Differential Diagnosis of Pulmonary Infiltrates in ICU Patients
The complexity of patients in the intensive care unit (ICU), together with the recent advances in radiographic images have led to new perspectives in the use of chest x-ray in the ICU. The American College of Radiology consensus committee recommends and maintains that chest x-ray in patients with cardiopulmonary disease or those receiving mechanical ventilation should be performed daily (5). Two approaches address the utility of daily radiologic studies of the chest in the ICU: one which proposes routine daily studies and the other in which radiologic studies are only performed when there is a change in clinical status or if the support equipment is implemented (1, 4, 6, 13).
Pulmonary infiltrates frequently develop in ICU patients (Table 1). Hospital-acquired pneumonia is one cause and occurs in 10 % to 30 % of the patients in the ICU (3, 17). Other possibilities include: atelectasia, cardiac failure, adult respiratory distress syndrome (ARDS), pulmonary fibrosis, embolism and pulmonary hemorrhage. Changes in the chest radiography depend on the level of positive pressure at the end of expiration (PEEP) and the inflation of the lung at the time of this action. Indeed the extent of the infiltrates may vary according to the level of PEEP often giving the false impression of resolving infiltrates.
Pneumonia may be suspected in patients with fever, leukocytosis, purulent secretions and the appearance of new or progressive pulmonary infiltrates on chest x-ray. Mechanical ventilation-associated pneumonia (VAP) is the most frequent nosocomial pneumonia reported in patients in the ICU, with an incidence varying from 10 % to 30 % and an estimated rate of 1 % to 3 % per day of mechanical ventilation (3, 17). Most episodes of VAP are caused by aspiration of the oropharyngeal content. Orotracheal intubation or tracheostomy facilitates the passage of bacteria from the oropharynx to the respiratory tract. Oropharnygeal colonization by Gram-negative bacteria generally occurs within the first 24 hours of ICU admission, with 40 % of the patients already being colonized the fifth day of hospitalization (17). Moreover, many patients receive enteral nutrition through a nasogastric tube, thereby facilitating aspiration from oropharyngeal or gastic flora. The clinical diagnosis of pneumonia in ventilated patients is difficult. The clinical manifestations are nonspecific. Chastre et al (2) found that only 38 % of the patients with fever, purulent tracheal secretions and persistent pulmonary opacity had pneumonia. Presence of new or progressive infiltrates on chest x-ray, especially in the presence of fever may be an important clue. No specific radiologic characteristic is diagnostic for pneumonia. Studies evaluating the precision of chest x-ray in the diagnosis of pneumonia have reported a specificity of 27 % to 35 %. Portable chest x-rays have even lower sensitivity and specificity. Wunderink (18) performed a study evaluating radiologic and autopsy findings and identified some radiologic signs, although none had a diagnostic performance of greater than 68%. Air bronchogram, especially if only one, was a very specific sign. Bilateral opacifications and the silhouette sign are often found but are nonspecific. Surprisingly, worsening of a previous radiologic opacity only corresponded to pneumonia in one third of the cases.
Pulmonary infiltrates that disappear within a few days may be due to pulmonary edema, atelectasia or aspiration, while radiologic changes of resolution in pneumonia evolve more slowly.
Atelectasis is relatively frequent in the ICU and is associated with general anesthesia and prolonged surgery. It is more common during the postoperative period, especially in patients undergoing thoracic or upper abdominal surgery. The incidence is increased if the patient is a smoker, has persistent pulmonary disease or is obese. Most of the infiltrates which appear within the first 48 hours after surgery are due to atelectasis especially if aspiration has been ruled out (3).
Numerous factors may be involved in the appearance of atelectasis: diminution in consciousness, pain, tracheal intubation, mechanical ventilation, non productive cough, and alteration of mucociliary secretion clearance. Edema of the mucosa or the presence of secretions obstructing airflow and the reduction in pulmonary surfactant promotes alveolar collapse.
Atelectasis is most often seen in the lower left pulmonary lobe (66 %) compared with the lower right lobe (22%) (5, 7, 17). Radiologic manifestations may include normal chest x-ray, slight loss of volume without a visible infiltrate (microatelectasias), parallel or oblique lineal opacities to the diaphragm, single or multiple rounded or irregular opacities, lobar consolidation or even total pulmonary collapse.
In a well penetrated film the presence of air bronchogram in the retrocardiac area, loss of pulmonary outline or the lateral edge of the descending aorta indicates the presence of a pulmonary infiltrate. Loss of volume of the lower left lobe may be observed on repletion of the vessels of air bronchogram in the retrocardiac area and by the elevation of the hemidiaphragm. Accentuated volume loss is indicated by displacement of the mediastinum towards the left, descent of the left hilum or the visualization of a triangular opacity behind the cardiac silhouette. The upper lobe may appear to be hyperlucent.
The absence of an air bronchogram suggests endobronchial lesions (often a mucous tamponade), while its presence suggests other causes (3, 17). Atelectasis without obvious signs of volume loss are difficult to differentiate from pneumonia. However, atelectasis may fluctuate or resolve within a short time (generally hours). The correct diagnosis is based on the observation of rapid changes in size, shape or localization of the infiltrates on serial chest x-rays, especially after initiating respiratory physiotherapy. Atelectasis which persist in patients after the fourth day post-cardiovascular surgery is often complicated with pneumonia (17).
Portable chest x-ray has poor sensitivity in detecting opacities at the pulmonary bases. Moreover, when atelectasis is identified, the extent of the involvement is often underestimated.
Aspiration is common in ICU patients since the patients frequently present with depression of the central nervous system, neuromuscular diseases and gastrointestinal tract disease (carcinoma, esophageal narrowing). In addition, endotracheal and nasogastric tubes are also present.
The clinical manifestations include chemical pneumonitis, pulmonary infections, ARDS, and acute airway obstruction. The patient may occasionally be asymptomatic. Fever, cyanosis, hypotension or tachypnea can also occur. The patients may improve within 1 or 2 days or the process may evolve into pneumonia or even ARDS.
The radiologic findings indicating aspiration vary and depends on the quantity aspirated. The appearance of new bilateral infiltrates may suggest aspiration, especially with basal localization in erect patients or in the upper lobes of supine patients (7). The presence of any localized patchy infiltrate may be a manifestation of aspiration.
In patients with underlying disease such as ARDS, cardiac insufficiency or massive atelectasis, aspiration will be difficult to diagnose. The evolution of the infiltrates is a great help in establishing the diagnosis. A pulmonary infiltrate which clears within 2 to 3 days is a common finding in aspiration (5). On the other hand, the progression of patchy infiltrates accompanied by deterioration in gas exchange suggests the appearance of ARDS.
The chest x-ray can detect pulmonary edema but does not specify the cause. The most specific sign is the presence of opacification of patches in the air space. It is observed in 58% of the patients with edema due to an increase in permeability and in only 13% of those with hydrostatic edema.
Cardiogenic Pulmonary Edema: An increase in the pulmonary vasculature with a cephalic disposition may be observed in patients with cardiogenic pulmonary edema. However, these findings are characteristic of chronic cardiac insufficiency but are not a reliable manifestation of cardiac failure in ICU patients since the x-ray is performed in a supine position. Posteriorly, dilatation of the vessels, interstitial edema (vascular and peribronchial thickening), Kerley lines, may be seen in the periphery. These radiologic changes are indistinguishable from the interstitial edema due to infection by cytomegalovirus or Pneumocystis carinii (9).
Progression of cardiogenic edema shows diffuse alveolar edema with a central distribution which is also difficult to distinguish from diffuse infection or pulmonary hemorrhage. The change in the pattern of pulmonary edema after several days is of great help at the time of evaluating whether the infiltrate is due to infection or is of hydrostatic origin. The presence of cardiomegaly, septal thickening and pleural effusion is frequently found in congestive heart failure. In these patients the opacification is of central localization. The distribution of the edema may be altered by the position of the patient due to the effect of gravity and by the presence of underlying pulmonary disease. These characteristics are usually absent in noncardiogenic pulmonary edema.
NonCardiogenic Pulmonary Edema: Noncardiogenic pulmonary edema due to an increase in permeability may be secondary to smoke inhalation or toxic metals, drowning, fat embolism, heroin intoxication, uremia, aspiration, neurologic alterations and allergic reactions (5).
The most representative entity of noncardiogenic pulmonary edema is ARDS. This syndrome occurs because of an alteration in pulmonary vascular permeability. It is characterized by a diminution in pulmonary compliance, pulmonary capillary wedge pressure less than 18 mm Hg and poor response to oxygen administration, all being triggered by known causes of respiratory distress (sepsis, shock, pneumonia, trauma, aspiration, pancreatitis).
Pulmonary infiltrates are not seen during the first 12 hours and generally appear at 24 hours. In the initial stages noncardiogenic pulmonary edema is indistinguishable from pneumonitis or cardiac failure (generally cardiomegaly, vascular redistribution or pleural effusion are not present) with progressive interstitial edema and presents a diffuse distribution in the two lungs. This corresponds with the exudative phase of ARDS when large quantities of protein-rich fluid are deposited in the alveolar space, secondary to the reduction in surfactant. In more advanced stages the edema is replaced by pulmonary fibrosis due to the action of the type II pneumocytes and the formation of hyaline membranes conferring the appearance of an interstitial infiltrate. This pattern may persist during weeks or months after the patient improves. The fibrosis established in this stage is irreversible (15).
In the initial stages, CT shows bilateral areas of diffuse opacification, in non uniform patches with aerial bronchogram. In the exudative phase the distribution of pulmonary infiltrates is independent of the severity, while in the last phase confluent severity-dependent opacities may be observed accompanied by atelectasias of cephalocaudal and dorsoventral distribution.
ARDS may be distinguished from cardiac failure by the absence of redistribution and vascular thickening. Bronchial dilatation is a frequent finding in patients with ARDS. Bronchiectasias is generally associated with groved glass opacifications. In the acute phase of distress this groved glass represents interstitial filling and the alveoli by inflammatory fluid and edema. Bronchial dilatation in this setting is generally reversible.
The presence of cysts or subpleural bullae may be found in most patients with ARDS, appearing during the first week and being associated with prolonged mechanical ventilation and barotraumas.
Pleural effusion is a frequent entity in ICU patients, especially in postoperative patients. The effusion is usually an exudate. They may resolve within one or two weeks. Pleural effusion may be confused with subphrenic abscesses but the latter generally appear two weeks after surgery with abdominal symptoms and/or fever (5).
Pleural effusion which rapidly increases in size may be hemorrhagic or infectious. The acute appearance of a loculated pleural effusion in an unusual localization suggests bleeding of the thoracic wall. Patients with pneumonia frequently present parapneumonic pleural effusion and empyema. Although the characteristics of the pleural fluid cannot be determined radiologically, the possibility of empyema should be ruled out in patients with pleural effusion and unexplained fever. If the effusion is loculated, the most probable diagnosis suspicion is empyema, although this must be confirmed by study of the fluid.
Postpericardiotomy syndrome should be considered in patients with pleural effusion of more than 3 days during the postoperative period following heart surgery (5). A frequent cause of pleural effusion is the placement of catheters with subsequent perforation of the pleura.
Ultrasonography is the method most frequently used in diagnostic thoracocentesis and for the placement of drainage catheters, having a low incidence of complications. It can detect lobar or loculated effusions. However, ultrasound does not allow the evaluation of the pulmonary parenchyma or the airways, so the CT provides information with respect to the size of the collection and the presence of loculation. The placement of drains guided by CT is successful in 72 % of the cases of loculated collections, with a low incidence of complications. Moreover, it reduces the complications of pulmonary abscesses or bronchopleural fistulas (8, 16).
Pulmonary embolism may be observed following surgery or in patients who are bed-ridden for prolonged periods. Predisposing conditions include history of embolic disease, malignancy, venous disease, COPD, and the use of oral contraceptives (3).
The chest x-ray may be normal or show nonspecific changes such as an elevation of the hemidiaphragm, atelectasia or diffuse infiltrates of peripheral localization. These infiltrates may vary and disappear rapidly due to edema or hemorrhage and when they persist this is due to infarction. Some frequent radiologic findings in these patients are those of wedge-shaped opacity or an increase in the size of the pulmonary artery. This may be accompanied by small to moderate, unilateral, pleural effusions in 50% of the cases, with remission within a few days (5, 10).
The diagnostic yield of spiral CT for pulmonary embolism depends on the size and localization of the embolism. The sensitivity and specificity for involvement of the principal or segmentary arteries are from 85 % to 90 % and > 90 %, respectively. The sensitivity is lower (50 % - 75 %) for the involvement of subsegmentary arteries. However, the use of pulmonary angiography is not better than chest CT at this level (16). In a study performed comparing the risk of pulmonary embolism following angiography by negative CT, the negative predictive value of angiography by spiral CT was of 98 % for patients with pulmonary disease and 100 % for individuals without respiratory disease (14).
Pulmonary emboli may present in diverse scenarios: a patient presenting a normal chest x-ray or with alveolar infiltrates which improves after several days, a patient with a chest x-ray suggestive of ARDS which may evolve to pulmonary fibrosis, or massive fat embolism without radiologic findings.
Pulmonary hemorrhage may be due to a focal pulmonary involvement, diffuse pulmonary disease or an alteration in coagulation. This entity is more frequent in neutropenic patients, particularly with leukemia and in certain immunological diseases such as systemic lupus erythematosus, Wegener granulomatosis and idiopathic pulmonary hemosiderosis. It is also frequent in certain renal diseases such as Goodpasture or rapidly progressive glomerulonephritis. Moreover, it may be observed in patients who have undergone bone marrow transplantation (7).
When the hemorrhage is due to a focal pulmonary process, chest x-ray and bronchoscopy are adequate for diagnosis. In patients with diffuse pulmonary disease, angiography may be necessary.
Drug-Induced Pulmonary Lesions
Many drugs may produce adverse effects in the lung. The diagnostic criteria include a history of drug exposure, radiologic alterations, histologic evidence of pulmonary lesion and the exclusion of other possibilities of pulmonary involvement (infections, radiation, thromboembolism, oxygen toxicity, worsening of pre-existing pulmonary disease) (12). In clinical practice, lung biopsy is seldom performed and the diagnosis is based on the clinical and radiologic manifestations. High resolution computerized tomography (HRCT) is superior to chest x-ray, since it not only helps to achieve the diagnosis but also allows follow up to be performed. The pharmacological groups associated with pulmonary toxicity include chemotherapy agents, antibiotics, antiinflammatories, and cardiovascular drugs. The most frequent histologic patterns include: diffuse alveolar damage, acute or chronic alveolar hemorrhage, pneumonitis due to hypersensitivity, non specific interstitial pneumonia, organizing pneumonia and eosinophilic pneumonia.
Diffuse Alveolar Damage
This histologic finding is observed in ARDS. It is an nonspecific manifestation of acute pulmonary lesion which may occur in severe infection, trauma, aspiration, collagen diseases, illegal drug use and acute interstitial pneumonia. Diffuse alveolar damage is the most frequent histologic finding associated with drug-induced pulmonary lesion, being associated with cyclophosphamide, bleomycin, melphalan, amiodarone, methotrexate, aspirin, narcotics, bisulfan, paclixatel and docexatel (11, 12). The acute stage of diffuse alveolar damage is characterized by edema and the formation of hyaline membrane while in the chronic phase alveolar repair is carried out by type II pneumocytes and fibroblasts. Fibrosis may progress and may produce honeycomb-like lesions. Chest x-ray shows homogeneous or heterogenic opacities of medium and inferior distribution.
The HRCT demonstrates consolidation in the air pace and ground glass opacities involving lung-dependent regions. In the exudative phase, the opacities are bilateral and in patches and afterwards distortion of the architecture with bronchiectasias may be observed.
This lesion is seldom seen. The drugs most commonly associated with alveolar hemorrhage are anticoagulants, amphotericin B, methotrexate, amiodarone, carbamazepine, propylthiouracil, cyclophosphamide, phenytoin, nitrofurantoin, and penicillamine. Bilaterial ground glass images are observed on HRCT (11, 12).
Pneumonitis by Hypersensitivity
This is an infrequent form of pulmonary lesion. It is most often associated with the use of methotrexate, cyclophosphamide, fluoxetine, amitriptyline, and paclitaxel. The histopathologic and radiologic findings are indistinguishable from pneumonitis by hypersensitivity secondary to organic antigen inhalation. Histologic findings show bronchiolitis, noncaseating granulomas and bronchocentric interstitial pneumonia poorly defined centrolobular nodules may be observed (12).
Nonspecific Interstitial Pneumonitis
This is the most common finding in drug-induced lesions, as well as in collagen diseases and in pneumonia by hypersensitivity. The following drugs may be associated with this lesion: amiodaraone, methotrexate, nitrofurantoin, bleomycin and hydrochlorthiazide. From a tomographic point of view, bilateral patches are found in images of ground glass, reticular opacities and bronchiectasis of basal localization. Histologic studies show zones of interstitial inflammation, hyperplasia of type II pneumocytes and fibrosis (11, 12).
Organizing Pneumonia (BOOP)
Bronchiolitis obliterans organizing pneumonia (BOOP), also known as cryptogenic organizing pneumonia, may be idiopathic or secondary to other conditions such as collagen diseases, aspiration, organ transplantation, radiotherapy, AIDS, hematologic diseases and drugs.
The drugs associated with BOOP include: bleomycin, gold salts, cyclophosphamide and methotrextate. Other drugs frequently associated with BOOP include amiodarone, nitrofurantoin, penicillamine, acebutolol, phenytoin, carbamazepine and sulphasalazine (11, 12). This pneumonia is generally reversible following drug withdrawal or steroid therapy. In amiodarone toxicity the areas of consolidation may have high attenuation at CT often accompanied by high attenuation in the liver or spleen.
Organizing pneumonia is characterized by proliferation of fibroblasts in the bronchioles, alveolar ducts and adjacent alveoli. On chest x-ray, diffuse homogeneous pulmonary infiltrates may be observed. On HRCT, consolidation may be found. Most of the patients present with bilateral ground glass images.
This type of pneumonia is characterized by the presence of eosinophils and macrophages in the alveoli, with thickening of the septa by infiltration. It has been described with the use of penicillamine, sulphasalazine, nitrofurantoin, non steroid antiinflammatories, amiodarone, bleomycin, phenytoin, antidepressants, beta blockers and hydrochlorothiazide.
The diagnosis of eosinophilic pneumonia is based on the presence of pulmonary infiltrates, more frequently of peripheral distribution and in the upper lobes. Eosinophilia in peripheral blood, pulmonary biopsy or bronchoalveolar lavage is common (11, 12). On HRCT consolidated ground glass images may be seen in the periphery and upper lobes (12).
Table 1. Differential Diagnosis of Pulmonary Infiltrate in Patients in the ICU
Table 2 Drug Induced Lesions Causing Pulmonary Infiltrates
1. Chahine-Malus N, Stewart T, Lapinsky SE, Marras T, Dancey D, Leung R et al. Utility of routine chest radiographs in a medical-surgical intensive care unit: a quality assurance survey. Crit Care 2001; 5(5):271-275. [PubMed]
2. Chastre J, Fagon JY, Soler P, Bornet M, Domart Y, Trouillet JL et al. Diagnosis of nosocomial bacterial pneumonia in intubated patients undergoing ventilation: comparison of the usefulness of bronchoalveolar lavage and the protected specimen brush. Am J Med 1988; 85(4):499-506. [PubMed]
3. Goodman LR. Cardiopulmonary disorders in the critically ill. In: Goodman LR, Putman CE, editors. Intensive care radiology: imaging of the crically ill. W.B. Saunders, 1983: 61-113. [PubMed]
13. Singh N, Falestiny MN, Rogers P, Reed MJ, Pularski J, Norris R et al. Pulmonary infiltrates in the surgical ICU: prospective assessment of predictors of etiology and mortality. Chest 1998; 114(4):1129-1136. [PubMed]
14 Tillie-Leblond I, Mastora I, Radenne F, Paillard S, Tonnel AB, Remy J et al. Risk of pulmonary embolism after a negative spiral CT angiogram in patients with pulmonary disease: 1-year clinical follow-up study. Radiology 2002; 223(2):461-467. [PubMed]