American Journal of Respiratory and Critical Care Medicine

Community-acquired pneumonia (CAP) in the elderly has a different clinical presentation than CAP in other age groups. Confusion, alteration of functional physical capacity, and decompensation of underlying illnesses may appear as unique manifestations. Malnutrition is also an associated feature of CAP in this population. We undertook a study to assess the clinical and nutritional aspects of CAP requiring hospitalization in elderly patients (over 65 yr of age). One hundred and one patients with pneumonia, consecutively admitted to a 1,000-bed teaching hospital over an 8-mo period, were studied (age: 78 ± 8 yr, mean ± SD). Nutritional aspects and the mental status of patients with pneumonia were compared with those of a control population (n = 101) matched for gender, age, and date of hospitalization. The main symptoms were dyspnea (n = 71), cough (n = 67), and fever (n = 64). The association of these symptoms with CAP was observed in only 32 patients. The most common associated conditions were cardiac disease (n = 38) and chronic obstructive pulmonary disease (COPD) (n = 30). Seventy-seven (76%) episodes of pneumonia were clinically classified as typical and 24 as atypical. There was no association between the type of isolated microorganism and the clinical presentation of CAP, except for pleuritic chest pain, which was more common in pneumonia episodes caused by classical microorganisms (p = 0.02). This was confirmed by a multivariate analysis (relative risk [RR] = 11; 95% confidence interval [CI]: 1.7 to 65; p = 0.0099). The prevalence of chronic dementia was similar in the pneumonia cohort (n = 25) and control group (n = 18) (p = 0.22). However, delirium or acute confusion were significantly more frequent in the pneumonia cohort than in controls (45 versus 29 episodes; p = 0.019). Only 16 patients with pneumonia were considered to be well nourished, as compared with 47 control patients (p = 0.001). Kwashiorkor-like malnutrition was the predominant type of malnutrition (n = 65; 70%) in the pneumonia patients as compared with the control patients (n = 31; 31%) (p = 0.001). The observed mortality was 26% (n = 26). Pleuritic chest pain is the only clinical symptom that can guide an empiric therapeutic strategy in CAP (typical versus atypical pneumonia). Both delirium and malnutrition were very common clinical manifestations of CAP in our study population.

Community-acquired pneumonia (CAP) in the elderly has a different clinical presentation than CAP in other age groups. Usually, the clinical picture in elderly patients is incomplete and both fever and cough can be absent. Furthermore, confusion, alteration of functional physical capacity, and decompensation of underlying illness may appear as a single manifestation. Additionally, malnutrition associated with CAP is a very common finding in persons of advanced age (1). The incomplete clinical picture of CAP in the elderly may be associated with a delay in establishing the diagnosis and, consequently, in starting adequate antibiotic therapy. Delay in diagnosis and treatment may contribute to the higher observed death rate in the elderly population with CAP (2-4). However, even when the diagnosis of pneumonia is straightforward, empiric antibiotic therapy is difficult to adjust, with available information suggesting that clinical markers do not help to distinguish classical CAP from that caused by atypical microorganisms (5-9). The aim of the present study was to assess the clinical characteristics of hospitalized elderly (over 65 yr of age) patients with CAP, with special emphasis on the nutritional aspects and mental status of these patients, as well as to determine the association between typical and atypical clinical presentations of CAP with the different microbial etiologies of this disease.


We studied 101 patients aged 65 yr or more, admitted consecutively to our hospital (a 1,000-bed teaching hospital) because of CAP (pneumonia cohort) between November 1993 and June 1994. This is the same study population as described in a previous report (1). Patients admitted to the hospital during the month previous to the beginning of the study and those with immunosuppression (patients with acquired immunodeficiency syndrome [AIDS] or those receiving chemotherapy and/or corticosteroids [equivalent prednisone dose: 20 mg/ d]) were excluded. For the analysis of mental status and nutritional factors associated with the development of CAP, each case patient was paired with a control subject without pneumonia arriving at the emergency room within ± 2 d of the patient; and who was of the same gender and age (± 5 yr if over 65 yr) and was admitted to the hospital. An inclusion criterion for control patients was the absence of any pneumonia episode during the 3 yr preceding the study. Patients were examined during the first 24 h of their arrival in the emergency room and also throughout their hospital stay. Patients were managed according to the decisions of the attending physician. A follow-up visit was done 30 d after each patient's entry into the study by one of the investigators. Thus, a matched case–control design was used to assess the association of mental status and nutritional aspects with the development of pneumonia, and a cohort study, including only patients who developed pneumonia, was done to describe the clinical characteristics of the disease. For the purpose of the study, the inclusion criteria for pneumonia were the presence of new chest radiographic infiltrates on admission, and at least one of the major or two of the minor criteria described previously (7). Major criteria were cough, sputum production, or a history of fever (37.8° C); minor criteria were dyspnea, pleuritic chest pain, pulmonary consolidation on examination, and leukocyte count > 12,000/ml.


Two serial blood cultures and serologic examinations were performed on all patients. Respiratory samples were obtained whenever possible by means of sputum collection (n = 47), bronchoalveolar lavage (BAL) (n = 8), protected specimen brush (n = 7), transthoracic needle aspiration (n = 3), and pleural fluid aspiration (n = 2). Sputum samples could not be obtained from 54 patients. According to previous reported criteria (1), the etiology of pneumonia was considered as definite in 43 cases.

Data Collection

Data recorded prospectively for both bases and controls were age, gender, residency, physical activity (as defined subsequently), alcohol intake (g/d), smoking habit (pack-yr), prior mental status and mental status upon admission, number of hospital admissions during the preceding year, number and type of underlying disease states, nutritional state (complete nutritional assessment), and Simplified Acute Physiology Score (SAPS) (11). Additionally, the following laboratory values were recorded: hemoglobin, leukocyte count, and creatinine.

In the pneumonia cohort, we examined the clinical presentation on admission for fever, asthenia, anorexia, chills, pleuritic pain, cough, sputum production, dyspnea, and delirium. Physical examination included the presence or absence of crepitations and rhonchi on auscultation. Radiographic infiltrates were evaluated as alveolar or interstitial and unilateral or bilateral.

The following end-point variables were described for all patients according to a previous report (12): shock, renal failure, admission to the intensive care unit (ICU), need for mechanical ventilation, progression of chest X-ray infiltrates, pleural effusion, empyema, cavitation, duration of hospital stay, nonresolution of pneumonia, and death.

Causal microorganisms were categorized according to microbial etiology into classical (Streptococcus pneumoniae, Moraxella catarrhalis, and gram-negative bacilli) and atypical (Chlamydia pneumoniae, Coxiella burnetti, and Mycoplasma pneumoniae). Streptococcus viridans and mixed buccal flora were not considered for our classification. Legionella spp. was considered an atypical microorganism in the present study.

Clinical presentation was classified as: (1) typical (suggesting a classical microbial etiology); or (2) atypical (suggesting an atypical microbial etiology) according to prior definitions (13).

Nutritional Assessment

The quality of feeding was evaluated with a simple questionnaire answered by the patient or a relative in cases of mental confusion. The following variables were recorded for all patients and controls: body weight, height, mid-arm perimeter, and triceps skinfold thickness, as well as protein-nutrition parameters (lymphocyte count, hemoglobin, total protein, albumin, prealbumin, transferrin, zinc, and retinol-binding protein [RBP]). Triceps skinfold, thickness, mid-arm perimeter, and serum albumin were considered as representative of body fat, muscle protein, and visceral protein, respectively. This allowed the classification of patients into the following grades of protein-energy malnutrition used at our institution: (1) well nourished; (2) fat-deficit marasmus; (3) muscle-deficit marasmus; (4) mixed-deficit marasmus; (5) kwashiorkor-like; (6) fat deficit/mixed malnutrition; (7) muscle deficit/mixed malnutrition; and (8) combined mixed malnutrition. Anthropometric tables validated for the Barcelona city population were employed (15).

Other Definitions

Physical activity was estimated by means of the Karnofsky scale (16). Patients were considered to be bedridden if they had a performance scale score > 3 (16). Diagnosis of chronic obstructive pulmonary disease (COPD) was based on clinical criteria and on spirometric values (if available before or after hospital admission).

Delay of diagnosis was arbitrarily considered when there was a deferral of more than 72 h. Chest-radiographic progression of disease was defined as an increase in the size of a pulmonary opacity by 50% or more within 48 h of admission (13). Nonresolution of pneumonia was defined as the persistence of (> 30 d), but not necessarily a change in chest X-ray infiltrates despite proper antibiotic treatment (17). These patients were subsequently followed until complete resolution of their chest X-ray infiltrates.

The patients' mental status was evaluated through a history obtained from spouses or other family members, and from caregivers when necessary. A history of dementia, cognitive decline, and previous episodes of delirium was pursued. Previous mental impairment was considered when other causes of global cognitive impairment persisted for more than 1 mo. Delirium was defined as an acute, reversible state of confusion (18, 19), and this definition included two patterns of altered mental status: acute confusion, and deterioration in baseline cerebral function when there was previous mental impairment. Clinical history, physical examination, and laboratory findings were used for distinguishing chronic mental alterations from delirium in cases of possible overlap.

Statistical Analysis

For the comparison of means, we used the nonparametric Mann– Whitney U test for continuous variables that departed from a normal distribution. Otherwise, we used Student's t test, and the chi-square test (Fisher's exact test when needed) for discrete data. To determine variables that could predict etiology (classical versus atypical microorganisms), we used a logistic regression technique and categorized continuous variables. The adjusted risk for developing pneumonia caused either by classical or atypical microorganisms was estimated for the following variables: pleuritic chest pain (yes/no) and serum protein level (< 60 or > 60 mg/dl). These variables were selected according to their p value by univariate analysis (0.10), and because they were deemed biologically plausible. In order to optimize the threshold that would discriminate two risk groups for each variable in relation to length of hospital stay, we used the formula for threshold computation based on the median value (50th percentile). All reported p values were two-tailed. The Statistical Package for Social Sciences (SPSS/PC+, Chicago, IL) software package was used for the statistical analyses.

General Characteristics and Underlying Diseases

One-hundred and one patients had pneumonia and were considered as cases, and an equal number without pneumonia were used as controls. The general characteristics and underlying conditions of the pneumonia cohort are outlined in Table 1. The patients' age was 78 ± 8 yr (mean ± SD). Thirty-six patients had previous normal physical activity as determined by a Karnofsky score > 80. The most frequent underlying disease states in the pneumonia cohort were cardiac disease in 38 cases, COPD in 30 cases, arterial hypertension in 29 cases, and neurologic diseases in 26 cases. Among the 30 patients with COPD, 19 (63%) had spirometric values for a FVC of 52 ± 29% predicted (mean ± SD); FEV1 of 41 ± 26% predicted, and FEV1/FVC of 77 ± 15%. Twenty-three (77%) of the 30 COPD patients were receiving regular treatment; five were receiving long-term oxygen therapy, and two had been mechanically ventilated during their last previous hospital stay. The mean number of exacerbations of COPD was 2 ± 1 per year. There were 17 (17%) patients with non-insulin-dependent diabetes mellitus. Four (4%) patients had liver diseases. Five patients had neoplasia (two lung cancer, two digestive malignancy, and one skin cancer).


Age, yr78 ± 8
Gender, M/F67/34
 Nursing home15 (15%)
 Family79 (79%)
 Alone 7 (7%)
Previous physical activity
 Normal, Karnofsky score 100–8036 (36%)
 Diminished, Karnofsky score 70–6027 (27%)
 Very limited, Karnofsky score ⩽ 5038 (38%)
 Bedridden12 (12%)
Hospitalization during previous year32 (32%)
Alcohol intake, g/d10 ± 24
Smoking habit, pack-yr27 ± 34
No. of underlying diseases1 ± 1
Type of underlying illness
 COPD30 (30%)
 Cardiac38 (38%)
 Neurologic disorders26 (26%)
 Diabetes mellitus17 (17%)
 Neoplasia 5 (5%)
 Renal failure 4 (4%)
 Liver diseases 4 (4%)
SAPS on admission10 ± 3

Definiton of abbreviations: COPD = chronic obstructive pulmonary disease; SAPS = simplified acute physiology score.

*Values are means ± SD.

Clinical Features

The duration of the symptoms of pneumonia before the hospitalization was 6 ± 5 d (mean ± SD). The main clinical features of the pneumonia episodes are shown in Table 2. The main presentations were dyspnea in 71 patients, cough in 67, and fever in 64. Nineteen patients did not present cough, pleuritic chest pain, or purulent sputum. The association of cough, dyspnea, and fever was observed in 31 patients. Seven patients presented without a clear clinical picture (i.e., without another sign or symptom) such as a confusional state, dyspnea alone, asthenia, or anorexia. Seventy-six (76%) patients had a leukocyte count > 10.000/mm3.


Symptoms on Admissionn(%)Auscultatory Findingsn(%)Radiographic Infiltrate Patternn(%)Radiographic Infiltrate Localizationn(%)
Cough67(67)Rhonchi12 (12)Interstitial 6 (6%)Bilateral30(30%)
Fever64(64)Pulmonary consolidation11(11)Mixed13(13%)
Asthenia58(58)Bronchial breathing 2 (2)
Purulent sputum52(52)Diminished breath sounds 1 (1)
Anorexia50 (50)Normal 9 (9)
Altered mental state45(45)
Pleuritic chest pain34(34)

The chest-radiographic infiltrate pattern on admission was alveolar in 82 cases and unilateral in 71. The most frequent auscultatory finding was crepitations (n = 65).

The analysis of the mental states of patients prior to the episode of pneumonia showed that chronic dementia was similar in both the pneumonia cohort (n = 25) and in the control group (n = 18; p = 0.22). However, delirium at admission was significantly more frequent in the pneumonia cohort than in controls at 45 patients versus 29 patients, respectively (p = 0.019). Table 3 shows the previous mental states and the mental states on admission of both patients and controls.


CasesControlsp Value
Previous mental state
 Chronic dementia25(25)* 18(18)0.2289
Mental state on admission
 Delirium45(45)* 29(29)0.019
 Usual mental state56(55)72(71)

*p = 0.003 for chronic dementia (25 of 101 patients) versus delirium (45 of 101 patients).

Diagnosis of pneumonia was established during the first medical visit in 71 patients. In the remaining 40 patients the initial diagnoses were acute bronchitis in 17, viral upper-respiratory-tract infection in six, and decompensation of underlying diseases in seven. In 62 patients diagnosis was considered to have been delayed > 72 h from the beginning of symptoms. Causes of delay were patients themselves in 29 cases, medical practice in 28 cases, and family members in five cases. There was no association between mortality and the different causes of delay.

Relationship Between Clinical Findings and Etiology

Etiologic diagnosis was established in 43 (42%) of the cases. The etiologic agents were S. pneumoniae 19 (36%) cases, C. pneumoniae in nine (17%) cases, C. burnetti in six (11%) cases, Legionella pneumophila in three (6%) cases, M. pneumoniae in two (4%) cases, Pseudomonas aeruginosa in one (2%) case, Proteus mirabilis in one (2%) case, S. viridans in one (2%) case, and M. catarrhalis in one (2%) case. Blood cultures were positive in 10 cases, all yielding S. pneumoniae.

Seventy-seven pneumonia episodes were clinically classified as typical and 24 as atypical. There was no association between the type of isolated microorganism and the clinical presentation. When comparing classical microorganisms (S. pneumoniae, M. catarrhalis, and gram-negative bacilli) with atypical microorganisms, no differences were found in terms of the clinical presentation except for pleuritic chest pain, which was more frequent in pneumonia episodes caused by classical microorganisms (p = 0.02) (Table 4). The multivariate analysis adjusted for the presence of pleuritic chest pain, and total serum proteins demonstrated that chest pain was the only independent factor predicting a classical etiology (RR = 11; 95% CI: 1.7 to 65; p = 0.0099). Also, there were no differences between classical and typical microorganisms with regard to radiographic presentation, anthropometric measurements, or laboratory values (Table 4).


VariableClassicalAtypicalp Value
Age, yr74 ± 8      77 ± 7n.s.
Gender, M/F9/915/5n.s.
Signs and symptoms
 Fever15 (84%)13 (65%)n.s.
 Asthenia12 (67%)13 (65%)n.s.
 Anorexia 8 (44%)10 (50%)n.s.
 Chills 8 (44%) 4 (20%)n.s.
 Chest pain10 (56%) 3 (15%)0.02
 Cough14 (78%)12 (60%)n.s.
 Purulent sputum 9 (50%)10 (50%)n.s.
 Dyspnea13 (72%)13 (65%)n.s.
 Delirium 8 (44%) 5 (25%)n.s.
Chest radiographic pattern
 Alveolar17 (94%)15 (75%)0.18
 Interstitial1 (6%) 5 (25%)n.s.
Anthropometric measurements
 Weight, kg       60 ± 10            61 ± 10n.s.
 Mid-arm perimeter, cm25 ± 4      25 ± 3n.s.
 Triceps skinfold thickness, mm12 ± 7      13 ± 7n.s.
Laboratory findings
 Hb, g/L 13 ± 3      13 ± 2n.s.
 Leukocyte count, × 109/L13 ± 8      16 ± 7n.s.
 Lymphocyte count, × 109/L        0.9 ± 0.8  1.4 ± 0.960.12
 Serum protein, g/L60 ± 8      65 ± 80.06
 Serum albumin, g/L30 ± 5      32 ± 60.15
 Prealbumin, mg/dl14 ± 9      12 ± 7n.s.
 Zinc, μg/dL51 ± 8      50 ± 18n.s.
 Transferrin, mg/dl212 ± 77     207 ± 68n.s.
 Retinol binding protein, mg/dl       2.9 ± 2.4   2.6 ± 1.1n.s.

Nutritional Aspects

The results of nutritional assessment, including anthropometric measurements and biochemical results, are shown in Table 5. Patients with pneumonia weighed less (62 ± 12 kg) than controls (66 ± 17 kg) (p = 0.043). Total serum protein, albumin, and prealbumin levels were significantly higher in control patients (p = 0.009).


CasesControlsp Value
Anthropometric measurements
 Height, cm 165 ± 7 165 ± 8 0.872
 Weight, kg  62 ± 12  66 ± 17 0.043
 Mid-arm perimeter, cm  25 ± 4  25 ± 4 0.848
 Triceps skinfold, mm  12 ± 8  13 ± 10 0.980
Laboratory findings
 Lymphocytes1,130 ± 9031,310 ± 850 0.148
 Serum proteins, g/L  62 ± 8  65 ± 6 0.009
 Serum albumin, g/L  30 ± 5  35 ± 5 0.000
 Prealbumin, mg/dl  15 ± 12  19 ± 11 0.030
 Transferrin, mg/dl 204 ± 69 216 ± 61 0.273
 Retinol binding protein, mg/dl   3 ± 2   3 ± 2 0.574
 Zinc, mg/dl  50 ± 21  56 ± 31 0.211
 Well nourished15 (16%)47 (47%)0.0000
 Fat-deficit marasmus 0 6 (6%)
 Muscle-deficit marasmus 2 (2%) 5 (5%)
 Mixed marasmus 0 1 (1%)
 Kwashiorkor-like65 (70%)31 (31%)0.0000
 Fat-deficit/mixed malnutrition 5 (6%) 6 (6%)
 Muscle-deficit/mixed malnutrition 4 (4%) 3 (3%)
 Combined mixed malnutrition 2 (2%) 1 (1%)

Using triceps skinfold thickness, midarm perimeter, and serum albumin, we classified 93 cases and 100 controls as well nourished or as falling within some grade of malnutrition. Only 15 patients with pneumonia were considered to be well nourished, as compared with 47 control patients, (p < 0.001). The remaining 78 pneumonia patients had some type of malnourishment. Kwashiorkor-like malnutrition was the predominant type in the pneumonia patients as compared with the controls (n = 65 [70%] versus n = 31; p < 0.001). The quality of feeding prior to hospital admission was considered regular or poor in 27% of the CAP cases as compared with only 15% of control subjects (p = 0.037). We did not show a significant correlation of nutritional level with either the presence of fever or the serum albumin level (p = 0.087), or with the concentration of RBP (p = 0.376).


Twenty patients presented progression of their chest-radiographic infiltrates, nine showed pleural effusion, four showed empyema, and two showed cavitation. Thirty-three patients had respiratory failure defined as PaO2 levels < 60 mm Hg, of whom 10 needed ICU admission and six required mechanical ventilatory support on routine basis. Eight patients suffered from septic shock.

The mortality rate in this study was 26% (26 of 101 CAP patients). In another study of the same population (1) we found that the number of affected lobes (> 3 lobes), respiratory rate (> 30/min), shock, previous bedridden status cyanosis, diminished physical activity, absence of fever (< 37° C), suspicion of aspiration, swallowing disorders, serum creatinine (> 1.4 mg/dL), PaO2 /Fi O2 ratio (< 200), deteriorated mental status, leukocyte count (⩾ 14.9 × 109/L), and cough were factors significantly associated with mortality. The multivariate analysis in the same study selected bedridden status, temperature (⩽ 37° C), and number of affected lobes (⩾ 3) as variables significantly related to death. None of the remaining clinical or nutritional variables was statistically significant. Nine of the 26 patients who died presented with pneumonia within the first 72 h of admission. Of these nine patients, five (55%) were considered to have had a delayed diagnosis of pneumonia. However, we did not find any relationship between delayed diagnosis and death.

We analyzed 41 clinical and nutritional variables and their relation to the length of hospital stay. None of these variables showed statistically significant differences in relation to the length of stay. The length of hospital stay in relation to a selection of the most relevant clinical and nutritional variables is given in Table 6. The length of hospital stay in survivors (n = 75) was 12 ± 12 d (mean ± SD; range: 1 to 60 d).


Present (n)Absent (n)p Value*
Clinical variables
 Aspiration12 ± 12 (10) 8 ± 28 (65)0.38
 Quality of life, good/poor11 ± 10 (67)15 ± 19 (8)0.47
 Chest pain10 ± 10 (48)12 ± 12 (27)0.44
 Bedridden 9 ± 4 (4)12 ± 12 (71)0.62
 Fever, < 37° C11 ± 14 (22)12 ± 10 (53)0.73
 Respiratory rate, > 30/min12 ± 10 (49)12 ± 11 (26)0.96
 Previous mental state, altered13 ± 12 (14)11 ± 11 (61)0.54
 No. of affected lobes, > 3 lobes12 ± 14 (4)13 ± 12 (71)0.87
 Delay in diagnosis, > 72 h12 ± 12 (48)12 ± 12 (27)0.94
 ICU admission13 ± 12 (4)10 ± 10 (71)0.56
 Mechanical ventilation13 ± 12 (3)10 ± 11 (72)0.64
Nutritional parameters
 Quality of feeding, good/poor12 ± 14 (59)12 ± 11 (16)0.97
 Serum albumin, < 10 g/L13 ± 12 (20)10 ± 10 (55)0.27
 Nutritional prognostic index, < 4813 ± 12 (27)11 ± 11 (48)0.46
 Serum prealbumin, < 10 mg/dl11 ± 12 (27)11 ± 7 (48)0.84
 Total serum proteins, < 45 g/L13 ± 11 (3)10 ± 11 (72)0.64
 Retinol binding protein, < 2.6 mg/dl13 ± 12 (24)10 ± 10 (51)0.25
 Transferrin, 200 mg/dl12 ± 10 (27)12 ± 12 (48)0.94
 Triceps skinfold thickness, 14 mm12 ± 7 (16)13 ± 7 (59)0.61

*p Values calculated from non parametric Mann-Whitney test.

Nonresolution of pneumonia was considered to have occurred in 18% (18 of 101) of the cases. All of these subjects were subsequently followed, and alternative diagnoses of radiographic infiltrates were always discarded.

The main findings of our study were that: (1) The clinical picture of CAP in the elderly is usually incomplete; in addition, there is a wide variety of signs and symptoms, with no association among them. (2) A primary diagnosis of pneumonia was not considered as such in 40 cases during the first medical visit, and the definite diagnosis of pneumonia was delayed for more than 72 h in 62 patients. (3) Delirium at admission was very common (45%) in patients with pneumonia, and significantly more frequent than in controls. (4) We did not find an association between the clinical presentation (typical versus atypical) and etiology. However, pleuritic chest pain was found to be an independent risk factor for pneumonia caused by “classical” microorganisms. (5) Malnutrition was more common in elderly patients with pneumonia than in controls.

Elderly patients hospitalized because of CAP constitute a special population, since they commonly have underlying illnesses, prior neurologic disturbances, and nutritional and/or immunologic deficits. CAP in this population has peculiar clinical characteristics. For instance, not all the signs and symptoms of pneumonia are present in all cases. The clinical presentation may consist only of an alteration of the patient's general condition, confusion, or decompensation of underlying disease. Harper and Newton (3) found that the classical constellation of CAP in the elderly of cough, fever, and dyspnea was absent in 56% of patients (9). However, complete absence of symptoms occurred in only 10% of their patients (9). Venkatesan and colleagues found that 96% of CAP patients had at least one respiratory symptom (4). Moreover, they found that the most common sign of infection was leukocytosis. In our study, a complete absence of clinical data was shown in 7% of the cases. Again, the combination of cough, fever, and dyspnea was absent in 69% of our patients. Seventy-six percent of our patients had leukocytosis.

The frequency of fever in elderly patients with CAP ranges from 33% to 60% (2, 4). In our study, 36% of the patients had no fever. Differences in this finding among studies could be due to variability in immune function, the frequency and severity of chronic underlying diseases, and/or the use of different tools to measure temperature. Kauffman and colleagues (20) observed that a poor febrile response due to a decreased release of interleukin-1 (IL-1) was related to low serum albumin and RBP concentrations in protein–energy-malnourished patients. In our study, although the absence of fever and low serum albumin were common findings, we could not find a significant correlation between fever and the different thresholds of serum albumin and RBP levels.

Our study affirms the difficulties met when diagnosing pneumonia in the elderly, and reminds the physician to consider the clinical ambiguity of pneumonia presentations in this type of population. Probably, pneumonia in the elderly is underdiagnosed with use of the classical criteria. Therefore, other factors, such as acute mental alterations or malaise, should be kept in mind when diagnosing pneumonia in this population. As alluded to earlier, clinical suspicion of CAP in the elderly is not an easy task. Some authors have suggested that delay in diagnosis and initiation of appropriate treatment may contribute to the higher death rate in the elderly population with CAP (2). In the present study, 30% of the patients were initially misdiagnosed as not having pneumonia. Furthermore, diagnosis was delayed for more than 72 h in 62% of cases. Although early diagnosis of pneumonia probably provides the best hope for averting morbidity and mortality in the highly vulnerable elderly population, we did not find that it had a clear relationship with outcome. However, given that there was no definitive diagnosis of pneumonia established at an earlier time, we do not know whether some patients' disease evolved from severe bronchitis into pneumonia. This could explain the lack of association between delay in diagnosis and outcome.

Elderly patients with pneumonia commonly present with neurologic symptoms. Delirium should be differentiated from other, previous causes of global cognitive impairment such as dementia, depression, and functional psychosis (18). These latter conditions are commonly seen in elderly patients. The term “delirium” refers to an acute, reversible state of confusion. In our study we did not observe differences in pneumonia in patients versus controls with regard to prior mental status (Table 3). However, elderly patients with pneumonia more commonly showed delirium (45%) than did controls (29%), and acute alterations in mental status (i.e., delirium) was significantly more common than chronic mental alterations. The prevalence of delirium in hospitalized elderly patients is reported as being as high as 50% (18). Venkatesan and coworkers (4) and Harper and associates (3) found a prevalence of 47% and 15%, respectively, of mental alteration in elderly patients with CAP at admission. However, they did not discriminate between chronic mental alteration and delirium. We suggest that our study provides a better understanding of delirium as an expression of mental alteration induced by pneumonia in the elderly population. A practical conclusion of clinical interest is that an effort has to be made to take chest radiographs, upon arrival at the hospital, of all elderly patients with delirium.

A former study by our group found that S. pneumoniae was the most common etiology of pneumonia in documented episodes in the elderly (1). In addition, 17% of cases were due to atypical microorganisms such as C. pneumoniae, M. pneumoniae, and C. burnetti. In the present study, clinical manifestations did not discriminate between classical and atypical microorganisms. Other studies attempting to elucidate etiology from clinical parameters have had poor results in patients hospitalized for CAP (5-8). However, we have shown that pleuritic chest pain was an independent factor for predicting causation of CAP by classical microorganisms, especially S. pneumoniae. Recently, Olaechea and coworkers (21), studying ICU-admitted patients with CAP, described a scoring system that could be of help in predicting the etiology of the disease and hence the initial treatment of these patients. This scoring system used the combination of four variables: acute physiology and chronic health evaluation II (APACHE II), serum sodium and phosphorus, and the duration of symptoms. The model described by Olaechea and coworkers showed a sensitivity of 90% and a specificity of 72% for predicting typical pneumonia. The failure to predicting etiology in CAP through the use of clinical markers alone leads to broad-spectrum antibiotic coverage policies. Unfortunately, antibiotic therapy may have important adverse effects, particularly in elderly patients. Therefore, we think that predictive-scoring modeling is warranted in the elderly population in order to elicit a specific etiology and thereby design a better initial empiric management regimen.

It appears that the observed age-related increase in both morbidity and mortality in CAP is not due to age per se (1), but rather to the interactions of immune competence, systemic diseases, and nutritional factors. The degree of risk for acquiring CAP would be related to the amount of overlap among these factors. Malnutrition may play an important role in favoring the development of CAP in the elderly. In a prior study, we observed that a low serum albumin level was an additional risk factor for developing CAP in the elderly (1). Overall, nutritional deficits have been reported in 35% of the elderly population (22). Kwashiorkor occurs in 20% to 60% of hospitalized elderly patients, and also in 5% to 10% of elderly subjects in the community (23). In our study, we found high rates of malnutrition in both cases of CAP (86%) and controls (54%), as assessed with the classification system described earlier. Furthermore, the frequency of kwashiorkor found in our control patients (31%) fits with that in previously reported studies (24). However, this figure was significantly lower than that observed in elderly patients with pneumonia (70%). Unfortunately, physicians are usually unaware that their elderly patients are malnourished. Accordingly, this risk factor, if properly diagnosed, is amendable to medical intervention.

Mortality associated with CAP in the elderly is still high. In our study, the mortality rate was 26%, a figure that fits with that in previous reports (24). Prognostic indicators of mortality in the population analyzed in the present study have been reported before (1). Of particular interest was the finding that absence of fever was associated with poor prognosis in the multivariate analysis. Other studies have documented that elderly individuals who present with indistinct features of pneumonia have an increased mortality. For instance, Venkatesan and coworkers have also reported that apyrexia was a prognostic indicator of mortality (24). The mean length of hospital stay among survivors in our study was 12 d. This figure is lower than the average stay of 21 d reported in previous studies (24). We also analyzed the length of hospital stay in relation to several clinical and nutritional variables. We did not find any relationship between these variables and the length of stay. However, the present study was not initially designed to examine this type of outcome, and the sample size for some of the variables was limited.

In summary, our study can contribute to a better understanding of the clinical picture of CAP in the elderly, and may help the early detection and management of pneumonia in this population.

The authors are indebted to Dr. Robert Rodriguez-Roisin for the revision of the manuscript.

Supported by the Fundació Privada Clı́nic per a la Recerca Biomèdica/Comissió Interdepartamental per a la Recerca i Tecnologia (CIRIT); GRQ94-9103; SEPAR Infecciones 1994; and Fondo de Investigación Sanitaria de la Seguridad Social (FIS) 95/1903.

1. Riquelme R., Torres A., El-Ebiary M., Puig de la Bellacasa J., Estruch R., Mensa J., Fernández-Solá J., Hernández C., Rodriguez-Roisin R.Community-acquired pneumonia in the elderly: a multivariate analysis of risk and prognostic factors. Am. J. Respir. Crit. Care Med154199614501455
2. Andrews J., Chandrasekaran P., McSwiggan D.Lower respiratory tract infection in an acute geriatric male ward: a one-year prospective surveillance. Gerontology301984290296
3. Harper C., Newton P.Clinical aspects of pneumonia in the elderly veteran. J. Am. Geriatr. Soc371989559872
4. Venkatesan P., Gladman J., Macfarlane J. T., Barer D., Berman P., Kinnear W., Finch R. G.A hospital study of community acquired pneumonia in the elderly. Thorax451990254258
5. Granados A., Podzamczer D., Gudiol F., Manresa F.Pneumonia due to Legionella pneumophila and pneumococcal pneumonia: similarities and differences on presentation. Eur. Respir. J21989130134
6. Pedro-Botet M. L., Sabria-Leal M., Haro M., Rubio C., Gimenez G., Sopena N., Tor J.Nosocomial and community-acquired Legionella pneumonia: clinical comparative analysis. Eur. Respir. J8199519291933
7. Farr, B. M., D. L. Kaiser, B. D. Harrison, and C. K. Connolly. 1989. Prediction of microbial aetiology at admission to hospital for pneumonia from the presenting clinical features. Thorax 1031–1035.
8. 1993. Guidelines for the initial management of adults with community-acquired pneumonia: diagnosis, assessment of severity, and initial antimicrobial therapy. Am. Rev. Respir. Dis. 148:1418–1426.
9. Granton J. T., Grossman R. F.Community acquired pneumonia in the elderly patients. Clin. Chest Med141993537553
10. Fang G. D., Fine M. J., Orloff J., Arisumi D., Yu V. L., Kapoor W., et al.New and emerging etiologies for community-acquired pneumonia with implications for therapy: a prospective multicenter study of 359 cases. Medicine691990307316
11. Le Gall J. R., Loirat P., Alperovitch A., et al.A simplified acute physiology score for ICU patients. Crit. Care Med121984975977
12. Celis R., Torres A., Gatell J. M., Almela M., Rodriguez-Roisin R., Agusti A., VidalNosocomial pneumonia: a multivariate analysis of risk and prognosis. Chest931988318324
13. Torres A., Serra-Batlles J., Ferrer A., Jimenez P., Celis R., Cobo E., et al.Severe community-acquired pneumonia. Epidemiology and prognostic factors. Am. Rev. Respir. Dis1441991312318
14. Gasull M. A., Cabre E., Vilar L., Montserrat A.Protein- energy malnutrition: an integral approach and a simple new classification. Human Nutrition: Clin. Nutr381984419431
15. Alastrue Vidal, A., M. Rull Lluch, I. Camps Ausas, C. Ginesta Nus, M. R. Melus Moreno, and J. A. Salva LacombeNeuvas normas y consejos en la valoración de los parámetros antropométricos en nuestra población: ı́ndice adiposo-muscular, ı́ndices ponderales y tablas de percentiles de los datos antropométricos útiles en una valoración nutricional. Med. Clin. (Barcelona)911988223236
16. World Health Organization. 1980. International Classification of Impairment Disabilities and Handicaps: a Manual or Classification Relating to Consequences of Diseases. World Health Organization, Geneva.
17. Kirtland S. H., Winterbauer R. H.A slowly resolving chronic and recurrent pneumonia. Clin. Chest Med121991303318
18. Rummans T., Evans J., Krahn L., Fleming K.Delirium in elderly patients: evaluation and management. Mayo Clin. Proc701995989998
19. American Psychiatric Association. 1994. Diagnostic and Statistical Manual of Mental Disorders, 4th ed. American Psychiatric Association, Washington, DC. 123–133.
20. Kauffman C., Jones P., Kluger M.Fever and malnutrition: endogenous pyrogen/interleukin-1 in malnourished patients. Am. J. Clin. Nutr441986449452
21. Olaechea P. M., Quintana J. M., Gallardo M. S., Insausti J., Maravi E., Alvarez B.A predictive model for the treatment approach to community-acquired pneumonia in patients needing ICU admission. Intensive Care Med22199612941300
22. Chandra R. K.Nutritional regulation of immunity and risk of infection in old age. Inmunology641989141147
23. Hoffman N.Diet in the elderly: needs and risk. Med. Clin. North Am771993745755
24. Venkatesan P., Gladman J., MacFarlane J. T., Barer D., Kinnear W., Finch R. G.A hospital study of community acquired pneumonia in the elderly. Thorax451990254258
Correspondence and requests for reprints should be addressed to Antoni Torres, M.D., Servei de Pneumologia i Al.lérgia Respiratória, Hospital Clı́nic, Villarroel 170, 08036 Barcelona, Spain.

Dr. Riquelme was a Research Fellow from the Centro Médico de Especialidades Respiratorias, Puerto Montt, Chile.


No related items
American Journal of Respiratory and Critical Care Medicine

Click to see any corrections or updates and to confirm this is the authentic version of record