Hermanussen · Arctic medical research 1995 · uncontrolled before-and-after trial · n=11

Acute and chronic effects of winter swimming on LH, FSH, prolactin, growth hormone, TSH, cortisol, serum glucose and insulin.

Cited 41 times in the scientific literature.

Level 4 - case-series / case-control

Uncontrolled pre-post study (self-experiment) without a separate control group

PubMed 7710600 · record verified 2026-08-29

What was done

In a 2.5-month self-experiment conducted between October 1992 and January 1993, 11 healthy students (5 female, 6 male, aged 24 to 29 years) and their teachers engaged in winter swimming at least once weekly for 2 to 10 minutes in the southern Baltic Sea (water temperatures ranging from 6.8 degrees C to 2.0 degrees C). Venous blood samples were collected before immersion, and at 30 and 60 minutes post-immersion, on both the first day (untrained baseline) and the final day (cold-adapted state) of the 2.5-month period to measure LH, FSH, prolactin, growth hormone, TSH, cortisol, glucose, and insulin.

What was found

Acutely, cold immersion increased TSH from 0.96 mU/l to 1.42 mU/l (p < 0.01) before adaptation and from 0.93 mU/l to 1.43 mU/l (p < 0.01) after adaptation, followed by a decrease at 60 minutes (p < 0.01). Serum cortisol rose from 99 ng/ml to 133 ng/ml in the untrained state and from 101 ng/ml to 137 ng/ml (p < 0.05) in the trained state at 30 minutes, decreasing thereafter (p < 0.01). Prolactin showed mild acute decreases, whereas LH, FSH, and growth hormone were unchanged. Blood glucose acutely increased by 12 mg/dl (p < 0.01) in untrained participants, but this acute rise disappeared after training. Chronically, basal insulin concentrations dropped by nearly 50% and basal prolactin concentrations roughly doubled after 2.5 months.

Why it matters

The study provides early exploratory human data indicating that acute cold-water immersion stimulates the pituitary-thyroid and adrenal stress axes, and suggests that chronic adaptation may alter resting metabolic and endocrine baselines, particularly insulin and prolactin regulation.

Limits

The study is a small, non-randomized self-experiment without a non-swimming control group, making it impossible to separate cold adaptation from seasonal, circadian, nutritional, or exercise-related confounders. Exact participant numbers beyond the 11 students are unclear, exposure duration varied widely (2 to 10 minutes), and the abstract is truncated without exact variance measures or numerical baseline-to-endpoint comparisons for chronic insulin and prolactin changes.

Cited by