DEHYDRATION AND DEATH DURING FEBRILE EPISODES IN THE NURSING-HOME

OBJECTIVE: To determine the incidence of early hypernatremic dehydrati on among residents of a nursing home care unit (NHCU) presenting with significant febrile episodes (FE). DESIGN: Prospective cohort analytic study. FE were defined as temperature (T) > 100 degrees F oral (o) or 101 degrees F rectal (r) for r 24 hours.. SETTING: NHCU in a Veterans Administration hospital. PATIENTS: A total of 130 residents of the NH CU were monitored for FE during a 4-month study period. MAIN OUTCOME M EASURES: Blood urea nitrogen (BUN)/creatinine (Cr) (abnormal greater t han or equal to 25) and serum sodium (Na)(abnormal greater than or equ al to 146 mmol/L) were drawn within 24-48 hours of the onset of all FE ; documentation of impaired oral intake (OI) by staff; necessity of tr ansfer to acute medical wards and mortality were recorded. RESULTS: Th ere were 48 FE among 42 residents (39 M, 3 F; mean age 75 +/- 11.3). M aximum recorded T during the FE ranged from 100.1 degrees F-102.2 degr ees F o and 101.2 degrees F-105.3 degrees F r. Laboratory values were available for 40/48 FE. Twenty-three percent (9/40) had elevated BUN/C r ratios, 25% (10/40) had elevated serum Na, and 12.5% (5/40) had both . In patients noted to have impaired OI (n = 11) as documented by staf f, increased serum Na or BUN/Cr ratio was observed in 82% (9/11). A ra ndom control group of 37 nonacutely ill, nonfebrile NHCU residents (33 M, 4 F; mean age 75 +/- 10.1) having routine annual laboratory tests revealed only 1 resident (age 95) with an elevated Na of 146 and BUN/C r ratio of 26 and 1 resident with an increased BUN/Cr ratio of 28. Non e of the controls had any staff documentation of impaired OI. Of the 5 deaths in the febrile group with laboratory data (total deaths = 6; 1 4%), 100% had either elevated serum Na and/or elevated BUN/Cr ratios, and 80% (4/5) had both. Comparing the febrile group with controls, BUN /Cr ratios were found to be significantly elevated in the febrile grou p (P < 0.05). Serum sodium values were also significantly elevated in the febrile group (P < 0.01). CONCLUSIONS: Staff documentation of impa ired OI was highly associated with either elevated serum Na or increas ed BUN/Cr ratios. These data show that many older NHCU patients with s ignificant fevers often have early impaired OI and laboratory evidence of dehydration. These data indicate that staff should institute appro priate monitoring for dehydration at the time of earliest detection of fever in this population.