Drug-induced diabetes insipidus is always of the nephrogenic type, i.e. unresponsiveness of the kidneys to the action of antidiuretic hormone. This condition is easily diagnosed by measuring urinary concentrating capacity during a thirst test (e.g. 12 hours of water deprivation) or by administration of a modified antidiuretic hormone, desmopressin, to demonstrate the renal unresponsiveness. Drug-induced nephrogenic diabetes insipidus is not a common disorder except in patients receiving treatment with lithium salts for affective disorders where it may affect about 10% of patients treated long term (15 years). Drug-induced nephrogenic diabetes insipidus caused by other drugs usually occurs in critically ill patients in intensive care units receiving a multitude of drugs dominated by antimicrobials and cytostatics. A search of the World Health Organization's adverse effect database revealed 359 reports of drug-induced diabetes insipidus. Lithium was the most common cause (159 reports) followed by foscarnet (15) and clozapine (10). Treatment is symptomatic in most patients and the offending drug should be stopped. If urine volumes exceed 4 L/day, treatment with thiazides and amiloride has been advocated, and nonsteroidal anti-inflammatory drugs, such as indomethacin, may be tried in severe cases. Prevention of lithium-induced nephrogenic diabetes insipidus is an important aspect of the treatment of affective disorders. In patients treated long term it appears to be only partly reversible upon lithium discontinuation. Close monitoring of the treatment aiming at 12-hour trough value of 0.4 to 0.6 mmol/L is recommended. Yearly measurement of the urinary volume/day is effective in making both the patient and the physician aware of the development of the drug-induced nephrogenic diabetes insipidus. The condition is a serious adverse effect because of the risk of developing dehydration and aggravation of drug intoxications.