| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
Professor and Chairman, Department of Medicine, New Jersey Medical School.
Correspondence to: Waldemar G. Johanson, Jr., MD, MPH, FCCP, Department of Medicine, MSB 1506, New Jersey Medical School, 185 South Orange Avenue, Newark, NJ 07103-2714
No one can doubt the importance of a new biomedical treatment that reduces mortality from some specified, highly lethal condition, and such discoveries are greeted with appropriate enthusiasm. Conversely, if a reduction in mortality is not demonstrated following the application of a new treatment or diagnostic technique, the newcomer is perceived to be of suspect value or is rejected out of hand. Mortality has great appeal as an outcome; it is readily measured, there is general agreement as to its presence or absence, and once an outcome is ascertained, subjects do not cross over to another category or group. On the other hand, mortality is applicable to only a limited range of treatments and conditions, at least within the usual periods of observation, and researchers have always used other end points to demonstrate benefit or the lack of it.
Nosocomial pneumonia has always posed a vexing problem in this regard. The condition is difficult to diagnose accurately. Antemortem clinical features correlate poorly with pathologic findings in the lungs at autopsy. Despite numerous studies, there is little agreement even on the interpretation of histologic changes in the lungs of patients who die of suspected pneumonia because of widely varying microbiologic results associated with specific pathologic features.1 ,2 ,3 ,4 ,5 ,6 ,7 ,8 Results of bronchoscopic techniques, including BAL, use of the protected specimen brush, and numerous modifications of these approaches, have been reported endlessly, usually in small numbers of patients from a single institution, so that widely applicable conclusions cannot be drawn. In any case, well-intended investigators have reached opposing conclusions on the usefulness of invasive sampling. How can this be, when the time course of bacterial populations inoculated into the lungs and the associated histopathologic changes in the absence of antibiotic therapy have been known in general terms for nearly 100 years and in detail for at least 50 years? The overwhelming confounding factor in almost all such studies is the virtually universal presence of antimicrobial therapy in intensive care patients receiving mechanical ventilation. Additional factors include the state of the patient's immune defenses and the continuous re-inoculation of the airways with the pathogenic bacteria that occurs in intubated patients. Given these confounders, it is small wonder that the value of diagnostic techniques in nosocomial pneumonia has been difficult to demonstrate.
In this issue of CHEST (see page 1076), Heyland and colleagues have chosen a novel outcome to examine for the beneficial effects of bronchoscopy in patients suspected of having nosocomial pneumonia. Rather than fall prey to the several pitfalls described, they studied the effect of bronchoscopy on the physicians caring for the patient: confidence in the diagnosis, comfort level with the management plan, and belief that pneumonia was or was not present. Having the results of bronchoscopy in hand caused the treating physicians to be more certain of the diagnosis and more comfortable with the treatment. Following bronchoscopy, physicians were more likely to discontinue antibiotics or restrict the number administered.
From the data presented, one cannot tell whether this newfound confidence was appropriate, since 93% of patients were receiving antibiotic therapy at the time of the procedure. When bronchoscopy results were known, antibiotics were discontinued for nearly 20% of patients, but this represents only one half of those (37%) who had sterile cultures. Whether this decision was correct or not is imponderable given our present state of knowledge. Although the authors report reduced mortality among patients undergoing bronchoscopy compared to control patients, little weight should be given to that observation since control patients were those suspected of having pneumonia who did not undergo the intended bronchoscopy. The reasons that led their physicians to forego bronchoscopy may have been the reasons for the greater mortality in these patients. One should not focus on that observation because the study was not designed to test a mortality outcome difference. The key finding is that bronchoscopy is useful in patients with suspected pneumonia because the results appear to reduce uncertainty among physicians. It is the role of researchers and educators to make certain that such assurance is warranted.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
