Comparative Effectiveness of Less vs. More Frequent Hydrocortisone Dosing in Septic Shock

Abstract

OBJECTIVES:

Guidelines recommend hydrocortisone as an adjunctive treatment in septic shock, but the optimal dosing regimen is unknown. A national shortage of hydrocortisone in 2023 prompted a change in institutional practice for hydrocortisone administration from 50 mg every 6 hours to 100 mg every 12 hours in an effort to reduce waste and conserve vials, creating an opportunity to evaluate the comparative effectiveness of these two regimens. The primary efficacy outcome was time to shock resolution, and secondary outcomes evaluated in this study were mortality, renal replacement therapy (RRT), medication costs, and maximum vasopressor dose attained.

DESIGN:

Single-center, retrospective cohort study.

SETTING:

ICUs in a quaternary academic medical center.

PATIENTS:

Adult patients admitted to an ICU with septic shock, defined by mean arterial pressure less than 65 mm Hg despite adequate fluid resuscitation and need for vasopressor infusion, who were treated with hydrocortisone for shock between October 24, 2022, and October 12, 2023.

INTERVENTIONS:

Treatment with hydrocortisone 50 mg every 6 hours or 100 mg every 12 hours.

MEASUREMENTS AND MAIN RESULTS:

One hundred thirty-eight patients were included in this retrospective chart review from October 24, 2022, to October 12, 2023. Data for 61 patients in the 50 mg every 6 hours group and 77 patients in the 100 mg every 12 hours group were collected and analyzed. In adjusted competing risk models, hydrocortisone regimen was not associated with differences in time to shock resolution (sub-hazard ratio [sub-HR] 0.95 [95% CI, 0.59–1.54]), ICU mortality (sub-HR 1.59; 95% CI, 0.89–2.84), in-hospital mortality (1.35; 95% CI, 0.81–2.26), or time to RRT (sub-HR 1.01; 95% CI, 0.45–2.31). In addition, the hydrocortisone dose regimen was not associated with differences in maximum vasopressor dose attained (mean difference in norepinephrine equivalent, 0.16 µg/kg/min; 95% CI, –0.26 to 0.58 µg/kg/min). The less frequent dosing resulted in cost savings of $446.10 (95% CI, 253.95–638.25) per patient treated with the more intensive but less frequent hydrocortisone dosing regimen.

CONCLUSIONS:

A less frequent hydrocortisone dosing regimen was not associated with differences in time to shock resolution. Studies of the comparative effectiveness of different corticosteroid dosing regimens for septic shock are needed.

KEY POINTS

Question: Does hydrocortisone 50 mg every 6 hours have comparative effectiveness with hydrocortisone 100 mg every 12 hours when administered as stress dose steroids for patients experiencing septic shock?

Findings: In 138 critically ill sepsis patients, hydrocortisone given 50 mg every 6 hours as compared with 100 mg every 12 hours was not associated with differences in time to shock resolution, renal replacement therapy, maximum vasopressor utilization, or mortality.

Meanings: A less frequent dosing regimen of hydrocortisone may yield equivalent clinical outcomes when compared with a more frequent regimen, but would result in substantially less medication waste, nursing time, and cost.

Hydrocortisone plays a critical role in maintaining hemodynamic homeostasis during septic shock. Its glucocorticoid effect reduces the body’s dysregulated, inflammatory response to infection, reduces nitric oxide production, increases vasopressor responsiveness, and replaces the diminished cortisol consequent of a suppressed hypothalamus-pituitary-adrenal axis. Hydrocortisone administration during septic shock has been consistently associated with a shorter time to shock reversal (1–4). Based on these and other data, the Surviving Sepsis Campaign and Society of Critical Care Medicine recommend empiric administration of stress dose steroids to patients experiencing septic shock. Although a dose of hydrocortisone 200 mg/d is recommended, the optimal regimen is unknown (5). A common regimen is 50 mg every 6 hours, but as hydrocortisone is supplied in 100 mg single-use vials, this approach results in significant medication waste, increased cost, and hospital staff workload (6).

METHODS AND MATERIALS

Study Design, Setting, and Ethics

This was a single-center, retrospective cohort study conducted at Baylor St. Luke’s Medical Center, an 850-bed quaternary care hospital in Houston, TX. We examined septic shock patients admitted to the ICUs between October 24, 2022, and October 12, 2023, who received hydrocortisone 100 mg every 12 hours or 50 mg every 6 hours. The study was conducted in accordance with the ethical standards of the CommonSpirit Institutional Review Board and with the Helsinki Declaration of 1975. The Comparative Effectiveness of Less vs. More Frequent Hydrocortisone Dosing in Septic Shock study was approved by the CommonSpirit Health Institutional Review Board on April 3, 2024 (IRBNet ID: 2115081-1). The authors designed the study, gathered and analyzed the data, and vouch for the completeness and accuracy of the data and the analysis.

Inclusion and Exclusion Criteria

Patients were eligible for inclusion if they displayed clinical evidence of infection with septic shock, defined by mean arterial pressure (MAP) less than 65 mm Hg necessitating vasopressor therapy and serum lactate level greater than 2 mmol/L in the absence of hypovolemia (7) before the initiation of hydrocortisone as stress dose steroids. Patients were excluded if they were younger than 18 years old, were of a protected population (pregnant, incarcerated), or received any form of corticosteroid 24 hours before the initiation of the studied hydrocortisone regimen.

Data Collection

The following patient data were recorded via electronic medical record: general patient characteristics, demographic data, interventions, including type and dose of vasopressor, amount of crystalloid fluid given before hydrocortisone administration, renal replacement therapy (RRT), and corticosteroid regimen administered. Laboratory data collected, when available, included serum lactate, cortisol levels, and cosyntropin test values. Appropriateness of antimicrobial therapy was not extracted as part of our dataset. Although we agree that patient outcomes are influenced by this factor (had a strong causal relationship to the outcome), we did not think this was causally related to the type of hydrocortisone regimen chosen, as the latter was principally influenced by timing, and therefore was not a significant confounder.

Study Definitions and Patient Management

The degree of illness for sepsis patients was measured at the time of shock diagnosis via the National Early Warning Score (NEWS) (8). ICU mortality was also estimated in all patients using the Acute Physiology and Chronic Health Evaluation (APACHE) II score at the same time (9). Active hydrocortisone treatment was defined from initiation of the studied stress dose corticosteroid regimen (hydrocortisone 50 mg IV every 6 hr, hydrocortisone 100 mg IV every 12 hr) to the beginning of tapering or discontinuation.

Primary and Secondary Outcomes

The primary outcome was time to resolution of shock, defined as maintenance of MAP greater than 65 mm Hg without vasopressor support for 24 hours. We additionally examined ICU and in-hospital mortality, maximum vasopressor requirements measured in norepinephrine equivalent (NEE) (10), and RRT among those not already on dialysis.

Statistical Analysis

Patient characteristics and outcomes are summarized using mean with sd (for normally distributed variables), median with 25th and 75th percentiles (for skewed variables), and frequency with percentages. T-test and Wilcoxon rank sum tests (for continuous normally distributed and skewed variables), and chi-square and Fisher exact tests (for categorical variables without and with rare cells) were used to compare characteristics and outcomes by hydrocortisone dose regimen.

Multivariable Fine-Gray competing risk models estimated the time to shock resolution and time to RRT with death as a competing risk, and multivariable Cox proportional hazards models estimated the time to ICU death and in-hospital death. Discharge from the ICU was considered a censoring event for the model of ICU death, and discharge from the hospital was considered a censoring event for all other outcomes. Of note, we considered patients who transitioned to hospice as having died. Although the choice of hydrocortisone regimen was principally driven by time period, on the premise that some physicians may override recommendations based on patient acuity, we adjusted for factors related to illness severity and likelihood of death: age, sex, body mass index, comorbid cirrhosis, heart failure, or kidney disease, and nearest lactate, vasopressor dose, and total fluids received before initiation of hydrocortisone.

A sample size calculation was not performed as the sample size was constrained by the timing of the hydrocortisone regimen.

Differences in cost for each regimen were calculated based on the current average wholesale price of $25.60, obtained from Micromedex REDBook (IBM, 2023, Ann Arbor, MI) (11).

Statistical analysis was performed using Stata v.18 (StataCorp LLC, 2023, College Station, TX). A two-sided p value of less than 0.05 was used to define statistical significance.

RESULTS

Baseline Characteristics

A total of 138 septic shock patients received scheduled hydrocortisone during the study period (Fig. 1). The average patient age was 61 years, 45% were male, and 44% were White. The median total body weight was 87 kg, and the median BMI was 31 kg/m². The average MAP and heart rate at the initiation of stress dose steroids were 65 mm Hg and 94 beats per minute (Table 1).

Figure 1.

This flowchart describes the retrospective patient screening and exclusion during the chart review process.

TABLE 1.

Patient Characteristics

Characteristic	50 mg Every 6 hr (n = 61)	100 mg Every 12 hr (n = 77)	p
Age, yr (mean, sd)	62.4 (14.8)	60 (12.5)	0.303
Gender, n (%)			0.864
Male	27 (44.3)	36 (46.8)
Female	34 (55.7)	41 (53.2)
Ethnicity, n (%)			0.311
Non-Hispanic White	26 (42.6)	36 (46.8)
Non-Hispanic Black	21 (34.4)	18 (23.4)
Non-Hispanic other	4 (6.5)	3 (3.9)
Hispanic	10 (16.4)	20 (26)
Known comorbidities, n (%)
Cirrhosis	22 (36.1)	43 (55.8)	0.026
Chronic obstructive pulmonary disease	12 (19.7)	25 (32.5)	0.122
Diabetes mellitus	39 (63.9)	45 (58.4)	0.599
Heart failure	37 (60.7)	39 (50.6)	0.302
Kidney disease, n (%)			0.685
Chronic dysfunction	20 (32.8)	20 (26)
End-stage renal disease with renal replacement	14 (23)	21 (27.3)
Body mass index (kg/m2) (median, IQR)	28.4 (22.9, 35.1)	31 (25, 37.3)	0.106
Acute Physiology and Chronic Health Evaluation II score (mean, sd)	9.6 (3.4)	10.6 (3.8)	0.096
National Early Warning Score (mean, sd)	8 (3.3)	8.1 (3.6)	0.870
Lactate (mmol/L) (median, IQR)	4.5 (2.9, 7.4)	4.3 (2.6, 5.7)	0.741
Total crystalloid 24 hr before vasopressor, mL (median, IQR)	1750 (500, 3500)	1000 (250, 2500)	0.123
Albumin given, n (%)	29 (47.5)	37 (48.1)	0.95
Time from vasopressor initiation to hydrocortisone initiation, hr (median, IQR)	11.2 (3.6, 33)	11.2 (3.7, 35.5)	0.956
Vasopressor dose at hydrocortisone initiation, norepinephrine equivalent, mg/kg/min (median, IQR)	0.2 (0.2, 0.3)	0.3 (0.1, 0.4)	0.382
Cortisol, µg/dL (median, IQR)
Random (n = 84/138)	14.9 (10.9, 24.5)	13.6 (11.4, 18.6)	0.471
Change with cosyntropin (n = 36/138)	5.5 (4.3)	7.1 (5.4)	0.343

IQR = interquartile range.

Clinical Characteristics of Separate Treatment Arms

During the study period, 61 patients received hydrocortisone 50 mg every 6 hours IV, and 77 patients received hydrocortisone 100 mg every 12 hours IV. Among the comorbidity data collected, patients with known cirrhosis were more likely to have been treated with hydrocortisone 100 mg bid (36% vs. 56%) (p = 0.026). With respect to the severity of illness, the mean NEWS and APACHE II scores collected at sepsis diagnosis were similar between the two groups. The average MAP and lactate collected at the time of hydrocortisone initiation were both clinically similar between the two study groups (Table 1).

Association of Hydrocortisone Regimen With Resolution of Shock

The median total duration of hydrocortisone treatment at “stress dose steroid” doses was 2.6 days for 50 mg every 6 hours vs. 2.0 days for hydrocortisone 100 mg every 12 hours (p = 0.164), the median total duration of hydrocortisone administration (including taper) was 4.4 days vs. 4.0 days (p = 0.043), and the median total hydrocortisone received including taper was 700 mg vs. 625 mg (p = 0.266), respectively. The median time from vasopressor initiation to hydrocortisone initiation was approximately 11 hours in both groups (p = 0.956).

There was no statistically significant difference in the time to shock resolution when comparing hydrocortisone 50 mg IV every 6 hours to hydrocortisone 100 mg IV every 12 hours. The sub-hazard ratio (sub-HR) for shock resolution was 0.95 (95% CI, [0.59–1.54]) (p = 0.847) (Tables S1 and S2, https://links.lww.com/CCX/B548) (Table 2).

TABLE 2.

Association of Hydrocortisone Regimen With Study Outcomes

Outcome	Unadjusted	p	Adjusteda	p
	Hazard or Sub-hazard Ratio (95% CI), Associated With 100 mg twice daily vs. 50 mg four times daily
Shock resolution	0.87 (0.57 to 1.32)	0.510	0.95 (0.59 to 1.54)	0.847
Renal replacement	0.99 (0.52 to 1.89)	0.984	1.01 (0.45 to 2.31)	0.976
ICU mortality	1.48 (0.88 to 2.50)	0.14	1.59 (0.89 to 2.84)	0.12
In-hospital mortality	1.32 (0.81 to 2.14)	0.262	1.35 (0.81 to 2.26)	0.249
	Mean Difference Associated With 100 mg Twice Daily vs. 50 mg Four Times Daily
Maximum vasopressor dose (mg/kg/min norepinephrine equivalent)	0.23 (–0.26 to 0.73)	0.352	0.16 (–0.26 to 0.58)	0.454

^aAdjusted for age, gender, body mass index, comorbidities (e.g., diabetes mellitus, hypertension, heart failure, cirrhosis), National Early Warning Score, Acute Physiology and Chronic Health Evaluation II score, lactate, and volume of fluids received.

Association of Hydrocortisone Regimen With Secondary Clinical Outcomes

Of the 138 participants, 60 died in the initial ICU encounter, and an additional 9 died during the hospitalization. Eighty-seven patients were not already receiving dialysis at the time of hydrocortisone initiation, and of these, 36 were started on RRT. The hydrocortisone regimen was not associated with time to ICU mortality (sub-HR 1.59; 95% CI, 0.89–2.84), in-hospital mortality (sub-HR 1.35; 95% CI, 0.81–2.26), or initiation of RRT (sub-HR 1.01; 95% CI, 0.45–2.31). There was no significant difference in maximum vasopressor attained (mean difference in NEE, 0.16 µg/kg/min; 95% CI, –0.26 to 0.58 µg/kg/min).

Association of Hydrocortisone Regimen With Medication Costs

The per-patient cost of hydrocortisone was $681.21 (95% CI, 470.61–891.8) for the every 6 hour regimen and $235.11 (95% CI, 189.04–281.18) for the every 12 hour regimen, resulting in a cost savings of $446.10 (95% CI, 253.95–638.25) per patient treated with the more intensive but less frequent hydrocortisone dosing regimen (Table S2, https://links.lww.com/CCX/B548).

DISCUSSION

In a single-center retrospective cohort of patients admitted to an ICU with septic shock, less frequent provision of hydrocortisone at the same total daily dose of 200 mg was not associated with differences in time to shock resolution and associated clinical outcomes, including mortality, need for RRT, or maximum severity of shock. To our knowledge, this is the first study that has evaluated the comparative effectiveness of these dosing regimens.

The dosing regimen selected in the ANNANE trial dosing interval of every 6 hours, was influenced by hydrocortisone’s relatively short half-life of 2 hours; however, patients experiencing septic shock have a reduced glucocorticoid clearance rate (12). The pharmacokinetic difference between our studied regimen and the historical regimens may be clinically insignificant. The onset of IV hydrocortisone action is immediate, and higher doses may be beneficial for patients’ severe adrenal insufficiency secondary to a combination of septic shock and comorbidities that may predispose a patient to low serum cortisol levels. In comparison to the ADjunctive coRticosteroid trEatment iN criticAlly ilL Patients With Septic Shock (ADRENAL) and Corticosteroid Therapy of Septic Shock - Corticus (CORTICUS) trials, our study’s time to shock reversal was similar and resulted in approximately three days. Further, the vasopressor requirements cited to be the threshold for starting corticosteroids in our study closely matched the thresholds stated in the Annane and Activated Protein C and Corticosteroids for Human Septic Shock (APROCCHSS) trial protocols (1–4).

National shortages (13) have compelled healthcare systems to find alternative methods to provide adjunctive steroids for patients experiencing septic shock. Expert guidance suggests alternative regimens including IV glucocorticoids combined with enteral fludrocortisone at equipotent doses administered at the same dosing intervals referenced in the 2021 Surviving Sepsis guidelines (14). However, among a myriad of other metabolic abnormalities, patients experiencing septic shock receiving vasopressors often experience inadequate splanchnic perfusion to intestinal organs secondary to the shunting of blood to vital organs, negatively impacting the absorption and metabolism of enteral medications such as fludrocortisone (15). Considering the evidence supporting hydrocortisone’s ability to reverse shock when compared with other corticosteroids alone, there is a paucity of literature validating the efficacy and safety of using alternative corticosteroids with an orally administered mineralocorticoid as adjunctive treatment for septic shock (1–4, 16).

We found that the reduced dosing regimen reduced costs. While not explicitly measured, we expect that such a regimen would also be associated with parallel reductions in waste and increased staff efficiency through less nursing demand needed to administer a more frequent dosing regimen. However, some limitations are noted. First, we did not examine whether empiric antibiotics started during episodes of shock were ultimately appropriate based on microbiological data, which, if unbalanced between hydrocortisone arms, would differentially bias the outcome. Second, because the recommended hydrocortisone dosing regimen was changed hospital-wide, there is potential for secular biases. However, participants appeared to be similar with respect to comorbidities, and we performed comprehensive adjustment for known confounders. Second, because data were obtained from chart review, there may be misclassification of some variables, although this would likely not be differential between the two groups. Finally, as a single-center study, our results may not be generalizable to other systems and settings, and our sample size was limited to the period of hydrocortisone shortage, preventing comparisons of more granular differences in inpatient outcomes.

CONCLUSIONS

Among patients with septic shock who initiated stress dose corticosteroids, hydrocortisone administered as 100 mg every 12 hours compared with 50 mg every 6 hours was not associated with differences in clinical outcomes but was associated with reduced cost. Future studies are needed to evaluate for differences in inpatient outcomes and adverse events with different steroid dosing regimens.

Supplementary Material

cc9-7-e1316-s001.pdf ^{(121.5KB, pdf)}