INTRODUCTION
Septic shock is a highly fatal disorder with mortality rates around 40% to 50% (1, 2). After appropriate fluid resuscitation and antibiotic administration, initiating agents that will augment hemodynamics is integral to supportive care (3). These agents can include continuous infusion vasopressors and corticosteroids. The mechanism of action of corticosteroids in patients with septic shock is complex and multifaceted and primarily relies on their anti-inflammatory properties and supplementation in a state of relative adrenal suppression (4). The use of corticosteroids in septic shock remains controversial, but the 2016 Surviving Sepsis Campaign (SSC) Guidelines and the 2017 Guidelines for Critical Illness-Related Corticosteroid Insufficiency suggest using hydrocortisone 200 mg daily in patients with septic shock who remain hemodynamically unstable after fluid resuscitation and vasopressor therapy (3, 5). Of note, these are listed as weak recommendations with low and moderate quality of evidence, respectively, because of contradictory results in the literature (3, 5–13). The disparity in results from clinical trials may be due to significant differences in how and when hydrocortisone was administered. The studies in which a benefit was seen more often had hydrocortisone initiated early after shock onset (6, 7, 9) and studies in which hydrocortisone was initiated 24 h after shock onset did not detect a benefit (8). Prior, small, observational studies have evaluated the timing of hydrocortisone initiation with beneficial results associated with earlier initiation (14–16). However, these results have yet to be replicated in large clinical trials. This study aimed to compare vasopressor duration and mortality in patients with septic shock who received hydrocortisone based on timing after shock onset. The hypothesis was that patients who received hydrocortisone earlier in their course of septic shock would confer more benefit from their utilization.
METHODS
This was a retrospective, observational, database study conducted at 10 medical, surgical, and neurosciences intensive care units (ICUs) at the Cleveland Clinic Main Campus, a large academic tertiary care referral center in Cleveland, Ohio. Patients admitted from January 2011 through November 2017 were screened for inclusion. Patients were included if they were over the age of 18, had septic shock requiring continuous infusion norepinephrine for at least 12 h, and received at least two consecutive doses of IV hydrocortisone after the initiation of vasopressors. Patients were excluded if they received multiple systemic corticosteroids concurrently, with the exception of fludrocortisone, or had missing electronic medical record data with which to evaluate outcomes. Septic shock was determined through the evaluation of clinical criteria based on the United States Centers for Disease Control and Prevention definition of an Adult Sepsis Event (17, 18) which was defined as having a presumed infection (with blood cultures obtained and ≥4 qualifying antibiotic days) requiring norepinephrine with a lactate >2.0 mmol/L. The use of hydrocortisone for septic shock is guided by physician preference and is not protocolized at our institution, lending to appropriate evaluation retrospectively as the timing of its initiation is relatively random. In this evaluation, hydrocortisone was administered as intermittent IV bolus doses and at dose of 50 mg or 100 mg. The primary objective was to compare vasopressor duration between patients initiated on hydrocortisone between 0 and 6 h after shock onset, 6 and 12 h after shock onset, 12 and 24 h after shock onset, 24 and 48 h after shock onset or >48 h after shock onset. Secondary outcomes included determining the association with hydrocortisone timing on vasopressor duration with multivariable linear regression and determining association with timing on ICU mortality with multivariable logistic regression. A subgroup analysis comparing patients who had hydrocortisone initiated in the first 24 h of shock onset and those who had it initiated after the first 24 h of shock onset was conducted. The cutoff of 24 h was chosen because the clinical trials that showed a benefit with the use of hydrocortisone utilized therapy within 24 h or sooner (6, 7). Shock onset was defined as the time of continuous infusion norepinephrine initiation. Vasopressor duration was evaluated as days alive and free from all vasoactive agents (19). This was calculated by adding the number of days the patient was alive and off of all vasoactive agents starting at the time of the last vasoactive agent discontinuation for that patient's hospital encounter. For example, if a patient was off vasopressors from days 5 to 10 but required vasopressors from day 12 to 20 yet survived the hospital encounter, there would be no credit for the days 10 to 12 off vasopressors. If patients died while receiving vasopressors, they were counted as 0 days alive and free from vasopressors.
Data is resulted as mean ± standard deviation or number (percent) unless otherwise specified. The aforementioned timing cohorts of patients were statistically compared with the Kruskal–Wallis test. To compare the two timing cohorts of < 24 h and >24 h in the subgroup analysis, the Chi-square or Fisher exact test was used for nominal data and the Student t test was used for continuous data. The primary outcome was days alive and free from vasopressors. Secondary outcomes included ICU mortality and ICU and hospital length of stay. A multivariable linear regression model was conducted for days alive and free from vasopressors adjusting for severity of illness (lactate, norepinephrine dose at corticosteroid initiation, and APACHE III score) and other variables with a P value <0.05 on univariate analysis. A similarly adjusted multivariable logistic regression model was conducted for ICU mortality. Variables included within the multivariable regression models were assessed for colinearity utilizing correlation statistics and variance inflation factors. A multivariable linear and logistic regression evaluating the association between hydrocortisone initiation <24 h and >24 h on days alive and free from vasopressors and ICU mortality, respectively, were also conducted. Logistic regression models were assessed for goodness of fit with the Hosmer–Lemeshow goodness of fit test. All analyses were performed based on an overall significance level of 0.05, StataIC 14 (StataCorp LLC, College Station, Tex).
RESULTS
A total of 1979 patients were screened for inclusion into this evaluation and 509 were excluded (Fig. 1). Of the 1,470 patients included, 567 (38.6%) received hydrocortisone between 0 and 6 h after shock onset, 231 (15.7%) between 6 and 12 h, 260 (17.7%) between 12 and 24 h, 195 (13.3%) between 24 and 48 h, and 217 (14.8%) >48 h after shock onset. The average age of included patients was 61 ± 15 years (Table 1) and the rate of patients with liver disease and immune suppression varied between timing groups. APACHE III scores at ICU admission were 106 ± 35 for 0 to 6 h, 108 ± 36 for 6 to 12 h, 102 ± 30 for 12 to 24 h, 104 ± 34 for 24 to 48 h, and 99 ± 31 for >48 h. Lactate concentrations at the time of hydrocortisone initiation varied between groups (4.6 ± 4.0 for 0–6 h; 4.9 ± 4.5 for 6–12 h; 3.9 ± 3.6 for 12–24 h; 3.9 ± 3.7 for 24–48 h and 3.1 ± 3.0 for >48 h; P < 0.01). Norepinephrine dose at the time of hydrocortisone initiation also differed between groups (23.6 ± 20.2 mcg/min for 0–6 h; 26.6 ± 23.0 for 6–12 h; 25.0 ± 23.8 for 12–24 h; 23.7 ± 22.3 for 24–48 h; 19.2 ± 19.5 for >48 h).
;)
Fig. 1: Patient inclusion tree. EMR indicates electronic medical record; HC, hydrocortisone; NE, norepinephrine.
Table 1 -
Baseline characteristics based on timing of hydrocortisone receipt
|
TotalN = 1470 |
HC 0–6 hN = 567 |
HC 6–12 hN = 231 |
HC 12–24 hN = 260 |
HC 24–48 hN = 195 |
HC >48 hN = 217 |
P value |
| Age, years |
61 ± 15 |
62 ± 15 |
61 ± 14 |
60 ± 15 |
60 ± 14 |
61 ± 13 |
0.19 |
| Male gender |
804 (54.7) |
320 (56.4) |
116 (50.2) |
141 (54.2) |
113 (58.0) |
114 (52.5) |
0.43 |
| Weight, kg |
88.7 ± 32.3 |
85.6 ± 25.3 |
91.1 ± 38.9 |
89.0 ± 31.5 |
93.6 ± 37.3 |
89.4 ± 36.2 |
0.38 |
| Race/ethnicity |
| White |
1,028 (69.9) |
403 (71.1) |
169 (73.2) |
178 (68.5) |
137 (70.3) |
141 (65.0) |
0.26 |
| African American |
353 (24.0) |
134 (23.6) |
48 (20.8) |
65 (25.0) |
47 (24.1) |
59 (27.2) |
|
| Asian |
17 (1.2) |
3 (0.5) |
2 (0.9) |
3 (1.2) |
1 (0.5) |
8 (3.7) |
|
| Other |
44 (3.0) |
17 (3.0) |
7 (3.0) |
8 (3.1) |
7 (3.6) |
5 (2.3) |
|
| Unknown |
28 (1.9) |
10 (1.8) |
5 (2.2) |
6 (2.3) |
3 (1.5) |
4 (1.8) |
|
| Comorbidities |
| COPD |
308 (21.0) |
121 (21.3) |
46 (19.9) |
62 (23.9) |
37 (19.0) |
42 (19.4) |
0.68 |
| AIDS |
10 (0.7) |
3 (0.5) |
0 (0) |
2 (0.78) |
1 (0.5) |
4 (1.8) |
0.20 |
| Diabetes mellitus |
432 (29.4) |
165 (29.1) |
66 (28.6) |
89 (34.2) |
53 (27.2) |
59 (27.2) |
0.41 |
| Liver disease |
302 (20.5) |
82 (14.5) |
44 (19.1) |
56 (21.5) |
47 (24.1) |
73 (33.6) |
<0.01 |
| Immune suppression |
274 (18.6) |
144 (25.4) |
35 (15.2) |
38 (14.6) |
29 (14.9) |
28 (12.9) |
<0.01 |
| Leukemia/Myeloma |
78 (5.3) |
33 (5.8) |
14 (6.1) |
12 (4.6) |
10 (5.1) |
9 (4.2) |
0.84 |
| Non-Hodgkins Lymphoma |
38 (2.6) |
12 (2.1) |
4 (1.7) |
6 (2.3) |
6 (3.1) |
10 (4.6) |
0.29 |
| No chronic health comorbidities |
359 (24.4) |
134 (23.6) |
72 (31.2) |
58 (22.3) |
47 (24.1) |
48 (22.1) |
0.13 |
| ICU location |
| Medical |
1,003 (68.2) |
366 (64.6) |
151 (65.4) |
192 (73.9) |
132 (67.7) |
162 (74.7) |
0.02 |
| Surgical |
334 (22.7) |
144 (25.4) |
59 (25.5) |
52 (20.0) |
41 (21.0) |
38 (17.5) |
|
| Neurosciences |
133 (9.1) |
57 (10.1) |
21 (9.1) |
16 (6.2) |
22 (11.3) |
17 (7.8) |
|
| Cortisol testing |
203 (13.8) |
52 (9.2) |
20 (8.7) |
36 (13.9) |
37 (19.0) |
58 (26.7) |
<0.01 |
| APACHE III |
104.3 ± 33.6 |
106.0 ± 35.0 |
107.5 ± 35.9 |
102.1 ± 30.1 |
103.9 ± 33.6 |
99.3 ± 30.9 |
0.10 |
| APS |
88.7 ± 32.1 |
90.2 ± 33.9 |
92.6 ± 33.8 |
87.1 ± 29.0 |
88.2 ± 31.5 |
83.1 ± 28.9 |
0.05 |
| Lactate at HC start, mmol/L |
4.2 ± 3.9 |
4.6 ± 4.0 |
4.9 ± 4.5 |
3.9 ± 3.6 |
3.9 ± 3.7 |
3.1 ± 3.0 |
<0.01 |
| MAP, mmHg |
66.4 ± 14.6 |
66.4 ± 14.9 |
67.5 ± 15.9 |
64.9 ± 13.6 |
66.1 ± 12.9 |
67.1 ± 14.8 |
0.63 |
| NE dose at HC initiation, mcg/min |
23.7 ± 21.6 |
23.6 ± 20.2 |
26.6 ± 23.0 |
25.0 ± 23.8 |
23.7 ± 22.3 |
19.2 ± 19.5 |
<0.01 |
| Other vasopressors when HC start |
| 0 |
695 (47.3) |
298 (52.6) |
86 (37.2) |
102 (39.2) |
87 (44.6) |
122 (56.2) |
<0.01 |
| 1 |
634 (43.1) |
210 (37.0) |
121 (52.4) |
125 (48.1) |
89 (45.6) |
89 (41.0) |
|
| 2 |
127 (8.6) |
50 (8.8) |
22 (9.5) |
31 (11.9) |
18 (9.2) |
6 (2.7) |
|
| 3 |
14 (1.0) |
9 (1.6) |
2 (0.9) |
2 (0.8) |
1 (0.5) |
0 (0) |
|
| HC dose in first 24 h, mg |
216.1 ± 63.8 |
215.8 ± 62.9 |
214.8 ± 55.1 |
217.5 ± 64.7 |
210.3 ± 54.8 |
221.7 ± 79.6 |
0.73 |
| Time from NE to HC, h |
26.1 ± 47.4 |
2.5 ± 1.7 |
8.8 ± 1.8 |
17.2 ± 3.4 |
33.3 ± 6.7 |
110.3 ± 78.3 |
– |
Data detailed as mean ± SD or n (%) unless otherwise indicated; P values derived from Kruskal–Wallis test for multiple independent samples.AIDS indicates acquired immune deficiency syndrome; APACHE III, acute physiology and chronic health evaluation III; APS, acute physiology score; COPD, chronic obstructive pulmonary disease; HC: hydrocortisone; ICU, intensive care unit; MAP, mean arterial pressure; NE, norepinephrine.
On univariate analysis, days alive and free from vasopressors did not significantly differ amongst the timing groups (median 3.3 days for 0–6 h; 1.9 for 6–12 h; 1.9 for 12–24 h; 0 for 24–48 h; 0 for >48 h; P = 0.39); similarly, ICU mortality did not differ (43% vs. 47% vs. 46% vs. 48% vs. 48%; P = 0.58). ICU and hospital lengths of stay were longer in the cohorts who had hydrocortisone started later (Table 2). On multivariable linear regression (Table 3), after adjustment for ICU location, lactate, presence of cortisol testing, norepinephrine dose, and APACHE III score, the timing of hydrocortisone was significantly associated with more days alive and free from vasopressors when comparing initiation within 0 to 6 h with over 48 h (beta-coefficient 2.8 [95% CI 0.8–4.7]), 6–12 h with >48 h (beta-coefficient 2.5 [95% CI 0.2–4.7]), and 12–24 h with >48 h (beta-coefficient 2.3 [95% CI 0.2–4.5]). On multivariable logistic regression, after adjustment for the same characteristics, timing of hydrocortisone receipt was significantly associated with reduced ICU mortality when comparing hydrocortisone receipt within 0–6 h from shock onset to >48 h after shock onset (OR 0.6 [95% CI 0.4–0.8]).
Table 2 -
Clinical outcomes based on timing of hydrocortisone receipt
|
TotalN = 1470 |
HC 0–6 hN = 567 |
HC 6–12 hN = 231 |
HC 12–24 hN = 260 |
HC 24–48 hN = 195 |
HC >48 hN = 217 |
P value |
| Days alive and free from vasopressors at Day 28 |
1.3 (0–25.7) |
3.3 (0–26.2) |
1.9 (0–25.2) |
1.9 (0–25.7) |
0 (0–25.4) |
0 (23.4) |
0.39 |
| HC duration, days |
5.4±6.9 |
5.2 ± 8.0 |
5.5 ± 5.8 |
5.1 ± 4.7 |
6.0 ± 6.9 |
5.9 ± 6.9 |
<0.01 |
| ICU Mortality |
671 (45.65) |
244 (43.0) |
108 (46.8) |
120 (46.2) |
94 (48.2) |
105 (48.4) |
0.58 |
| Hospital Mortality |
768 (52.24) |
275 (48.5) |
121 (52.4) |
137 (52.7) |
107 (54.9) |
128 (59.0) |
0.10 |
| ICU LOS, days |
13.0 ± 15.0 |
11.1 ± 15.3 |
11.6 ± 12.4 |
13.2 ± 15.6 |
13.6 ± 12.5 |
18.9 ± 16.8 |
<0.01 |
| Hospital LOS, days |
22.4 ± 22.8 |
21.2 ± 26.2 |
20.6 ± 18.8 |
22.7 ± 21.0 |
22.5 ± 19.9 |
27.3 ± 21.3 |
<0.01 |
| MV duration, days |
9.3 ± 11.4 |
7.9 ± 10.9 |
8.0 ± 9.2 |
9.7 ± 11.6 |
9.0 ± 8.9 |
14.7 ± 15.0 |
<0.01 |
Data detailed as mean ± SD or n (%) unless otherwise indicated; P values derived from Kruskal–Wallis test for multiple independent samples.HC indicates hydrocortisone; ICU, intensive care unit; LOS, length of stay; MV, mechanical ventilation.
Table 3 -
Multivariable models
| Linear Regression - Dependent variable: days alive and free of vasopressors |
| Evaluation of timing based on categories |
Subgroup analysis of initiation within 24 h |
| Timing of HC |
Beta coefficient (95% CI) |
P value |
Timing of HC |
Beta coefficient (95% CI) |
P value |
| 0–6 h vs. >48 h |
2.75 (0.84–4.65) |
0.005 |
<24 h vs. >24 h |
1.90 (0.53–3.27) |
0.007 |
| 6–12 h vs. >48 h |
2.47 (0.23–4.71) |
0.03 |
|
|
|
| 12–24 h vs. >48 h |
2.31 (0.16–4.45) |
0.04 |
|
|
|
| 24–48 h vs. >48 h |
1.36 (−0.94–3.67) |
0.25 |
|
|
|
| Logistic regression - Dependent variable: ICU mortality |
| Evaluation of timing based on categories |
Subgroup analysis of initiation within 24 h |
| Timing of HC |
OR (95% CI) |
P value |
Timing of HC |
OR (95% CI) |
P value |
| 0–6 h vs. >48 h |
0.59 (0.41–0.85) |
0.004 |
<24 h vs. >24 h |
0.69 (0.53–0.90) |
0.006 |
| 6–12 h vs. >48 h |
0.59 (0.38–0.91) |
0.02 |
|
|
|
| 12–24 h vs. >48 h |
0.75 (0.50–1.13) |
0.17 |
|
|
|
| 24–48 h vs. >48 h |
0.82 (0.53–1.28) |
0.39 |
|
|
|
Adjusted for ICU location, lactate, presence of cortisol testing, norepinephrine dose, and APACHE III score.HC indicates hydrocortisone; ICU, intensive care unit; OR, odds ratio.
In the post hoc subgroup analysis comparing patients who received hydrocortisone within 24 h of shock onset to those who received it after 24 h, there were baseline differences in APACHE III scores (105.4 ± 34.0 within 24 h vs. 101.5 ± 32.3 after 24 h; P = 0.048), lactate (4.5 ± 4.0 mmol/L vs. 3.5 ± 3.4; P < 0.001) and norepinephrine dose (24.6 ± 21.8 mcg/min vs. 21.3 ± 21.0; P = 0.009) (Table 4). Patients who had hydrocortisone started within 24 h of shock onset had more days alive and free from vasopressors (median 2.3 days vs. 0 days; P = 0.08) and lower in-hospital mortality (50.4% vs. 57.0%; P = 0.02), but no difference in ICU mortality. ICU and hospital LOS were shorter in patients who had hydrocortisone started within 24 h of shock onset (10.1 ± 13.0 days vs. 13.7 ± 12.9; P < 0.001 and 16.3 ± 20.6 vs. 18.7 ± 15.9; P = 0.37, respectively). On multivariable linear regression (Table 3), after adjustment for ICU location, lactate, presence of cortisol testing, norepinephrine dose, and APACHE III score, receipt of hydrocortisone within 24 h was significantly associated with more days alive and free from vasopressors when comparing to initiation after 24 h from shock onset (beta-coefficient 1.9 [95% CI 0.5–3.3]). On multivariable logistic regression, after adjustment for ICU location, lactate, presence of cortisol testing, norepinephrine dose, and APACHE III score, receipt of hydrocortisone within 24 h was significantly associated with reduced ICU mortality when comparing to hydrocortisone receipt after 24 h from shock onset (OR 0.7 [95% CI 0.5–0.9]).
Table 4 -
Subgroup analysis of hydrocortisone receipted within 24 h from shock onset
|
TotalN = 1,470 |
Early HC(<24 h from NE)N = 1,058 |
Late HC(>24 h from NE)N = 412 |
P value |
| Age, years |
61 ± 15 |
61 ± 15 |
60.6 ± 13.7 |
0.32 |
| Male gender |
804 (54.7) |
577 (54.5) |
227 (55.1) |
0.85 |
| Weight, kg |
88.7 ± 32.3 |
87.6 ± 30.4 |
91.4 ± 36.8 |
0.047 |
| ICU location |
| Medical |
1,003 (68.2) |
709 (67.0) |
294 (71.4) |
0.13 |
| Neurosciences |
133 (9.1) |
94 (8.9) |
39 (9.5) |
|
| Surgical |
334 (22.7) |
255 (24.1) |
79 (19.2) |
|
| APACHE III |
104.3 ± 33.6 |
105.4 ± 34.0 |
101.5 ± 32.3 |
0.048 |
| APS |
88.7 ± 32.1 |
90.0 ± 32.7 |
85.6 ± 30.2 |
0.02 |
| Lactate at HC start, mmol/L |
4.2 ± 3.9 |
4.5 ± 4.0 |
3.5 ± 3.4 |
<0.001 |
| Cortisol |
203 (13.8) |
108 (10.2) |
95 (23.1) |
<0.001 |
| MAP, mm Hg |
66.4 ± 14.6 |
66.3 ± 14.8 |
66.6 ± 13.9 |
0.68 |
| NE dose at HC initiation, mcg/min |
23.7 ± 21.6 |
24.6 ± 21.8 |
21.3 ± 21.0 |
0.009 |
| Other vasopressors when HC start |
| 0 |
695 (47.3) |
486 (45.9) |
209 (50.7) |
0.021 |
| 1 |
634 (43.1) |
456 (43.1) |
178 (43.2) |
|
| 2 |
127 (8.6) |
103 (9.7) |
24 (5.8) |
|
| 3 |
14 (1.0) |
13 (1.2) |
1 (0.2) |
|
| Time from NE to HC, h |
26.1 ± 47.4 |
7.5 ± 6.5 |
73.9 ± 68.7 |
– |
| Outcomes |
| Days alive and free from vasopressors at Day 28 |
1.3 (0–25.7) |
2.31 (0–25.8) |
0 (0–24.8) |
0.082 |
| ICU mortality |
671 (45.7) |
472 (44.6) |
199 (48.3) |
0.20 |
| Hospital mortality |
768 (52.2) |
533 (50.4) |
235 (57.0) |
0.022 |
| ICU LOS, days∗
|
11.1 ± 13.1 |
10.1 ± 13.0 |
13.7 ± 12.9 |
<0.001 |
| Hospital LOS, days∗
|
17.0 ± 19.4 |
16.3 ± 20.6 |
18.7 ± 15.9 |
0.037 |
| MV duration, days∗
|
8.4 ± 10.2 |
7.5 ± 9.8 |
10.8 ± 10.7 |
<0.001 |
| HC duration |
5.4 ± 6.9 |
5.2 ± 6.9 |
6.0 ± 6.9 |
0.076 |
Data detailed as mean ± SD or n (%) unless otherwise indicated.
∗Evaluated time after shock onset.APACHE III indicates acute physiology and chronic health evaluation III; APS, acute physiology score; HC, hydrocortisone; ICU, intensive care unit; LOS, length of stay; MAP, mean arterial pressure; MV, mechanical ventilation; NE, norepinephrine.
DISCUSSION
Despite being more critically ill at baseline, patients who received hydrocortisone earlier had better clinical outcomes when compared to patients who received hydrocortisone later after shock onset. Specifically, days alive and free from vasopressors were increased, ICU and hospital mortality rates were lower, and ICU and hospital lengths of stay were shorter. After adjustment for multiple variables, including severity of illness (lactate, norepinephrine dose, and APACHE III score), starting hydrocortisone within 24 h of shock onset was significantly associated with increased days alive and free from vasopressors when compared to starting hydrocortisone >48 h after shock onset. Additionally, having hydrocortisone started within 12 h of shock onset was associated with reduced odds of ICU mortality. When evaluating the subgroup of patients before and after 24 h from shock onset, patients who had hydrocortisone initiated within 24 h had improved hospital mortality and more days alive and free from vasopressors despite being more severely ill at baseline.
The true mechanism and reason that early corticosteroid initiation, in this patient population, is beneficial is ultimately unknown. However, one evaluation, similar to the current study, found early initiation of hydrocortisone (within 9 h of shock onset), was independently associated with a favorable outcome (OR 0.40 [95% CI 0.20–0.81]) (15). Importantly, authors also found that production of tumor necrosis factor-α, a cytokine heavily involved in the host response to infection and implicated in the vasodilation occurring in sepsis, was found to be lower in patients who had hydrocortisone initiated early (15). It is likely that the immunomodulatory effects of hydrocortisone are a key contributing factor to the benefit of early initiation. Additionally, other interventions in septic shock have been shown to provide increased effectiveness with earlier initiation including antibiotics (20) and resuscitation (21), enforcing the idea that initiating therapeutic interventions in patients with sepsis and septic shock is integral.
The findings of the current trial correlate with prior evaluations of the timing of hydrocortisone initiation which favored early initiation (14–16). However, the timing of hydrocortisone initiation in randomized controlled trials is not consistent and this variable may be a contributor to the discrepant results seen. In 2002, Annane et al. (7) randomized 299 patients with septic shock to receive hydrocortisone 50 mg every 6 h and fludrocortisone 50 mcg for 7 days. To be included, patients had to be randomized within 3 to 8 h from shock onset and overall, the average time from vasopressor initiation to hydrocortisone receipt was 4.1 ± 3.4 h. Patients who received hydrocortisone and fludrocortisone in this study had shorter vasopressor duration (HR 1.54, 95% CI 1.10–2.16) and a trend toward improved ICU mortality (OR 0.61, 95% CI 0.37–1.02), with a significant mortality benefit in patients who were non-responders to the corticotropin stimulation test (OR 0.50, 95% CI 0.28–0.89) (7). In 2008, the Corticosteroid Therapy of Septic Shock (CORTICUS) study randomized 499 patients with septic shock to receive hydrocortisone 50 mg every 6 h followed by a taper over 11 days (8). To be included, patients had to have developed shock within the prior 72 h after randomization. The mean time from shock onset or vasopressor initiation to hydrocortisone initiation was not detailed; however, 79% of the hydrocortisone patients received it within 12 h. No differences in clinical outcomes were seen between patients who received hydrocortisone and those who received placebo (8). In comparing these studies, the focus for the disparate results has centered on differences in severity of illness of the included patients (22–24). However, the timing of hydrocortisone initiation has often been overlooked as a key factor contributing to discrepant findings.
Two more recent studies evaluating the use of hydrocortisone in patients with septic shock, the Adjunctive Corticosteroid Treatment in Critically Ill Patients with Septic Shock (ADRENAL) and the Activated Protein C and Corticosteroids for Human Septic Shock (APROCCHS) trials, also showed disparate results regarding mortality (6, 9). The APROCCHS study utilized hydrocortisone 50 mg every 6 h with fludrocortisone for 7 days and included 1,241 patients who had septic shock for less than 24 h (6). Patients randomized to receive hydrocortisone and fludrocortisone had reduced rates of death from any cause at 90 days (43.0% vs. 49.1%; P = 0.03) and more vasopressor-free days at day 28 (17 ± 11 days vs. 15 ± 11; P < 0.01). The timing from shock onset to hydrocortisone initiation in this study is not detailed, but was within at least 24 h (6). The ADRENAL study randomized 3,658 patients to receive a 200 mg per day of a continuous infusion of hydrocortisone for 7 days (9). There were no differences in mortality between patient groups. However, patients who received hydrocortisone had faster shock resolution (median 3 days [IQR 2–5] vs. 4 [2–9]; P < 0.01), as well as quicker time to ICU discharge (median 10 days [IQR 5–30] vs. 12 [6–42]; P < 0.01) compared with those who received placebo (9). The mean time from shock onset to randomization was 20.9 ± 91.9 h; the time from shock onset to hydrocortisone initiation was not specifically detailed. When evaluating patients based on the timing from shock onset to randomization, the only difference in death at 90 days was seen in patients who were randomized within 6 to 12 h from shock onset (OR 0.71, 95% CI 0.54–0.94) (9). It is important to note that in ADRENAL, no loading dose of hydrocortisone was administered. Because only a continuous infusion was started, it is anticipated that the time to steady state of hydrocortisone would be delayed, thus delaying the time to achieve therapeutic effect. Again, the disparate results seen in APROCCHS and ADRENAL could be due to differences in the timing of hydrocortisone initiation, the method of hydrocortisone administration, and/or the difference in underlying severity of illness in APROCCHS compared with ADRENAL.
In comparison with the aforementioned studies, the current evaluation included a severely ill patient population with a baseline APACHE III score of 104, a baseline norepinephrine requirement of 23.7 mcg/min at the time of hydrocortisone initiation, and an overall ICU mortality rate of 45.7%. This study found a temporal relationship between hydrocortisone use and beneficial outcomes, suggesting that if hydrocortisone is to be initiated in patients with septic shock, it should be initiated within at least the first 24 h after shock onset, and ideally within the first 12 h. The findings of the current study highlight the fact that the first 12 h of shock onset appear to be the most integral time to initiate hydrocortisone if it is to be utilized. Although clinical trials that have shown a benefit to hydrocortisone use have initiated it within the first 24 h after shock onset (6, 7), in clinical practice, hydrocortisone use is typically reserved for salvage therapy, in line with guideline suggestions (3), and the average time to its initiation in this study was 26.11 ± 47.37 h. Results of this study showed that use after 24 h does not appear to confer any benefit. Future randomized studies should focus on the timing of hydrocortisone initiation, ensuring initiation within the first 12 h from shock onset.
The significant differences in the various clinical trials evaluating hydrocortisone in patients with septic shock, including the timing of hydrocortisone initiation, the hydrocortisone dosing regimen, fludrocortisone use, as well as the severity of illness of the included patients have resulted in disparate recommendations for hydrocortisone's utility in patients with septic shock. The current evaluation adds to the literature by specifically analyzing the timing of the initiation of hydrocortisone, in a large patient population, which may be a significant factor determining the benefit of hydrocortisone in this population. This study is not without limitations: it is a retrospective evaluation that relies on accurate documentation in the medical record and runs the risk of intervention bias in which sicker patients may have received hydrocortisone earlier. The decision to initiate hydrocortisone, and at what time point, was solely based on provider discretion. Additionally, there may be confounders for which this study could not account based on the limited dataset.
CONCLUSION
This evaluation identified that, in patients with vasopressor-dependent septic shock who received hydrocortisone, initiation of hydrocortisone within the first 12 h appears to confer more benefit than initiation after 12 h. As this was a retrospective evaluation, further prospective data is warranted to definitively determine the exact time point at which hydrocortisone should be initiated. Based on the findings of this study, timing of hydrocortisone initiation in patients with septic shock appears to be crucial and hydrocortisone should be started within the first 12 h after shock onset.
Acknowledgment
The authors thank Eric Vogan, MSPH, for his assistance with data extraction from the electronic medical record.
REFERENCES
1. Kadri SS, Rhee C, Strich JR, Morales MK, Hohmann S, Menchaca J, Suffredini AF, Danner RL, Klompas M. Estimating ten-year trends in
septic shock incidence and mortality in United States Academic Medical Centers Using Clinical Data.
Chest 151 (2):278–285, 2017.
2. Vincent JL, Jones G, David S, Olariu E, Cadwell KK. Frequency and mortality of
septic shock in Europe and North America: a systematic review and meta-analysis.
Crit Care 23 (1):196, 2019.
3. Rhodes A, Evans LE, Alhazzani W, Levy MM, Antonelli M, Ferrer R, Kumar A, Sevransky JE, Sprung CL, Nunnally ME, et al. Surviving
sepsis campaign: international guidelines for management of
sepsis and
septic shock: 2016.
Intensive Care Med 43 (3):304–377, 2017.
4. Annane D.
Corticosteroids for severe
sepsis: an evidence-based guide for physicians.
Ann Intensive Care 1 (1):7, 2011.
5. Annane D, Pastores SM, Rochwerg B, Arlt W, Balk RA, Beishuizen A, Briegel J, Carcillo J, Christ-Crain M, Cooper MS, et al. Guidelines for the Diagnosis and Management of Critical Illness-Related Corticosteroid Insufficiency (CIRCI) in Critically Ill Patients (Part I): Society of Critical Care Medicine (SCCM) and European Society of Intensive Care Medicine (ESICM) 2017.
Crit Care Med 45 (12):2078–2088, 2017.
6. Annane D, Renault A, Brun-Buisson C, Megarbane B, Quenot JP, Siami S, Cariou A, Forceville X, Schwebel C, Martin C, et al. Hydrocortisone plus fludrocortisone for adults with
septic shock.
N Engl J Med 378 (9):809–818, 2018.
7. Annane D, Sebille V, Charpentier C, Bollaert PE, Francois B, Korach JM, Capellier G, Cohen Y, Azoulay E, Troche G, et al. Effect of treatment with low doses of hydrocortisone and fludrocortisone on mortality in patients with
septic shock.
JAMA 288 (7):862–871, 2002.
8. Sprung CL, Annane D, Keh D, Moreno R, Singer M, Freivogel K, Weiss YG, Benbenishty J, Kalenka A, Forst H, et al. Hydrocortisone therapy for patients with
septic shock.
N Engl J Med 358 (2):111–124, 2008.
9. Venkatesh B, Finfer S, Cohen J, Rajbhandari D, Arabi Y, Bellomo R, Billot L, Correa M, Glass P, Harward M, et al. Adjunctive glucocorticoid therapy in patients with
septic shock.
N Engl J Med 378 (9):797–808, 2018.
10. Rochwerg B, Oczkowski SJ, Siemieniuk RAC, Agoritsas T, Belley-Cote E, D’Aragon F, Duan E, English S, Gossack-Keenan K, Alghuroba M, et al.
Corticosteroids in
sepsis: an updated systematic review and meta-analysis.
Crit Care Med 46 (9):1411–1420, 2018.
11. Volbeda M, Wetterslev J, Gluud C, Zijlstra JG, van der Horst IC, Keus F. Glucocorticosteroids for
sepsis: systematic review with meta-analysis and trial sequential analysis.
Intensive Care Med 41 (7):1220–1234, 2015.
12. Sligl WI, Milner DA Jr, Sundar S, Mphatswe W, Majumdar SR. Safety and efficacy of
corticosteroids for the treatment of
septic shock: a systematic review and meta-analysis.
Clin Infect Dis 49 (1):93–101, 2009.
13. Minneci PC, Deans KJ, Eichacker PQ, Natanson C. The effects of steroids during
sepsis depend on dose and severity of illness: an updated meta-analysis.
Clin Microbiol Infect 15 (4):308–318, 2009.
14. Ibarra-Estrada MA, Chavez-Pena Q, Reynoso-Estrella CI, Rios-Zermeno J, Aguilera-Gonzalez PE, Garcia-Soto MA, Aguirre-Avalos G. Timing, method and discontinuation of hydrocortisone administration for
septic shock patients.
World J Crit Care Med 6 (1):65–73, 2017.
15. Katsenos CS, Antonopoulou AN, Apostolidou EN, Ioakeimidou A, Kalpakou GT, Papanikolaou MN, Pistiki AC, Mpalla MC, Paraschos MD, Patrani MA, et al. Early administration of hydrocortisone replacement after the advent of
septic shock: impact on survival and immune response.
Crit Care Med 42 (7):1651–1657, 2014.
16. Park HY, Suh GY, Song JU, Yoo H, Jo IJ, Shin TG, Lim SY, Woo S, Jeon K. Early initiation of low-dose corticosteroid therapy in the management of
septic shock: a retrospective observational study.
Crit Care 16 (1):R3, 2012.
17. Centers for Disease Control and Prevention. Hospital toolkit for adult
sepsis surveillance. Available at:
https://www.cdc.gov/sepsis/pdfs/Sepsis-Surveillance-Toolkit-Mar-2018_508.pdf. Accessed February 1, 2020.
18. Rhee C, Dantes R, Epstein L, Murphy DJ, Seymour CW, Iwashyna TJ, Kadri SS, Angus DC, Danner RL, Fiore AE, et al. Incidence and trends of
sepsis in US hospitals using clinical vs claims data, 2009-2014.
JAMA 318 (13):1241–1249, 2017.
19. Russell JA, Lee T, Singer J, De Backer D, Annane D. Days alive and free as an alternative to a mortality outcome in pivotal vasopressor and
septic shock trials.
J Crit Care 47:333–337, 2018.
20. Kumar A, Roberts D, Wood KE, Light B, Parrillo JE, Sharma S, Suppes R, Feinstein D, Zanotti S, Taiberg L, et al. Duration of hypotension before initiation of effective antimicrobial therapy is the critical determinant of survival in human
septic shock.
Crit Care Med 34 (6):1589–1596, 2006.
21. Rivers E, Nguyen B, Havstad S, Ressler J, Muzzin A, Knoblich B, Peterson E, Tomlanovich M. Early Goal-Directed Therapy Collaborative G. Early goal-directed therapy in the treatment of severe
sepsis and
septic shock.
N Engl J Med 345 (19):1368–1377, 2001.
22. Suffredini AF. A role for hydrocortisone therapy in
septic shock?
N Engl J Med 378 (9):860–861, 2018.
23. Annane D. Why my steroid trials in
septic shock were “positive”?
Crit Care Med 47 (12):1789–1793, 2019.
24. Venkatesh B, Cohen J. Why the adjunctive corticosteroid treatment in critically ill patients with
septic shock (ADRENAL) trial did not show a difference in mortality.
Crit Care Med 47 (12):1785–1788, 2019.
