Criteria abstracted from The Users' Guide to Medical Literature, from the Health Information Research Unit and Clinical Epidemiology and Biostatistics, McMaster University

Highlighted lines and questions below provide links to the pertinent description of criteria in The EBM User's Guide, now available at the Canadian Centres for Health Evidence

Article Reviewed:

Please visit the new Evidence Based Journal Club Reviews

Absolute and relative adrenal insufficiency in children with septic shock.

Pizarro CF, Troster EJ, Damiani D, Carcillo JA.

Crit Care Med. 2005 Apr;33(4):855-859. [abstract]

Reviewed by Angela A. Hsu MD, Michael J. Bell MD, and Christiane Corriveau MD, Children's National Medical Center, Washington, DC

Review posted August 24, 2005

I. What is being studied?

Study objective:

To evaluate the incidence of absolute and relative adrenal insufficiency and determine the relationship between adrenal function, the development of catecholamine-resistant septic shock, and outcome in the pediatric population.

Study design

A prospective blinded cohort study at a single institution. The authors, however, did not describe the blinding procedure in the article. Also, it was not clear if the patients with adrenal insufficiency were treated or not, and how they blinded the treatments.

II. Are the results of the study valid?

Note: These questions follow from Randolph AG et al. Understanding articles describing clinical prediction tools. Crit Care Med 1998;26:1603-1612. [abstract]

1. Was a representative and well-defined sample of patients at a similar point in the course of the disease? Was follow-up sufficiently long and complete?

Fifty-seven consecutive children admitted to the PICU with septic shock were enrolled. The American College of Critical Care Medicine (ACCM) guidelines were used for the diagnosis of septic shock. Patients were excluded if there was a history of HIV infection, adrenal insufficiency, and those receiving any steroid therapy or etomidate within the week prior to diagnosis. The risk of mortality, MOSF scores on admission, and the duration of shock indicated that the enrolled patients represented a wide range of severity of illness.

Seventy-four percent of the patients enrolled had chronic disease, with the majority of these being oncologic in nature. Given that steroids are a part of many chemotherapy regimens, the incidence of adrenal insufficiency may have been overestimated in this setting even if steroid therapy was not given within the week of diagnosis. This potential bias may be clarified if the duration since these patients' last glucocorticoid therapy was presented. It has been shown in children that the duration of glucocorticoid-induced inhibition of the hypothalamic-pituitary-adrenal axis may recover in as little as 1 month following discontinuation of their glucocorticoid therapy (1).

Patients were followed until death or discharge from the PICU. The exact cause of death was not explicitly reported in this study.

One of the objectives of this paper was to determine outcome in association with the presence or absence of adrenal insufficiency. It would have been interesting if the authors included which patients received steroid therapy and what dose did they give. Did treatment affect the overall mortality, the duration of shock, duration of inotrope support, PICU length of stay, or hospital stay?

2. Were all potential predictors included?

All patients enrolled in the study had a baseline cortisol level determined and a short corticotropin test was performed using 250 mcg of corticotropin within the first 24 hours after diagnosis of septic shock. Serum cortisol levels were obtained 30 and 60 minutes after the corticotropin dose. The peak cortisol level was defined as the maximum cortisol level at either 30 or 60 minutes and the amount increase was determined as the peak cortisol level minus the baseline level. Based on these values, the patients were classified into 4 groups.

• Group 1 (Absolute adrenal insufficiency): Baseline cortisol < 20 mcg/dL and increase ≤ 9 mcg/dL
• Group 2 (Relative adrenal insufficiency): Baseline cortisol ≥ 20 mcg/dL and increase ≤ 9 mcg/dL
• Group 3 (Adequate adrenal response with elevated baseline cortisol): Baseline cortisol ≥ 20 mcg/dL and increase > 9 mcg/dL
• Group 4 (Adequate adrenal response without an elevated baseline cortisol): Baseline cortisol < 20 mcg/dL and increase > 9 mcg/dL

These patients were also classified into 3 groups according to their need for cardiovascular support according to the guidelines established by the American College of Critical Care Medicine (ACCM).

• Group A (Fluid-responsive shock): Shock that is reversed by 60 cc/kg of isotonic crystalloid or colloid fluid resuscitation in 1 hr
• Group B (Fluid refractory Dopamine/Dobutamine-responsive shock): Shock that is reversed by 60 cc/kg of isotonic crystalloid or colloid fluid resuscitation and ≤ 10 mcg/kg/min of Dopamine and/or Dobutamine infusion
• Group C (Catecholamine-resistant shock): Shock that persisted despite the use of epinephrine or norepinephrine

The study demonstrated chronic disease, multi-organ dysfunction, and catecholamine-resistant shock predicted mortality outcome. It would be useful to stratify the group into different underlying chronic diseases. A larger study would be needed to identify subgroups of patients (i.e., patients with cancer, after various surgeries, and after organ transplant) that may be at higher risk of adrenal insufficiency.

3. Did the investigators test the independent contribution of each predictor variable?

Univariate and multiple logistic regression models were constructed using dependent variables of shock response to catecholamines and survival. A cut-off point of p < 0.1 in the univariate analysis was used for the multiple regression analyses. The stepwise backward selection procedure was used. Variables were kept in the model if p < 0.05 and/or when it was considered to be a confounder.

4. Were outcome variables clearly and objectively defined?

The main outcome variables (adrenal insufficiency, shock, and mortality) were clearly and objectively defined in this paper. All patients enrolled were classified into 3 groups of shock based on ACCM criteria. Patients were also classified according to their adrenal function (see above) using criteria determined by the authors after review of the current literature. As mentioned above, the blinding procedure used in this study was unclear. A description of the methods used to blind the clinicians involved in the diagnosis and treatment of the study patients may strengthen the validity of this study.

III. What are the results?

1. What is(are) the prediction tool(s)?

One of the main outcomes of this study was to determine the incidence of absolute and relative adrenal insufficiency in children. As discussed in this paper, the incidence of adrenal insufficiency is variable as a result of the multiple definitions of adrenal insufficiency that currently exist in the literature. According to the criteria of adrenal insufficiency used by the authors (see above), the incidence of adrenal insufficiency in this population was 44% (95% CI, 31.1%-56.9%). The overall mortality was 39%. In the 8 patients with fluid-responsive shock, the mortality was 0%. The mortality was higher in the Dopamine/Dobutamine-responsive shock group (17%, 2 of 12 patients). As expected, the mortality was highest in the catecholamine-resistant shock group (54%, 20 of 37 patients).

Clinical Characteristics in the 4 adrenal function groups (from Table 3, pg 857)

Variables	Absolute Adrenal Insufficiency (Group 1, n=10)	Relative Adrenal Insufficiency (Group 2, n=15)	Cortisol ≥ 20 mcg/dL and adequate response (Group 3, n=15)	Cortisol < 20 mcg/dL and adequate response (Group 4, n=17)	p value
Risk of mortality, %	17.6 (0.56-98.6)	10 (1.63-81.9)	10.8 (2.36-88.8)	6.7 (0.36-94.1)	.76
Admission MOSF score	2.5 (0-6)	2 (1-3)	2 (0-4)	1 (0-3)	.11
MOSF during PICU stay	3.5 (1-7)	2 (1-5)	2 (0-4)	1 (0-4)	.02
Duration of PPV, days	3.5 (0-25)	4 (0-14)	4 (0-17)	2 (0-46)	.92
Time shock, days	3.5 (1-11)	4 (0-9)	3 (0-12)	2 (0-16)	.86
PICU LOS, days	7 (1-37)	8 (0-20)	12 (0-18)	6 (1-49)	.68
Baseline cortisol, mcg/dL	15.3 (6-18)	37.8 (22-91)	35.1 (23-51)	14.4 (7-19)	< 0.001
Peak cortisol, mcg/dL	18.5 (6-23)	36 (23-51)	60.5 (37-80)	31 (22-52)	< 0.001

MOSF: multiple organ system failure score, PPV: positive pressure ventilation, LOS: length of stay

All of the patients with absolute adrenal insufficiency had catecholamine resistant shock and all patients with fluid-responsive shock demonstrated a corticotropin response > 9 mcg/dL (Chi-square test for trend was significant, p = 0.009). The Chi-square analysis for trend was not significant (p = 0.08) between overall mortality and adrenal function. MOSF, chronic disease, and catecholamine-resistant shock predicted death (p < 0.05) using univariate analysis. In the stepwise regression analysis, chronic disease and multiple organ failure at admission predicted outcome (p < 0.05).

2. How large is the likelihood of the outcome event(s) in a specified period of time?

Using adrenal insufficiency as a predictor for adverse outcomes, the investigators combined absolute and relative insufficiency (group 1 and 2) compared with patients with normal adrenal function (group 3 and 4).

Outcomes	Relative Risk	95 % CI	p value
Catecholamine-resistant shock	1.88	1.26-2.79	0.01
Death	1.72	0.97-3.06	0.12

This means that patients admitted for septic shock with evidence of adrenal insufficiency are twice as likely to develop catecholamine-resistant shock (RR = 1.88). Although mortality rate for catecholamine-resistant shock was very high (54%) in compared with fluid-responsive (0%) and catecholamine-responsive shock (17%), adrenal insufficiency per se does not predict mortality.

Adrenal function had no relation to ventilator days, duration of shock, MOSF score during PICU stay and length of PICU stay.

3. How well does the model categorize patients into different levels of risk?

Chi-square test for trend demonstrated a significant association between inotrope/vasopressor requirement group and different levels of adrenal function (p < 0.009).

Adrenal function	% with catecholamine resistant shock
Absolute adrenal insufficiency	100
Relative adrenal insufficiency	80
Cortisol ≥ 20 mcg/dL and adequate response	60
Cortisol < 20 mcg/dL and adequate response	30

The relationship between adrenal function and prognosis did not reach statistical significance. However, those patients with adrenal insufficiency (absolute or relative) revealed a trend for poorer prognosis.

• Poor Prognosis (50% and 53% mortality): Group 1 and Group 2, respectively
• Intermediate Prognosis (33% mortality): Group 3
• Good Prognosis (24% mortality): Group 4

4. How confident are you in the estimates of risk?

In this relatively small sample size, the estimated risk for developing catecholamine-resistant shock with associated adrenal insufficiency is 1.88 with a 95% CI 1.26-2.79. Given the lower limit of the confidence interval approaching one (no difference) in this sample size, confirmation of this study's findings in a larger study of pediatric patients with septic shock is warranted before widespread application of these findings in clinical practice.

IV. Will the results help me in caring for my patients?

1. Does the tool maintain its prediction power in a new sample of patients?

This was not tested in this study.

2. Are your patients similar to those patients used in deriving and validating the tool(s)?

Although the study population is critically ill children, there was a disproportionate amount of children with chronic illness (74%) with multiple organ failure at admission. Overall mortality rates in this study are higher (39%) than recent studies in pediatric patients (9%) with septic shock (2). Mortality was independently predicted by chronic illness or multiple organ failure (p < 0.05), not adrenal insufficiency. Findings in this patient population with co-morbid conditions may not reflect the general pediatric patient presenting with septic shock.

3. Will the results lead directly to selecting or avoiding therapy?

The results of this study reveal that the incidence of adrenal insufficiency in this pediatric patients population diagnosed with septic shock is relatively high (44% with a 95% CI, 31.1-56.9%) based on a cortisol response of ≤ 9 mcg/dL after a short corticotropin test. The study population included a majority of patients with chronic diseases that may have predisposed them to adrenal insufficiency as defined in this study. However, these findings confirm the existence of adrenal insufficiency in critically ill children previously reported. As a result of this data, we believe that a corticotropin stimulation test may improve the identification of critically ill children with relative adrenal insufficiency.

There was no mention in this study of physiologic or supraphysiologic replacement of corticosteroids. Effective clinical therapies cannot be determined by this study. To date, no studies have demonstrated an effect on mortality of response to corticotropin or hydrocortisone therapy. A larger clinical trial will be needed to determine if treating adrenal insufficiency would improve overall outcome.

4. Are the results useful for reassuring or counseling patients?

These results can help us categorize our patient into various groups based on their severity of adrenal insufficiency and cardiovascular support requirements. Since the mortality rates differ greatly depending on the amount of cardiovascular support, this information can help guide us in discussing prognosis with the family members of our patients.

References

1. Cunha Cde F, Silva IN, Finch FL. Early adrenocortical recovery after glucocorticoid therapy in children with leukemia. J Clin Endocrinol Metab. 2004 Jun;89(6):2797-802. [abstract]
2. Carcillo JA, Fields AI, Task Force Committee Members. Clinical practice parameters for hemodynamic support of pediatric and neonatal patients in septic shock. Crit Care Med 2002;30:1365-1378. [abstract]

 

Comments

Back to the EB Journal Club Index