[Community-acquired pneumonia among the elderly: differences between patients living at home and in nursing homes]

The etiology, presentation, and prognosis of community-acquired pneumonia (CAP) among nursing home residents are believed to differ from those of other groups. However, few Spanish studies have confirmed those assumptions or studied regional differences in CAP etiology.

A prospective study which included all patients over 65 years of age admitted to our hospital with CAP was carried out over a period of 18 months (2002-2003). We examined clinical, analytical, and radiographic characteristics paying particular attention to functional status--using the Eastern Cooperative Oncology Group (ECOG) scale and Barthel and Karnofsky indices--and comorbidity. Two blood cultures, a Legionella antigen test in urine, and serology for atypical bacteria were used for the etiologic diagnosis; bacterial cultures of respiratory samples were also used in certain cases.

Ninety-one patients, 25 of whom were nursing home residents, were enrolled. The nursing home residents were older than the other patients (mean [SD] age of 82 [4] compared with 73 [5]; P=.0001) and had greater comorbidity (P=.0001)--with a significantly greater presence of diabetes mellitus, cerebrovascular disease, congestive heart failure, and dementia. They also had a poorer functional status (ECOG, 2.09 [0.9] compared with 0.93 [1.1], P=.001; Barthel Index, 19 [33] compared with 77 [35], P=.001; Karnofsky In-dex, 51 [17] compared with 78 [23], P=.001). Regarding clinical characteristics, significant differences were found for respiratory rate (39 [11] compared with 27 [7] breaths/min; P=.001), blood pressure (69.5 [20] compared with 79.2 [18] mm Hg; P=.029), and temperature (36.6 [1.2] compared with 37.7 [1.1] degrees C; P=.001). CAP patients from nursing homes presented a greater number of affected lobules in chest x-rays (P=.004), more hypoxemia, acidosis, anemia, hypoalbuminemia, and greater scores of urea and creatinine. Fine Scale scores were also greater (134 [26] compared with 95 [28]; P=.001) as was mortality (7/25 compared with 3/66; P=.005). Few patients had an etiologic diagnosis and no significant differences were observed between the groups. The variable that predicted mortality in elderly patients in this series, according to stepwise logistic regression, was urea (adjusted R2=0.452).

In our sample population, nursing home residents were older, had greater comorbidity, and severe functional impairment. Under these circumstances the severity of CAP increases and becomes an important cause of mortality despite the fact that the etiologic agents do not appear to differ from those of the other patients.