key: cord-0005789-oyhpwut7 authors: Oppert, Michael; Reinicke, Albrecht; Gräf, Klaus-Jürgen; Barckow, Detlef; Frei, Ulrich; Eckardt, Kai-Uwe title: Plasma cortisol levels before and during "low-dose" hydrocortisone therapy and their relationship to hemodynamic improvement in patients with septic shock date: 2000-11-18 journal: Intensive Care Med DOI: 10.1007/s001340000685 sha: 39ae2989aa863b09f46c44eaa86dbbd26d8c969b doc_id: 5789 cord_uid: oyhpwut7 Objectives: To compare cortisol levels during "low-dose" hydrocortisone therapy to basal and ACTH-stimulated endogenous levels and to assess whether clinical course and the need for catecholamines depend on cortisol levels and/or pretreatment adrenocortical responsiveness. Design and setting: Prospective observational study in a medical ICU of a university hospital. Patients: Twenty consecutive patients with septic shock and a cardiac index of 3.5 l/min or higher, started on "low-dose" hydrocortisone therapy (100 mg bolus, 10 mg/h for 7 days and subsequent tapering) within 72 h of the onset of shock. Measurements and results: Basal total and free plasma cortisol levels ranged from 203 to 2169 and from 17 to 372 nmol/l. In 11 patients cortisol production was considered "inadequate" because there was neither a response to ACTH of at least 200 nmol/l nor a baseline level of at least 1000 nmol/l. Following the initiation of hydrocortisone therapy total and free cortisol levels increased 4.2- and 8.5-fold to median levels of 3587 (interquartile range 2679–5220) and 1210 (interquartile range 750–1846) nmol/l on day 1, and thereafter declined to median levels of 1310 nmol/l and 345 nmol/l on day 7. Patients with "inadequate" steroid production could be weaned from vasopressor therapy significantly faster, although their plasma free cortisol concentrations during the hydrocortisone treatment period did not differ. Conclusions: (a) During proposed regimens of "low-dose" hydrocortisone therapy, initially achieved plasma cortisol concentrations considerably exceed basal and ACTH stimulated levels. (b) Cortisol concentrations decline subsequently, despite continuous application of a constant dose. (c) "Inadequate" endogenous steroid production appears to sensitize patients to the hemodynamic effects of a "therapeutic rise" in plasma cortisol levels. Abstract Objectives: To compare cortisol levels during ªlow-doseº hydrocortisone therapy to basal and ACTH-stimulated endogenous levels and to assess whether clinical course and the need for catecholamines depend on cortisol levels and/or pretreatment adrenocortical responsiveness. Design and setting: Prospective observational study in a medical ICU of a university hospital. Patients: Twenty consecutive patients with septic shock and a cardiac index of 3.5 l/min or higher, started on ªlow-doseº hydrocortisone therapy (100 mg bolus, 10 mg/h for 7 days and subsequent tapering) within 72 h of the onset of shock. Measurements and results: Basal total and free plasma cortisol levels ranged from 203 to 2169 and from 17 to 372 nmol/l. In 11 patients cortisol production was considered ªinadequateº because there was neither a response to ACTH of at least 200 nmol/l nor a baseline level of at least 1000 nmol/l. Following the initiation of hydrocortisone therapy total and free cortisol levels in-creased 4.2-and 8.5-fold to median levels of 3587 (interquartile range 2679±5220) and 1210 (interquartile range 750±1846) nmol/l on day 1, and thereafter declined to median levels of 1310 nmol/l and 345 nmol/l on day 7. Patients with ªinadequateº steroid production could be weaned from vasopressor therapy significantly faster, although their plasma free cortisol concentrations during the hydrocortisone treatment period did not differ. Conclusions: (a) During proposed regimens of ªlow-doseº hydrocortisone therapy, initially achieved plasma cortisol concentrations considerably exceed basal and ACTH stimulated levels. (b) Cortisol concentrations decline subsequently, despite continuous application of a constant dose. (c) ªInadequateº endogenous steroid production appears to sensitize patients to the hemodynamic effects of a ªtherapeutic riseº in plasma cortisol levels. adrenal insufficiency per se can lead to a high output circulatory failure that resembles septic shock [3] . In addition, animal experiments [4] and clinical observations [5, 6, 7, 8] have shown that diminished steroid levels during sepsis are associated with adverse prognosis. In experimental settings in humans the administration of hydrocortisone before or during endotoxin challenge significantly reduces the inflammatory response [9] . On the other hand, some studies have shown that mortality is higher with increased levels of cortisol in patients with sepsis and septic shock [10, 11] . A recent prospective cohort study in 189 patients with septic shock found that patients with high cortisol levels and a reduced response to adrenocorticotropic hormone (ACTH, corticotropin) have the worst outcome [12] . Several attempts to improve the outcome of septic patients with high doses of steroids (approx. 2±8 g methylprednisolone in 24 h) have failed [13, 14, 15, 16, 17] . One meta-analysis concluded that steroids at such doses may even be harmful since they appear to increase the mortality in patients with overwhelming infection [16] . In contrast, two open studies [18, 19] and two recent randomized controlled trials [20, 21] have shown ªlowdoseº steroid therapy to have beneficial effects on hemodynamics and outcome in patients with septic shock, using no more than 300 mg hydrocortisone daily administered either as bolus injections of 100 mg three times daily or as continuous infusion. This work greatly stimulated the interest in using corticosteroids for septic patients, but little is known about the cortisol concentrations achieved by such a ªlow-doseº regimen and their relationship to endogenous basal and ACTH-induced cortisol levels. In fact, a number of studies have considered similar moderate amounts of hydrocortisone as a ªsupraphysiologicalº [18, 20] , a ªphysiologicalº [19] , a ªreplacementº [22] , or a ªstressº dose [21] . Moreover, it remains uncertain whether the response to hydrocortisone depends on endogenous steroid secretion. While some investigators believe that hemodynamic improvement after moderate doses of hydrocortisone detects adrenal insufficiency [18] , a recent randomized trial by Bollaert et al. [20] suggested that the effectiveness of hydrocortisone is unrelated to adrenocortical function. To obtain further insight into the pharmacological and pathophysiological basis of the proposed ªlowdoseº steroid treatment in septic patients we performed a prospective observational study in 20 medical intensive care patients with septic shock receiving continuous hydrocortisone therapy (10 mg/h). Total and free cortisol levels were determined under basal conditions, after ACTH stimulation, and during the subsequent course of hydrocortisone therapy to compare the levels under therapy with the endogenous production and to assess whether clinical course and hemodynamic response dif-fer in patients with and without ªinadequateº endogenous production. Adult patients with septic shock were included who were being treated in the medical intensive care unit of the Virchow Klinikum. Demographic data, the focus of sepsis, and causative organisms are presented in Table 1 . According to the American College of Chest Physicians/Society of Critical Care Medicine Consensus Committee [1] septic shock was defined as sepsis with hypotension of 90 mmHg or less or a drop of 40 mmHg or more despite adequate fluid resuscitation along with the presence of perfusion abnormalities. Furthermore, patients were only included with evidence of a causative organism or focus of infection and a cardiac index 3.5 l min ±1 m ±2 or greater. Patients with pancreatitis, infection with human immunodeficiency virus, and those in whom withholding maximal therapy was considered were excluded. Informed consent was obtained from next of kin. All patients were monitored hemodynamically with an arterial, a central venous, and a Swan-Ganz pulmonary artery catheter. After inclusion in the study dopamine, when given in daily doses above 240 mg, was switched to noradrenaline and dobutamine when possible depending on cardiac output and peripheral vascular resistance in order to make the needed doses of vasopressors comparable. Tapering of catecholamines and fluid expansion was guided by hemodynamic and pulmonary function to optimize tissue perfusion and gas exchange. Vasopressor therapy was usually titrated to achieve a mean arterial pressure of 70 mmHg or higher and a cardiac index of 3 l min ±1 m ±2 . The pulmonary artery occlusion pressure was aimed to be between 15 and 18 mmHg. Within 72 h following the onset of septic shock a ªshortº corticotropin test was performed in all patients [23, 24] . For this plasma samples were drawn before and 30 min after the administration of 0.25 mg of 1,24-corticotropin (Synacthen; CIBA, Switzerland; corresponding to 25 IU ACTH) for measurement of basal and stimulated cortisol levels. Cortisol production was defined as ªadequateº when the baseline level was greater than 500 nmol/l (to convert values for cortisol to mg/dl, divide by 27.5) and the increase after ACTH was greater than 200 nmol/l, or the baseline level was already above 1000 nmol/l (see ªDiscussionº). Following the corticotropin test all patients were given a bolus injection of 100 mg hydrocortisone (Pharmacia & UpJohn, Germany), followed by a continuous infusion. This infusion was given at a rate of 10 mg/h for 7 days, was reduced to 6 mg/h on day 8, and then reduced by 2 mg/h per day until it was discontinued on day 10. Plasma samples for determination of plasma cortisol and transcortin levels were drawn daily between 6 and 8 a.m. Measurements were performed en bloc, and the results were therefore not available to the physicians treating the patients. Plasma cortisol was measured by solid-phase radioimmunoassay (Biermann, Bad Nauheim, Germany). Plasma transcortin was de-termined using a competition radioimmunoassay (Medgeniox Diagnostics, Fleurus, Belgium). The concentration of free cortisol was calculated from total cortisol and transcortin concentrations using the following formula: free cortisol (mmol/l) = (Z 2 +0.00128 C±Z) ±1/2 , with Z = 0.0167+0.182 (T±C) [C = total cortisol (mmol/l); T = transcortin (mmol/l)] [25] . For comparison of differences between groups the Mann-Whitney U test was used. Fisher's exact test was used to test for dependencies between groups. A p value less than 0.05 was considered significant. All statistics were computed using SPSS for Windows, version 7.0. Unless otherwise indicated values are presented as medians with the interquartile range in parenthesis. Basal and stimulated cortisol levels of the 20 patients studied are given in Table 1 . Both basal and stimulated cortisol levels varied considerably, ranging from 203 to 2169 and from 269 to 2253 nmol/l, respectively. Cortisol production was considered ªadequateº in 9 patients because their cortisol level increased by at least 200 nmol/l following synacthen (nos. 8, 13, 14, 17, 20) , which most investigators consider as a normal response [23] , or the pre-ACTH concentration was already above 1000 nmol/l (nos. 12, 15, 16, 19 ; see ªDis-cussionº). In the remaining 11 patients endogenous cortisol production was defined as ªinadequateº because there was neither a response to ACTH of at least 200 nmol/l nor a baseline concentration of more than 1000 nmol/l. In one female patient (no. 2) cortisol levels were strikingly low. She had no evidence, however, of preexisting adrenal insufficiency, and after recovery her endogenous basal cortisol level increased to 545 nmol/l (day 14). The median basal levels of total cortisol were 780 nmol/l (601±883) in patients with ªinadequateº and 1068 nmol/ l (706±1769) in patients with ªadequateº cortisol production (p < 0.05; Table 2 ). The concentrations of free cortisol, as derived from the determination of transcortin and total cortisol levels, were 263 nmol/l (154±381) and 104 (79±154) in both groups (p < 0.01) and thus amounted to about 24.6 % and 13.3 % of the total cortisol concentration. It should be noted that the formula does not include albumin, which also binds cortisol at a lower affinity. Although this may lead to a slight overestimation of the fraction of free cortisol, plasma albumin levels were low (2.85 g/dl, 2.62±3.53) and did not change significantly during the course of the study. Following ACTH administration the increment in median total cortisol level was 16.1 % and 5.1 %, respectively. Pa- tients in the two groups did not differ with respect to their demographic data, disease severity, as assessed by APACHE II score levels, organ dysfunction indices, the type of causative organisms, hemodynamics, or catecholamine doses. Plasma cortisol levels following hydrocortisone therapy The time course of total and free cortisol levels during the hydrocortisone treatment period in patients with and without ªadequateº endogenous cortisol production is illustrated in Fig. 1 Free cortisol levels that were calculated from total cortisol and transcortin concentrations rose 8.5-fold to a median of 1210 nmol/l (750±1846) following the initiation of hydrocortisone therapy and did not differ be-tween patients with and those without ªadequateº endogenous cortisol production (Fig. 1, lower panel) . Before the initiation of hydrocortisone therapy patients on renal replacement therapy had lower total and free cortisol levels [620 (545±803) and 95 (79±196) vs. 945 (853±1503) and 262 (129±337) nmol/l; p < 0.04). However, during hydrocortisone therapy cortisol levels did not differ between patients receiving and those not receiving hemodialysis or hemofiltration, with the exception of day 2, when free cortisol was even higher in patients on renal replacement therapy [1057 (802±1481) vs. 430 (290±924) nmol/l; p < 0.01]. Thus, as expected from the molecular size of cortisol, blood purification did not seem to contribute to differences or to the reduction in cortisol levels with prolonged therapy. Hemodynamics and outcome following hydrocortisone therapy In all but two patients catecholamines could be reduced during the observation period. Although systolic and mean arterial blood pressure blood did not differ between patients with and without ªinadequateº cortisol production (Table 3) , the rate of reduction in noradrenaline did differ. Figure 2 illustrates that patients with ªinadequateº cortisol production required 15 % (5±25 %) of the initial noradrenaline dose on day 2. In contrast, in patients with ªadequateº cortisol production noradrenaline could only be reduced by 50 % (21±88 %; p < 0.01). In addition, patients with ªinadequateº cortisol production were free of vasopressor sup- port significantly earlier than patients with ªadequateº cortisol production (Table 3 ). There was, however, no difference in survival or the course of inflammatory parameters between the two groups ( Table 3) . To identify patient characteristics other than endogenous cortisol production that determine hemodynamic improvement during hydrocortisone therapy patients were divided into two groups according to whether the catecholamine dose could be reduced by at least 70 % within 48 h. Fourteen of the 20 patients investigated ful-filled this criterion. As shown in Table 4 , the only detectable difference between patients with and without rapid weaning from catecholamines was their endogenous cortisol production; 11 of 14 patients with rapid hemodynamic improvement, but none of the 6 patients with continued high catecholamine dependence, showed inadequate endogenous cortisol production. Accordingly, free and total cortisol levels before the initiation of hydrocortisone therapy in patients with early improvement were significantly lower. During the course of corticosteroid therapy, however, cortisol levels were very similar in the two groups. There was also no difference between the groups with respect to patient demographics, the disease severity score, or the course of inflammatory parameters. The pretreatment cortisol status did not differ between survivors (n = 11) and nonsurvivors (n = 9). Respective levels for total cortisol were 800 (560±1662) and 883 nmol/l (700±955). Free cortisol levels also did not differ between the two groups [103 (79±376) vs. 187 (115±242) nmol/l]. The endogenous cortisol production was considered ªinadequateº in 6 of 11 survivors and in 5 of 9 nonsurvivors. This study confirms that endogenous plasma cortisol concentrations are increased in patients with septic shock, but that the degree of increase is highly variable. Only one of 20 patients had a post-ACTH cortisol plasma level less than 500 nmol/l, which supports the concept that absolute adrenocortical deficiency is rare in critically ill patients [5, 6, 7, 26, 27, 28, 29, 30, 31] . Whether increases in cortisol levels are appropriate for the severity of illness or reflect functional hypoadrenalism is less clear. The rapid ACTH test is widely used as a simple method to identify adrenocortical hyporesponsiveness, but controversy exists as to how diagnostic criteria should be derived from the basal cortisol level, the stimulated cortisol level or their difference [24] . Three [6, 7, 32] of four previous studies [6, 7, 28, 32] using a rapid ACTH test to investigate adrenocortical function in septic shock used an increment of at least 200±250 nmol/l to define ªadequateº responsiveness. However, in healthy controls, large unselected patient populations and also in patients with septic shock [7, 24] (Table 1) cortisol increments are inversely related to basal levels. As discussed previously [24, 29, 31, 33, 34] , it appears likely that disproportionately low increments from high basal cortisol concentrations are due to maximal endogenous ACTH and therefore not indicative of adrenocorticoid hyporesponsiveness. We have thus arbitrarily defined endogenous cortisol production as ªadequateº in patients not only when they responded significantly to ACTH, but also when their baseline level was above 1000 nmol/l. Using these criteria endogenous cortisol production appeared to be ªinadequateº in approximately one-half of our patients. It should be noted, however, that an elevation in plasma cortisol levels above 1000 nmol/l in septic patients may reflect not only increased production but could also indicate decreased hepatic cortisol clearance [35, 36] . Following cortisol therapy with a similar dose as in previous trials (340 mg/day) [18, 19, 20, 21 ], peak concentrations of total cortisol were almost threefold higher than the post-ACTH level in patients considered to have an ªadequateº endogenous response. These peak concentrations corresponded to a 6.2-fold and 9.1-fold increase in free cortisol concentrations in patients with and without ªadequateº endogenous production. Clearly therefore this dose must be considered as supraphysiological. Interestingly, although the median total cortisol level remained above 1000 nmol/l thereafter, the levels of both total and free cortisol declined progressively during the course of therapy despite the administration of a constant dose of 10 mg/h. A similar, albeit less pronounced reduction in cortisol levels during ªlow-doseº hydrocortisone therapy has also been reported by Briegel et al. [22] . However, in this study the dose was tapered after only 1±5 days. Therefore, in contrast to our observations, the decline could be attributed to the reduction in steroid dose. Even if one assumes that in the present investigation the administration of ACTH and the initial bolus injection of 100 mg contributed to the peak concentration on day 1, and that endogenous cortisol production was subsequently suppressed, these two effects cannot entirely explain the reduction in median level from 3587 to 1310 nmol/l between days 1 and 7. Moreover, it has been shown that cortisol production in septic shock is in fact not suppressible by steroids [34] . It is possible therefore that the metabolism of hydrocortisone changes during the treatment period. Earlier studies have shown that cortisol extraction from blood is decreased, and that its half-life is prolonged during septic shock [35, 36] . The observed decline in cortisol levels during the treatment period could reflect a reversal of these alterations. Our intention in this study was not to compare the efficacy of cortisol therapy to untreated controls. It is nevertheless noteworthy that inotropes could be reduced rather quickly after initiation of the hydrocortisone infusion. The reduction in catecholamine dose was comparable [21] or even faster than in previous studies using similar doses of cortisol [18, 20] , which may be related to the fact that we started therapy earlier during the septic course. In contrast to previous reports [6, 7, 28], mortality in patients with ªinadequateº endogenous cortisol was not higher (Table 2) , and survivors and nonsurvivors did not have significantly different steroid levels Fig. 2 Time course of noradrenaline needs in patients with and without adequate endogenous cortisol production (CP). Two days after the beginning of ªlow-doseº hydrocortisone therapy patients with inadequate endogenous production needed significantly less vasopressor support before hydrocortisone therapy. This observation is in accordance with the results of a recent randomized controlled trial by Bollaert et al. [20] and is compatible with the growing evidence that ªlow-doseº therapy reverses the putative adverse risk factor of inappropriate cortisol production [18, 19, 20, 21, 22] . Although patients with ªadequateº and ªinadequateº endogenous production were clinically indistinguishable at the onset of therapy (Table 2) , and did not have significantly different levels of free cortisol during hydrocortisone therapy (Fig. 1) , those with ªinadequateº endogenous response showed a significantly faster reduction in their vasopressor support (Fig. 2) . In addition, the only detectable difference between patients with and without rapid hemodynamic improvement was their basal cortisol status (Table 4) . Similarly Brie-gel et al. [18] found that 6 of 14 patients weaned from catecholamines within 48 h of hydrocortisone therapy had lower cortisol levels before steroid treatment. In addition, Annane et al. [32] reported that patients in septic shock with inadequate as compared to adequate cortisol production had lower baseline responsiveness to noradrenaline but showed a more marked improvement in vasopressor sensitivity in response to hydrocortisone bolus. It appears therefore that the greatest benefit of hydrocortisone therapy is achieved in patients with blunted endogenous production, although the exact mechanisms of this sensitization remain to be elucidated. In contrast to the hemodynamic changes the febrile response and the course of C-reactive protein and procalcitonin levels or white blood cell counts did not depend on the pretreatment cortisol production (Table 3) . Thus, although it has been suggested that hydrocortisone treatment reduces indicators of the acute-phase response [22] , these effects appear not to be directly related to the hemodynamic effects. We also found no association between the hemodynamic response and survival rates (Table 4 ), but larger controlled trials are necessary to exclude or verify effects on such outcome parameters. In conclusion, there are three main findings of this study. Firstly, during proposed regimes of ªlow-doseº hydrocortisone therapy plasma cortisol concentrations are achieved initially which considerably exceed basal and ACTH stimulated levels. Thus in order to achieve a substitution of deficient endogenous production, doses even lower than those used so far may be sufficient. Secondly, during continuous administration of hydrocortisone, cortisol levels decline, suggesting that changes in cortisol metabolism have a significant impact on its plasma concentration. Thirdly, ªinadequateº endogenous steroid production appears to sensitize patients to the hemodynamic effects of a rise in plasma cortisol levels. A 3-level prognostic classification in septic shock based on cortisol levels and cortisol response to corticotropin The effects of high-dose corticosteroids in patients with septic shock A controlled clinical trial of high-dose methylprednisolone in the treatment of severe sepsis and septic shock Effect of high-dose glucocorticoid therapy on mortality in patients with clinical signs of systemic sepsis Corticosteroid treatment for sepsis: a critical appraisal and metaanalysis of the literature Steroid controversy in sepsis and septic shock: a metaanalysis Contribution of cortisol deficiency to septic shock Abrupt hemodynamic improvement in late septic shock with physiological doses of glucocorticoids Stress doses of hydrocortisone reverse hyperdynamic septic shock: a prospective, randomized, double blind Low-dose hydrocortisone infusion attenuates the systemic inflammatory response syndrome Screening for adrenocortical insufficiency with cosyntropin Rapid adrenocorticotropic hormone test in practice Clinical use of unbound plasma cortisol as calculated from total cortisol and corticosteroid binding globulin Adrenocortical function: an indicator of severity of disease and survival in chronic critically ill patients Spectrum of serum cortisol response to ACTH in ICU patients Adrenal insufficiency occurring during septic shock: incidence, outcome, and relationship to peripheral cytokine levels Plasma cortisol levels in patients with septic shock A comparison of the adrenocortical response during septic shock and after complete recovery Adrenocortical function during septic shock Gajdos P (1998) impaired pressor sensitivity to noradrenaline in septic shock patients with and without impaired adrenal function reserve Adrenal insufficiency Hypercortisolism in septic shock is not suppressible by dexamethasone infusion Comparative studies on adrenal cortical function and cortisol metabolism in healthy adults and in patients with shock due to infection Secretion and metabolism of cortisol after injection of endotoxin