PFOS Exposure and Depressive Symptoms Rarely Moved in Tandem After Adjustment
A new NHANES analysis in adults aged 18 to 25 years found little evidence that serum PFOS exposure, or the balance between its linear and monomethyl-branched isomers, independently tracks with depressive symptoms. The study focused on a narrow but clinically relevant question: whether internal burden and isomer composition of PFOS relate to the Patient Health Questionnaire-9, a widely used screening measure for depression severity.
The analytic sample included 169 participants, with 147 complete cases for the core adjusted models. Median total PFOS was 1.70 ng/mL, and the median linear fraction was 72.2%. Depressive symptom burden was not rare; the weighted prevalence of PHQ-9 scores of 10 or higher was 22.6%. Even with that background prevalence, the exposure-outcome signal stayed small after accounting for covariates.
A doubling of total PFOS at fixed composition corresponded to a change of -0.49 PHQ-9 points, with a 95% confidence interval from -1.65 to 0.68. The linear-to-branched ratio at fixed total corresponded to 0.21 points, with a confidence interval from -1.94 to 2.36. Rescaling across the reported interquartile ranges yielded similarly modest differences: -0.58 points for total PFOS and 0.13 points for composition. Those estimates sit close to null and do not support a clinically persuasive relationship in this young adult sample.
What the Joint Model Did and Did Not Support
The modeling strategy separated two biologically distinct features of exposure: how much PFOS circulated in serum and how that burden was partitioned between isomer classes. In principle, that distinction matters because isomers can differ in protein binding, distribution, and elimination. A serum ratio, however, is still only a proxy for what reaches vulnerable tissues, including the brain.
Measured this way, PFOS composition did not behave like a stable toxicologic fingerprint for depressive symptoms. The inverse continuous-score association weakened after adjustment, and the screening odds ratio largely lost precision. Those patterns suggest either a weak association, substantial confounding, or insufficient information in this sample to resolve a small effect. The data also make it difficult to argue for equivalent biologic activity across PFOS isomers on the basis of serum composition alone.
Excluding three observations with an isomer below detection shifted both joint coefficients, which signals sensitivity to sparse exposure data at the lower end of measurement. In a biomonitoring analysis with a small sample, that kind of instability is expected. It also means any attempt to infer a dose-response gradient from the observed ratio requires caution.
Why Depressive Symptoms Remain a Plausible Endpoint for PFOS Research
Experimental work has already placed PFOS on a biologically credible path toward neuropsychiatric effects. Oxidative stress, inflammatory signaling, and synaptic disruption all remain plausible mechanisms. Those pathways are relevant to mood regulation because they can alter neuronal plasticity, monoaminergic signaling, and the inflammatory tone that has been linked to depressive symptom expression in other settings.
Still, biological plausibility is not the same as human-level effect estimation. PFOS circulates bound to proteins, and different isomers may clear at different rates. A serum composition ratio may therefore reflect transport and elimination dynamics more than the actual dose delivered to brain tissue. In practical terms, the measured exposure metric may be too indirect to cleanly map onto the neurobiology of mood.
The age group also matters. Young adults represent a population in which depressive symptoms can fluctuate for many reasons unrelated to PFAS exposure, including sleep disruption, psychosocial stress, substance use, and other environmental exposures not fully captured in a single cross-sectional survey. Any weak exposure signal can be diluted quickly when the outcome has many competing drivers.
Cross-Sectional NHANES Data Set the Boundaries of Interpretation
This analysis used publicly available, de-identified NHANES data from August 2021 through August 2023 and drew on the DEMO_L, PFAS_L, DPQ_L, BMX_L, and COT_L files. The survey-weighted design strengthens population relevance, but the structure is still cross-sectional. Exposure and outcome are measured at one point in time, which prevents temporal ordering and limits any claim about causality.
A single blood measurement can also miss important exposure history. PFOS is persistent, yet mood symptoms evolve over time, and the biologically relevant window may not line up with the survey visit. That mismatch can blur true relationships even when a compound has a real effect.
The preprint status adds another layer of caution. The findings have not yet been certified by peer review, and the authors themselves frame the results as insufficient to establish an independent composition association or equivalent isomer effects. For now, the strongest conclusion is negative rather than positive: the available data do not show a clear, independent link between PFOS isomer composition and depressive symptoms in this cohort.
The Next Test Will Require Better Exposure Resolution
Future work will need larger longitudinal datasets with isomer-resolved exposure and toxicokinetic measurements. That design could distinguish whether serum ratios actually mirror tissue delivery, whether elimination differs meaningfully by isomer, and whether depressive symptoms emerge after a lag rather than in parallel with biomonitoring values.
Without those data, the current analysis remains a useful but limited snapshot. It shows that total PFOS and its isomer composition can be modeled separately in relation to mood outcomes, yet the observed associations are small, imprecise, and sensitive to analytic choice. The unresolved issue is not whether PFOS exists in the circulation of young adults; it is whether the serum ratio carries enough biological meaning to predict depressive symptoms in a way that survives stronger measurement and a longer follow-up window.