| 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 |
|||||||||
* From the Department of Medicine (Drs. Janz, Madan, and Schnader), Wright State University School of Medicine, and Department of Medicine (Drs. Janz, Madan, and Schnader), Dayton VA Medical Center, Dayton, OH; Department of Medicine (Dr. Marini), University of Minnesota, and Department of Medicine (Dr. Marini), Regions Hospital, St. Paul, MN; Department of Medicine (Dr. Summer), Louisiana State University School of Medicine, New Orleans, LA; Department of Medicine (Dr. Meduri), University of Tennessee School of Medicine, Memphis, TN; Department of Medicine (Dr. Smith), University of California at San Diego, and the VA Healthcare System, San Diego, CA; Department of Medicine (Dr. Epler), Brigham & Women’s Hospital, and Department of Medicine (Dr. Epler), Harvard Medical School, Boston, MA.
Correspondence to: Jeff Schnader, MD, CM, FCCP, Chief, Division of Pulmonary & Critical Care Medicine, Wright State University School of Medicine, Dayton VA Medical Center (111), 4100 W Third St, Dayton, OH 45428; e-mail: jeff.schnader{at}med.va.gov
Key Words: acute interstitial pneumonitis • ARDS • barotrauma • corticosteroid therapy • Hamman-Rich syndrome • liquid ventilation • mechanical ventilation • positive end-expiratory pressure • protective ventilation • surfactant
 |
Introduction |
|---|
 TOP Introduction Case PresentationResponse of the Pulmonary...Hospital CourseComments of the Treating...AutopsyCommentaryReferences |
|---|
This case conference addresses difficult management problems without singularly correct decisions; its object is not necessarily to seek consensus. Defining the exact issues, formulating rationales for decision making, and committing to the decisions themselves are all equally important in this presentation. This is a real case in which the decisions were made by the "treating pulmonologist" or "treating intensivist" without input from the other participants. The "responses of pulmonary experts" are given only with the knowledge of the case presentation up to the moment at which each expert gave his or her decision and without the knowledge of any of the other opinions rendered. The last "commentary" opinions are given only with the knowledge of the "case presentation" and the remarks of the "treating pulmonologist (or intensivist)" but without the knowledge of any of the other opinions rendered. Although the "commentary" opinions are the last in the sequence of this presentation, they are not necessarily offered as definitive solutions to the problems posed in the case. The reader is the ultimate arbiter in this presentation of decision-making alternatives.
|
Case Presentation |
|---|
|
TOPIntroductionCase PresentationResponse of the Pulmonary...Hospital CourseComments of the Treating...AutopsyCommentaryReferences |
|---|
|
The patient was transferred to the medical ICU, and on the fourth hospital day, bronchoscopy demonstrated normal bronchial mucosa, minimal secretions, and no endobronchial lesions. Stains and cultures of bronchial washings were negative for Pneumocystis, acid-fast bacilli, fungi, bacteria, and viruses. However, because of complications during the procedure, including unstable respiratory parameters, BAL and transbronchial biopsy were judged too dangerous and were not performed. The results of the HIV-1 test were negative, and cotrimoxazole and prednisone were discontinued. The erythrocyte sedimentation rate (ESR) was 78 mm/h (normal, 0 to 15 mm/h). The antinuclear antibody (ANA) titer was 1:80 with a cytoplasmic pattern; the antineutrophilic cytoplasmic antibody test was negative. Serum IgG and IgA levels were 2,598 mg/dL (normal, 564 to 1,765 mg/dL) and 616 mg/dL (normal, 85 to 385 mg/dL), respectively; serum IgM, C3, and C4 levels were normal. Antiglomerular basement membrane antibodies were not present.
Respiratory rates increased to 32 to 40 breaths/min requiring an FIO2 of 0.60 to maintain an SaO2 of > 90%. The chest radiograph showed worsening infiltrates (Fig 2 ). On the sixth hospital day, a right-heart catheter was placed to confirm that the infiltrates were noncardiogenic in nature (see Table 1 for hemodynamic data). The patient was afebrile. The pulmonary artery catheter was removed on hospital day 8.
|
|
Several hours after the initiation of mechanical ventilation, the patient’s BP fell to 89/48 mm Hg, prompting fluid administration, 5 µg/kg/min of dopamine, and 5 µg/kg/min of dobutamine. A right-heart catheter was inserted for hypotension (see Table 2 for hemodynamic data). On a VT of 800 mL and PEEP of 14 cm H2O, the peak airway pressure was 84 cm H2O; plateau pressure was 82 cm H2O; respiratory dynamic and static compliances were 11.4 and 11.8 mL/cm H2O, respectively. Arterial blood gases were pH 7.25; PaCO2, 55 mm Hg; PaO2, 70 mm Hg; and SaO2, 90%.
|
|
Response of the Pulmonary Experts |
|---|
|
TOPIntroductionCase PresentationResponse of the Pulmonary...Hospital CourseComments of the Treating...AutopsyCommentaryReferences |
|---|
The condition that initiated this patient’s problems cannot be confidently established from the limited data at hand. Medical, occupational, smoking, and social histories are not available. The 3 months of exertional dyspnea prior to hospitalization, in association with reduced lung volumes, generalized infiltrates that initially predominated in the lower lung fields, and streaky hemoptysis suggest an underlying disease such as accelerated interstitial pneumonitis (Hamman-Rich syndrome), aspiration pneumonitis, embolism or thrombosis, bronchiolitis obliterans organizing pneumonia (BOOP), limited Wegener’s granulomatosis, parenchymal hemorrhage, or congestive heart failure. Elevations of LDH, ESR, serum immunoglobulins, and antinuclear antibodies are nonspecific markers of inflammation that do not give insight to the underlying pathophysiology. Although this hypertensive patient might be at risk for diastolic dysfunction, echocardiographic and pulmonary artery catheterization data argued strongly against a primary role for systolic myocardial dysfunction or for impairment of the left-sided cardiac valves. Extensive arborizing air bronchograms are evident, a radiographic sign that argues further against heart failure as the primary problem.1
The possibility of underlying immune compromise and secondary infection with Pneumocystis or another opportunistic organism was considered, but dismissed on the strength of a negative HIV-1 test and unproductive bronchial washings. A truly negative HIV-1 test makes AIDS unlikely, but not impossible, especially if the virus were acquired very recently or if the immune compromise results from an uncommon strain. We are not told the total lymphocyte count, nor the complete leukocyte differential. In view of the patient’s sexual history and elevated LDH, I would have ordered a CD4 count and a helper/suppressor ratio if the total lymphocyte count were subnormal. In an immune-compromised patient, invasive aspergillosis is a devastating disease that can present in this fashion and often eludes diagnosis.
Normal serum indicators of renal function, as well as normal hemoglobin levels and cardiac filling pressures, suggest a problem largely confined to the lungs. Pulmonary neoplasia (eg, adenocarcinoma or alveolar cell carcinoma) can present with a subacute course and extensive pulmonary infiltration without involving extrapulmonary vital organs.
It is surprising that despite extensive bibasilar infiltrates, the patient’s profound hypoxemia initially responded to modest concentrations of inspired oxygen without the need for continuous positive airway pressure. Reduced hypoxic drive to breathe and, consequently, less work of breathing and metabolic stress, could have amplified the benefit of supplemental oxygen. Widely varying and extremely elevated values for cardiac output (up to 17 L/min), at a time when the patient did not require mechanical ventilation, suggest a hypermetabolic condition (eg, liver disease, sepsis) or the possibility of measurement error. Such artifacts may arise from tricuspid regurgitation or, more rarely, from direct transseptal flow of venous blood and thermal indicator to the left circuit under the influence of elevated pulmonary arterial pressure (an oxygen-induced reduction of pulmonary artery pressure and intracardiac shunt might help to explain the unexpectedly dramatic effect of inspired oxygen on arterial oxygenation).
How would my management have differed? Although it is always risky to conjecture from a distance, I feel certain that I would have intubated the patient, if necessary, to allow lung lavage to be conducted safely. A thin-section CT and/or spiral CT with IV contrast might also have helped to classify the nature of the pulmonary infiltrates. If these results proved nondiagnostic, I would have strongly considered a videofluoroscopic or even open biopsy to establish the diagnosis. With a reliable diagnosis established, high-dose corticosteroid therapy and/or antifungal treatments could have been undertaken or withheld with more confidence.
Whatever the underlying process, I would not have allowed airway pressures to rise as high as those described. Although a patient with a stiff chest wall may tolerate relatively high pressures without alveolar overdistention, end-inspiratory plateau pressures exceeding 35 cm H2O should be applied only with caution. Laboratory experiments unequivocally demonstrate that high tidal inflation pressures cause hemorrhagic edema, disrupt functional lung units, and may predispose to bacteremia.2 Moreover, in clinical practice, peak airway pressures that exceed 60 cm H2O are associated with a high incidence of life-threatening alveolar rupture (barotrauma).3 Two recent clinical studies suggest that limiting tidal alveolar pressures to values below those normally associated with total lung capacity (35 cm H2O in most patients) may reduce morbidity. A multicenter trial of the low-VT, pressure-limiting approach to ARDS ventilation was recently terminated after a 25% reduction in mortality was documented using this strategy.4 Another study placed emphasis on re-aerating collapsed lung tissue without over-stretching the lung parenchyma.5 To achieve both objectives, high PEEP levels were used, VT was reduced, and hypercapnia was accepted as a consequence. Recruitment maneuvers (sustained applications of moderately high transpulmonary pressure) were often used to help establish airway patency. Data from that study emphasized the central importance of maintaining adequate end-expiratory lung volume—a more important determinant of survival than "driving pressure" (a direct correlate of VT). There is also a growing consensus that prone positioning, a highly effective method for recruitment, often improves oxygenation in this setting and may help avert ventilator-induced lung injury.6 This benefit may accrue because prone positioning sustains relatively high regional transpulmonary forces and establishes airway patency in the well-perfused dorsal areas.
In the case at hand, I would have initiated a strategy geared toward establishing and preserving patency of the open lung without exceeding plateau pressures of ~35 cm H2O. I generally allow hypercapnia but keep pH > 7.20. Therefore, after intubation, completion of diagnostic bronchoscopy, and initiation of a high PEEP/low VT strategy, I would have attempted recruiting maneuvers and repeated them after prone positioning. If the biopsy did not demonstrate invasive fungi or otherwise contraindicate their use, corticosteroids may have helped this patient, perhaps at very high doses. Although the documentation for steroid use is currently limited to the important, yet unconfirmed, work of Meduri and colleagues,7 my own experience is in agreement that corticosteroids are indeed helpful in the later phases of ARDS.
Warren R. Summer, MD, FCCP; New Orleans, LA
This case presents with a wide range of etiologic possibilities
that could be responsible for the subacute illness and acute decline.
Despite a negative HIV-1 enzyme-linked immunosorbent assay, infection
with human T-cell lymphotrophic virus (HTLV)-1 is still possible with
resultant opportunistic infection or noninfectious lung injury
(eg, lymphocytic interstitial pneumonitis, Kaposi’s
sarcoma, etc). Of course, unusual presentations of common infections or
diffuse alveolar damage (DAD) in an immunocompetent patient is also a
possibility. The rapid decline in this patient on day 9 may be related
to progression of the primary disease due to lack of appropriate
therapy, discontinuation of prednisone, or a nosocomial infection.
Despite a lack of evidence supporting improved outcome with invasive
diagnostic interventions in groups of patients with acute lung disease,
I would have attempted a repeat BAL on day 5 or even video-assisted
thoracoscopic surgery with biopsy on day 6 because a specific diagnosis
is likely to help direct therapy in a puzzling individual patient.
However, in the ICU, we often have to focus on a patient’s compromised
cardiopulmonary status and initiate aggressive therapy with no clear
diagnosis.
Optimum management of ARDS is still problematic, and the American-European Consensus Conference on ARDS has outlined several excellent goals.8 The findings of the recent National Institutes of Health (NIH) ARDS Network Clinical Trial has firmly demonstrated that low VT (6 mL/kg) and presumably lung inflation below the upper inflection point of the volume-pressure curve reduces mortality by > 25%.4 Other experimental data and clinical trials have demonstrated benefit with PEEP at or above the lower inflection point.9 In this patient, a VT of 800 mL is probably overdistending many alveolar units with potential ventilator-induced lung injury. Although we don’t know the "best" PEEP for any given individual, I take a practical approach to reach my gas exchange goals. I try to adjust PEEP according to the static volume-pressure curve characteristics. I begin ventilation with low VT (~6 mL/kg), a PEEP of 5 cm H2O, and an FIO2 of 1.0. Then I raise the PEEP 2 to 4 cm H2O every 5 min until SaO2 exceeds 90% or until the plateau pressure (Pplat) rises to ~30 cm H2O. This usually results in recruitment of lung units without overdistention and usually allows a reduction in FIO2 to < 0.75, avoiding oxygen toxicity. When these targets have been achieved, I usually increase the PEEP by another 2 to 4 cm H2O, looking for an additional large increase in PaO2 (which would suggest that more recruitable lung is present) without exceeding a Pplat of 35 cm H2O. The usual PEEP is 14 cm H2O. A recent clinical study suggested that overdistention may be occurring in some patients at low PEEP despite an improvement in SaO2.10 However, maintaining low VT and not allowing Pplat to exceed our target will likely prevent significant ventilator-induced lung injury. If the PaO2 is still < 55 mm Hg or if the FIO2 is > 0.75, a slight prolongation of the inspiratory time to increase mean airway pressure may further improve PaO2 into the low 60s, which would be acceptable. The Pplat of 82 cm H2O in this patient is much too high. The calculated static compliance is extremely low because he is being ventilated above his upper inflection point with excessive VT.11 His compliance is likely to be better if VT is only 6 mL/kg. Low VT may compromise gas exchange; however, a slow elevation of PCO2 appears well-tolerated in most patients. Low VT may also produce oxygenation difficulties, but these can be overcome (as mentioned above) and mortality will be reduced.
I acknowledge that sepsis is a concern in this patient, but my major concern at this juncture is not sepsis (although it is perhaps present). Rather, my greatest concern rests with the ventilator-induced elevations in intrathoracic pressure to which this patient has been subjected, and that may explain some of the more worrisome hemodynamic findings that others might attribute to sepsis. Although very stiff lungs will not always transmit ventilator-induced positive airway pressures to the pleural space (which could decrease cardiac output), the combination of high PEEP, large inflation pressures, and diuresis can result in decrements in cardiac output,12 as seen in this patient. Lung overinflation may impinge on the cardiac fossa, leading to decreased venous return, and also may promote ventricular interdependence by physiologically restricting the heart. Thus, the relatively low cardiac indexes in this patient—4.8 and 3.0 L/min/m2 (see Table 2 )—probably reflect a ventilator-induced effect. However, in spite of this attenuation of cardiac index, it is still clinically acceptable at this point, and I would not try to increase venous return through fluid administration (in view of the risk of increasing lung edema and worsening gas exchange). I would rather initially treat the patient’s altered hemodynamics by manipulating the ventilator in order to reduce mean airway pressure, principally through reductions in VT.
In summary, major attention must be paid to oxygen delivery and avoiding ventilator-induced lung injury. If hemodynamic stability cannot be achieved (mean arterial pressure in this patient was 84 mm Hg) by VT and pleural pressure reduction, the use of norepinephrine should be instituted and treatment with dopamine and dobutamine stopped. Treatment of nosocomial infection with broad coverage for methicillin-resistant Staphlococcus aureus and Pseudomonas is also indicated on day 9. Optimizing this patient’s cardiopulmonary status may allow for further diagnostic evaluation.
|
Hospital Course |
|---|
|
TOPIntroductionCase PresentationResponse of the Pulmonary...Hospital CourseComments of the Treating...AutopsyCommentaryReferences |
|---|
On the 10th hospital day, subcutaneous emphysema was noted on physical examination, and the chest radiograph and CT scan showed diffuse alveolar infiltrates, extensive pneumomediastinum, pneumoperitoneum, and a small right pneumothorax (Fig 3 ). A right chest tube was placed. Temperature dipped to 94.0°F, PaO2 fell to 43 mm Hg, and creatinine rose to 2.6 mg/dL. Methylprednisolone was begun empirically, and midazolam and vecuronium infusions were discontinued. In spite of this, the patient remained comatose: he was unresponsive to stimuli and diffusely flaccid, and lacked light, corneal, and oculocephalic reflexes. On hospital day 12, arterial blood gases revealed a pH of 7.22, PaCO2 of 44 mm Hg, PaO2 of 59 mm Hg, an oxygen saturation of 88% on a VT of 800 mL, a respiratory rate of 22/min, an FIO2 of 1.00, and 20 cm H2O of PEEP. On day 15, a left chest tube was inserted for a tension pneumothorax (Fig 4 ). On the following day, a nuclear brain scan was performed and revealed no cerebral blood flow. The regional organ procurement organization was consulted; however, the patient was refused as an organ donor. After the family was consulted, the patient was removed from life support and died. [Editor’s note: The decisions against bronchoscopic lavage, transbronchial biopsy, and open lung biopsy in this case were not made by the authors.]
|
|
|
Comments of the Treating Intensivist in this Case |
|---|
|
TOPIntroductionCase PresentationResponse of the Pulmonary...Hospital CourseComments of the Treating...AutopsyCommentaryReferences |
|---|
Over the last several years, animal and human studies have documented the existence of ventilator-induced lung injury.13 14 15 Currently advocated "lung-protective strategies" for ventilatory management include the use of VTs of 5 to 7 mL/kg, the avoidance of peak airway pressures of > 40 to 50 cm H2O, Pplats of > 30 to 35 cm H2O, inverse ratio ventilation, permissive hypercapnia, and PEEP titrated to the inflection point on the static pressure-volume curve.9 Rather than constructing a pressure-volume curve, we titrated the PEEP to optimize oxygen delivery and assure adequate arterial oxygenation. Several methods for optimizing PEEP therapy have been suggested, but none have proven to be superior to other methods.8 16 My major concern has generally been arterial oxygenation. At the time of this case, none of the above-mentioned "advocated approaches" to mechanical ventilation in ARDS had been conclusively shown to be superior to more conventional, volume-controlled methods, eg, VTs of 10 to 12 mL/kg. Because of the conflicting literature, my own approach has been to use a conventional method of mechanical ventilation with PEEP therapy, unless airway pressures are consistently high, eg, Pplat of > 30 to 35 cm H2O or peak airway pressures of > 40 cm H2O. When they are high, I prefer to use smaller VTs (eg, 6 to 8 mL/kg) and, if needed, inverse-ratio ventilation.17
However, a recent multicenter ARDS study was prematurely stopped when patients ventilated with 6 mL/kg VTs had a 25% reduction in mortality compared with patients ventilated with 12 mL/kg VTs.4 Although the study data were not available when we managed this patient, use of smaller VTs in ARDS may prove to be superior to more conventional VTs. However, in our patient, we found that a PEEP of 17 to 20 cm H2O accomplished our above-stated goals, with FIO2 ranging from 70 to 100%. Because of the very high airway pressures and eventual development of barotrauma, the level of PEEP was not increased further, and the goal of lowering the FIO2 was sacrificed. Although oxygen toxicity is well described in the literature, its clinical occurrence, particularly in the setting of ARDS, is uncertain.18 Because the clinical presentation of oxygen toxicity is similar to that of ARDS, its contribution to morbidity in this setting is confusing. The FIO2 in our patient ranged from 70 to 100% for 7 to 8 days. Despite these high FIO2 levels for a relatively long time, I do not think our patient developed oxygen toxicity, because his PaO2 values and chest radiograph stabilized and showed mild improvement in the days prior to his being disconnected from life support.
The initial VT of 800 mL (approximately 9.5 mL/kg) was reduced in an attempt to decrease the high airway pressures. However, as the VT was reduced, the patient became more hypoxemic. Repeated attempts to lower the VT were uniformly associated with reductions in our ability to oxygenate the patient. Consequently, the VT was maintained at the initial setting of 800 mL. We tried inverse-ratio ventilation, but there was no significant improvement in either oxygenation or ventilation. Some controlled trials have shown that as many as 75% of patients with ARDS have improvement in oxygenation when moved from a supine to a prone position.6 Prone positioning was considered, but because of the technical difficulties of positioning the patient, it was abandoned. Despite the inherent technical difficulties, a trial of prone positioning might have improved gas exchange, but we did not feel this would have ultimately altered the progression of the disease or the patient’s hospital course.
Although corticosteroids have not been proven to be helpful in the early stages of ARDS,19 there is data suggesting that their use during the fibroproliferative phase of ARDS may be of benefit.7 Based on this data and the severity of the disease in this case, we empirically began corticosteroids 7 to 8 days after the onset of ARDS (hospital day 10). Unfortunately, we were unable to demonstrate any advantage with their use in this patient. I am not currently convinced that the data supports the use of corticosteroids in all or most cases of ARDS, although the rationale for their use is appealing.
|
Autopsy |
|---|
|
TOPIntroductionCase PresentationResponse of the Pulmonary...Hospital CourseComments of the Treating...AutopsyCommentaryReferences |
|---|
|
|
Commentary |
|---|
|
TOPIntroductionCase PresentationResponse of the Pulmonary...Hospital CourseComments of the Treating...AutopsyCommentaryReferences |
|---|
The patient had several factors indicating the presence of a rapidly progressive interstitial lung disease: (1) failure to resolve spontaneously over a period of 3 months with recent clinical deterioration; (2) severe hypoxemia at rest; (3) diffuse bilateral interstitial infiltrates on chest radiograph; and (4) echocardiographic evidence of pulmonary hypertension with right atrial enlargement. The immunologic profile revealed several abnormalities (elevated ESR, ANA, serum IgG and IgA levels) usually not observed in ARDS. In the presence of a nondiagnostic incomplete bronchoscopy (no BAL or transbronchial biopsies obtained), I would have subjected the patient to open lung biopsy early in the hospital course. The differential diagnosis included several conditions with a similar clinicoradiographic presentation, such as usual interstitial pneumonia, desquamative interstitial pneumonia, diffuse alveolar hemorrhage syndrome (history of hemoptysis), acute eosinophilic lung disease, rapidly progressive interstitial pneumonitis associated with connective tissue diseases, hypersensitivity pneumonitis, and so forth.
Recently it has been appreciated that a short course of glucocorticoids
may increase proinflammatory cytokine (eg, TNF-
) release
in response to a later inflammatory signal for a period of 
6 days
after steroid discontinuation.21
Theoretically, the rapid
deterioration observed in this patient in association with a new
suspected infection might have been related to an amplification of the
inflammatory response from prior use of a short course of
glucocorticoids. Conversely, recent randomized trials indicate a
beneficial response to prolonged (7 to 12 days)
glucocorticoid administration in patients with septic
shock.22
23
In this patient, methylprednisolone treatment for presumed unresolving ARDS was initiated on hospital day 10 and was not associated with a physiologic improvement. This response is not surprising, as acute interstitial pneumonia (AIP) is a condition unlikely to be responsive to glucocorticoids once acute respiratory failure develops.24 The histologic findings of AIP resemble the organizing stage of DAD. In our institutional experience, the histologic features of AIP resemble those of ARDS patients who failed to respond to methylprednisolone treatment, and include dense acellular fibrosis, loss of alveolar architecture, and presence of arteriolar subintimal fibroproliferation. In our experience, the following variables were significantly different among responders and nonresponders: presence of preserved alveolar architecture (p = 0.045), myxoid-type alveolar and interstitial fibrosis (p = 0.045), coexistent intraluminal bronchiolar fibrosis (p = 0.0045), and absence of arteriolar subintimal fibroproliferation (p = 0.045).25
I believe prolonged methylprednisolone administration in ARDS patients failing to improve in the first week of mechanical ventilation is the only treatment intervention shown to be associated with an effective and sustained biological and physiologic improvement,7 26 although it is not universally effective. Late institution (ie, day 20 of ARDS) has been associated with a 50% failure rate.7 Furthermore, we have observed that morbidly obese patients frequently do not respond to treatment (unpublished data).
We have underscored the fact that during methylprednisolone treatment, infections develop in the absence of fever, and that a surveillance program is indispensable for detecting and treating infections that can otherwise be easily missed. Not uncommonly, we have observed a dampening of patient recovery in the presence of infection that is reversed following treatment of the infection.
Robert M. Smith, MD, CM, FCCP; San Diego, CA
It should probably be said up front that it is likely that no
therapeutic intervention would have had a major impact on the outcome
in this case. Nevertheless, there are a number of points that are
worthy of discussion. The epidemic of AIDS and our improved
understanding of the manifestations of severe HIV infection have
altered practice patterns significantly and largely for the better.
Along these lines, the possibility of HIV-associated illness has led us
to place other possible etiologies lower on the differential diagnosis.
In this particular case, the probability of Pneumocystis
carinii pneumonia or other HIV-associated diseases likely prompted
a more leisurely approach to diagnosis than would otherwise have been
pursued. By the fourth or possibly the fifth hospital day, in the
setting of a negative HIV-1 test and negative bronchial washings, a
more aggressive diagnostic approach may have been warranted. A narrow
window of opportunity at this time may have allowed a thoracoscopic
lung biopsy that was not possible later. Similarly, a high-resolution
CT scan with thin cuts of the lung parenchyma may have provided
additional diagnostic information. An early diagnosis of AIP may have
prompted high-dose corticosteroid therapy and possibly cytotoxic
therapy at that time. If the risks of obtaining a diagnostic tissue
sample were believed to outweigh the benefits (as the clinicians in
this case may have believed), I believe I would have instituted
high-dose corticosteroid therapy earlier in the patient’s course.
Unfortunately, only 20/20 hindsight makes this decision simple. At the
bedside, it is never that straightforward. As this case illustrates,
the alternative of expectant observation may be equally fraught with
risk, particularly in the setting of rapidly progressive infiltrates
with marked derangement of function.
Once intubated, the patient meets physiologic criteria for ARDS.27 There is no clear consensus of what defines the term acute, and time limits in various clinical trials vary. In this patient, although the history of dyspnea dates back 3 months, there is no information about the duration of the pulmonary infiltrates and so it is unclear if a diagnosis of ARDS is truly appropriate. At a minimum, the time course of the patient’s presentation is atypically prolonged and suggests that he may not respond or behave in ways that are the norm for ARDS. Nevertheless, there is no diagnostic test or assay that confirms or denies the presence of this syndrome, and recent lessons from the management of ARDS should apply. We know that ARDS is associated with intense inflammation in the lung and the eventual onset of a dramatic fibrotic response. A variety of therapies directed specifically at that inflammation (eg, early corticosteroid administration19 ) have been attempted without success. Direct instillation of pulmonary surfactant into the airways appears promising,13 but I do not believe this preliminary data is sufficient to support surfactant administration outside of controlled clinical trials. In contrast, recent information from animal studies and preliminary information from large-scale clinical trials now confirms that ventilator management can markedly influence outcome in patients with ARDS.
Substantial animal data demonstrate that positive-pressure ventilation with airway pressures as low as 30 cm H2O causes diffuse injury to the alveolar-capillary membrane.14 15 This is accompanied by influx of inflammatory cells and morphologic changes identical to those seen in patients with ARDS.28 This injury appears related to repetitive stretch of the alveolar-capillary membrane rather than pressure per se, as efforts to limit excursion of the lung minimize the injury. For this reason, application of PEEP in these animal models (presumably by limiting the decrease in the volume of alveoli in exhalation) protects against this injury.14 These observations lead to the obvious questions of whether these findings apply to ventilation of patients with lung injury and whether they apply to ventilator management strategies in general.
A recent NIH ARDS Network trial examined a low-volume vs a moderate-volume ventilation strategy. In this study, subjects with ARDS were randomly assigned to receive ventilation with either 6 mL/kg or 12 mL/kg VT. These volumes were calculated based on lean body weight and were closer to 5 and 11 mL/kg actual body weight in the study subjects. PEEP and FIO2 were adjusted based on algorithm and were comparable in both groups. Respiratory rate was adjusted to maintain minute volume at prerandomization levels. Pplat limits of 30 and 50 cm H2O were set for the small and large VT groups and algorithms were used to reduce VT if these limits were exceeded. Although pH and PCO2 levels were comparable in the two groups, PO2 levels were marginally but significantly lower in the low-volume group. Despite this, the study was stopped before it reached its target patient accrual because of excess mortality in the subjects receiving larger VT (40%) compared with those receiving small VT (30%).4 29 I believe this study, together with similar results in smaller trials,5 conclusively demonstrates that ventilator strategies resulting in higher airway pressures and larger VT can damage human lungs as well as those of small or large animals. The occurrence of this survival benefit despite somewhat worse oxygenation parameters suggests that optimization of oxygenation may not be a sufficient goal for patients with ARDS and perhaps not for ventilated patients in general.
In this particular patient, I would have opted for a lung-protective strategy with initiation of low VT breaths from the outset. To achieve this, I prefer a pressure-cycled ventilation strategy limiting the driving pressure to 20 cm H2O with an inspiratory time of 0.8 to 1.0 s, adding PEEP to a minimum of 8 to 10 cm H2O, and incrementing to achieve PaO2 of > 50 mm Hg. However, acceptance of the principles of evidence-based medicine suggests that a strategy directly paralleling that used in the NIH trial may be preferable. I have not found inverse-ratio ventilation helpful except in very rare circumstances, and would not have used it here. Whether any of these approaches would have been successful in this unfortunate individual is not possible to predict.
Gary R. Epler, MD, FCCP; Boston, MA
Acute infectious pneumonia or a pneumonia superimposed on a
chronic process are appropriate considerations for a patient with 3
weeks of worsening dyspnea, cough, crackles, severe hypoxemia, and
lower lung radiographic infiltrates. The fulminant course that then
developed during the first 5 days of hospitalization suggests an
infectious or septic-related ARDS. The findings also suggest a
noninfectious catastrophic illness such as rapidly progressive BOOP or
AIP. A lung biopsy would have been beneficial at this stage, especially
for the discovery of an inflammatory process that could respond to
corticosteroid therapy, but it was reported that the patient was too
unstable.
During a recovery stage of a viral or atypical bacterial pneumonia, some patients may develop an "organizing pneumonia" process. This is an unlikely occurrence in this patient because the illness relentlessly progressed in severity and the process did not respond to corticosteroid therapy. The organizing pneumonia generally occurs during a stabilization phase and almost always responds to corticosteroid therapy. Lung tissue at autopsy, 12 days after hospitalization, indicated a diagnosis of AIP.
The clinical findings are typical of AIP. In an early report of eight patients24 and a later report of 29 patients,30 most had symptoms for < 3 weeks. Many had an associated febrile or flu-like illness. The mean age was 30 years in the earlier study and 50 years in the later study. Male and female patients were equally affected. Nonproductive cough was common. Crackles were reported in 10% and hemoptysis occurred in one patient. These findings are indistinguishable from fulminant or rapidly progressive BOOP.31 32
The chest radiograph in this patient was typical for AIP. Among a group of nine patients,33 the films showed bilateral airspace opacification in all of them, lower lungs in two. The chest CT scans showed ground-glass attenuation in all patients. Subpleural honeycombing was seen in three patients. Among 14 patients with AIP, the high-resolution CT scans showed increased attenuation without traction bronchiectasis in the exudative phase and increased attenuation with traction bronchiectasis and honeycombing in the fibrotic phase.34 Air bronchograms were prominent in this patient, suggestive of an infectious pneumonia, although this finding occurs in BOOP as well. Air bronchograms were not mentioned in the studies of the radiographic features of AIP,34 35 although a case report of a 63-year-old woman with AIP indicated nodules with an air bronchiologram.35
The histologic hallmark of AIP is interstitial fibrosis and edema associated with type II pneumocyte hyperplasia.24 There is extensive fibroblast proliferation and little collagen deposition. Findings of hyaline membranes reflect early findings in DAD.30 No histologic features are prognostically useful. AIP differs from BOOP because the former is associated with DAD, hyaline membranes, and interstitial fibrosis, while the latter is associated with granulation tissue organization in the bronchioles and alveoli. Corticosteroid therapy is considered the best treatment option, although most patients do not respond. The prognosis is poor. Survival rates range from a low of 14%24 33 to a high of 41%.30
In summary, AIP is a distinct respiratory entity. It differs from acute eosinophilic pneumonia, fulminant BOOP, nonspecific interstitial pneumonia, desquamative interstitial pneumonia, and from usual interstitial pneumonia. Synonyms include acute diffuse interstitial fibrosis and the Hamman-Rich syndrome. Among the four patients described by Hamman and Rich, three were typical of AIP.30 The disorder is characterized by rapid-onset acute respiratory failure and poor response to corticosteroid therapy.
|
Footnotes |
|---|
Received for publication August 17, 1999. Accepted for publication August 18, 1999.
|
References |
|---|
|
TOPIntroductionCase PresentationResponse of the Pulmonary...Hospital CourseComments of the Treating...AutopsyCommentaryReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. F. Isik, I. Kati, I. Bayram, and H. Ozbek A new agent for treatment of acute respiratory distress syndrome: thymoquinone. An experimental study in a rat model Eur. J. Cardiothorac. Surg., August 1, 2005; 28(2): 301 - 305. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. R. Epler Bronchiolitis Obliterans Organizing Pneumonia Arch Intern Med, January 22, 2001; 161(2): 158 - 164. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS |
