytical technique capable of detecting metals and metalloids at extraordinarily low concentrations by ionizing atoms in a plasma torch and sorting the resulting ions by their mass-to-charge ratio.

The statistical analysis, performed with GraphPad Prism 9.0 software, used unpaired t-tests to compare the two groups, with a threshold of P less than 0.05 for significance. Out of the full panel of elements, only four showed statistically significant differences between patients and controls: copper, lithium, calcium, and phosphorus. And even those differences, on close inspection, were modest and clinically unconvincing. Serum copper averaged 897.8 plus or minus 164.2 nanograms per milliliter in the alopecia group versus 922.3 plus or minus 188.2 in the controls, both comfortably inside the reference interval of 240 to 2000 nanograms per milliliter. Serum lithium, measured at 1.20 versus 1.35 nanograms per milliliter, sat far below the upper limit of 200 nanograms per milliliter in both groups.

Calcium and phosphorus told a similar story of technical significance without clinical meaning. Some individuals in the control group had serum calcium values exceeding the reference interval of 80 to 100 micrograms per milliliter, but the proportion of out-of-range results was similar in both groups, indicating no meaningful association with hair loss. Phosphorus levels remained within the reference interval of 25 to 200 micrograms per milliliter in more than 99 percent of participants in both arms of the study. In other words, the elements that reached statistical significance did so on the strength of a very large sample size, which can detect tiny shifts that would go unnoticed in smaller cohorts, rather than because of any dramatic biological divergence.

Perhaps most striking was what did not differ at all. Zinc, the element most frequently blamed for hair loss in popular nutrition discourse, showed no statistically significant difference between patients and healthy controls. Neither did magnesium, iron, manganese, molybdenum, nickel, selenium, chromium, or cobalt. This null result carries particular weight because zinc and iron have been the focus of numerous earlier investigations, some of which reported lower serum levels in people with various forms of hair loss, fueling a persistent belief that supplementation could slow or reverse the process. The new data, drawn from a far larger patient population, undercut that assumption for androgenetic alopecia specifically.

The authors draw a direct practical conclusion: because serum trace element levels in patients with androgenetic alopecia fall within reference intervals and do not differ meaningfully from those of healthy individuals, routine screening of serum trace elements in these patients is not recommended. That recommendation, if adopted widely, could reshape clinical practice. Dermatologists frequently order panels of micronutrient tests when patients present with pattern hair loss, partly to reassure anxious patients and partly because historical literature suggested deficiencies might contribute. Each panel costs money, delays diagnosis, and can lead to unnecessary supplementation, which is not always harmless. Excess selenium, for example, has been linked in animal studies to changes in hair structure, and indiscriminate zinc supplementation can interfere with copper absorption.

The biology of why trace elements became suspects in the first place is not entirely unreasonable. Zinc participates in hundreds of enzymatic reactions, including processes central to the hair follicle cycle, and severe zinc deficiency is a documented cause of hair loss in conditions such as acrodermatitis enteropathica. Iron deficiency anemia is associated with diffuse hair shedding. Copper is a cofactor for enzymes involved in cross-linking keratin, and experimental work has explored copper-binding peptides in tissue remodeling. Selenium-dependent enzymes protect follicles from oxidative stress. The follicle, one of the most metabolically active structures in the human body, plausibly depends on an adequate mineral supply, which is precisely why the absence of measurable serum differences in this study is informative: the problem in androgenetic alopecia is not supply but signaling.

That signaling problem is hormonal and genetic. Follicles in the scalp of susceptible individuals express androgen receptors and the enzyme 5-alpha reductase, which converts testosterone into dihydrotestosterone. In these follicles, dihydrotestosterone binding to the androgen receptor triggers a cascade that shortens the growth phase of the hair cycle and progressively miniaturizes the follicle. Recent research has even traced how miniaturization depletes follicular stem cells through mechanosensory signals involving calcium flux and inflammatory mediators, an intracellular pathway entirely independent of how much calcium circulates in the serum. Approved treatments, including topical minoxidil and oral finasteride, act on vascularization and hormone conversion respectively, not on nutrition, and their efficacy reflects the true nature of the disease.

None of this means nutrition is irrelevant to hair health in general. Diffuse shedding can accompany crash dieting, malabsorption disorders, or genuine deficiencies, and clinicians should still investigate those contexts. But the distinction matters: androgenetic alopecia is a genetically programmed, hormone-driven process, and this study provides the strongest evidence yet that a blood test for minerals will not reveal anything useful about it. For the millions of people spending money on mineral supplements in the hope of regrowing a receding hairline, the message from this large, carefully measured dataset is blunt and liberating at once. The minerals are there. The hair loss has other causes, and future progress will come from targeting the hormonal and molecular machinery of the follicle itself rather than the chemistry of the bloodstream that feeds it.

Subject of Research: Serum trace element concentrations in patients with androgenetic alopecia compared with healthy controls

Article Title: A study on the serum trace element concentration in patients with androgenetic alopecia

Article References: Sun, H., Luo, J., & Xie, C. (2026). A study on the serum trace element concentration in patients with androgenetic alopecia. Archives of Dermatological Research, 318(1), Article 486. https://doi.org/10.1007/s00403-026-04967-z

Image Credits: AI Generated

DOI: 10.1007/s00403-026-04967-z

Keywords: androgenetic alopecia, trace elements, serum copper, zinc, hair loss, inductively coupled plasma mass spectrometry, micronutrients, dermatology, clinical screening, mineral supplements, hair follicle, case-control study

The study’s central innovation lies not in inventing new algorithms but in applying them to a problem that has resisted conventional clinical workflows for decades. The researchers drew their training data from the National Epidemiologic Survey on Alcohol and Related Conditions-III, a nationally representative US survey that includes both detailed diagnostic interviews and a rich array of background variables resembling the kinds of information captured in real-world health records. This design allowed the team to create an EHR-like dataset in which every person had a ground-truth diagnostic status, something almost never available in actual medical systems, where definitive diagnostic data on substance use is sparse and often contaminated by treatment-seeking behavior.

Because the survey contained standardized DSM-5 diagnostic assessments, the researchers could label each respondent as having or not having an untreated alcohol use disorder, opioid use disorder, or cocaine use disorder. The models were trained to predict these labels using only features that plausibly map onto data already present in health records: demographic characteristics, socioeconomic indicators, healthcare utilization patterns, and other routinely collected variables. Crucially, the team performed much of the modeling in two configurations, one excluding and one including psychiatric comorbidity information, so that they could assess how much predictive signal persists when mental health diagnostic data is unavailable, a common situation in primary care settings.

The technical pipeline reflects careful attention to the pitfalls that have undermined many machine learning studies in medicine. The authors benchmarked several algorithm families, including logistic regression, balanced random forests, extreme gradient boosting, and Gaussian naive Bayes, and tuned hyperparameters through grid search with nested cross-validation to prevent optimistic bias. They avoided relying on area under the receiver operating characteristic curve as the sole performance metric, citing well-documented critiques that this measure can be misleading for imbalanced classification problems like rare disorder detection. Instead, they optimized models for precision and recall, the quantities that matter most clinically: how many flagged patients actually have an untreated disorder, and how many true cases are successfully caught.

The headline result is striking. When models were optimized to balance precision and recall, the number of people who would need to be screened to identify one untreated case of alcohol use disorder fell dramatically relative to universal screening. In several scenarios, the reduction exceeded an order of magnitude, meaning that a healthcare system using the model as a triage tool could find the same number of hidden cases by conducting structured assessments on a small fraction of the population. For a public health problem as large as untreated alcohol use disorder, which affects tens of millions of Americans, such efficiency gains translate into enormous savings of clinician time, screening costs, and patient burden.

Interpretability was another pillar of the study. Rather than treating the models as opaque black boxes, the researchers used Shapley additive explanations, a game-theoretic attribution method that decomposes each prediction into the contributions of individual features. This approach, rooted in the Shapley value framework from cooperative game theory, allows clinicians to see exactly why a particular patient was flagged: which combination of demographic factors, utilization patterns, and other variables pushed the risk estimate upward. In an era when healthcare algorithms face legitimate scrutiny over transparency and accountability, building explanation into the tool from the ground up is a deliberate design choice with practical consequences for clinical trust and adoption.

The findings for opioid and cocaine use disorders were more preliminary but still encouraging. Because these disorders are less prevalent in the general population than alcohol use disorder, the number of positive cases available for training was small, limiting the statistical power of the exploratory models. The authors are candid about this constraint: the opioid and cocaine models showed promise but were limited by sample size. This honesty matters, because the field of clinical machine learning has repeatedly been criticized for overstating the readiness of tools trained on thin data. The study’s framing of these results as exploratory signals rather than validated instruments reflects a more mature approach to model development.

The broader context of this research includes a growing literature applying machine learning to addiction science. Previous studies have used neuroimaging data to classify nicotine dependence, resting-state functional connectivity to predict smoking status, and behavioral markers such as impulsivity dimensions to identify cocaine dependence. Others have built predictive models for opioid use disorder, including the Veterans Health Administration’s Stratification Tool for Opioid Risk Mitigation, which was developed to improve opioid safety and prevent overdose and suicide. What distinguishes the new work is its focus on untreated disease in a general population sample and its grounding in data types that real health systems actually collect at scale, rather than specialized research measurements.

The clinical deployment pathway the authors envision is clinical decision support. Integrated into electronic health record systems, such a model could continuously evaluate incoming patient data and flag individuals who warrant a structured substance use assessment, in the way that suicide risk prediction models have been trialed in large health systems. The literature on computerized decision support suggests these tools can produce absolute improvements in care when they are well designed and properly implemented, though the track record is mixed and depends heavily on workflow integration, alert fatigue, and clinician trust. A screening nudge that reduces the number of assessments needed per case found by more than tenfold is precisely the kind of efficiency that could make universal detection ambitions feasible in under-resourced primary care.

The study also engages seriously with the ethical dimensions of psychiatric prediction. Screening for stigmatized conditions raises questions about privacy, labeling, and the possibility that algorithmic flags could lead to discrimination if mishandled. The authors acknowledge concerns about bias in artificial intelligence applications for mental health, noting the field-wide calls to assess and mitigate algorithmic bias, and they emphasize that interpretability is essential for accountability when automated tools influence clinical decisions. Questions about who takes responsibility for algorithmic recommendations, how patients consent to being flagged, and how flagged individuals are approached with sensitivity to stigma all remain live issues that must be resolved before deployment. The treatment gap this technology targets is enormous and costly: untreated substance use disorders impose substantial burdens on individuals, families, and health systems, and most people who meet diagnostic criteria never seek care. By showing that routinely collected, EHR-like data carries enough signal to identify these hidden cases efficiently and transparently, the study offers a concrete technical foundation for closing a gap that has persisted through decades of conventional screening efforts, while making clear that careful validation in real health systems will be the decisive next step.

Subject of Research: Machine learning detection of untreated substance use disorders from electronic health record-like data

Article Title: Supervised machine learning enables identification of people with untreated substance use disorders using EHR-like data

Article References: Supervised machine learning enables identification of people with untreated substance use disorders using EHR-like data. (n.d.). https://doi.org/10.1038/s44220-026-00717-2

Image Credits: AI Generated

DOI: 10.1038/s44220-026-00717-2

Keywords: machine learning, substance use disorders, electronic health records, alcohol use disorder, opioid use disorder, cocaine use disorder, clinical screening, clinical decision support, interpretability, Shapley values, NESARC-III, early detection