Pneumocystis jiroveci pneumonia

Objective: To examine outcome and associated factors of acute respiratory failure (ARF) in non-HIV-related Pneumocystis pneumonia (PCP) in patients admitted to a medical ICU between 1995 and 2002.

Design: A retrospective review of medical records and an APACHE (acute physiology and chronic health evaluation) III database.

Setting: Academic tertiary medical center.

Results: We identified 30 patients with non-HIV-related PCP and ARF. In-hospital, 6-month, and 1-year mortality rates were 67%, 77%, and 80%, respectively. Median age was 63.5 years. Median APACHE III score on day 1 was 65.5. Median ICU and hospital lengths of stay were 13 days and 21 days, respectively. All seven patients having a pneumothorax died. All but one patient had an elevated lactate dehydrogenase level (median, 563 U/L). The diagnosis was made using BAL in 28 patients and by transbronchial biopsy in the remaining 2 patients. All patients were immunosuppressed (eight were receiving corticosteroids, seven were receiving chemotherapy, and the remainder received both). Median immunosuppressive prednisone-equivalent dose was 40 mg (median length of treatment, 4.5 months). Not a single patient received PCP prophylaxis. All but one patient required intubation and invasive positive pressure ventilation (PPV). Hospital mortality was associated with high APACHE III scores on day 1 (p = 0.05), intubation delay (p = 0.03), length of PPV (p = 0.003), and development of pneumothorax (p = 0.033). Logistic regression analysis demonstrated that association of intubation delay with hospital mortality persisted after adjusting for severity of illness (p = 0.03).

Conclusions: Among patients with ARF secondary to non-HIV-related PCP, poor prognostic factors include high APACHE III scores, intubation delay, longer duration of PPV, and development of pneumothorax. None of the patients in this series received PCP prophylaxis prior to the development of pneumonia.

Key words: acute respiratory failure; ICU; Pneumocystis pneumonia

Abbreviations: APACHE = acute physiology and chronic health evaluation; ARF = acute respiratory failure; LDH = lactate dehydrogenase; LOS = length of stay; MV = mechanical ventilation; NIPPV = noninvasive positive pressure ventilation; PBW = predicted body weight; PCP = Pneumocystis pneumonia; PED = prednisone-equivalent dose; PEEP = positive end-expiratory pressure; PPV = positive pressure ventilation. TBBX = transbronchial biopsy. TMP/ SMZ = trimethoprim/sulfamethoxazole; VT = tidal volume

**********

The increased incidence of Pneumocystis pneumonia (PCP) observed between 1981 and 1995 was largely related due to the burgeoning AIDS epidemic. (1,2) Recent advances in diagnosis, prophylaxis, and treatment of HIV infection have subsequently resulted in decreased numbers of PCP cases among patients with HIV infection, and better outcomes have been reported for those with PCP. (3,4) Mansharamani et al (5) showed that after the peak of HIV-related PCP cases in 1987, the number of cases declined almost 80% by 1995. The proportion of cases requiring hospitalization declined as well, and the overall mortality improved substantially. (4,5) Despite this, the high mortality of patients requiring mechanical ventilation remained unchanged, ranging from 50 to 60%. (5) Over this same period of time, patients with non-HIV-related PCP showed at least stable, if not increased, hospitalization, ICU admission, and mortality rates. (5-7) The overall mortality was reported to be 39%, and mortality for intubated patients with non-HIV-related PCP was 59% in one large series. (5)

It is well established that the clinical course of PCP differs in patients with or without HIV infection. (1,8-9) Non-HIV-related PCP presents as an acute fulminant pneumonia with abrupt onset of respiratory failure, while HIV patients experience a more insidious course. Furthermore, the organism burden and inflammatory response in lungs are markedly different between these two groups, and this contributes to the observed difference in outcomes, (8-10) However, despite these differences, the previous Literature (5,11-14) on both HIV- and non-HIV-related PCP prior to 1995 indicate that the mortality of patients with acute respiratory failure (ARF) requiring mechanical ventilation (MV) does not differ widely, and ranges between 40% and 60%. In a retrospective analysis (15) of 116 non-HIV-related PCP cases at our institution between 1985 and 1991, the in-hospital mortality of patients who had ARF (n = 50) was 66%.

In order to examine the outcome and associated factors of ABF due to non-HIV-related PCP at our institution in more recent years, we retrospectively collected data for a consecutive series of non-HIV-related PCP patients requiring positive pressure ventilation (PPV) and admission to the ICU between 1995 and 2002. Clinical, laboratory, and radiologic features were examined, as were markers of disease severity, mechanical ventilation parameters, and patient outcomes.

MATERIALS AND METHODS

We performed a retrospective analysis of the Mayo Clinic Rochester medical record registry and APACHE (acute physiology and chronic health evaluation) III database on all consecutive patients with microbiologically confirmed diagnosis of PCP, who were admitted to the medical ICU requiring treatment for ARF with PPV. Patients were excluded from the study if they had suspected but unconfirmed PCP, if they had known previous history of PCP prior to the current admission to the hospital and ICU, if they had mild disease and therefore did hot require ICU admission, or if they had positive HIV test results or were believed to clinically be at high risk for AIDS in the absence of confirmation of HIV. Clinical data abstracted included the following: general demographic information; APACHE III score; hospital and ICU length of stay (LOS); overall hospital mortality; 6-month and 1-year mortality rates; laboratory analysis; radiology; microbiology; means of diagnosis; underlying immunosuppressive condition, including medications, PCP prophylaxis, antibiotic and steroid therapy; mechanical ventilation; and hospital discharge location.

All patients had a PCP diagnosis established by identification of organisms on BAL or transbronchial biopsy (TBBX). The doses of corticosteroids used for immunosuppression and as adjunctive therapy for PCP were converted to a prednisone-equivalent dose (PED) based on the standardized relative potency criteria36 Laboratory data collected in addition to the APACHE data included lactate dehydrogenase (LDH) and CD4 counts. The microbiology data reflected positive Gram stain or culture specimens during or shortly (1 week) after ICU stay. MV data included invasive vs noninvasive PPV, time of intubation relative to the time of ICU admission (on or off noninvasive PPV) and MV settings: positive end-expiratory pressure (PEEP), and tidal volumes (VTS) reported as actual volumes and in milligrams per kilogram of predicted body weight (PBW).

Statistical analysis was done using statistical software (JMP, Version 5.1; SAS Institute; Cary, NC). For continuous variables, median values and ranges were reported. Association of independent variables with hospital mortality was assessed by using nonparametric Wileoxon test and [chi square] test. Logistic regression analysis was performed in order to assess whether delayed intubation was associated with hospital mortality when adjusted for severity of illness (APACHE III score); p < 0.05 was considered statistically significant.

RESULTS

Demographic Features

A search of the Mayo Clinic medical record registry from January 1, 1995, to December 31, 2002, yielded 153 patients with microbiologically confirmed PCP. Sixty-five patients were excluded because they did not require ICU admission. Therefore, 88 records were examined for other study criteria. Of those, 33 patients were round to have their first episode of PCP associated with ARF requiring PPV and ICU management. Three patients were found to be HIV positive and were excluded from the study. The final cohort consisted of 30 patients with ARF secondary to non-HIV-related PCP. Of these 30 patients, 17 were femme. The median age of the patients in the cohort was 63.5 years (range, 25 to 79 years). The median APACHE III score on day 1 was 65.5 (range, 23 to 118), and the median ICU LOS and hospital LOS were 13 days (range, 1 to 43 days) and 21 days (range, 3 to 60 days), respectively (Table 1).

Immunosuppressive Conditions and Prophylaxis

All patients were immunosuppressed prior to the development of PCP. Eight patients had received systemic corticosteroids only. Seven patients were treated with cytotoxic chemotherapy not containing corticosteroids, and the remainder had received a combination of both corticosteroids and chemotherapy (Table 1). Of the 23 patients who received corticosteroids (with or without chemotherapy), 19 had complete documentation in the medical records of the dose and duration of therapy. The median PED and duration of corticosteroid use prior to diagnosis of PCP was 40 mg (range, 5 to 200 mg) and 4.5 months (range, 12 days to 25 months), respectively. Hematologic malignancies, inflammatory diseases, and solid malignancies were the primary diagnoses present in the majority of patients, accounting for 90% of the underlying immunosuppressive conditions (Table 1). No patient was receiving prophylactic medications for PCP prior to onset of pneumonia. Only one patient had transiently received trimethoprim/sulfamethoxazole (TMP/SMZ) prophylaxis, which was discontinued because of a skin rash. No alternative way of prophylaxis was prescribed subsequently.

Laboratory and Radiologic Features

The diagnosis of Pneumocystis jiroveci infection was established by calcofluor white staining of BAL material in 28 of these patients. Four of these patients had additional confirmation of the Pneumocystis organisms, two by TBBX and two on postmortem examination. The diagnosis for the remaining two patients was established by TBBX alone. In addition to Pneumocystis, other microorganisms were further detected either by Grain stains or culture specimens in 15 patients over the course of ICU stay or within 1 week thereafter (Table 2). The presence of microorganisms on Grain stains or cultures did not show evidence of association with increased mortality (p = 0.16). Positive respiratory specimen results were deemed to be secondary to microorganism colonization or latent infection rather than active disease.

Laboratory data including LDH determinations and CD4 counts were available from 18 patients and 6 patients, respectively. The median LDH was 563 U/L (range, 104 to 1,529 U/L), with only one patient having an LDH value within the normal range. The median CD4 count was 70 cells/[micro]L (range, 9 to 223 cells/[micro]L), and only one patient had CD4 count level > 150 cells/[micro]L (Table 1).

As expected, chest radiographic examinations generally revealed diffuse interstitial and alveolar infiltrates on hospital admission. Subsequent chest radiographs in all patients identified seven patients with pneumothorax over the course of their ICU stay. None of these patients survived the hospitalization.

Treatment and Ventilatory Support

Twenty-nine of the 30 patients were treated with TMP/SMZ; of these, one patient individual also received primaquine and clindamycin for 2 days. The remaining patient was allergic to TMP/SMZ, and therefore received pentamidine. All patients were additionally treated with adjunctive corticosteroids. The median PED of corticosteroids used in treatment for ARF related to PCP in these patients without AIDS was 65 mg (range, 10 to 100 mg).

Seventeen patients were placed on noninvasive positive pressure ventilation (NIPPV) upon ICU admission, 13 in the form of continuous positive airway pressure, and 4 in the form of bilevel positive airway pressure (BiPAP; Respironics; Murrysville, PA). Complete data on NIPPV was available for 16 of these 17 patients. The median duration of NIPPV prior to intubation was 35 h (range, 7 to 132 h). Four of the 16 patients who received NIPPV survived until hospital discharge (median duration of NIPPV, 10.5 h). The median duration of NIPPV was 49 h in 12 of the 16 patients who died (Table 3).

Twenty-nine patients required intubation and invasive PPV for a median of 5 days (range, 1 to 42 days). Complete data on MV was available for 27 of 29 intubated patients. The median time from ICU admission until intubation was 20 h (range, 0.1 to 192 h). For those who survived to hospital discharge (10 of 28 patients), the median time from ICU admission to intubation was 4 h (range, 0.1 to 96 h); for those who died (18 of 28 patients), intubation occurred a median of 48 h after admission (range, 0.1 to 192 h) [Table 3]. The median PEEP and VTS received by these patients on day 1 were 5 mm Hg (range, 5 to 15 mm Hg) and 10.5 mL/kg PBW (range, 5.5 to 15.2 PBW), respectively. The median ratio of Pa[O.sub.2] to fraction of inspired oxygen on day 1 was 96 (range, 49 to 439), with only one patient having a ratio > 200. The median PEEP and VT required over the first 3 days were 8.3 mm Hg (range, 5 to 13.3 mm Hg) and 10 mL/kg of PBW (range, 5.5 to 15.7 mL/kg of PBW), respectively.

Patient Outcomes

Twenty patients did not survive to discharge from the hospital, yielding a mortality of 66.6%. Of the 10 patients who were discharged from the hospital, 3 died within 6 months and another died within a year. Accordingly, the 6-month and 1-year mortality rates were 77% and 80%, respectively. Of the 10 patients who survived until hospital discharge, 6 were discharged to home, 2 were discharged to short-term rehabilitation facilities, and 2 were discharged to nursing homes. Of the six patients who were discharged home, rive were alive after 1 year. Of the two patients who were discharged to rehabilitation facilities, one died within 6 months and the other was alive at 1 year. Both patients who were discharged to nursing home died within 6 months.

Statistical analyses were performed to identify clinical features associated with mortality. Using univariate analyses, clinical features associated with mortality are presented in Table 3. Clinical features significantly associated with death from PCP include elevated APACHE III score on day 1 (p = 0.05), delay in intubation (p = 0.033), duration of PPV (p = 0.003), and the subsequent development of pneumothorax (p = 0.033). Logistic regression analysis confirmed that the delay in intubation was associated with higher hospital mortality after adjusting for severity of illness (p = 0.03; odds ratio, 1.04 for each hour of delay; 95% confidence interval, 1.01 to 1.09).

DISCUSSION

Our data indicate that ARF related to PCP in patients without HIV infection remains a serious illness with substantial hospital mortality. Compared with a similar study cohort from our institution 10 years earlier, (15) the mortality remained unchanged (66%). We further round that mortality was associated with elevated APACHE III score on the day of ICU admission, delay in intubation, duration of MV, and the development of pneumothorax in these patients.

Similarly, Forrest et al (13-14) showed, in a retrospective review, that APACHE II and modified multisystem organ failure scores were both significant predictors of outcome in patients with AIDS-related PCP and respiratory failure. Our study further documents that APACHE III scores are predictive of mortality among non-HIV patients with PCP. The APACHE III model has previously been demonstrated to consistently prediet overall morality across diagnoses in two large, prospective, multicenter North American series (37,668 patients and 116,340 patients). (17,18) An Australian study (19) on 3,398 patients further demonstrated that the APACHE III could perform well as an independent assessment of probable mortality.

Delay in intubation, longer duration of PPV, and development of pneumothorax were also associated with mortality in our cohort. Bedos et al (20) prospectively evaluated a similar cohort of patients with HIV-related PCP and respiratory failure. Their study revealed that delay of MV beyond 3 days, duration of MV for > 5 days, and development of pneumothorax were predictive of patient's death within 3 months of ICU admission. (20) Among patients with HIV-associated PCP and both delayed MV and pneumothorax, the predicted probability of 3-month death was dose to 100% in that study. In another review (21) of patients with AIDS with respiratory failure due to PCP requiring MV, those with a pneumothorax also had a mortality of 100%. In our series of ARF in patients with PCP in settings other than HIV, pneumothorax developed in seven patients, and none of them survived. If remains uncertain whether pneumothorax simply represents a marker of severe disease and poor prognosis or whether it independently worsens disease severity and mortality, as previously suggested. (22) Because all patients with pneumothorax in our study died, logistic regression analysis was not feasible to verify if pneumothorax was associated with hospital mortality after adjusting for severity of illness.

Similar to patients with HIV infection and ARF due to PCP, (20) delayed institution of MV in our patients was associated with increased mortality. For those who survived to hospital discharge, the median rime from ICU admission to intubation was 4 h, compared to 48 h in those patients who died. In addition, for patients initially managed with NIPPV, the length of time from onset of NIPPV until intubation and institution of PPV was also significantly associated with outcome. Thus, prudent and earlier intubation and PPV may also be of benefit in the management of patients with ARF associated with PCP in patients without HIV.

Our current study clearly indicates that similar prognostic factors may be applied to both HIV- and non-HIV-related PCP patients with ARF. The clinical differences observed in milder cases of HIV- and non-HIV-related PCP appear to become much less significant once the disease has progressed to ARF. In this setting, both prognostic factors and mortality are very similar regardless of HIV status.

The Acute Respiratory Distress Syndrome Network study2a demonstrated decreased mortality in patients receiving MV with reduced VT (6 mL/kg PBW), as compared to traditional MV strategies. Although this study was published in May of 2000, protective lung ventilation strategies were not immediately applied nationwide. As a result, we are unable to determine whether these patients might have benefited from such a ventilatory strategy. The observation that all our patients except one had a ratio of Pa[O.sub.2] to fraction of inspired oxygen < 200 and received MV with a median VT of 10.5 mL/kg, with relatively low PEEP, suggests that these individuals may have been candidates for lower-VT protective lung ventilation. Furthermore, these patients also were predisposed to ventilator-associated lung injury. The combination of PCP-related lung injury and ventilator-associated lung injury might in part explain the high mortality observed in this cohort. As a strict protocol for ventilator management in patients with acute lung injury or ARDS was initiated at our institution in early 2003, the future evaluation of a similar cohort may help determine beneficial role for reduced VT ventilation in non-HIV-associated PCP.

Although CD4 counts have not been routinely measured in patients with PCP without HIV infection, among these patients, six had CD4 counts measured with a median of 70 cells/[micro]L. In another study by Mansharamani et al, (24) the median CD4 eount was 61 cells/[micro]L, and > 90% of patients had CD4 counts < 300 cells/[micro]L at the rime of non-HIV-related PCP diagnosis. Moreover, approximately 40% of the patients receiving long-term corticosteroid therapy deemed "at risk" for PCP had CD4 counts < 300 cells/[micro]L, while the CD4 levels in patients not considered at risk for PCP were similar to those in healthy subjects. (24) Thus, in immunosuppressed non-HIV patients, CD4 counts < 200 cells/[micro]L might also help identify individuals at particularly high risk for PCP, which in turn should help guide prophylaxis. However, it should be noted with a word of caution that occasional patients with hematologic malignancies and other non-HIV-related immune suppressed conditions have been observed to acquire PCP with entirely normal or even elevated CD4 levels.

In our study cohort, no patient received PCP prophylaxis, in spite of receiving a median PED of 40 mg for a median duration of 4.5 months. One fourth of our patients received [less than or equal to] 30 mg of the PED for [less than or equal to] 2 months. In the previous study by Yale and Limper (15) at our institution 10 years earlier, the median PED daily dose was 30 mg for a median duration of 12 weeks. One fourth of patients received as little as 16 mg or less of prednisone for 8 weeks. Even though our current study did not address the true incidence of PCP among immunosuppressed patients, it showed that 20 patients died in period of 6 years from a disease that was potentially preventable. With this in mind and the concerns that resistance to TMP/SMZ may emerge with increased reservoir for infection, it is prudent that the primary PCP prophylaxis is strongly considered for all immunosuppressed non-HIV patients, particularly those receiving > 16 mg/d prednisone for > 2 months, those individuals receiving cytotoxic chemotherapy, and those treated with other immunosuppressive regimens. (15,25)

Comparing the immunosuppressive conditions associated with non-HIV PCP in the current series to the earlier study reported from our institution reveals that there has been a remarkable drop in the incidence of PCP among transplant recipients. The previous study by Yale and Limper (15) reported that PCP developed in 25% of that series in the setting of solid-organ or bone marrow transplantation, which represented the second most common condition associated with PCP in settings other than AIDS. Interestingly, however, only 8% of the current series of individuals with PCP and ARF occurred in patients with organ transplantation (one patient). The incidence of PCP among all other predisposing conditions remained proportionally unchanged over the intervals of these two studies. The decrease in the incidence of PCP with ARF was attributed to a routine PCP prophylaxis among organ transplant recipients in more recent years. (7)

After learning the results of this study, performance-improvement protocol was implemented at our institution with a target of 100% PCP prophylaxis among all immunosuppressed non-HIV patients receiving prolonged systemic corticosteroid therapy, and/or cytotoxic chemotherapy. Moreover, early intubation and institution of mechanical ventilation on admission to the ICU is now recommended for patients with ARF due to non-HIV-related PCP.

Potential confounding features of this study should also be noted. The majority of these are inherent within the retrospective design employed in this study. Although virtually all of the reported data were charted prospectively as part of the documentation requirements for clinical care and our quality assurance programs (APACHE III database), it remains possible that some important variables may not have been recorded. However, we believe this to be unlikely, since the data set analyzed for each significant prognostic factors was entirely complete (100% data availability for APACHE III scores, duration of PPV and pneumothorax) or nearly complete (97% of data regarding the timing of intubation was available). Thus, only the secondary analyses might have been affected by missing data, but we also believe this to be not likely. In addition, with only 30 patients in this cohort, sufficient power might not have been present to demonstrate other potential relationships in clinical features and mortality.

In conclusion, our data suggest that ARF due to non-HIV-related PCP remains a serious illness with high mortality. Among our patients, poor prognostic features include elevated APACHE III scores, intubation delay, and longer duration of PPV, as well as the development of pneumothorax. The delay in intubation on admission to the ICU is associated with higher hospital mortality rate even after adjusting for severity of illness. These findings in non-HIV-related PCP patients are consistent with similar findings in HIV/AIDS-assoeiated PCP documented in the literature. All patients in our cohort failed to receive PCP prophylaxis, which should be strongly considered for all non-HIV immunosuppressed patients receiving prolonged systemic corticosteroid therapy, and/or cytotoxic chemotherapy.

REFERENCES

(1) Kovacs JA, Hiemenz JW, Macher AM, et al. Pneumocystis carinii pneumonia: a comparison between patients with the acquired immunodeficiency syndrome and patients with other immunodeficiencies. Ann Intern Med 1984; 100:663-671

(2) Safrin S. Pneumocystis carinii pneumonia in patients with the acquired immunodeficiency syndrome. Semin Respir Infect 1993; 8:96-103

(3) HIV/AIDS surveillance supplemental report. Vol. 9. No. 3. 2003 Atlanta, GA: Centers for Disease Control and Prevention

(4) Morris A. Wachter RM, Luce J, et al. Improved survival with highly active antiretroviral therapy in HIV-infected patients with severe Pneumocystis carinii pneumonia. AIDS 2003; 17:73-80

(5) Mansharamani NG, Garland R, Delaney D, et al. Management and outcome patterns for adult Pneumocystis carinii pneumonia, 1985 to 1995: comparison of HIV-associated cases to other immunocompromised states. Chest 2000; 118:704-711

(6) Sepkowitz KA. Pneumocystis carinii pneumonia in patients without AIDS. Clin Infect Dis 1993; 17(Suppl 2):S416-S422

(7) Sepkowitz KA, Brown AE, Armstrong D. Pneumocystis carinii pneumonia without acquired immunodeficiency syndrome: more patients, same risk. Arch Intern Med 1995; 155:1125-1128

(8) Limper AH, Offord KP, Smith TF, et al. Pneumocystis carinii pneumonia: differences in lung parasite number and inflammation in patients with and without AIDS. Am Rev Respir Dis 1989; 140:1204-1209

(9) Thomas CF Jr, Limper AH. Pneumocystis pneumonia: clinical presentation and diagnosis in patients with and without acquired immune deficiency syndrome. Semin Respir Infect 1998; 13:289-295

(10) Jacobs JA, Dieleman MM, Cornelissen EI, et al. Bronchoalveolar lavage fluid cytology in patients with Pneumocystis carinii pneumonia. Acta Cytol 2001; 45:317-326

(11) Delclaux C, Zahar JR, Amraoui G, et al. Corticosteroids as adjunctive therapy for severe Pneumocystis carinii pneumonia in non-human immunodeficiency virus-infected patients: retrospective study of 31 patients. Clin Infect Dis 1999; 29:670-672

(12) Pareja JG, Garland R, Koziel H. Use of adjunctive corticosteroids in severe adult non-HIV Pneumocystis carinii pneumonia. Chest 1998; 113:1215-1224

(13) Forrest DM, Djurdjev O, Zala C, et al. Validation of the modified multisystem organ failure score as a predictor of mortality in patients with AIDS-related Pneumocystis carinii pneumonia and respiratory failure. Chest 1998; 114:199-206

(14) Forrest DM, Zala C, Djurdjev O, et al. Determinants of short- and long-term outcome in patients with respiratory failure caused by AIDS-related Pneumocystis carinii pneumonia. Arch Intern Med 1999; 159:741-747

(15) Yale SH, Limper AH. Pneumocystis carinii pneumonia in patients without acquired immunodeficiency syndrome: associated illness and prior corticosteroid therapy. Mayo Clin Proc 1996; 71:5-13

(16) Haynes RC Jr, Murad F. Adrenocorticotropic hormone; adrenocortical steroids and their synthetic analogs; inhibitors of adrenocortical steroid biosynthesis. In: Gilman AG, Goodman LS, Gilman A, eds. Goodman and Gilman's the pharmacologic basis of therapeutics. 6th ed. New York, NY: Macmillan, 1980; 1466-1496

(17) Zimmerman JE, Wagner DP, Draper EA, et al. Evaluation of APACHE III predictions of hospital mortality in an independent database. Crit Care Med 1998; 26:1317-1326

(18) Sirio CA, Shepardson LB, Rotondi AJ, et al. Community-wide assessment of intensive care outcomes using a physiologically based prognostic measure. Chest 1999; 115:793-801

(19) Cook DA. Performance of APACHE III models in an Australian ICU. Chest 2000; 118:1732-1738

(20) Bedos JP, Dumoulin JL, Gachot B, et al. Pneumocystis carinii pneumonia requiring intensive care management: survival and prognostic study in 110 patients with human immunodeficiency virus. Crit Care Med 1999; 27:1109-1115

(21) Wachter RM, Luce JM, Safrin S, et al. Cost and outcome of intensive care for patients with AIDS, Pneumocystis carinii pneumonia, and severe respiratory failure. JAMA 1995; 273: 230-235

(22) Murry CE, Schmidt RA. Tissue invasion by Pneumocystis carinii: a possible cause of cavitary pneumonia and pneumothorax. Hum Pathol 1992; 23:1380-1387

(23) The Acute Respiratory Distress Syndrome Network. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. N Engl J Med 2000; 342: 1301-1308

(24) Mansharamani NG, Balachandran D, Vernovsky I, et al. Peripheral blood CD4 + T-lymphocyte counts during Pneumocystis carinii pneumonia in immunocompromised patients without HIV infection. Chest 2000; 118:712-720

(25) Thomas CF Jr, Limper AH. Pneumocystis pneumonia. N Engl J Med 2004; 350:2487-2498

* From the Divisions of Pulmonary and Critical Care Medicine, Departments of Internal Medicine Mayo Clinic Rochester, MN (Drs. Gajic, Limper and Aksamit); and Mayo Clinic, Jacksonville, FL (Dr. Festic).

Financial support was provided by Mayo Clinic and Mayo Foundation.

Manuscript received October 11, 2004; revision accepted January 18, 2005.

Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (www.chestjournal. org/misc/reprints.shtml).

Correspondence to: Emir Festic, MD, Mayo Clinic 4500 San Pablo Rd, Jacksonville, FL 32224; e-mail: festic.emir@mayo.edu

COPYRIGHT 2005 American College of Chest Physicians
COPYRIGHT 2005 Gale Group