Differentiating pulmonary embolism from an exacerbation of chronic obstructive pulmonary disease (COPD) can be difficult because the conditions share similar clinical signs and symptoms. In a patient with COPD, therefore, how well can PE be diagnosed with commonly used methods—a clinical probability estimate, spiral computed tomographic (CT) angiography, D–dimer analysis, or ventilation–perfusion (V/Q) scintigraphy? Very well, according to a prospective, multicenter study.

Hartmann et al evaluated the effect of the presence of COPD on the accuracy of diagnostic procedures for PE in 627 consecutive patients; 91 (15%) patients had confirmed COPD, and 536 did not. The prevalence of PE in the two groups was 29% and 31%, respectively.

The presence of COPD did not affect the diagnostic accuracy of the clinical probability estimate made by the treating physician before objective testing was performed. Nor did it alter the accuracy of spiral CT angiography, D–dimer analysis, or V/Q scintigraphy. The sensitivity and specificity of these three techniques in patients with and without COPD are shown in Table 1. Neither age nor the presence of congestive heart failure influenced the performance of spiral CT angiography or the D–dimer test, but both factors increased the number of nondiagnostic V/Q scan results. The authors note that nondiagnostic V/Q scan results are more likely to occur in the presence of COPD, thus decreasing the procedure’s cost–effectiveness. Nevertheless, V/Q scintigraphy is still a valuable, noninvasive test for PE when screening patients with COPD.

Even moderate weight loss can reduce the severity and progression of sleep–disordered breathing (SDB), a population–based, prospective cohort study indicates. The alternative—nasal continuous positive airway pressure therapy—is burdensome and probably impractical for patients with mild or asymptomatic SDB, the study authors suggest.

Peppard et al studied 690 randomly selected subjects enrolled in the Wisconsin Sleep Cohort Study to determine the association between weight gain and increased SDB severity, as well as between weight loss and decreased SDB severity.

The primary outcome measures were the influence of a change in weight on the percentage change in the apnea–hypopnea index (AHI, the number of apnea and hypopnea events per hour of sleep) and on the possibility of developing moderate–to–severe SDB (defined as an AHI of 15 or more). The subjects were evaluated at baseline and at four–year follow–up.

After adjustments for sex, age at baseline, body mass index, and smoking behavior, weight change correlated positively with change in the AHI. Compared with no change in weight, a 10% increase was associated with a 32% increase in the AHI, and a 10% decrease was associated with a 26% decrease in the AHI. The authors’ analysis also showed that subjects who had a 10% weight gain had an odds ratio of 6.0 for the development of moderate–to–severe SDB.

Peppard et al caution that their findings may not apply to patients who experience weight changes in excess of 20%.

Peppard PE, Young T, Palta M, et al. Longitudinal study of moderate weight change and sleep-disordered breathing. JAMA. 2000;284:3015– 3021.

Patients being treated with warfarin often have an elevated international normalized ratio (INR) and are at increased risk for hemorrhage. Often, low doses of vitamin K are given in this situation, but the efficacy of such treatment in asymptomatic patients has never been firmly proved. A multicenter, randomized Canadian study now confirms that low doses of vitamin K can safely and effectively lower INR elevations in patients taking warfarin.

Crowther et al studied 89 asymptomatic patients who were being treated with warfarin and had INRs of 4.5 to 10.0. Of these patients, 45 were randomized to 1 mg of oral vitamin K and 44 to placebo. Warfarin administration was stopped in all patients. The main outcome measurement was the INR the day after treatment. Additional follow–up was obtained through telephone interviews or clinic visits at one and three months.

On the day after vitamin K administration, 25 (56%) of the vitamin K–treated patients and nine (20%) of the controls had INRs between 1.8 and 3.2. At three months, two vitamin K patients (4%) and eight controls (17%) reported having had bleeding episodes that required transfusion or hospital admission. Thrombotic episodes occurred in one vitamin K–treated patient (myocardial infarction) and one placebo control (deep venous thrombosis).

Of the 14 persons who died during follow–up, eight had been treated with vitamin K. Eleven of the 14 died of cancer, one of multisystem organ failure, one of aortic stenosis, and one of unknown causes.

Crowther MA, Julian J, McCarty D, et al. Treatment of warfarin–associated coagulopathy with oral vitamin K: a randomised controlled trial. Lancet. 2000;356:1551– 1553.

Ventilator circuit compliance is particularly important in determining the actual volume of air delivered to the lungs of infants and children. Measuring airway pressure and expired tidal volume (VT) at the expiratory valve does not account for the compliance of the ventilator circuit or for uncontrolled variations in the circuit setup. In mechanically ventilated infants, therefore, delivered VT should be measured by a pneumotachometer placed at the airway.

Cannon and associates reached this conclusion after studying 98 conventionally ventilated infants and children. In all cases, VT was measured by the ventilator as well as by a pneumotachometer (coupled to a respiratory mechanics monitor) that was placed between the ventilator circuit and the endotracheal tube. In addition, the authors estimated effective VT by using mathematical formulas designed to correct for compliance in the ventilator circuit.

In 70 infants (mean age, 2.8 months), the mean VT measured by the pneumotachometer was 39.4 mL, significantly lower than the ventilator’s VT reading of 70.4 mL or the calculated effective VT of 59.2 mL. Among 28 children (mean age, 7.3 years), the mean VT measured by the pneumotachometer, 135.3 mL, was also significantly lower than the ventilator VT reading of 185.4 mL but similar to the calculated effective VT of 167.8 mL.

Analysis of data for specific infants showed that the correlations between the various types of VT measurements were usually poor. Among the children, the correlations between these measurements were much better.

The authors caution that neonates may experience significant adverse consequences, including lung injury, hypoxia, and hypercapnia, from the delivery of imprecise tidal volumes. Conversely, Cannon and associates suggest that accurate delivery of a suitable volume may minimize barotrauma and volutrauma and decrease intrathoracic pressures, thereby avoiding adverse cardiovascular and neurologic effects.

Cannon ML, Cornell J, Tripp–Hamel DS, et al. Tidal volumes for ventilated infants should be determined with a pneumotachometer placed at the endotracheal tube. Am J Respir Crit Care Med. 2000;162:2109– 2112.

Bronchogenic cysts are congenital lesions that generally manifest within the first few decades of life; initial presentation after age 50 is unusual. Most bronchogenic cysts can be easily diagnosed with nonenhanced computed tomography (CT); however, in about one third of cases, the true nature of the lesion can be difficult to identify. The use of contrast material during CT or the addition of magnetic resonance imaging (MRI) helps differentiate cysts with soft-tissue attenuation from mediastinal neoplasia.

McAdams et al retrospectively studied 68 patients, from newborn to age 72, with histopathologic evidence of bronchogenic cyst. Sixty–two patients had undergone CT, 23 had had T1–weighted MRI, and 18 had had T2-weighted MRI.

All but four of the 62 cysts visualized with CT were in the mediastinum. Of the 58 mediastinal cysts, 38 could be accurately diagnosed as bronchogenic cysts. In the remaining 20 cases, confident identification could not be made because of internal heterogeneity, streak artifact, or other causes.

Nine of the 20 cysts that could not be diagnosed with CT were, however, visualized with MRI. In each case, evidence of markedly increased signal intensity on T2–weighted images confirmed the cystic nature of the lesion.

Only 38 of 66 patients (58%) whose clinical histories were available were symptomatic at presentation. The authors recommend that all symptomatic cysts be resected regardless of the patient’s age and that asymptomatic cysts be removed in young patients, who are at low surgical risk. Watchful waiting is the preferred course in the management of asymptomatic adults and high–risk patients.

McAdams HP, Kirejczyk WM, Rosado–de–Christenson ML, et al. Bronchogenic cyst: imaging features with clinical and histopathologic correlation. Radiology. 2000;217:441-446.

Scheduled changes in the predominant antibiotic class used empirically to treat suspected gram–negative bacterial infections may reduce the incidence of inadequate therapy, reported the authors of a prospective study. This approach may be of greatest benefit to the most severely ill patients.

Kollef et al evaluated 3,668 critically ill patients who were treated in one urban teaching hospital during three consecutive time periods of six, six, and five months. During each time period, a specific class of antibiotics was used empirically to treat proven or suspected gram–negative bacterial infections:

• Period 1: Ceftazidime, a third–generation cephalosporin (n = 1,323).

• Period 2: Ciprofloxacin, a fluoroquinolone (n = 1,243).

• Period 3: Cefepime, a fourth-generation cephalosporin (n = 1,102).

Throughout the study, each patient’s physician had the authority to prescribe vancomycin and/or aminoglycoside antibiotics as needed. The primary outcome measure was inadequate treatment of nosocomial infections.

During the study, infections were identified in 1,371 patients. The infections were nosocomial in 517 patients; 125 patients had both community–acquired and nosocomial infections.

The incidence of inadequate antimicrobial therapy decreased from 6.1% to 4.5% during the study’s course, largely because of a significant decrease in the incidence of inadequate treatment of gram–negative bacterial infections (from 4.4% to 1.6%). In–hospital mortality was similar during each of the study periods (15.6%, 16.4%, and 16.2%, respectively) despite a significant increase in illness severity. However, for the most severely ill patients (those with an APACHE II score of 15 or higher), in–hospital mortality was significantly lower during period 3 than during periods 1 and 2.

Multivariate analysis showed that inadequate antimicrobial treatment of nosocomial infection was the most important risk factor for in–hospital mortality.

Kollef MH, Ward S, Sherman G, et al. Inadequate treatment of nonsocomial infections is associated with certain empiric antibiotic choices. Crit Care Med. 2000;28:3456-3464.

Return to table of contents