Interictal serum levels of serotonin and plasma and mononuclear cell concentrations of beta-endorphin were measured in 20 juvenile patients (13 suffering from migraine without aura and 7 from episodic tension-type headache) before and after 3 months of L-5-hydroxytryptophan treatment (5 mg/kg/day) and compared with a control group of 17 headache-free healthy subjects. While no significant differences in serum serotonin levels emerged between the three groups (migraine 104.6 +/- 26 micrograms/L, tension-type headache 90.7 +/- 26.2 micrograms/L, controls 96 +/- 32.9 micrograms/L), significantly lower plasma and mononuclear cell concentrations of beta-endorphin were found in both patient groups by comparison with the healthy controls (beta-endorphin in plasma: migraine sufferers 16.2 +/- 4.2 pmol/L [P < 0.05], tension-type headache subjects 14.5 +/- 1.7 pmol/L [P < 0.001] vs controls 21.3 +/- 4.6 pmol/L and respectively, beta-endorphin in mononuclear cells: migraine sufferers 110.5 +/- 16.4 pmol/10(6) GB/L [P < 0.001], tension-type headache subjects 142.3 +/- 22.7 pmol/10(6) GB/L [P < 0.001] vs controls 359.3 +/- 31.6 pmol/10(6) GB/L). No differences emerged between the two clinical forms of headache for the plasma and mononuclear cell concentrations of beta-endorphin. After L-5-hydroxytryptophan treatment, serum serotonin and both plasma and mononuclear cell beta-endorphin levels tended to be higher, though not significantly so, than prior to treatment, and the clinical score (frequency x intensity of headache attacks) was significantly lower in both headache groups than at the baseline. This study supports the theory that opiate analgesic system function is abnormally low in juvenile primary headache as in adults, and confirms that administering serotoninergic precursor drugs increases beta-endorphin, even in the peripheral blood, and may favorably affect clinical symptoms.