Correction: Association of Adrenal Function and Disease Severity in Community-Acquired Pneumonia

due to spaces in the text being lost during the typesetting process. The publisher apologizes for these errors. Please download the PDF again to view the corrected article. The originally published, uncorrected article and the republished, corrected article are provided here for reference. Copyright: ß 2014 The PLOS ONE Staff. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.


Introduction
Community-acquired pneumonia (CAP) is the mostcommon sepsis-defining illness, and its treatment and prognosis hasnot changed since the 1950ies, while in-hospital mortality remainedabout 12% [1][2][3]. To challengeits high mortality and morbidity and for appropriate allocation of healthcareresources, rapid and early risk stratification in CAP is necessary. An acceptedscoring system for severity assessment in CAP is the pneumonia severity index (PSI) 4]. However, its complexity is high, jeopardizing itsdissemination and implementation in everyday's practice.
One of the first measurable physiopathologic response to infectionis the activation of the hypothalamo-pituitary-adrenal (HPA)axis via stimulation of the central noradrenergic stress system by cytokinesand other mediators released upon inflammation [9].In this context, we have previously shown that free and total cortisol arepredictive parameters for outcome in CAP [10].
Other surrogates of adrenal function like dehydroepiandrosterone (DHEA)also change upon inflammation [7]. Under healthycondition, DHEA secretion is synchronized with cortisol in response to adrenocorticotropichormone (ACTH) and corticotropin-releasing hormone (CRH),while in critical illness, a remarkable decrease of DHEA and DHEA-Sulfate (DHEAS)secretion has been observed [11]. The physiologicalrole of DHEA and DHEAS is poorly understood, but there are clear indicationsthat it modulates the immune response and influences both proandanti-inflammatory cytokine release [8,[11][12][13][14].
In sepsis, cortisol to DHEA ratio was put forward as a prognosticmarker [15]. Arlt et al.observed an increased cortisol to DHEA ratio in the most severely ill as comparedto the least ill patients. Both cortisol and cortisol to DHEA ratio were ofa similar prognostic accuracy [15].
In CAP, only one study evaluated cortisol and DHEA/-Slevels for outcome prediction, but only with a relatively small sample size (n = 58),and without including assessment of CAP severity [7].We herein evaluated the prognostic value of cortisol, DHEA and DHEAS and ofcortisol/DHEA-, cortisol/DHEAS-and DHEA/ DHEAS-ratios,respectively, compared to the PSI in hospitalized patients with CAP.

Study Design and Setting
For the purpose of this study, we measured adrenal functionparameters in remaining plasma samples of a prospective randomized trial inpatients with CAP that was conducted from November 2003 through February 2005at the University Hospital of Basel, a tertiary care hospital in Switzerland.The design of the original study has been described elsewhere [6] (ISRCTN04176397).The primary aim of the original study was to evaluate antibiotic durationby procalcitonin guidance as compared to standard recommended guidelines.A predefined secondary endpoint was the prognostic value of adrenal functionin CAP in comparison to the PSI. Adrenal function at study entry was assessedin 179 patients.

Participants
Adult patients (.18 years) with CAP as their principaldiagnosis on admission were eligible for the study. CAP was defined as a newinfiltrate on chest x-ray accompanied by one or several acquired respiratorysymptoms and signs (cough, sputum, dyspnea, temperature of 38uC ormore, abnormal auscultatory findings, leukocytes .10 G/l or ,4 G/l)and the absence of a hospital stay within 14 days of admission. Exclusioncriteria were cystic fibrosis, active tuberculosis, HIV infection with a CD4-cellcount below 200, neutropenia of less than 500 G/l, chemotherapywith neutropenia between 500-100 G/l and an expected decreasebelow 500 G/l, and immunosuppression after organ transplantation.Baseline assessment included clinical data, vital signs, comorbidities, routineblood tests and the PSI. A follow-up telephone interview was performedafter 6 weeks.

Ethics Statement
Ethical approval was obtained from the local ethics committee (ethicscommittee Basel EKBB), and all included participants or their legal representativesgave written informed consent before inclusion into the study. This studyadhered to the consolidate STROBE standards for the reporting of observationaltrials [16].

Measurement of serum DHEA, DHEAS and cortisol
Serum samples were collected on admission and batchmeasured.DHEA was analyzed at Bioanalytica AG, Luzern, Switzerland, by a routinelyavailable ELISA-assay (IBL Interna-

Statistical analysis
Calculations were made with STATA 12.1 (StataCorp LP,College Station, TX, USA) or Graph Pad Prism 5.0 (GraphPad Software,Inc., La Jolla, CA, USA). Correlation of hormones and PSI was calculated by Spearmanrank correlation. For multi-group comparisons, Kruskal-Wallisanalysis was performed. We calculated a logistic regression model adjustedfor age and gender to assess associations of hormones and mortality after6 weeks. Receiver operating characteristics (ROC) were performedand area under the curve (AUC) was calculated to assess the overallpredictive accuracy of hormones. Finally, for graphical display, we also assessedtime to death in Kaplan Meier curves; and used log rank tests to comparegroups stratified by the highest decile of hormone levels.

Discussion
Our findings show that cortisol, DHEAS and their differentratios significantly correlate with CAP severity, while DHEA and cortisolare predictive for mortality. Thus, not only cortisol, but adrenal functionin general and its possible intra-adrenal shift from DHEAS to cortisolproduction in severe pneumonia is an important factor for outcome and survivalin CAP.
The finding that high cortisol levels predict bad outcome andmortality confirms previous data [6]. Serum cortisolis a marker of stress and shows the degree of the activation of the HPA axisthereby reflecting the severity of illness, with a gradual increase of cortisollevels at a greater degree of illness.
In patients with a high severity of CAP, we found not onlyhigher cortisol levels, but also lower levels of DHEAS compared to patientswith a low CAP severity. Furthermore, in non-survivors, cortisol andDHEA levels were higher as compared to survivors.
A similar dissociation between cortisol and DHEAS levels hasbeen observed in sepsis and in stroke [5,8,15].This discrepancy between low levels of DHEAS and high levels of cortisol issurprising as both hormones are mainly secreted by the adrenal cortex. Thismay indicate an intra-adrenal shift from DHEAS towards the potentiallylife-saving cortisol production during critical illness [11].
Low DHEAS levels were not only proposed as predictive markerfor stroke and sepsis but were an even more sensitive marker than cortisolfor the diagnosis of so-called relative adrenal insufficiency 5], today called critical illness-related corticosteroidinsufficiency (CIRCI). Further studies are required to investigatethis question.
The pathophysiological mechanism why DHEAS is low in the presenceof high cortisol in severely ill patients is currently unknown. A recent studyfound a biological activity for DHEAS, but not for DHEA, being able to enhancethe activity of human neutrophils. Therefore, diminished levels of DHEAS couldhave adverse effects, especially in relation to susceptibility to bacterialinfection [17]. It yet has to be established towhat extent these endocrine changes impact or counteract the immunosuppressiverole of cortisol.
In addition, we observed not only a dissociation between cortisoland DHEAS but also between cortisol and DHEA levels in patients with progressiveCAP severity. Arlt et al. suggested the cortisol/DHEAratio as a novel prognostic marker in septic shock, as they found a similar,significant dissociation between cortisol and DHEA in non-survivorsand between the least ill and severest ill patients despite increased DHEAlevels [15].
In our study we were able to confirm these data in CAP, aswe found an elevated cortisol to DHEA ratio within increasing severity ofCAP. Although this ratio missed significance for mortality prediction, therewas a tendency for a higher cortisol to DHEA ratio. The mechanism for thisimbalance of adrenal hormone synthesis is unknown.
So far, only one study evaluated DHEA and DHEAS levels in CAP 7]. Thereby, Kolditz et al.found elevated levels of cortisol, DHEA and an increased DHEA/DHEAS ratioin patients who were clinically instable 72 hours after enrolment,but only cortisol was predictive for mortality [7].Our data support these results. We additionally show that adrenal functioncorrelates with disease severity, and that DHEAS levels decrease with progressiveCAP severity, while levels of cortisol increase. Furthermore, we show thatDHEA is prognostic for mortality in CAP.

Limitations
There are limitations to our study. First, it is a secondaryanalysis of remaining blood samples. However, we analyzed a predefined secondaryendpoint in prospectively randomized trial patients. Second, DHEA/S levelswere measured at the time of presentation, that is, at different times duringthe day, not taking into account the fact that DHEA/S-levels showa certain diurnal and individual variability [18,19].However, a standardized DHEA/S level at the same time of day in all patientswould most probably has had a higher prognostic accuracy. Third, DHEAS issecreted in an age-dependent fashion, with maximum levels during thethird decade and very low levels in old age. Therefore, we included age andgender in a multivariate model. Another way to adjust for age and gender isthe method of Kolditz et al., which for our dataprovided only little differences in Odds Ratios [7].Fourth, a correlation with cytokines to support our findings has not beenperformed.

Conclusion
In conclusion, this study found that adrenal function measuredby DHEAS and cortisol is prognostic for severity in CAP, while DHEA and cortisolare predictors for mortality.