Data collection
Senior stroke nurses identified patients on admission to hospital and recorded baseline details, including: demographic data; risk factors; prestroke medication; and prior history (i.e. renal impairment). Once the patient had consented to the study, ward nurses took a daily morning urine sample at first voiding or from a catheter for 10 days (or until death/discharge if sooner). If patients were incontinent of urine and not catheterized nurses squeezed a sample of urine from the pad or bedding if possible. We collected an early morning urine sample to prevent confounding by normal diurnal variation in urine as urinary concentration rises during the night and is usually at its highest between 4 AM and 8 AM in normal subjects (Buchsbaum & Harris 1971). This occurs because of reduced fluid intake at night; a nocturnal dip in blood pressure; increase in Antidiuretic Hormone (ADH) and recumbency. We also aimed to avoid any confounding due to morning medications, such as diuretics, which might cause patients to pass dilute urine while becoming dehydrated.