Low-dose hydrocortisone improves shock reversal and reduces cytokine levels in early hyperdynamic septic shock* Michael Oppert, MD; Ralf Schindler, MD; Claudia Husung; Katrin Offermann; Klaus-Jürgen Gräf, MD; Olaf Boenisch, MD; Detlef Barckow, MD; Ulrich Frei, MD; Kai-Uwe Eckardt, MD

Objectives: To investigate the effect of low-dose hydrocortisone on time to shock reversal, the cytokine profile, and its relation to adrenal function in patients with early septic shock. Design: Prospective, randomized, double-blind, single-center study. Setting: Medical intensive care unit of a university hospital. Patients: Forty-one consecutive patients with early hyperdynamic septic shock. Interventions: After inclusion and a short adrenocorticotropic hormone test, all patients were randomized to receive either low-dose hydrocortisone (50-mg bolus followed by a continuous infusion of 0.18 mg/kg body of weight/hr) or matching placebo. After shock reversal, the dose was reduced to 0.06 mg/kg/hr and afterward slowly tapered. Severity of illness was estimated using Acute Physiology and Chronic Health Evaluation II score and Sequential Organ Failure Assessment score. Measurements and Main Results: Time to cessation of vasopressor support (primary end point) was significantly shorter in

S

eptic shock is the most important cause of death in noncoronary intensive care units. Despite improvement in many aspects of supportive care, septic shock is still associated with a mortality rate !50% (1–3). Although our understanding of the pathophysiology of septic shock has improved over the past decades, many drugs used to improve organ function or survival failed to show a benefit (4, 5). Glucocorticoids have potent antiinflammatory effects and were the first drugs tested in large randomized clinical

*See also p. 2683. From the Department of Nephrology and Medical Intensive Care (MO, RS, CH, KO, OB, DB, UR) and Department of Gastroenterology (K-JG), Charite Universitätsmedizin Berlin, Humboldt University, Berlin, Germany; and Department of Nephrology and Hypertension (K-UE), University of Erlangen-Nuremberg, Erlangen, Germany. The authors have no financial interests to disclose. Copyright © 2005 by the Society of Critical Care Medicine and Lippincott Williams & Wilkins DOI: 10.1097/01.CCM.0000186370.78639.23

Crit Care Med 2005 Vol. 33, No. 11

hydrocortisone-treated patients compared with placebo (53 hrs vs. 120 hrs, p < .02). This effect was more profound in patients with impaired adrenal reserve. Irrespective of endogenous steroid production, cytokine production was reduced in the treatment group with lower plasma levels of interleukin-6 and a diminished ex vivo lipopolysaccharide-stimulated interleukin-1 and interleukin-6 production. Interleukin-10 levels were unaltered. Adverse events were not more frequent in the treatment group. Conclusions: Treatment with low-dose hydrocortisone accelerates shock reversal in early hyperdynamic septic shock. This was accompanied by reduced production of proinflammatory cytokines, suggesting both hemodynamic and immunomodulatory effects of steroid treatment. Hemodynamic improvement seemed to be related to endogenous cortisol levels, whereas immune effects appeared to be independent of adrenal reserve. (Crit Care Med 2005; 33:2457–2464) KEY WORDS: hydrocortisone; shock; septic; adrenal function; cytokine; hemodynamics; organ failure

trials in septic patients. However, several attempts using high-dose regimes (approximately 2– 8 g of methylprednisolone per day) failed to improve outcomes (6 – 8). A meta-analysis concluded that steroids given in such doses may even be harmful, as they appear to predispose patients to overwhelming infections (9). The findings that severe sepsis and septic shock may be associated with absolute or relative adrenal insufficiency (10) and steroid receptor resistance (11) renewed the interest in steroid treatment in septic patients. A novel approach of “low-dose” regimes (up to 300 mg of hydrocortisone [HC] per day) was found to have beneficial effects on hemodynamics and outcome (1, 3, 12). These effects have mainly been attributed to a sensitization of the vasculature to vasopressor agents (13). There is some evidence also that this low-dose regimen has effects on the immune response (14). In a recent crossover study, Keh et al. (15) found that patients under HC therapy showed an attenuated cytokine response. Interleukin (IL)-6 and IL-8 levels were decreased,

whereas human leukocyte antigen-DR expression was unchanged. An unresolved question is whether the indication for low-dose HC therapy should be based on endogenous cortisol levels and their responsiveness to adrenocorticotropic hormone (ACTH). The hemodynamic response to steroids was found to be more profound in patients with relative adrenal insufficiency than in patients with adequate adrenal reserve (3, 16). Also, patients with inadequate adrenal function have a worse outcome (17). Whether the potential immunomodulation by low-dose HC is related to endogenous adrenal function is yet unknown. To investigate the impact of endogenous steroid responsiveness as inferred from a short synacthen test on the hemodynamic and the cytokine response, we performed a single-center double-blinded trial in which patients in early hyperdynamic shock were randomized to receive low-dose HC or placebo within 24 hrs after the onset of septic shock. The purpose was to assess time to shock reversal and cytokine response as measured by 2457

circulating IL-6 and IL-10 levels. As circulating cytokines may be produced by immune and nonimmune cells, including smooth muscle cells and endothelial cells, they reflect a relatively broad response. To measure the response of immune cells more directly, we also investigated the ex vivo response of peripheral blood cells to lipopolysaccharide (LPS) stimulation.

PATIENTS AND METHODS Patients Adult patients treated at our intensive care unit were prospectively enrolled if they met the following criteria: a) two or more of the following: tachycardia !90 beats/min, temperature !38.5°C or " 36°C, leukocytosis of !12,000/nL or !10% immature cells, respiratory rate of !20 per minute, or mechanical ventilation; b) evidence or strong clinical suspicion of infection (polymorphs in normally sterile body fluid, except blood), positive culture or Gram-negative stain of a normally sterile body fluid, clinical focus of infection (e.g., purulent sputum or tracheal aspirate); c) arterial systolic blood pressure "90 mm Hg for !1 hr despite adequate fluid resuscitation as demonstrated by a central venous pressure of !10 mm Hg and/or a pulmonary artery occlusion pressure of !15 mm Hg; d) a cardiac index !3.5 L/min/m2; and e) need for vasopressor support. Time to inclusion after the onset of septic shock had to be "24 hrs. Etomidate was used in seven patients within the last 7 days before inclusion in the study. Four of these patients were later classified as nonresponders (see Results). Excluded were patients who were pregnant, were known to be HIV positive, were recipients of organ transplants, or had a contraindication to or formal indication for steroids. Furthermore, all patients were excluded who received glucocorticosteroids (including inhaled steroids) 4 wks before this episode. The protocol was approved by the local ethics committee, and written informed consent was obtained by the patient or a legal guardian and/or the next of kin.

Protocol Patients were randomly assigned to receive low-dose steroid treatment or matching placebo. Randomization was performed by the pharmacist with prepared nontransparent envelopes. After randomization, a 250-#g short synacthen test was performed. For this, blood was drawn at baseline and 60 mins after the injection of 250 #g of ACTH. Syringes were prepared by the pharmacist and contained either 500 mg of hydrocortisone in 50 mL of

2458

normal saline or saline only. After the short synacthen test, a bolus injection of 50 mg was given, followed by a continuous infusion of 0.18 mg/kg of body weight/hr, or the patient received an equal volume of placebo. After cessation of vasopressor support for !1 hr (shock reversal), the volume of the study drug was reduced corresponding to a dose of 0.06 mg of HC/kg of body weight/hr for 24 hrs. It was then further reduced by 0.02 mg/kg/hr every day. Patients were followed for 28 days. Conventional therapy consisted of fluid resuscitation with colloids or crystalloids to achieve a central venous pressure !10 mm Hg and a pulmonary artery occlusion pressure !15 mm Hg. Norepinephrine was titrated to achieve a mean arterial pressure of !70 mm Hg and a systemic vascular resistance of 700 –1000 dyne·sec·cm$5. In case inotropes were needed, dobutamine was added; dopamine was not used. Patients with acute renal failure needing renal replacement therapy were treated with continuous venovenous hemofiltration for the duration of septic shock. All patients were sedated using a regime including midazolam and fentanyl.

Analytical Methods Cytokines. For cytokine measurements, blood was drawn at time of inclusion and at 8:00 am on days 1, 3, 5, 7, and 14 after inclusion. For measurement of cytokine levels (IL-6 and IL-10), plasma samples were frozen at $20°C. For the ex vivo production, 500 #L of blood was added to 500 #L of saline, 50 units of heparin, and LPS (18). After incubation at 37°C for 24 hrs, the blood sample was frozen at $20°C. After thawing IL-6 and IL-10, plasma levels as well as ex vivo stimulated IL-1 and IL-6 production were determined by enzyme-linked immunosorbent assay technique using 96-well plates (Maxisorp, Nunc, Denmark). Primary and biotinylated antibodies against IL-1, IL-6, and IL-10 were purchased from R&D Systems (Wiesbaden, Germany). All cytokine measurements were determined after the end of the study en bloc. To also test the effect of exogenously added HC on IL-6 production, blood was drawn from eight healthy volunteers. Ex vivo stimulation and measurement of IL-6 production were performed as mentioned previously. HC was added to blood in different doses (2500 nmol/L and 10,000 nmol/L) and at different time points (with LPS and 3 hrs before LPS). C-reactive protein. Plasma concentrations of C-reactive protein were measured by an immunoturbidimetric method using reagents from Greiner (Flacht Germany), adapted to a Dax 72 (Bayer, München, Germany) or a Hitachi 911 (Boehringer, Mannheim, Germany) random access clinical chemistry analyzer. The normal value is "0.6 mg/dL. Cortisol. Plasma cortisol was measured by a solid phase radioimmunoassay, wherein 125Ilabeled cortisol competes for binding with the

cortisol in the patient’s sample for a fixed time (Biermann, Bad Nauheim, Germany).

Statistics: The primary end point was time to cessation of vasopressor support. Secondary end points were cytokine response, 28-day survival, and the Sequential Organ Failure Assessment (SOFA) score. The statistical calculation was performed in an intent-to-treat analysis. Kaplan Meier curves were calculated for the time to cessation of vasopressor support (censored for subjects who died). To compare serial measurements, nonparametric analysis of variance was used. Groups were compared using the chi-square test. Unless otherwise indicated, values are presented as medians with the interquartile range in parenthesis. All statistics were computed using the SPSS software, version 11.0 (SPSS, Chicago, IL).

RESULTS Patient Characteristics, Cortisol Levels, and Serum Cortisol Response to ACTH Figure 1 shows the number of patients screened, randomized, and treated in each arm. Baseline characteristics of those patients included are given in Table 1. Patients in the treatment group tended to be slightly older compared with placebo. Platelet counts were lower in the placebo group. In all other aspects, both groups were comparable, especially severity of illness as expressed by the Acute Physiology and Chronic Health Evaluation II score and organ dysfunction as expressed by the SOFA score (Table 1). Median cortisol levels at baseline and after ACTH stimulation were similar between the placebo and treatment group. Sixty-three percent of all patients showed an increment of "250 nmol/L after the injection of 0.25 mg of ACTH and were thus classified as ACTH nonresponders (19, 20). The proportion of nonresponders was equal in hydrocortisoneand placebo-treated patients. Albumin levels at time of adrenal testing were comparable between those patients classified as ACTH responders and ACTH nonresponders (2.9 mg/dL [2.3–3.2] vs. 2.7 mg/dL [2.6 –2.9], respectively). Cortisol levels are shown in Table 2.

Hemodynamics and Outcome As shown in Figure 2 (left panel), time to vasopressor withdrawal was markedly Crit Care Med 2005 Vol. 33, No. 11

Figure 1. Flow chart of screening, randomization, and analysis of patients in the study. HC, hydrocortisone.

shorter in the treatment group compared with placebo (median 53 hrs [35–96] vs. 120 hrs [48 –197]; p " .02, log rank test). On day 2 after initiation of study medication, nine patients in the treatment group and 19 patients in the placebo group were still in shock (p % .04, chisquare test). After vasopressor support was stopped and study medication was tapered, some patients experienced a shock rebound. This was seen in three patients in the HC group and in five patients in the placebo group (p % .68, chi-square test). Shock reversal on day 7 was similar between both groups (73% vs. 79%, p % .73 chi-square test). A nonsignificant reduction in mortality (39% in HC-treated vs. 48% in placebo treated patients, p % .6; log rank test) was observed.

Cytokine Response IL-6 plasma levels are shown in Figure 3 (upper panel). At baseline, the values were highly elevated. During the course of the disease they gradually fell to near normal levels on day 14. In the HC group, this decline occurred significantly faster. Crit Care Med 2005 Vol. 33, No. 11

From days 1 to 5, the levels were significantly lower compared with placebo (p " .001, nonparametric analysis of variance). The levels of the anti-inflammatory cytokine IL-10 were elevated in both groups at baseline. They decreased over the study period to near normal values on day 14. There was no difference between the two study groups (Fig. 3, lower panel). The ex vivo LPS-stimulated IL-1 production at baseline was suppressed in both groups. Under HC, however, this suppression was prolonged compared with placebo, and the IL-1 production was lower until day 5 (p % .03, nonparametric analysis of variance). After day 5, IL-1 production was comparable in both groups. The ex vivo LPS-stimulated IL-6 production (Fig. 4, left panel) was comparable in both groups at baseline. After initiation of HC treatment, the IL-6 production fell and was significantly lower than the IL-6 production in placebo-treated patients from day 1 until day 5 (p % .02, nonparametric analysis of variance). After shock reversal and discontin-

uation of HC, the IL-6 production was similar in both groups. The courses of C-reactive protein and procalcitonin are given in Table 2. The measurement of the LPS-stimulated ex vivo IL-6 production in healthy volunteers showed that addition of HC resulted in a dose-dependent reduction of the IL-6 production. IL-6 levels were 19,910 (15,030 –25,330) pg/mL without the addition of HC, 3645 (2230 – 4665) pg/mL with 2500 nmol/L HC, and 3075 (1833– 1173) pg/mL with 10,000 nmol/L HC. Ex vivo pretreatment of the blood with HC and addition of LPS after 3 hrs resulted in a further decrease of IL-6 production (median 14,520 vs. 1430 vs. 725 pg/mL). The hemodynamic and cytokine effects of HC in patients classified as ACTH responders and nonresponders are given in Figure 2 (right panel) and Figure 4 (right panel). In nonresponders (n % 26), HC therapy resulted in a reduction in time to shock reversal (Fig. 2 right panel) compared with placebo (p % .06, log rank test). Among the ACTH responders (n % 15), median time to vasopressor withdrawal was not significantly different in the HC group compared with placebo (p % .9, log rank test). The ex vivo LPS-stimulated IL-6 production on day 1 after initiation of study medication was reduced in HC-treated ACTH responders and HC-treated ACTH nonresponders to a similar extent (Fig. 4, right panel). In the ACTH nonresponder group, the mortality of the treatment group was five of 12 patients (42%) compared with six of 14 (43%) in the placebo group. In the group of the ACTH responders, the mortality was two of six patients (33%) in the HC-treated patients compared with five of nine (55%) in the placebo group.

Organ Dysfunction Organ dysfunction was comparable between the two groups at baseline. The SOFA score (Table 2) from day 1 until day 5 tended to be lower in the HC group compared with the placebo group (p % .09, nonparametric analysis of variance). Apart from the hemodynamics, the resolution of organ failure was comparable between the two groups (Table 2).

Side Effects There was no difference in adverse events between the two treatment groups. The number of units of packed red blood 2459

Table 1. Patient baseline characteristics Hydrocortisone Age, yrs Gender, female/male APACHE II SOFA Site of infection, % Lung Abdomen Urinary tract Heart Other Gram-positive, n Gram-negative, n Mixed, n Cardiovascular dysfunction, median (IQR) CVP, mm Hg PAOP, mm Hg CI, L/m2 NE, mg/day Organ dysfunction, median (IQR) PaO2/FIO2 ratio Serum creatinine, mg/dL Serum bilirubin, mg/dL Platelets, G/L C-reactive protein, mg/dL Procalcitonin, ng/L Cortisol, nmol/L Baseline, median (IQR) After ACTH, median (IQR) Responder/nonresponder, n "552 nmol/L, n

Placebo

59 5/13 25 (19–30) 11 (10–13)

47 4/19 25.5 (19.8–29) 10 (8.5–13)

50 27 0 13 10 12 4 2

35 30 5 13 17 14 6 2

12 (8–14) 17 (12–19.5) 4.2 (3.4–4.9) 22 (10–62)

12 (10–19) 18 (16–19) 3.9 (3.0–4.4) 34 (18–56)

217 (130–284) 2.7 (1.2–5.6) 1.7 (1.2–4.9) 166 (49–323) 18.8 (15.2–23.5) 7.7 (3.7–83.3)

177 (160–243) 2.2 (1.1–3.1) 1.7 (1.2–4.7) 80 (32–162) 21.4 (17.1–27.4) 21.9 (5–62.2)

586 (404–697) 780 (512–1004) 6/12 4

659 (481–1006) 814 (568–1200) 9/14 4

cells as a surrogate marker for blood loss was equal in both groups. There was no increase in secondary infection under HC; neither was there a relevant rise in serum sodium. The blood glucose was comparable between the two treatment groups, with a trend, however, to higher amounts of insulin needed in the HC-treated patients (Table 2). Electrophysiologic studies on critical illness neuropathy and myopathy were not performed.

DISCUSSION

APACHE, Acute Physiology and Chronic Health Evaluation; SOFA, Sequential Organ Failure Assessment; IQR, interquartile range; CVP, central venous pressure; PAOP, pulmonary artery occlusion pressure; CI, cardiac index; NE, norepinephrine; ACTH, adrenocorticotropic hormone.

The main findings of this singlecenter study in patients with early septic shock are that a) relative adrenal insufficiency is common in this situation; b) HC leads to a significantly shorter duration of septic shock and a reduced production of proinflammatory cytokines; and c) the anti-inflammatory effects of HC occur irrespective of adrenal function, whereas the hemodynamic effects are more marked in patients with adrenal insufficiency. The finding that HC compared with placebo reduces time to shock reversal is consistent with previously reported data of two smaller studies and one large randomized controlled trial (1, 3, 12). The

Table 2. Evolution of hemodynamic variables, Sequential Organ Failure Assessment (SOFA) score, and laboratory tests during the first 5 days of the study in hydrocortisone (HC) and placebo (PL) treated patients Day 0 HR, beats/min HC PL MAP, mm Hg HC PL SVR, dyne ! sec ! cm$5 HC PL SOFA HC PL C-reactive protein, mg/dL HC PL PCT, ng/mL HC PL Serum glucose, mg/dL HC PL Insulin, IU/day HC PL Cortisol, nmol/L HC PL

Day 1

Day 2

Day 3

Day 4

Day 5

p Value

109 (99–120) 110 (92–130)

96 (83–115) 102 (90–124)

90 (66–98) 95 (86–108)

94 (80–100) 96 (88–108)

93 (78–103) 96 (88–119)

88 (80–107) 93 (85–107)

.3

68 (58–75) 70 (64–85)

80 (68–88) 75 (63–90)

82 (70–88) 81 (71–98)

72 (64–91) 77 (72–95)

70 (63–91) 75 (66–98)

74 (65–94) 78 (65–95)

.8

526 (396–787) 583 (558–979)

.4

6 (3.5–10) 8 (6.5–11.5)

.09

625 (495–687) 678 (512–944)

694 (545–1017) 591 (471–924)

797 (723–1027) 713 (575–861)

670 (516–929) 732 (532–935)

10 (8.5–13) 11 (10–13)

10 (8.3–11) 12 (9–14)

8 (6–12) 12 (8.8–13.3)

18.8 (15.2–23.5) 21.4 (17.1–27.4)

21.3 (16.7–33.5) 24.8 (19.2–30.0)

15.2 (10.7–23.2) 18.1 (11.4–22.5)

8.6 (3.7–12.2) 18.3 (9.8–25.2)

5.3 (2.7–9.4) 16.1 (5.9–20.5)

6.3 (3.0–12.1) 15.2 (3.9–21.8)

.02

7.7 (3.7–83.3) 21.9 (5–62.2)

7.1 (2.9–39.3) 11.3 (2.4–67.8)

23.3 (2.2–237) 4.3 (0.6–17.5)

23.2 (4.6–345) 2.4 (1–23.9)

Not measured

2 (0.5–13.9) 4.1 (0.8–2.9)

.04

137 (118–189) 138 (103–179)

139 (117–207) 152 (128–179)

115 (109–184) 143 (123–150)

140 (116–181) 139 (120–149)

108 (84–116) 131 (109–145)

122 (100–131) 122 (106–137)

.6

0 (0–30) 0 (0–15)

48 (0–102) 24 (0–56)

24 (0–102) 0 (0–45)

586 (404–697) 659 (481–1006)

2370 (1439–5128) 652 (443–923)

2532 (1538–3863) 790 (541–1015)

7 (5–12) 10.5 (8.3–13)

590 (287–1025) 692 (577–817)

36 (2–66) 0 (0–41) 1659 (1031–4133) 656 (576–988)

7 (5–9) 9 (7–12)

20 (0–31) 0 (0–44) 958 (644–1525) 680 (593–858)

17 (0–69) 0 (0–8) 697 (457–1471) 677 (551–906)

.3 ".001

HR, heart rate; MAP, mean arterial pressure; SVR, systemic vascular resistance; PCT, procalcitonin. Data presented as median (interquartile range); p value calculated using the nonparametric analysis of variance.

2460

Crit Care Med 2005 Vol. 33, No. 11

underlying mechanisms are unclear, but an improved vasopressor responsiveness of peripheral vessels may play an important role. This is supported by the observation in septic shock patients showing that HC can restore vascular responsiveness to catecholamines (13). Also, a decreased pressor response to norepinephrine in patients with impaired adrenal reserve has been described (13). Although improving hemodynamics through sensitization against catecholamines may be an important benefit, the present study also provides clear evidence that the application of HC interferes with more proximal mechanisms in the pathogenesis of septic shock, which include the inflammatory response. Thus, we found reduced plasma levels of IL-6, which is consistent with previously reported studies of lower proinflammatory cytokines under HC (14, 15, 21). It is of note that HC induced both a reduction in IL-6 plasma levels and a marked effect on the responsiveness of IL-6 production in blood samples of HC-treated patients stimulated ex vivo, which indicates direct effects on peripheral mononuclear cells. This inhibition of IL-6 production was most marked during those study days when HC levels were highest, suggesting a concentration-dependent mechanism.

Although we have not measured plasma catecholamines levels, they are supposed to be lower in samples from HC-treated patients, given the reduction in the administered dose. Since catecholamines have been shown to have no effect on IL-6 levels in septic patients both in vivo and ex vivo (21) and even decrease IL-6 in healthy volunteers (22, 23, 24), the reduced cytokine levels are unlikely to be the result of reduced vasopressor support. Furthermore, the experiments in blood from healthy volunteers indicate that the inhibitory steroid effect on cytokine production is neither dependent on the presence of sepsis nor due to priming of macrophages in vivo, since the ex vivo addition of HC to LPS-stimulated blood also resulted in a significant reduction of cytokine production. This observation is in line also with a previous study showing that the ex vivo addition of dexamethasone to whole blood resulted in a reduced production of IL-6 (25). IL-1 is another important proinflammatory cytokine. It is rapidly blocked by various inhibitors and antagonists and therefore difficult to measure in plasma (26). We could, however, demonstrate reduced IL-1 production after ex vivo LPS stimulation of whole blood in the HC-

treated septic patients. These findings illustrate another HC effect on inflammatory mediators. Despite evidence for a reduction in the proinflammatory response, we found no effect on the anti-inflammatory cytokine IL-10, which is in accordance with previously published results (14). Another study, however, found a slight but statistically significant reduction also in IL-10 levels of patients under low-dose HC (15). The measured human leukocyte antigen DR expression was only marginally changed under HC, and monocyte and granulocyte function was preserved (15). In summary, therefore, in early septic shock low-dose HC may reduce hyperinflammation, whereas anti-inflammation appears not to be affected. One important question is whether low-dose HC therapy is mainly a replacement therapy for those patients with impaired adrenal reserve, which might imply that patients without relative adrenal insufficiency would not benefit. There is, however, no uniform agreement on the definition of adrenal hyporesponsiveness in septic shock. Although the rapid ACTH test is being widely used as a simple method to identify adrenocortical hyporesponsiveness, controversy exists as to

Figure 2. Kaplan Meier Plot of the probability of being in septic shock for hydrocortisone (HC, broken line) and placebo (PL, full line) treated patients. Left panel, all patients; right panel, adrenocorticotropic hormone (ACTH) responders (thin line) and ACTH nonresponders (bold line), separately.

Crit Care Med 2005 Vol. 33, No. 11

2461

how diagnostic criteria should be derived from the basal cortisol level, the stimulated cortisol level, or their difference (10, 27, 28). In addition, recent data suggest that measurement of total cortisol may be misleading in evaluating the adrenal reserve in critically ill patients and that measurements of free cortisol are more reliable (29). Currently, however, an increment of total cortisol of "250 nmol/L after stimulation with ACTH seems to be the most widely accepted cutoff to define relative adrenal insufficiency (17, 19). Applying this criterion to the present study population, the inci-

dence of adrenal hyporesponsiveness was 63%, which is comparable to the French trial (3). It is known that medications such as etomidate may interfere with the adrenal axis. We therefore performed an analysis excluding the seven patients who received etomidate within 1 wk before inclusion. This revealed that the results are identical compared with those including all patients. Irrespective of the definition of adrenal hyporesponsiveness, we have previously shown that levels achieved using the so-called low-dose regime far exceed levels reached in patients considered to

Figure 3. Plasma interleukin (IL)-6 (upper panel) and IL-10 (lower panel) levels of hydrocortisone (filled squares) and placebo (open squares) treated patients. For IL-6, p " .001 days 0 –5. For IL-6, p " .001 days 1–5. Data are expressed as median (interquartile range).

have an intact endogenous response (16) and should therefore be classified as pharmacologic. We have also shown that calculated free cortisol levels were not different between patients classified as having an adequate or inadequate adrenal reserve according to the previously mentioned criteria (16). Nevertheless, one previous study, which also assessed the ACTH response, found a significant hemodynamic and survival benefit only in those patients classified as ACTH nonresponders (3). Although this study has been criticized (30, 31), our present investigation also demonstrates a trend for a beneficial effect of HC in ACTH nonresponders on hemodynamics. On the other hand, the inflammatory response did not show such a trend; that is, there was no obvious difference in reduction of cytokine levels between ACTH responders and nonresponders. Organ dysfunction as assessed by the SOFA score is associated with high rates of morbidity. It could be shown that it is possible to predict mortality not only from initial scores but also from the delta score after the initial 48 hrs or the mean score (32). In our study, the SOFA score in the treatment group tended to be lower compared with placebo during the first 5 days of the study. After day 5, the SOFA score was comparable between the two groups. One needs to be cautious in interpreting the mortality data. The study was neither planned nor powered to investigate this. Especially, conclusions cannot be drawn from the difference in mortality results between our study and the French trial.

Figure 4. LPS-stimulated ex vivo interleukin (IL)-6 production (left panel) in hydrocortisone (filled squares) and placebo (open squares) treated patients, p % .02 (nonparametric analysis of variance) and IL-6 production on day 1 in adrenocorticotropic hormone (ACTH) responders and nonresponders (right panel). Data are expressed as median (interquartile range).

2462

Crit Care Med 2005 Vol. 33, No. 11

T

reatment

with

low-dose hydrocortisone acceler-

ates shock reversal in early hyperdynamic septic shock.

The possible benefits of the widespread application of therapeutic interventions in critically ill naturally have to be balanced against potential side effects. In this respect, the present study did not reveal any hint of a relevant difference in the rate of secondary infections, gastrointestinal bleeding, electrolyte imbalance, or rebound shock. Special attention also needs to be drawn to blood glucose control. Since a recent study demonstrated a survival benefit in patients with tight glucose control (33), there is some concern about the potential of steroids to impair glucose tolerance. Although, to date, there is no evidence for an impact of hyperglycemia on mortality rate in patients with severe sepsis and septic shock, in the present study the aim was to keep blood glucose levels "150 mg/dL. In patients receiving HC, this was achieved by a slightly higher but not significantly different usage of insulin. It is possible that the continuous infusion of HC in this study (in contrast to bolus application in other trials) (1, 3) reduced the variation of glucose levels and made reaching blood glucose levels "150 mg/dL easier. A limitation of this study is the relatively small number of patients included. Also, patients in the placebo group had a lower platelet count, possibly indicating more severe disseminated intravascular coagulation. However, the treatment group tended to be older, and the Acute Physiology and Chronic Health Evaluation II score as well as the SOFA score were comparable. These imbalances therefore do not seem to account for the observed differences.

CONCLUSIONS The present single-center study adds additional evidence that so-called lowdose HC has beneficial effects in patients with septic shock and points out that these effects are not limited to hemodynamic improvement but also include the Crit Care Med 2005 Vol. 33, No. 11

underlying imbalance between pro- and anti-inflammation. Although the hemodynamic effect of HC seems to be related to endogenous cortisol production, the immune response under HC appears to be independent of adrenal function.

14.

15.

REFERENCES 1. Bollaert P, Charpentier C, Levy B, et al: Reversal of late septic shock with supraphysiologic doses of hydrocortisone. Crit Care Med 1998; 26:645– 650 2. Bollaert P, Fieux F, Charpentier C, et al: Baseline cortisol levels, cortisol response to corticotropin and prognosis in late septic shock. Shock 2003; 19:13–15 3. Annane D, Sebille V, Charpentier C, et al: Effect of treatment with low doses of hydrocortisone and fludrocortisone on mortality in patients with septic shock. JAMA 2002; 288: 862– 871 4. Eichacker PQ, Parent C, Kalil A, et al: Risk and the efficacy of anti-inflammatory agents: Retrospective and confirmatory studies of sepsis. Am J Respir Crit Care Med 2002; 166:1197–1205 5. Bernard GR, Wheeler AP, Russell JA, et al: The effects of ibuprofen on the physiology and survival of patients with sepsis. The Ibuprofen in Sepsis Study Group. N Engl J Med 1997; 336:912–918 6. Bone RC, Fisher CJ, Clemmer TP, et al: A controlled clinical trial of high-dose methylprednisolone in the treatment of severe sepsis and septic shock. N Engl J Med 1987; 317:653– 658 7. Sprung CL, Caralis PV, Marcial EH, et al: The effects of high-dose corticosteroids in patients with septic shock. N Engl J Med 1984; 311:1137–1143 8. The Veterans Administration Systemic Sepsis Cooperative Study Group. Effect of high-dose glucocorticoid therapy on mortality in patients with clinical signs of systemic sepsis. N Engl J Med 1987; 317:659 – 665 9. Cronin L, Cook DJ, Carlet J, et al: Corticosteroid treatment for sepsis: A critical appraisal and meta-analysis of the literature. Crit Care Med 1995; 23:1430 –1439 10. Rothwell PM, Udwadia ZF, Lawler PG: Cortisol response to corticotropin and survival in septic shock. Lancet 1991; 337:582–583 11. Molijn GJ, Spek JJ, van Uffelen JC, et al: Differential adaptation of glucocorticoid sensitivity of peripheral blood mononuclear leukocytes in patients with sepsis or septic shock. J Clin Endocrinol Metab 1995; 80: 1799 –1803 12. Briegel J, Forst H, Haller M, et al: Stress doses of hydrocortisone reverse hyperdynamic septic shock: A prospective, randomized, double blind, single center study. Crit Care Med 1999; 27:723–732 13. Annane D, Bellissant E, Sebille V, et al: Impaired pressor sensitivity to noradrenaline in septic shock patients with and without im-

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

paired adrenal function reserve. Br J Clin Pharmacol 1998; 46:589 –597 Briegel J, Jochum M, Gippner-Steppert C, et al: Immunomodulation in septic shock: Hydrocortisone differentially regulates cytokine responses. J Am Soc Nephrol 2001; 12(Suppl 17):S70 –S74 Keh D, Boehnke T, Weber-Cartens S, et al: Immunologic and hemodynamic effects of “low-dose” hydrocortisone in septic shock: A double-blind, randomized, placebo-controlled, crossover study. Am J Respir Crit Care Med 2003; 167:512–520 Oppert M, Reinicke A, Graf KJ, et al: Plasma cortisol levels before and during “low-dose” hydrocortisone therapy and their relationship to hemodynamic improvement in patients with septic shock. Intensive Care Med 2000; 26:1747–1755 Annane D, Sebille V, Troche G, et al: A 3-level prognostic classification in septic shock based on cortisol levels and cortisol response to corticotropin. JAMA 2000; 283: 1038 –1045 Montrucchio G, Bosco O, Del Sorbo L, et al: Mechanisms of the priming effect of low doses of lipopoly-saccharides on leukocytedependent platelet aggregation in whole blood. Thromb Haemost 2003; 90:872– 881 Cooper M, Stewart P: Corticosteroid insufficiency in acutely ill patients. N Engl J Med 2003; 348:727–734 Voort P van der, Gerritsen R, Bakker A, et al: HDL-cholesterol level and cortisol response to synacthen in critically ill patients. Intensive Care Med 2003; 29:2199 –2203 Kilger E, Weis F, Briegel J, et al: Stress doses of hydrocortisone reduce severe systemic inflammatory response syndrome and improve early outcome in a risk group of patients after cardiac surgery. Crit Care Med 2003; 31:1068 –1074 Bergmann M, Gornikiewicz A, Sautner T, et al: Attenuation of catecholamineinduced immunosuppression in whole blood from patients with sepsis. Shock 1999; 12:421– 427 Smits HH, Grunberg K, Derijk RH, et al: Cytokine release and its modulation by dexamethasone in whole blood following exercise. Clin Exp Immunol 1998; 111: 463– 468 Maes M, Lin A, Kenis G, et al: The effects of noradrenaline and alpha-2 adrenoceptor agents on the production of monocytic products. Psychiatry Res 2000; 96:245–253 DeRijk R, Michelson D, Karp B, et al: Exercise and circadian rhythm-induced variations in plasma cortisol differentially regulate interleukin-1 beta (IL-1 beta), IL-6, and tumor necrosis factor-alpha (TNF alpha) production in humans: High sensitivity of TNF alpha and resistance of IL-6. J Clin Endocrinol Metab 1997; 82:2182–2191 Granowitz EV, Santos AA, Poutsiaka DD, et al: Production of interleukin-1-receptor antagonist during experimental endotoxaemia. Lancet 1991; 338:1423–1424

2463

27. Moran JA, Chapman MJ, O’Fathartaigh MS, et al: Hypocortisolemia and adrenocortical responsiveness at onset of septic shock. Intensive Care Med 1994; 20:489 – 495 28. Marik PE, Zaloga GP: Adrenal insufficiency during septic shock. Crit Care Med 2003; 31:141–145 29. Hamrahian AH, Oseni TS, Arafah BM. Mea-

2464

surements of serum free cortisol in critically ill patients. N Engl J Med 2004; 350: 1629 –1638 30. Zijlstra JG, van der Horst IC, Tulleken JA, et al: Corticosteroids for patients with septic shock. JAMA 2003; 289:42 31. Mokhlesi B: Corticosteroids for patients with septic shock. JAMA 2003; 289:43

32. Ferreira FL, Bota DP, Bross A, et al: Serial evaluation of the SOFA score to predict outcome in critically ill patients. JAMA 2001; 286:1754 –1758 33. van den Berghe G, Wouters P, Weekers F, et al: Intensive insulin therapy in the critically ill patients. N Engl J Med 2001; 345: 1359 –1367

Crit Care Med 2005 Vol. 33, No. 11