Community-Acquired Pneumonia

Pneumonia remains one of the "captains of death," as Sir William Osler once called it. It is the leading cause of death from infection and is the sixth leading cause of death overall. In the United States, 4 million adults-1.5% of the population-develop community-acquired pneumonia each year. About 1% of all visits to internists are for pneumonia. The total economic burden is estimated to be about $5 billion a year. About 80% of patients are treated as outpatients, at a cost of about $1 billion; 10% of that cost is for antibiotics. In addition, more than 600 000 persons are hospitalized for pneumonia each year.

Low-Risk Patients with Pneumonia Were Identified by Using Simple Assessment

Fine MJ, Auble TE, Yealy DM, et al. A prediction rule to identify low-risk patients with community-acquired pneumonia. N Engl J Med. 1997; 336:243-50.

One of the most important questions facing physicians is, When should a patient with pneumonia be admitted to the hospital? Admission rates vary markedly from one geographic region to the next. Because physicians are often uncertain of the severity of illness, a subjective impression is commonly used to decide whether to hospitalize the patient or send him or her home with a prescription for oral antibiotics. Physicians tend to overestimate the risk for death from acute community-acquired pneumonia, so hospitalization rates are probably higher than necessary.

Fine and colleagues sought to develop a prediction rule to identify patients with pneumonia who are at low risk for death and therefore do not need hospitalization. In a retrospective review of 14 199 adult inpatients, the researchers developed a prediction rule that placed patients into one of five classes of risk for death within 30 days after hospitalization. The prediction rule assigned points on the basis of age and the presence of coexisting disease, abnormal physical findings, and abnormal laboratory findings at presentation. The authors then tested the rule on 38 039 inpatients and 2287 inpatients and outpatients in the Pneumonia Patient Outcomes Research Team cohort study.

The authors initially identified 21 independent risk factors based on demographic factors, coexisting diseases, abnormal vital signs, and simple laboratory and chest radiography findings. During the validation stage, they developed a two-step rule. The first step used data from the initial patient assessment to identify patients who needed no further workup and could be safely treated as outpatients. Demographic, clinical, laboratory, and radiographic findings were then given specific scores in order to stratify patients into various risks for complications and death. The two-step rule and results of the stratification are shown in Table 1 and Table 2.

This study shows that readily available data can be used to objectively stratify patients into categories of risk in order to decide who should be hospitalized. Patients in the lowest-risk class were defined strictly on the basis of initial history and physical examination findings alone; thus, even routine laboratory studies were unnecessary. Patients in the three lower-risk groups had mortality rates less than 1% and could probably be treated as outpatients. For the two higher-risk groups, hospitalization with parenteral antibiotics is a wise course.

This large study confirms data from two earlier studies that identified the following as important prognostic factors: old age, vital sign changes, previous heart disease, confusion, and an increase in blood urea nitrogen levels [4, 5]. The addition of oxygen saturation, a measure now available in most medical office practices, as an important factor may also help identify patients who may initially need hospitalization or who are deteriorating during outpatient therapy and will require hospitalization.

This two-step prediction rule may be awkward to use in most practices because it places a score on each risk factor, a task that requires some calculation. However, it does guide us in the acute care setting, and such rules may become commonplace as clinical sites become more computerized.

Guideline Compliance Was Costlier

Gleason PP, Kapoor WN, Stone RA, et al. Medical outcomes and antimicrobial costs with the use of the American Thoracic Society guidelines for outpatients with community-acquired pneumonia. JAMA. 1997; 278:32-9.

In 1993, the American Thoracic Society (ATS) developed guidelines for the treatment of community-acquired pneumonia [6]. These guidelines were derived from expert opinion and evidence that existed in 1993. The expert panel attempted to keep recommendations simple. For example, it recommended that patients who are younger than 60 years of age and have no comorbid conditions receive erythromycin or, if they cannot tolerate erythromycin, another macrolide. For patients unable to take any macrolide, tetracycline is recommended. For patients 60 years of age or older or those with a comorbid condition, the panel recommended a second-generation cephalosporin, trimethoprim-sulfamethoxazole, or a ß-lactam plus ß-lactamase inhibitor. For patients with a Legionella species as a possible pathogen, a macrolide is also recommended. For patients who required hospitalization, the panel recommended second- or third-generation cephalosporin or a ß-lactam plus ß-lactamase inhibitor, with or without a macrolide (depending on whether a Legionella species was thought to be present). If a Legionella species is identified, rifampin should be added. For the sickest patients, the panel recommended a macrolide, a cephalosporin, and an aminoglycoside.

This study, which was also reviewed in the 1998 Update in Geriatrics [2], compared medical outcomes and antimicrobial costs for patients whose antimicrobial therapy was consistent or inconsistent with the ATS guidelines. In a multicenter cohort study, Gleason and colleagues identified 864 immunocompetent adult outpatients with community-acquired pneumonia. About 63% of patients were 60 years of age or younger and had no comorbid conditions, and 37% were older than 60 years of age or had at least one comorbid condition. Antimicrobial therapy was classified as being consistent or inconsistent with the guidelines. Mortality, subsequent hospitalization, medical complications, symptom resolution, return to work and usual activities, health-related quality of life, and antimicrobial costs were compared among patients whose treatment was consistent with ATS guidelines and patients whose treatment was inconsistent with the guidelines.

The younger, healthier patients who were prescribed therapy consistent with ATS guidelines had threefold lower antimicrobial costs compared with patients treated with antimicrobial agents not recommended by ATS guidelines ($5.43 compared with $18.51). Medical outcomes did not differ between these groups. In the older, less healthy patients who were prescribed therapy consistent with ATS guidelines, antimicrobial costs were 10-fold higher than those in patients whose antimicrobial treatment was not consistent with ATS guidelines ($73.50 compared with $7.50); trends toward higher mortality and subsequent hospitalization were seen in the former group.

Compliance with the ATS guidelines was high in the young patients. In older patients, compliance rates decreased to as low as 20%, primarily because erythromycin was prescribed more frequently. No significant differences in medical outcomes were seen between patients whose treatment accorded with ATS guidelines and those whose treatment did not.

In summary, most young patients were prescribed therapy consistent with ATS guidelines. They had good outcome and a much lower cost of therapy. Most older patients were not treated according to ATS guidelines. Those who were treated by the guidelines incurred much higher costs and did not have a clinically significant improvement in outcomes. Both young and older patients treated with a macrolide alone had favorable outcomes.

This study does have some limitations. First, selection bias may have occurred because the study was observational. Second, a relatively small number of patients had a comorbid condition, and most patients were young. As a result, the data may lack statistical power. Third, because the ATS guidelines were actually published after the study began, data on compliance with the guidelines were not extensive.

Still, the clinical message is important: Guidelines must be tested. Partly as a result of this study, along with those of many studies from this group and others, the ATS plans to reevaluate its guidelines. I suspect that erythromycin may be recommended as initial therapy in all patients with community-acquired pneumonia.

Organizing Pneumonia

Bronchiolitis obliterans organizing pneumonia is a nonspecific pathologic abnormality characterized by granulation tissue within the lumen of the distal small airways, alveolar ducts, and air spaces. The clinical syndrome may be idiopathic, secondary to many causes (such as infections, drugs, and inhalation injuries), or associated with autoimmune diseases.

Idiopathic bronchiolitis obliterans organizing pneumonia (better known as cryptogenic organizing pneumonia) is now a well-defined clinicopathologic entity that has characteristic clinical and imaging features and can be identified by a characteristic pattern on lung biopsy specimens. The clinical presentation of cryptogenic organizing pneumonia often mimics that of community-acquired pneumonia. The onset is usually described as a flu-like syndrome with fever, malaise, fatigue, and cough. A persistent and usually nonproductive cough is the most common presenting symptom. Patients often experience dyspnea with exertion, and weight loss is common. Clinical recovery, usually with complete clinical and physiologic improvement and normalization of findings on chest radiography, occurs in two thirds of the patients treated with corticosteroids. Relapse may occur when the corticosteroids are withdrawn.

Organizing Pneumonia Was Categorized

Lohr RH, Boland BJ, Douglas WW, et al. Organizing pneumonia. Features and prognosis of cryptogenic, secondary, and focal variants. Arch Intern Med. 1997; 157:1323-9.

Lohr and colleagues described the clinical course and outcomes in patients with variants of organizing pneumonia. They retrospectively studied 74 patients with organizing pneumonia seen at the Mayo Clinic during a 10-year period. Initial features were obtained from medical records, chest radiographs and pathology specimens were reviewed, and clinical outcomes were obtained from medical records and patient questionnaires.

The investigators identified three clinical variants: cryptogenic organizing pneumonia (50% of patients); secondary organizing pneumonia associated with hematologic cancer, collagen vascular disease, or drugs (36% of patients); and focal organizing pneumonia (14% of patients). No difference was found between cryptogenic and secondary variants in type or severity of symptoms, signs, laboratory and pulmonary function test results, or radiologic or pathologic findings. Symptoms resolved more frequently in patients with cryptogenic organizing pneumonia than in those with secondary disease. Relapse was infrequent in both groups. Five-year survival rates were higher in patients with cryptogenic organizing pneumonia (73%) than in those with secondary disease (44%), and respiratory-related deaths were more frequent in patients with secondary organizing pneumonia. Focal organizing pneumonia was an asymptomatic localized rounded opacity in 10 patients (14%), most often detected on chest radiography and diagnosed on lung biopsy done for suspicion of lung cancer. No patients with focal organizing pneumonia required treatment, had relapse, or died of respiratory-related causes.

Cryptogenic organizing pneumonia most often was a symptomatic bilateral lung process that had an overall favorable prognosis if prolonged corticosteroid therapy was given. Patients with secondary organizing pneumonia had a high mortality rate when the disease was associated with predisposing conditions or drugs, but patients with asymptomatic focal organizing pneumonia had an excellent prognosis.

The clinical presentation of cryptogenic organizing pneumonia often mimics that of community-acquired pneumonia, and the disease can result in severe respiratory failure. A high index of suspicion is required because no specific features suggest this entity. The management (corticosteroid therapy) is at odds with management of diseases that mimic organizing pneumonia; thus, an aggressive diagnostic approach must be taken in patients whose pneumonia does not respond as expected. Diagnosis is made by using transbronchial or thoracoscopic lung biopsy.

Focal organizing pneumonia should be considered in patients with focal or migratory opacities on chest radiography. This entity may mimic hypersensitivity reactions, parasitic infection, allergic bronchopulmonary aspergillosis, or Churg-Strauss vasculitis.

Interstitial Lung Diseases

Interstitial lung diseases comprise more than 180 disorders that are characterized by inflammation and fibrosis of the alveolar spaces and walls. Some interstitial lung diseases are caused by drugs or occupational or environmental exposures, but in most cases, no specific cause can be found. Two diseases that cause idiopathic interstitial lung diseases are sarcoidosis and idiopathic pulmonary fibrosis.

Long-Term Steroid Therapy Helped Marginally in Sarcoidosis

Gibson GJ, Prescott RJ, Muers MF, et al. British Thoracic Society Sarcoidosis study: effects of long term corticosteroid treatment. Thorax. 1996; 51:238-47.

Sarcoidosis is a murky, ill-defined entity. It can involve any organ but is most worrisome when it affects the eyes, nervous system, heart, bone, and, of course, lungs. For patients with lung involvement, the criteria for diagnosing sarcoidosis are the presence of a compatible clinical illness, histopathologic confirmation of granulomatous disease, and exclusion of other illnesses that can cause a granulomatous lung disease. Biopsy specimens reveal noncaseating granulomas with giant cells. About half of patients present without symptoms; in these patients, the diagnosis is raised after abnormalities are identified on chest radiography. Up to 15% of cases begin acutely with the Loffler syndrome: erythema nodosum and bilateral hilar lymphadenopathy. Almost all of these patients have a favorable outcome. About 90% of patients with sarcoidosis will have disease in the chest at some point. The chest radiograph is a guide to the presence of disease, but it is not adequate for monitoring the stage or activity of the disease.

The incidence of sarcoidosis varies worldwide. In the United States, it is as high as 5 cases per 100 000 white persons and 40 cases per 100 000 African-American persons. Sarcoidosis is more common in Denmark and Sweden and less common in Japan, Brazil, and South Africa. Autopsy studies show that the incidence of sarcoidosis is actually several hundred-fold higher than the incidence of clinically diagnosed cases.

The marked variability in presentation and clinical course of sarcoidosis makes the development of treatment guidelines difficult. Current treatment is aimed at suppressing the inflammatory response, reducing the burden of granulomas, and preventing the development of fibrosis. Corticosteroids have been the most commonly used agents for the treatment of pulmonary sarcoidosis.

In a prospective, nonrandomized, multicenter study, Gibson and colleagues assessed the effects of corticosteroids on the natural history of sarcoidosis and sought to determine whether this therapy has long-term benefit. Patients with shadowing on chest radiographs due to sarcoidosis were recruited. After 6 months, patients whose condition had neither worsened nor improved were allocated, on the basis of their clinical findings, to receive either long-term ( 18 months) continuous corticosteroid therapy or selective use of corticosteroids for clinical exacerbations. Symptoms, radiographic appearance, and respiratory function were assessed periodically during the 5-year study.

Initially, 149 patients were recruited: Thirty-three required prednisolone for symptoms within 6 months, and 58 had radiologic improvement. The remaining 58 patients were then allocated to one of the two treatment groups. The group receiving long-term continuous corticosteroid therapy had 9% improvement in vital capacity compared with patients receiving corticosteroids selectively. However, continuous steroid treatment often had to be reinstituted or the dose had to be increased after unsuccessful attempts at weaning. Side effects of treatment were common but were usually mild, necessitating withdrawal in only two patients.

This study had several limitations. There was no placebo group or controls over the protocol, and quality of the tests was not measured in the various study centers. The patients were not randomly assigned, and the allocation procedure was not adequately described. The "most seriously" affected patients were treated separately, so it is unclear whether this sample represents all patients with sarcoidosis seen by the referring physicians. Finally, the study groups had important imbalances (sex distribution, presence of "fibrotic" opacities, differences in lung function, symptoms, number of untreated patients).

Nonetheless, at least two conclusions for clinicians can be drawn from this study. First, an observation period of approximately 6 months is important before corticosteroid treatment begins. About 40% of the patients who did not require steroids at study entry showed radiographic improvement within 6 months after study entry. Improvement was associated with younger age, hilar adenopathy, a recent history of erythema nodosum, and current cigarette smoking. Second, patients who received long-term steroid therapy had slightly better pulmonary function, but this benefit came at the cost of much longer courses of steroid therapy than had been expected. By the end of the study, steroid use in these patients was sixfold higher than use in patients taking steroids selectively. The benefit of long-term, continuous steroid treatment may not equal the risk.

Steroid-Induced Remission of Sarcoidosis Often Failed

Gottlieb JE, Israel HL, Steiner RM, et al. Outcome in sarcoidosis. The relationship of relapse to corticosteroid therapy. Chest. 1997; 111:623-31.

Another study on corticosteroid therapy for sarcoidosis also suggests that the benefits do not outweigh the risks. Gottlieb and colleagues assessed characteristics of patients who received corticosteroid therapy and compared them with characteristics of patients who had spontaneous remission of disease.

A total of 337 patients attending a sarcoidosis clinic were enrolled in this nonrandomized study. Approximately 35% of these patients had spontaneous remission of disease during the 4-year study. Another 31% received corticosteroid therapy and had induced remission, after which corticosteroid therapy was discontinued. The third group of patients (34%) was considered "recalcitrant" because corticosteroid therapy could not be stopped within 1 month as a result of severity of symptoms or lack of compliance.

Disease relapsed in 74% of patients with induced remission and 8% of patients with spontaneous remission. Relapse occurred with similar frequency in white persons and African-American persons (20% and 28%, respectively), despite a lower treatment rate in white patients (43% and 76%, respectively). However, sustained remission was maintained twice as frequently in white patients as in African-American patients (58% compared with 29%). Initial presentation with asymptomatic chest radiograph abnormalities, erythema nodosum, or peripheral adenopathy portended a favorable prognosis; 60% of such patients had sustained remission that lasted a mean of 130 months from the time of diagnosis. In contrast, patients who presented with musculoskeletal symptoms were nine times more likely to have relapse than were patients who presented without these symptoms, and those with symptoms caused by liver involvement were three times more likely to have relapse than those whose symptoms were not caused by liver involvement. About half of all relapses occurred between 2 and 6 months after discontinuation of steroid therapy, but late relapse was not unusual.

Relapse occurred frequently in patients with sarcoidosis who had been treated with corticosteroids. The obvious question is, Does corticosteroid therapy predispose patients with sarcoidosis to clinical relapse? This study does not answer this question, but an analysis of the data from this study and that by Gibson and colleagues indicates that caution is needed in the use of corticosteroids for sarcoidosis. The depression of inflammatory response by corticosteroids may make patients worse because inflammation may play a role in resolution of symptoms. Certainly, much remains to be understood about sarcoidosis, and new therapies are needed.

Cyclosporin A Was Not Superior to Steroids

Wyser CP, van Schalkwyk EM, Alheit B, et al. Treatment of progressive pulmonary sarcoidosis with cyclosporin A. A randomized controlled trial. Am J Respir Crit Care Med. 1997; 156:1371-6.

Theoretically, cyclosporin A, which exerts inhibitory effects on T cells, would seem to be an effective treatment for patients with sarcoidosis. In this open-label clinical trial, 37 patients with newly diagnosed sarcoidosis were assigned to receive prednisone, 20 mg/d with a planned taper, or prednisone, 20 mg/d with the same taper, plus cyclosporin A (5 to 7 mg/kg of body weight per day) for up to 18 months. Clinical evaluation and pulmonary studies were done at 3, 9, and 18 months.

In an intention-to-treat analysis, the two groups did not differ significantly in response to treatment. Patients treated with both drugs had many adverse effects, including elevated serum creatinine levels and twice the number of infectious complications seen in the other group. At present, methotrexate is the only drug used for sarcoidosis other than corticosteroids, but the data on the effectiveness of methotrexate remain scant.

Clinical History of Pulmonary Fibrosis Was Described

Johnston ID, Prescott RJ, Chalmers JC, et al. British Thoracic Society study of cryptogenic fibrosing alveolitis: current presentation and initial management. Fibrosing Alveolitis Subcommittee of the Research Committee of the British Thoracic Society. Thorax. 1997; 52:38-44.

Few epidemiologic data exist on the occurrence of pulmonary fibrosis. Recent findings suggest a much higher prevalence (up to 29 per 100 000 persons) than was previously assumed. It appears that worldwide mortality due to pulmonary fibrosis is increasing, particularly in elderly persons.

Idiopathic pulmonary fibrosis (also called cryptogenic fibrosing alveolitis) is one of the more common interstitial lung diseases of unknown cause. It is a well-defined clinical entity that occurs mostly in persons 40 to 70 years of age. Patients with confirmed idiopathic pulmonary fibrosis have no history of occupational or environmental exposure to an agent known to cause parenchymal lung disease and are free of other systemic illnesses, especially connective tissue diseases. The typical patient presents with breathlessness and bibasilar end-inspiratory dry rales on physical examination. The chest radiograph often shows diffuse bibasilar reticular opacities. Lung function tests usually show a restrictive process with a reduced diffusing capacity and abnormal gas exchange. For unknown reasons, the disease seems to begin at the lung bases and progress upward and inward.

Usual interstitial pneumonia is the histologic hallmark of idiopathic pulmonary fibrosis. It has a heterogenous appearance in the peripheral subpleural parenchyma that is characterized by alternating areas of normal lung, interstitial inflammation, fibrosis, and honeycomb change. The interstitial inflammation is usually patchy and consists of an alveolar septal infiltrate of lymphocytes, plasma cells, and histiocytes associated with hyperplasia of type 2 pneumocytes. The fibrotic zones are primarily composed of dense collagen, although scattered foci of proliferating fibroblasts (so-called fibroblastic foci) are a consistent finding. Response to treatment is uncommon, and average survival is 4 to 6 years after diagnosis.

In this questionnaire study of pulmonologists in the United Kingdom, investigators from the British Thoracic Society sought to determine the natural history and management of idiopathic pulmonary fibrosis. Participating physicians (n = 150) identified 588 patients (63% of whom were male) with the disease. This represents the largest cohort of patients studied to date. The mean age was 67 years, and the median duration of symptoms at presentation to the specialist was 9 months. Clubbing was more common in men than in women (54% compared with 40%), and 36% of patients had severe dyspnea at presentation. Forty-seven percent of patients had some history of dust exposure, often including asbestos, and these patients more frequently tended to smoke cigarettes, a risk factor that has been reported elsewhere [7]. Transbronchial biopsies had been done in 28% of patients, and open-lung biopsies had been done in 12%. Sixty percent of patients had no histologic studies. Those who underwent diagnostic procedures tended to be younger, had better lung function, and had been exposed to asbestos. At presentation, only 48% of patients received specific treatment; most treatment was initiated on the basis of the severity of clinical symptoms. By 2 years, 45% of patients had died.

Few tests were done to help identify these patients as having idiopathic pulmonary fibrosis. In the United States, lung biopsy is necessary to diagnose this condition. One should wonder whether many of these patients did, in fact, have idiopathic pulmonary fibrosis.

One reason the British pulmonologists were not more aggressive was their estimation that most patients had minor symptoms and were stable. Yet, the overall mortality was high. Their estimation raises an important point for the assessment of any patient suspected of having pulmonary disease. Examining a patient at rest can provide false assurance. It is important to at least assess patients after they walk down a hallway. Formal exercise tests can provide further important information on functional ability.

The strength of this study is the large number of patients evaluated. In addition, the study setting was referral practices (rather than specialist centers), and the data were obtained prospectively. The study's weakness is that clinical diagnoses were made despite the absence of formal validation of the diagnosis in many patients. Substantial numbers of these patients had a history of dust exposure. In the United States, this history would preclude the diagnosis of idiopathic pulmonary fibrosis.

We need to identify pulmonary fibrosis earlier and treat it more aggressively. Current treatment is problematic. Anti-inflammatory drugs, including corticosteroids and cyclophosphamide, are ineffective for most patients. However, about 20% of patients do have some response; these patients tend to have earlier-stage disease, are young, and have more inflammation seen on biopsy specimens. No excellent antifibrotic drug exists. A study of interferon-ß 1a is now under way, and data from Europe suggest that interferon- may have a role in treating these patients.

Dr. Roberts (Series Editor): Madrona Medical Group, 3199 Steller Court, Bellingham, WA 98226.