This open trial was conducted with inpatients of the University Departments of Munich (n = 4) and Freiburg (n = 1). Patients of both gender were included in the trial after checking in- and the inclusion/exclusion criteria had been examined. Inclusion criterion was the diagnosis of a current manic or hypomanic episode in bipolar or schizoaffective disorder as defined by the International classification of diseases, 10th revision (ICD 10) The diagnosis was based on the clinical interview. Exclusion criteria were concomitant medical illness, especially hints of a neurological disease, a history of allergic reaction to valproate or a related substances, inability to give informed consent, and age below 18 or above 65.

After giving informed consent, intravenous valproate treatment was started according to the following regimen. The required amount of valproate (Orfiril® solution for injection (Desitin), one ampoule containing 300 mg Na- valprote in 3 ml solution) was further diluted in 100 ml Ringer solution and administered over 20 minutes through a peripheral intravenous catheter, usually three times a day. Treatment was started on day one with either 1200 mg/d (2x 600 mg) or 1800 mg/d (3x600 mg) depending on the patients weight. On day two, the dosage was increased individually depending on the appearance of side-effects, especially hypersedation and ataxia. Adaptations of dosage during the following days were done in response to clinical efficacy or side effects, not in response to plasma levels. In 4/5 patients the medication was switched to the equivalent dosage of oral valproate in an extended release formulation on day 6. In one patient (patient 2), infusions were continued until day 12. Plasma levels of valproate were generally determined 12 h after the last application of valproate during the infusion period or oral application during the oral continuation treatment. Additional plasma levels were taken 2 hours after medication in order to compare peak- trough concentrations.The affective symptomatology was rated with the Bech- Rafaelson Mania rating Scale (BRMAS) and/ or the Hamilton Depression Rating scale, 21 -item version (HRS) on a daily basis at 9 a.m. before the first infusion or oral medication on that day.

In those patients with existing and ineffective mood stabilizing pre- treatment (patient 2 and 4), the drug was kept on the same dosage during the trial with close monitoring of serum concentrations to reduce the chance of effects of dosage adjustments.

The means ± SD of the daily dosages, plasma levels and response measured by the BRMAS in all 5 patients (4 responders, 1 non- responder) are depicted below

Patient and diagnosis	Current symptomatology	final dosage of valproate i.v. (mg/d)	plasma level on day 5 (mg/l)	BRMAS baseline	BRMAS day 1	BRMAS day 2	BRMAS day 5
1, BP I	EM	1800	77	41	39	17	10
2, BP I	EM	1800	75.4	43	32	18	9
BP I	MM	3600	87	32	*	8	4
BP I	MM,RC	2400	51.1	17	17	15	14
5 SA	MM	2400	56.8	18	7	4	3
mean (pat.1-5)± SD		2400±735	69.5±15	30.2±12.3	23.8±14.5	12.4±6.1	8±4.5

All but one manic patients showed a rapid and favorable response to i.v. valproate loading (initial dosage between 1200 and 1800 mg/d), building up sufficient blood levels which were maintained on consecutive oral treatment. The most pronounced response, measured as a sharp decline of the BRMAS score, happened already very early on treatment, usually after 24-48 h. Valproate i.v. was tolerated without problems, especially without gastrointestinal problems, otherwise common in oral loading therapy, and also led to a drastic reduction and tempering out of concomitant benzodiazepine treatment in two cases. Interestingly, one patient was previously non-responsive to oral valproate loading and all of them showed a drastic remission of mania with valproate blood levels even below or only slightly above 50 mg/l (blood taken 12 hours after last application). Valproate peak concentrations, measured 2 h after the i.v. application, were in the range between 95 and 150 mg/l.

Besides the efficacy and safety of this treatment regimen, these findings may imply that the difference of pharmacokinetics in i.v. loading with a quick saturation of plasma protein binding and building up high peak concentrations of valproate may be part of the beneficial action, even in patients previously non-responsive to oral valproate.

Grunze H, Erfurth A, Kammerer C, Amann B, Giupponi G, Walden J (1998) Intravenous valproate loading in acutely manic and depressed bipolar 1 patients. J Clin Psychopharmacol, in Druck

Keck PE Jr, McElroy SL, Nemeroff CB ( 1992) Anticonvulsants in the treatment of bipolar disorder. J Neuropsychiatry Clin Neurosci 4: 395-405

Keck PE Jr, McElroy SL, Tugrul KC, Bennett J (1993) Valproate oral loading in the treatment of acute mania. J Clin Psychiatry 54: 305 308

Wangemann M, Wolf C, Retzow A (1997) Replacement of oral valproate with intravenous valproate: a study on dose finding and bioavailability. Eur J Clin Res 9: 209-215