Bacteriuria detected in the screening of adult and elderly populations has been associated with an increased mortality rate, but it is not clear whether the increase is a result of the bacteriuria itself or of differences in age, concomitant disease, or both. We screened a representative sample of the elderly population of Göteborg, Sweden (n = 1966), for bacteriuria. The mean (+/- SD) age at the time of screening was 70 years +/- 2 months. The five-year mortality among women with bacteriuria was 13.4 percent, whereas that among women without bacteriuria was 9.4 percent. The nine-year mortality in the two groups of women was 23.9 and 23.3 percent, respectively (P not significant). When the women with indwelling catheters were excluded from the analysis, the five-year mortality was 9.0 and 9.2 percent, respectively. Men with bacteriuria had an increased frequency of cancer (27.3 vs. 5.8 percent at age 70; P less than 0.002) and a higher five-year mortality than the other men; however, among the men with bacteriuria but not cancer the mortality was not increased. We conclude that fatal diseases associated with bacteriuria may account for the increase in mortality among elderly patients with bacteriuria.

The endotoxin levels in serum of 377 72-year-old individuals were quantitated. The study population was a representative sample of this age group and was participating in a general study of health and disease among the elderly in Göteborg, Sweden. The endotoxin levels in serum were quantified by the chromogenic Limulus amebocyte lysate assay and were correlated with the health status and laboratory findings for each individual. The mean endotoxin levels (+/- 1 standard deviation) in men and women, when excluding four outliers, were 6.6 +/- 3.8 and 6.9 +/- 3.8 pg/ml, respectively. All included, 21.5% of individuals had endotoxin levels equal to or above the sensitivity limit of 10 pg/ml. Strong positive correlations were found between endotoxin levels and plasma triglycerides (P > 0.995) and between endotoxin levels and serum protein (P > 0.9875). The endotoxin activity also correlated with mean corpuscular hemoglobin concentration (P < 0.005, negative correlation), body mass index (P > 0.9875), and decreased appetite (P > 0.9875). A high alcohol consumption was associated with increased endotoxin levels (P = 0.995). There are no previous studies which examine endotoxin levels in serum samples from individuals representative of the population. This study showed that elderly individuals had the same mean level of endotoxin as has been found in other age groups. The increased endotoxin levels seen in heavy drinkers may be explained by a decreased ability of the liver to remove endotoxin. The correlations found between endotoxin and triglycerides, protein, mean corpuscular hemoglobin concentration, decreased appetite, and body mass index are discussed.

Screening for bacteriuria was performed between 1984 and 1988 in persons aged 72-79 years representative of the general population in Göteborg, Sweden. The frequency of bacteriuria (greater than or equal to 10(5)/ml) at a single screening was 6% and 16% at age 72 years and 6% and 14% at age 79 years for the screened men (n = 235 and 259) and women (n = 259 and 297), respectively. By repeated screening after one month and 30 months of those previously negative at age 72 years, an additional 4% and 3% of men and 3% and 7% of women with bacteriuria were detected. Bacteriuric persons were excluded from further screening and controlled by frequent cultures during several years, with careful monitoring of clinical interventions. The persistence of untreated bacteriuria was analyzed in relation to bacterial species and number in the untreated subgroup of bacteriuric individuals. Nine of 10 Escherichia coli (E. coli) with less than 10(6)/ml and 22/22 non-E. coli strains disappeared spontaneously. In contrast, 20/26 (77%, p less than 0.01) with greater than or equal to 10(6) E. coli/ml persisted. Of 17 persons with bacteriuria persisting at least 12 months, 16 were women and 16 had E. coli. Of 201 E. coli cultures obtained from this group, 94% had greater than or equal to 10(6)/ml, and 99% had greater than or equal to 5 x 10(5)/ml. The results indicate that screening for high counts (greater than 10(6)/ml) of E. coli most effectively detects persisting bacteriuria in the general elderly population.

Phosphorylcholine is an immunodominant determinant of pneumococcal teichoic acids. Antibodies to phosphorylcholine are naturally occurring in man and decline in amount with age. Since antibodies to phosphorylcholine are markers of the immune responsiveness to polysaccharides and since anti-polysaccharide antibodies are highly protective against most bacterial pneumonia we expected a higher rate of pneumonia in elderly individuals with low levels of antibodies to phosphorylcholine. The relationship between the levels of antibodies to phosphorylcholine and mortality was analyzed prospectively in a representative sample of elderly individuals. A significant anti-phosphorylcholine antibody response occurred in a subgroup of the probands. There was a strong association (p less than 0.0001) between high levels of antibodies to phosphorylcholine in the serum at 70 years of age and pneumonia related death up to 14 years later. A similarly strong association was not observed between mortality and the antibody titer to another naturally occurring polysaccharide antigen: the blood group B antigen. Furthermore, there was no association between mortality due to diseases other than pneumonia and the levels of antibodies to phosphorylcholine. The association between antibody levels and subsequent fatal pneumonia provides a means of detecting individuals at risk for pneumonia-related death.

It has been proposed that immunity declines with age. Most evidence for this hypothesis has been obtained from cross-sectional samples of unrelated populations that differ in age, antigen exposure, and morbidity. In the present study, the authors used serum samples collected repeatedly from the same persons in longitudinal studies. Two representative samples of the population in Göteborg, Sweden were obtained; the first was studied at ages 38, 50, and 62 years, and the second at ages 70, 75, 79, and 81 years, respectively. The phosphorylcholine determinant of pneumococcal teichoic acid and the B blood group determinant were selected as model polysaccharide antigens. The results demonstrate a consistent decline in individual antibody levels in the decades before age 70 years but not later. Antibodies to phosphorylcholine and blood group B were highly parallel, suggesting that the decline was a general phenomenon for antibodies to polysaccharide antigens.