Adjusted logistic model for reported reaction

Sample-size warning for kidney outcomes

Kidney outcomes are much less frequent than skin reactions. A stable adjusted kidney model may require several thousand responses and adequate events in both exposed and comparison groups. With sparse events, kidney results remain descriptive; non-significance must not be interpreted as evidence of safety.

Skin, hair and nail reactions

No data have yet arrived; this discussion will update automatically as responses accumulate.

Dose or duration patterns, onset after use, improvement after stopping and recurrence on re-use help identify cases warranting clinical assessment; none is sufficient alone to establish causation.

Skin lightening and kidney health

The lightening module is included because WHO warns that mercury-containing skin-lightening products may harm the kidneys, nervous system and skin. This product-specific concern must not be generalized to all cosmetics or all lightening products.

The comparison will appear after data arrive; kidney outcomes remain secondary and exploratory.

Hair smoothing and fumes

FDA guidance supports capturing heat, ventilation and acute symptoms because some smoothing products may release formaldehyde when heated and are associated with eye and respiratory irritation, headache, dizziness, nausea, chest pain and rash. Interpretation depends on the product and setting and does not automatically apply to every hair treatment.

Current treatment users in the sample: 0.

Confounding and causal interpretation

NIDDK identifies diabetes, hypertension, heart disease and family history as important kidney-disease risk factors. Any cosmetic–kidney association therefore requires adjustment for them; long-term pain medicines and industrial exposures are additional potential confounders.

Selective recall or behaviour change after illness can create reverse causation, while heavy users may share health or social characteristics that explain part of an observed association.

Current principal findings

• Current analytical sample: 0.
• Reported any reaction: 0.0%.
• Pre-specified high-exposure indicator: 0.0%.
• High unsafe-practice score: 0.0%.
• Reported kidney results/diagnoses: 0.0%.

Conclusion

The study identifies self-reported statistical associations between specified cosmetic-use patterns and selected health outcomes. Because the design is cross-sectional and exposure, diagnosis and timing are participant-reported, findings do not establish that cosmetics caused the observed conditions.

Recommendations

• Complete dermatologist, nephrologist, toxicologist, statistician and ethics review before formal launch.
• Conduct back-translation, cognitive interviews and a 50–100 response pilot.
• Pre-register primary outcomes, exposure definitions and the analysis plan before main recruitment.
• Add a separately consented nested clinical study to verify eGFR and urine albumin and, when justified, product testing or biomonitoring.

Limitations

Limitations include self-report, recall error, selection bias, lack of direct medical verification, possible ingredient misclassification, multiple testing, rare kidney outcomes and residual confounding. Inclusion of several countries does not imply population representativeness.