Printer Friendly

Bloodstream infections in patients older than eighty years.


During a period of 3 years in a University Hospital in Israel, 339 episodes of bacteraemia were observed in patients 80 years of age or older, and 658 episodes in patients 60-79 years of age. Patients older than 80 were more often residents of nursing homes, frequently had a history of a cerebrovascular accident, but were less often neutropenic. Twenty-four per cent of bacteraemia episodes in the very old were hospital acquired compared with 40% in the old patients. The most common source of bacteraemia was the urinary tract, 50% of episodes in the very old, and 34% of episodes in the old.

The percentage of episodes in which anaerobic bacteria were isolated was 5%, in the very old and 1% in the old, and the difference was significant when corrected for the sources of bacteraemia. All cases of community-acquired bacterial endocarditis in patients of 80 or over were caused by pathogens originating from the gut.

Thirty-five per cent of patients of 80 and over and 30% of patients aged 60-79 years died during hospitalization. Fatality was not associated with advanced age in the very old. Factors significantly and independently associated with fatality in both groups were a hospital-acquired infection, shock, low serum albumin, renal dysfunction and inappropriate antibiotic treatment.


Bloodstream infections are more common in older people[1-5], and carry a fatality of about 30%[1, 4, 6-8]. In developed countries the number of the very old is growing rapidly[9] and an increasing number of very old patients with bacteraemia are hospitalized and treated. Reports on the causes of bacteraemia and its fatality and morbidity in the elderly are contradictory[1, 7, 8, 10-13]. The purpose of the present study was to define factors characteristic of bloodstream infections in elderly patients. We compared the underlying disorders, presentation, pathogens and outcome in a large group of bacteraemic patients aged 80 and over with a group of bacteraemic patients 60-79 years of age.

Patients and Methods

Included in the present survey were all episodes of bacteraemia affecting patients 60 years of age or older during 1988-90 at the Beilinson Medical Center, Petah Tiqva, Israel. Patients aged 80 or older (the very old) were compared with patients aged 60-79 years (the old). The episodes were detected by daily surveillance of the microbiology laboratory records. Beilinson Medical Center is a 900-bed university hospital that serves an urban population of about 200000 as a first-line facility. It is also a referral centre for several hospitals in the vicinity.

Data were collected prospectively from the patient charts and the medical staff. We recorded each patient's age, sex, domicile, functional capacity before hospitalization, underlying disorders, cause of hospitalization and medications (with special reference to antibiotics, cytotoxic drugs and corticosteroids). For every episode of bacteraemia, we recorded the day and the ward in which it occurred, its presumptive source, the presence of urethral, intravascular or endotracheal devices, temperature, systolic and diastolic blood pressure, evidence of pneumonia on chest radiograph, total and differential white blood cell count, haemoglobin level, blood biochemistry and urinalysis results.

During follow-up, the antibiotic treatment administered and the results of cultures of blood, urine and samples from other sites were recorded. Data were gathered on complications of bacteraemia during hospital stay, duration of fever and of hospitalization, and outcome of hospitalization, whether discharge or demise.

Blood cultures: A 10-ml volume of venous blood was aseptically obtained and inoculated into a two-bottle set (6B aerobic and 7D anaerobic tryptic soy broth; Johnston Laboratory, Towson, MD). The bottles were incubated at 37 [degrees] C and tested on the Bactec-460 system (Johnston Laboratories, Cockeysville, MD) twice on the first day and daily thereafter for 7 days. The median number of sets obtained during one bacteraemia episode was three, with a range of one to nine. The median number of sets and the percentage of positive blood culture were similar in the study and control groups. Microorganisms were identified by standard methods. Organisms that are commonly recovered from the environment or the skin (mainly coagulase-negative staphylococci and aerobic Gram-positive rods) were judged to be contaminants unless the clinical findings, the results of cultures of material from other body sites, or the number of positive sets (two or more) indicated a high probability of true bloodstream infection.

Definitions: An episode of bacteraemia was considered as hospital-acquired if it occurred > 48 hours after admission to hospital. Neutropenia was defined as a count of [less than or equal to] 1.[10.sup.9] neutrophils/l, and hypothermia as a temperature on the day the blood cultures were taken of [less than or equal to] 36.5 [degrees] C.

Empirical antibiotic treatment was considered appropriate if the infecting micro-organism was subsequently found to be susceptible in vitro to the drug administered and if that antibiotic was given intravenously.

Septic shock was defined as a systolic blood pressure of < 90 mmHg with at least one of the following: clouded perception, oliguria (output of < 30 ml of urine/h), hypoxaemia [Po.sub.2], < 70 mmHg), acidosis (pH < 7.3), or findings compatible with disseminated intravascular coagulopathy (prothrombin time, < 50% relative to control; partial thromboplastin time, > 40 s; and elevation of fibrin split products or thrombocytopenia with less than 1 00 000 cells ml).

An episode of bacteraemia was designated as antedated by antibiotic, corticosteroid, or cytotoxic treatment if the drug(s) had been given for > 1 day during the preceding month.

Data analysis: In a univariate analysis we used the [X.sup.2] test (or the Fisher exact test for small numbers) to look for statistical significance of contingency tables. We employed the Cochran-Mantell-Haenszel (CMH) statistic to adjust comparisons for confounding factors[14]. Homogeneity of odds ratios across strata was evaluated by the Breslow-Day test[14].

As most of the continuous variables were not normally distributed, we used the Wilcoxon rank sum test for comparisons of continuous variables between two classes. The values of continuous variables were expressed as medians and ranges.

To look for significant and independent associations with a binary dependent variable, we used stepwise regression logistic analysis (LOGIST)[15]. We then used the regression coefficients and their standard deviations to compute multivariate adjusted odds ratios (ORs) with 95% confidence intervals (95% CIs).

For life-table analysis we used the LIFETEST procedure of SAS[16], and the statistical significance of comparisons between strata was tested by the log-rank test.


Clinical features and microbiology: Of 1735 episodes of bacteraemia detected in our hospital during the study years, 339 (20%) occurred in patients 80 years of age or older, and 656 episodes in patients 60 79 years of age (Figure 1). Univariate comparisons of the two groups regarding demographic data and underlying disorders are given in Table I. Hospital-acquired episodes constituted 40% of bacteraemias in patients 60-79 years of age, but only 24% in patients aged 80 and over (p = 0.000 1). The percentage of nursing-home residents was significantly higher in the very old, 36% compared with 14% in the old (p = 0.0001), and the functional capacity of patients older than 80 was significantly lower. A quarter of patients aged 80 years or more had an indwelling urinary catheter, compared with 8% of the younger patients (p = 0.0001). Very old patients had a significantly higher percentage of cerebrovascular accidents, dementia, decubitus ulcer and hypothyroidism, but the prevalence of ischaemic heart disease, hypertension and diabetes mellitus was similar to that in patients 60-79 years of age. Episodes of bacteraemia in the very old were less often antedated by treatment with corticosteroids or cytotoxic drugs, and only 1% of episodes occurred while the patient was neutropenic, compared with 4% of episodes in the old (p = 0.02).


The sources of bacteraemia are detailed in Table II. In patients aged 80 and over, the source of 50% of bacteraemia episodes was in the urinary tract, compared with 34% of episodes in patients 60-79 years of age (p = 0.0001). Very old patients had a smaller percentage of episodes of unknown origin and of intravenous line infections as the source of bacteraemia. The predominace was significant by the CMH statistic both when corrected for the presence of a urinary catheter, and for residence in a nursing home (data not shown).


The median temperature in patients aged 80 and older was lower than in patients 60-79 years of age, 38.5 [degrees] C compared with 38.7 [degrees] C (p = 0.02) but the percentage of hypothermic patients was similar. Patients aged 80 and over had higher values of blood urea nitrogen and creatinine, and lower values of albumin and cholesterol (Table III).


Comparisons of pathogens isolated from the blood in the two groups are given in Table IV (community-acquired episodes) and Table V (hospital-acquired episodes). Although the overall frequency of pathogens was similar, several points merit emphasis. The percentage of episodes in which an anaerobic pathogen was isolated in the patients aged 80 and over was 5%, compared with 1% of episodes in the old patients (p = 0.0001). The difference remained significant when corrected by the CMH statistic for the source of bacteraemia, residence in a nursing home and for the presence of a decubitus ulcer (OR 2.5, 95% CI 1.7-3.7). An anaerobic micro-organism was isolated from the blood of 29% of very old patients with infection of the extremities compared with 0% in the old; and in patients with bacteraemic infected decubitus ulcer, the percentages were 56% and 5%, respectively.


Of eight episodes of community-acquired bacterial endocarditis in very old patients, five were caused by group D streptococci, one by an enterococcus, two by Gram-negative bacteria and none by Streptococcus viridans. In patients aged 60-79 years, only one of 22 community-acquired episodes was caused by group D streptococci, two by enterococci, 12 by Streptococcus viridans, and none by Gram-negative bacteria. Four of the eight patients aged 80 or more with endocarditis had a solid malignancy. All isolates of Salmonella sp. were from patients 60-79 years of age. The percentage of polymicrobial episodes was 6% in both groups.

Treatment, complications and outcome: Appropriate empirical antibiotic treatment was given in 68% of patients aged 80 and over and in 63% of patients 60-79 years of age (p > 0. 1). Thirteen per cent of the very old and 9% of the old were diagnosed as suffering from septic shock (p = 0.08). The frequency of all other complications (adult respiratory distress syndrome, acute renal and liver failure, disseminated intravascular coagulopathy and secondary septic dissemination) was similar in the two groups.

Thirty-five per cent of patients 80 or older and 30% of patients aged 60-79 years died during hospitalization (p > 0.1). Figure 2 details the percentage of survival by days of hospitalization. By the end of the seventh day, 78% of the very old patients and 82% of the old patients were alive (p = 0.1, log-rank test). The median survival time was 29 days for the very old and 42 days for the old (p = 0.04, log-rank test).

Variables associated with a fatal outcome on univariate analysis in the two groups are detailed in Table VI. Fatality was associated with advanced age and with solid malignancies in patients aged 60-79, but not in the very old. A pseudomonal bloodstream infection was associated with increased fatality in the old, but not in the very old, probably because a considerable percentage of Pseudomonas sp. infections in patients aged 80 or over originated from the urinary tract.


The variables which were independently and significantly associated with a fatal outcome in each group, and in all patient8, are detailed in Table VII. Advanced age was not related to fatality when the analysis was conducted in all patients aged 60 or older, nor in patients aged 80 or older. In patients 60-79, an increment of 1 year of age was associated with an increase in the odds ratio for death of 1.1, 95% CI 1.0-1.2. A hospital-acquired infection, shock (or low systolic blood pressure on admission), low serum albumin, renal dysfunction and inappropriate antibiotic treatment were universally associated with fatality.

Patients 90 years of age or more: Forty-one patients were aged 90 or over. The median age in this group was 98 years, and 46% were residents of nursing homes. Only 14% had full functional capacity compared with 43% of patients aged 80-89, but the percentage of bedridden patients was not significantly higher than in patients aged 80-89 (18% and 13%, respectively). Twenty-two per cent of patients aged 90 and over had had a cerebrovascular accident in the past, compared with 15% of patients aged 80-89 (p = 0.007). The percentage of patients treated with antibiotics in the preceding month was 12% in patients aged 90 or more, and 32% in patients aged 80-89 (p = 0.01). Solid malignancies were known to exist in 15% of patients aged 90 or over and in 26 % of the octogenarians (p = 0.1). The sources of bacteraemia, percentage of nosocomial episodes, and the pathogens were identical in patients 80-89 years of age and patients aged 90 or more.

Appropriate antibiotic treatment was started in 63% of patients 90 and over and 68% of patients aged 80-89. The fatality was 24% in patients aged 90 and over, and 36% in patients aged 80-89 (p = 0.1).


During 1988-90 20% of all episodes of bacteraemia and fungaemia in our hospital occurred in patients aged 80 or older. Several features distinguished this group of very old patients with bloodstream infections from younger patients, aged 60-79. More than a third were residents of nursing homes, compared with only 14% of patients aged 60-79. Only 39% of the very old had full functional capacity. The main difference with regard to underlying disorders was in the frequency of past cerebrovascular accidents, 15% in octogenarians 33% in patients aged 90 or more, but only 9% in patients aged 60-79. A similar trend is evident in previous studies. In a group of 175 elderly patients with bacteraemia (mean age 80.3) seen in a geriatric ward by Sonnenblick et al.[10], 38% of patients suffered from |neuropsychiatric disorders', and cerebrovascular accident was a major underlying disease in other groups of elderly bacteraemic patients[7, 17]. Although the prevalence of malignant diseases was similar in the two groups, the very old were treated less frequently with corticosteroids or cytotoxic drugs, ;Ind neutropenia was less prevalent.

Urinary tract infections accounted for half of all episodes of bacteraemia in the very old, compared with only 34% of episodes in patients aged 60-79. The urinary tract is the most common source of bacteraemia in elderly people[1, 7, 8, 10-12, 17]. In the present study we were able to show that the increased frequency of urinary tract infection a8 a source of bacteraemia in the very old is independent of underlying disorders, and is evident in patients with and without urinary, catheters, and in hospitalas well as in community-acquired bacteraemias. An explanation for this can be found in the increasing prevalence of asymptomatic and symptomatic bacteriuria in elderly people[18], and in the higher frequency of bacteraemia in older patients with pyelonephritis[19, 20].

Anaerobic micro-organisms were isolated more frequently from the blood of patients aged 80 or older than from patients aged 60-79, and this difference was statistically significant both on univariate and on multivariate analysis. The predisposition of older patients to anaerobic bloodstream infections was evident for all sources of bacteraemia, but most of all in bacteraemic intra-abdominal infection and infected decubitus ulcers. Both sites are frequently infected with anaerobic bacteria, but an association of anaerobic bacteraemia with age was not reported in previous studies[21-23].

All episodes of community-acquired bacterial endocarditis in patients aged 80 or over were caused by bacteria originating in the gastro-intestinal tract. A similar finding was described by Terpenning et al.[24]. The association of endocarditis caused by bowel bacteria, and especially of Streptococcus bovis, with colon cancer is amply documented[25]. Half of our elderly patients with endocarditis had a colonic carcinoma. Thus the spectrum of endocarditis shifts in elderly patients from infections caused by Streptococcus viridans and Staphylococcus aureus toward Gram-negative, group D streptococcal and enterococcal infections.

The fatality within the first week and the total fatality during hospital stay was similar in the two age groups. In a life-table analysis (Figure 2), the curves of survival diverge after the first week, and patients older than 80 had a shorter median survival time. However, a life-table analysis may be misleading in this situation, as patients with less severe underlying conditions will be discharged earlier, and the percentage of deaths unrelated to the bacteraemic episode probably increased with time.

The risk of a fatal outcome increased with advanced age in patients 60-79 years of age, but not in patients aged 80 or over. In this group the gross fatality actually decreased with age, from 36% in octogenarians to 24% in patients aged 90 or over. Hospital-acquired infection, hypotension and septic shock, and inappropriate empirical antibiotic treatment were strongly associated with a fatal outcome in both groups. Similar associations were delineated by Meyers et al.[11]. Low albumin was significantly and independently associated with increased fatality, as in previous studies in which this correlation was sought[4, 17]. This finding may stem from a dual association. Sepsis and severe acute toxic states cause a rapid decrease in serum albumin[26-28], and low albumin level is a marker of increased mortality in the general population[29].

Neutropenia, an important predictor of death in the younger patients, was too rare to emerge as an influence on outcome in the very old. Hypothermia was not associated with higher fatality in our patients. The association of hypothermia and fatality was stressed in a recent survey[10], but was not found in other studies [1, 7, 8, 11].

A subgroup of interest is the small population of individuals who live to maximum age. Forty-one patients in the present group were aged 90 or more. The prevalence of malignant diseases in this group of patients was lower than in the younger age groups, and a history of cerebrovascular accident more prevalent. The fatality associated with bacteraemia was only 24% in this group, but the sources of bacteraemia and the infecting micro-organisms were similar to those in octogenarians. If extreme longevity is associated with specific characteristics, we were unable to demonstrate such factors related to bacteraemia.

In summary, the very old constitute a major and increasing percentage of patients treated because of bacteraemia and severe bacterial infections. With advanced age, the percentage of patients with curtailed functional capacity and patients who have suffered a cerebrovascular accident is increasing. On the other hand, neutropenia related to cytotoxic drugs is seen less often. Most episodes of bacteraemia in the very old are community acquired and originate from the urinary tract. The very old have a predisposition to anaerobic bloodstream infections, mostly of soft tissue and intra-abdominal origin. Micro-organisms causing endocarditis in our patients aged 80 and over were gut bacteria.

Fatality is not related to advanced age in the very old. The odds to survive are better in patients treated with appropriate empirical antibiotic treatment, regardless of age. Factors related to higher fatality in the very old and amenable to change are prevention of hospital-acquired infections, better management of hypotension and septic shock, and selection of appropriate empirical antibiotic treatment.


[1.] McCue JD. Gram-negative bacillary bacteremia in the elderly: incidence, ecology, etiology and mortality. J Am Geriatr Soc 1987;35:213-18. [2.] Mellors JW, Horwitz RI, Harvey MR, Horwitz SM. A simple index to identify occult bacterial infection in adults with acute unexplained fever. Arch Intern Med 1987;147:666-71. [3.] Leibovici L, Cohen O, Wysenbeek AJ. Occult bacterial infection in adults with unexplained fever: validation of a diagnostic index. Arch Intern Med 1990;150:1270-2. [4.] Leibovici L, Greenshtain S, Cohen O, Mor F, Wysenbeek AJ. Bacteremia iii febrile patients: a clinical model for diagnosis. Arch Intern Med 1991;151:1901-6. [5.] Saviteer SM, Sanisa GP, Rutala WA. Nosocomial infections in the elderly: increased risk per hospital day. Am J Med 1988;84:661-6. [6.] Ispahani P, Pearson NJ, Greenwood D. An analysis of community and hospital-acquired bacteraemia in a large teaching hospital in the United Kingdom. Q J Med 1987;63:427 40. [7.] Setia W, Serventi I, Lorenz P. Bacteremia in a long-term care facility. Arch Intern Med 1984;144:1633-5. [8.] Esposito AL, Gleckman RA, Cram S, Crowley M, McCabe F, Drapkin MS. Community acquired bacteremia in the elderly: analysis of one hundred consecutive episodes. J Am Geriatr Soc 1980;28:315-19. [9.] Editorial. Over ninety. Lancet 1991;338:858. [10.] Sonnenblick M, Carmon M, Rudenski B, Friedlander Y, van Dijk JM. Septicemia in the elderly: incidence, etiology and prognostic factors. Isr J Med Sci 1990;26:195-9. [11.] Meyers BR, Sherman E, Mendelson MH, et al. Bloodstream infections in the elderly. Am J Med 1989;86:379-84. [12.] Chattopadhyay B, Al-Zahawi M. Septicaemia and its unacceptably high mortality in the elderly.J Infect 1983;7:134-8. [13.] McCue JD. Improved mortality in Gram-negative bacillary bacteremia. Arch Intern Med 1985;145:1212-16. [14.] SAS Institute. SAS user's guide: statistics. Version 5 edn. Cary, NC; SAS Institute. 1986;420-4. [15.] Harrell FE. The LOGIST procedure. In: SAS Institute, SUGI: supplemental library user's guide. Version 5 edn. Cary NC: SAS Institute, 1986;269-93. [16.] The LIFETEST Procedure. In: SAS user's guide: statistics. version 5 edn. Cary, NC: SAS Institute. 1985;155-69. [17.] van Dijk JM, Rosin AJ, Rudenski B. Septicaemia in the elderly. Practitioner 1982;226:1439-43. [18.] Baldassarre JS, Kaye D. Special problems of urinary tract infection in the elderly. Med Clin North Am 1991;75:375-90. [19.] Gleckman R, Blagg N, Hilbert D, et al. Acute pyelonephritis in the elderly. South Med J 1982;75:551-4. [20.] Leibovici L, Greenshtain S, Cohen O, Wysenbeek AJ. Toward improved empiric management of moderate to severe urinary tract infections. Arch Intern Med 1992;152:2481-6. [21.] Brook I. A 12 year study of aerobic and anaerobic bacteria in intra-abdominal and post-surgical abdominal wound infections. Surg Gynecol Obstet 1989;169:387-92. [22.] Brook I. Anaerobic bacteria in suppurative genitourinary infections. J Urol 1989;141:889-93. [23.] Brook I. Anaerobic bacterial bacteremia: 12-year experience in two military hospitals. J Infect Dis 1989;160:1071-5. [24.] Terpenning MS, Buggy BP, Kauffman CA. Infective endocarditis: clinical features in young and elderly patients. Am J Med 1987;83:626-34. [25.] Panwalker AP. Unusual infections associated with colorectal cancer .Rev Infect Dis 1988;10:347-64. [26.] Deysine M, Liwblich N, Aufsen HA. Albumin changes during clinical septic shock. Surg Gynecol Obstet 1973;173:474-8. [27.] Kenny RA, Hodkinson HM, Cox ML, Caspi D, Pepys MB. Acute phase protein response to infection in elderly patients. Age Ageing 1984;13:89-94. [28.] Holcroft JW, Trunkey DD, Carpenter MA. Extravasation of albumin in tissues of normal and septic baboons and sheep. J Surg Res 1979;26:341-7. [29.] Phillips A, Shaper AG, Whincup PH. Association between serum albumin and mortality from cardiovascular disease, cancer, and other causes. Lancet 1989; ii: 1434-6.
COPYRIGHT 1993 Oxford University Press
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1993 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Leibovici, Leonard; Pitlik, Silvio D.; Konisberger, Hanna; Drucker, Moshe
Publication:Age and Ageing
Date:Nov 1, 1993
Previous Article:Dying when you're old.
Next Article:Quantification of proteinuria: a re-evaluation of the protein/creatinine ratio for elderly subjects.

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters