Published ahead of print on November 19, 2004, doi:10.1164/rccm.200406-808OC
© 2005 American Thoracic Society doi: 10.1164/rccm.200406-808OC
Hydrocortisone Infusion for Severe Community-acquired PneumoniaA Preliminary Randomized StudyOspedale di Trieste, Trieste; Ospedale Gradenigo-Torino; Ospedale Molinette di Torino; Arcispedale S. Anna di Ferrara; Ospedale di Crema; Ospedale di Paderno Dugnano-Milano; Institute of Respiratory Diseases, University of Milan, Milan, Italy; and Memphis Lung Research Program, Memphis, Tennessee Correspondence and requests for reprints should be addressed to Dr. Marco Confalonieri, Azienda Ospedaliero-Universitaria di Trieste, Strada di Fiume 447, 34100 Trieste, Italy. E-mail: marco.confalonieri{at}aots.sanita.fvg.it
We hypothesize that hydrocortisone infusion in severe community-acquired pneumonia attenuates systemic inflammation and leads to earlier resolution of pneumonia and a reduction in sepsis-related complications. In a multicenter trial, patients admitted to the Intensive Care Unit (ICU) with severe community-acquired pneumonia received protocol-guided antibiotic treatment and were randomly assigned to hydrocortisone infusion or placebo. Hydrocortisone was given as an intravenous 200-mg bolus followed by infusion at a rate of 10 mg/hour for 7 days. Primary end-points of the study were improvement in PaO2:FIO2 (PaO2:FIO2 > 300 or  100 increase from study entry) and multiple organ dysfunction syndrome (MODS) score by Study Day 8 and reduction in delayed septic shock. Forty-six patients entered the study. At study entry, the hydrocortisone group had lower PaO2:FIO2, and higher chest radiograph score and C-reactive protein level. By Study Day 8, treated patients had, compared with control subjects, a significant improvement in PaO2:FIO2 (p = 0.002) and chest radiograph score (p < 0.0001), and a significant reduction in C-reactive protein levels (p = 0.01), MODS score (p = 0.003), and delayed septic shock (p = 0.001). Hydrocortisone treatment was associated with a significant reduction in length of hospital stay (p = 0.03) and mortality (p = 0.009).
Key Words: community-acquired pneumonia • C-reactive protein • hydrocortisone • respiratory failure • severe sepsis Pneumonia is the leading cause of community-acquired infection requiring intensive care unit (ICU) admission (1). Despite advances in antimicrobial therapy and supportive measures, mortality for patients with severe community-acquired pneumonia admitted to the ICU remains high—22 to 54% (2). Between 58 and 87% of patients with severe community-acquired pneumonia admitted to the ICU develop respiratory failure and require mechanical ventilation, a factor associated with a higher mortality (2). Irrespective of severity of initial presentation, the development of sepsis-related complications (delayed septic shock, adult respiratory distress syndrome [ARDS], and extrapulmonary organ dysfunction) during ICU stay is associated with a significantly higher ICU mortality (57–100%) (3). Several studies have shown increased pulmonary and circulating inflammatory cytokine levels in patients with severe community-acquired pneumonia (4–7). Among patients admitted to the ICU, higher circulating inflammatory cytokine levels correlated with the presence of bilateral pneumonia (6), bacteremia (7), need for mechanical ventilation (7), and higher Acute Physiology and Chronic Health Evaluation (APACHE) II and multiple organ dysfunction syndrome (MODS) scores (4, 6). Among patients with severe community-acquired pneumonia, nonsurvivors, unlike survivors, exhibit persistent elevation of plasma interleukin (IL)-6 levels over time (7). Taken as a whole, the findings of these studies (4–7) indicate that, similar to patients with ARDS (8), degree and duration of the systemic inflammation have a strong effect on final outcome in patients with severe community-acquired pneumonia. Glucocorticoids, the most important natural inhibitors of inflammation, are not always effective in suppressing life-threatening systemic inflammation. The presence of systemic inflammation-induced tissue resistance to glucocorticoids and/or inadequate adrenal output might explain why older clinical trials found no efficacy with a time-limited course of massive doses of glucocorticoids (9, 10), while recent randomized studies have shown efficacy and safety with prolonged glucocorticoid treatment in low to moderate doses in patients with catecholamine-dependent septic shock (11–16), severe pneumocystis pneumonia (12), and unresolving ARDS (13). In these studies, patients randomized to prolonged glucocorticoid treatment, in contrast to control subjects, had a significant reduction in circulating inflammatory cytokine levels over time (14–16). Little information is presently available on the effect of prolonged glucocorticoid administration in patients with severe community-acquired pneumonia presenting without septic shock. In a recent retrospective study, Monton and coworkers (17) reported that, among patients with severe community-acquired pneumonia requiring mechanical ventilation, those who received methylprednisolone for 9 ± 7 days (most for bronchodilation) had an attenuated systemic and pulmonary inflammatory response and trended toward lower mortality (36 versus 67%; p = 0.37). We hypothesize that hydrocortisone administration initiated early in the course of severe community-acquired pneumonia attenuates pulmonary and systemic inflammation and leads to earlier resolution of pneumonia and a reduction in sepsis-related complications and mortality. For these reasons, we evaluated the efficacy and safety of prolonged hydrocortisone infusion in a prospective, randomized, double blind, placebo-controlled trial of patients with severe community-acquired pneumonia admitted to the ICU. Some of the results of this study have been previously reported in the form of an abstract (18).
Patients From July 2000 through March 2003, 46 eligible patients were enrolled in this randomized, double-blind, placebo-controlled trial, which was conducted at six hospitals in Italy. Patients were admitted to the Intensive Care Unit or Respiratory Intermediate Unit (RICU) of the following hospitals: Ospedale di Trieste, Ospedale Gradenigo (Torino), Ospedale Molinette (Torino), Arcispedale S. Anna (Ferrara), Ospedale di Crema, or Ospedale di Paderno Dugnano (Milano). A regional institutional review board approved the protocol, and written informed consent was obtained from all participants or their authorized representatives. The clinical coordinator was available throughout the study to answer investigators' questions regarding patients' eligibility and safety and the reporting of serious adverse events.
Selection Criteria Exclusion criteria included any of the following: (1) nosocomial pneumonia; (2) severe immunosuppression; (3) acute burn injury; (4) a preexisting medical condition with a life expectancy less than 3 months; (5) pregnancy; (6) a major gastrointestinal bleed within 3 months of the current hospitalization; or (7) a condition requiring more than 0.5 mg/kg/day of prednisone equivalent (i.e., acute asthma or chronic obstructive pulmonary disease [COPD]). Exit criteria included (1) active gastrointestinal bleeding event requiring transfusion of five units packed red blood cells (PRBC); (2) recovery of Candida spp. from multiple sites; or (3) development of a condition requiring prolonged glucocorticoid administration such as exacerbation of COPD or asthma.
Treatment Assignments
Study Aims We adopted the simplified MODS score (up to 6 points) proposed by Bernard and associates (21). The number of ventilator-free days was defined as the number of days after ventilation was discontinued up to Study Day 8. Shock was defined as requirement of vasopressors. ARDS was defined by consensus criteria (22); method of BAL, laboratory processing, and diagnostic criteria followed consensus guidelines (23). The threshold for diagnosing pneumonia by quantitative bacterial culture of the BAL was 104 cfu/ml (23).
Statistical Analysis
Data were analyzed using standard methods. In all cases, statistical significance was defined as a two-tailed test with an
After an interim analysis of data from 46 patients, enrollment was suspended because a significant difference was identified between the two groups for improvement in PaO2:FIO2 by Study Day 8 (p = 0.002) and hospital mortality (p = 0.009). Table E1 in the online supplement provides outcome data per each center. Unless specified, data are reported as placebo versus hydrocortisone groups. The two groups met similar study entry criteria: requirement of mechanical ventilation (19 versus 15), requirement of vasopressors for more than 4 hours (1 versus 2), respiratory rate greater than 30 breaths per minute (18 versus 17), PaO2:FIO2 < 250 (22 versus 23), chest radiograph showing multilobar involvement (23 versus 23). Nine patients (4 versus 5) met two minor criteria and no major criteria. Figure 1 shows progress through the phases of the trial. Two patients exited the study shortly after randomization when it was discovered that they met exclusion criteria. They did not receive the assigned treatment and were not included in the analysis.
Table 1 shows the clinical and physiological characteristics at study entry. Among the 34 patients requiring mechanical ventilation, noninvasive positive pressure ventilation (NPPV) was applied more frequently to those randomized to hydrocortisone (3 versus 8; p = 0.03). There were 33 patients with comorbidities (20 versus 13; p = 0.08). No significant difference was observed for type of underlying comorbidities (placebo versus hydrocortisone group): hypertension (8 versus 4), ischemic heart disease (6 versus 4) diabetes mellitus (5 versus 3), alcohol abuse (4 versus 1), chronic liver disease (3 versus 1), COPD (1 versus 2), chronic renal insufficiency (0 versus 2), and others (5 versus 1).
Information on identified pathogens and diagnostic methods is available in the online supplement. Patients received combination therapy in agreement with the 1993 ATS guidelines (19). Initial antibiotic classes and changes in antibiotics are available in the online supplement. At study entry, serum C-reactive protein (CRP) was higher in patients randomized to hydrocortisone (Table 1). On Study Day 8, a greater than 50% reduction in CRP from study entry was observed in all but two patients in the hydrocortisone group and in only five patients in the control group (p < 0.0001). By Study Day 8, the 20 patients with persistent elevation in CRP levels had a higher incidence of delayed septic shock (9 versus 0; p = 0.47). After randomization (Figure 2, top), a progressive reduction in CRP values was seen in the hydrocortisone group; a significant difference from the control group was seen on Study Day 8 (Table 2).
During the 7 days of treatment, the clinical response of the two groups diverged. In the control group, the single patient admitted with septic shock died by Study Day 3, whereas nine patients developed delayed septic shock that was complicated by ARDS in four. One additional control patient developed delayed septic shock after Day 8. In contrast, in the hydrocortisone group, the two patients admitted with septic shock were weaned off vasopressors by Days 3 and 5, and no other patient developed sepsis-related complications. Hydrocortisone treatment appeared to be associated with a protective effect against delayed septic shock (odds ratio = 0.03, confidence interval 0.0015–0.51; p = 0.0005, relative risk = 0.05).
At baseline, the hydrocortisone-treated group had a significantly lower PaO2:FIO2 (Table 1). As shown in Figure 2, a significant increase in PaO2:FIO2 from study entry was observed within the hydrocortisone group by Day 1 (p < 0.001), and within the placebo group by Day 5 (p = 0.03). Additional information is available in the Web repository. On Study Day 8, a PaO2:FIO2 Major complications occurring after study entry are shown in Table 3. Four control patients developed ARDS, one died by Day 7; two of the three survivors received prolonged methylprednisolone treatment for unresolving ARDS (13). Figure 3 (top) shows the survival curves of patients randomized to hydrocortisone and placebo. Survival to hospital discharge was 70% in the placebo group and 100% in the hydrocortisone group (p = 0.009). All nonsurvivors died before Study Day 28. Deaths were reported in four of the six participating centers (Table E1). All nonsurvivors required conventional mechanical ventilation and all had progression of radiographic densities on chest radiograph by Day 8. All but one patients developed pressor-dependent shock. The one patient without shock had Legionella pneumonia that progressed to ARDS. Causes of death included septic shock (4), ARDS (1), hypoxemic respiratory failure (1), and MODS (1).
All nonsurvivors received a macrolide antibiotic in combination with a third-generation cephalosporin (4), a quinolone (3), or an antipseudomonal penicillin (1). Three patients had a change in antibiotic for a specific laboratory finding (addition of vancomycin for Staphylococcus aureus in a patient with polymicrobial pneumonia and addition of rifampin in two patients with positive Legionella urinary antigen already receiving macrolide). Additional final outcome results are shown in Table 4. Time to removal of mechanical ventilation is shown in Figure 3 (bottom). After hospital discharge, one control patient was readmitted with recurrent pneumonia and died within 60 days of initial study entry.
This is the first randomized study evaluating the efficacy and safety of low-dose hydrocortisone infusion in patients with severe community-acquired pneumonia, the leading cause of community-acquired sepsis. The findings of this study support our original hypothesis that modulation of systemic inflammation with early introduction of prolonged low-dose glucocorticoid administration hastens resolution of pneumonia and prevents development of life-threatening sepsis-related complications. Patients randomized to a 7-day hydrocortisone infusion had a significant reduction in CRP levels over time and by Study Day 8 had, compared with control subjects, a significant improvement in PaO2:FIO2 and chest radiograph score, and a significant reduction in median CRP levels, MODS score, and incidence of delayed septic shock. Hydrocortisone treatment was well tolerated and led to a significant reduction in duration of mechanical ventilation, length of ICU/RIU and hospital stay, and increased survival to hospital discharge and to 60 days.
The role of glucocorticoid supplementation in patients with systemic inflammation is undergoing a cyclical reassessment stirred by a new understanding of the role of glucocorticoids in modulating inflammation and immunity (26) and the consistent positive results of recent randomized studies (10). It is now appreciated that at the cellular level, transcription factor nuclear factor-
We have recently shown that in patients with unresolving ARDS, normal PBL exposed to plasma samples collected during prolonged methylprednisolone treatment exhibited (1) a progressive increase in cytoplasmic binding of GR
The response to glucocorticoid treatment observed in our study and in recent randomized trials of patients with catecholamine-dependent septic shock (11, 15, 31–34) and severe pneumocystis pneumonia (12) contrasts with the negative findings of older trials. A careful assessment of the literature indicates that failure to detect a beneficial effect in older randomized studies evaluating a time-limited course of massive doses of glucocorticoids in sepsis and early ARDS was most likely related to the short duration of treatment (9, 10). Among a number of studies in support of this hypothesis (9), a randomized trial of patients with severe community-acquired pneumonia found that a single dose of hydrocortisone (10 mg/kg) before antibiotic administration had no effect on plasma TNF- Although dosage and duration of hydrocortisone in our study are similar to the one used in recent randomized studies in patients with septic shock, treatment was initiated much earlier in the course of sepsis in an attempt to hasten resolution of pneumonia and to prevent the development of sepsis-related complications. In this regard, treated patients improved gas exchange and radiologic findings more rapidly than control subjects, and none developed delayed septic shock or ARDS. The relationship between reduction in CRP values and absence of sepsis-related complications and death supports the notion that containment of systemic inflammation is an important priority in the management of patients with severe community-acquired pneumonia. The small number of patients in our study may have biased the estimate of the treatment effect on mortality, and a larger randomized trial is necessary to support the mortality findings of this trial. Study limitations are attributed primarily to the small sample size that resulted in a few imbalances among prognostic factors among treatment groups. These imbalances included unplanned differences in the number of patients receiving NPPV, and with co-morbidities (a trend), as well as an imbalance of treatment allocation at the various sites. Although there were an equal number of subject randomized to each treatment group, this distribution was not reflected at each center. The small sample limits our ability to control for these imbalances. Moreover, our study was designed in 1999 and did not include recent advancements made in the management of patients with severe sepsis or septic shock (36), including patients with severe pneumonia (37). In the recent study by Annane and collaborators (34), the 28-day mortality among patients with severe community-acquired pneumonia and septic shock was 45% and 65% for patients randomized to hydrocortisone (n = 47) and placebo (n = 54), respectively (odds ratio 0.44; 95% confidence interval 0.20–0.98; p = 0.044) (data graciously provided by Dr. D. Annane). In our study, however, only 1 of 11 control patients with septic shock received hydrocortisone. In a future randomized study, we recommend using a smaller block size to assure balanced treatment assignment among participation centers. In addition, it would be beneficial to use a dynamic allocation scheme that allows stratification among key prognostic factors such as initial mode of mechanical ventilation. The new trial should incorporate recent guidelines for the management of patients with severe sepsis and septic shock (36), evaluate adrenal reserve in patients with septic shock (34), record timing to initial antibiotic administration (37), and obtain serial measurements of circulating inflammatory cytokine levels over time. In conclusion, the finding of this small randomized study of a patients with severe community-acquired pneumonia support the original hypothesis that control of inflammation with prolonged low-dose hydrocortisone infusion hastens resolution of pneumonia and prevents the development of sepsis-related complications. In this study, a 7-day course of low-dose hydrocortisone infusion was associated with a significant reduction in duration of mechanical ventilation, hospital length of stay, and hospital mortality.
The authors acknowledge Drs. Amado Freire and David Armbruster for critical review of the manuscript.
This study was supported in part by the Assisi Foundation of Memphis. This manuscript has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org Conflict of Interest Statement: M.C. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; R.U. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; A.P. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; M.P. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; P.P. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; G.P. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; R.D.P. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; C.G. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; F.B. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; R.U. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; G.U.M. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form June 24, 2004; accepted in final form October 31, 2004
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
less than 250; (3) chest radiograph showing bilateral involvement or multilobar involvement; (4) systolic blood pressure less than 90 mm Hg; or (5) diastolic blood pressure less than 60 mm Hg. Major criteria included (1) requirement of mechanical ventilation; (2) increase in the size of opacities on chest radiograph of 50% or more at 48 hours; (3) requirement of vasopressors for more than 4 hours; or (4) serum creatinine 2 or more mg/dl. 
of 0.05. All statistical calculations were preformed using the SAS System for Windows (release 9.0; SAS Institute, Cary, NC). Additional information on the sequential design and all analyses is available in the online supplement. 
One patient blindly randomized to placebo met exclusion because he had hepatic cancer with a life expectancy less than 3 months. The patient did not receive placebo. Analyzed as randomized unless otherwise specified: 
One patient blindly randomized to placebo exited the study on Study Day 4 to receive hydrocortisone for refractory septic shock. Two patients blindly randomized to placebo received prolonged methylprednisolone treatment for unresolving ARDS after Study Day 8.
B (NF-