Arterial bicarbonate may be a useful indicator of inadequate cortisol response in children with catecholamine resistant septic shockMB Maralihalli, CT Deshmukh
Department of Pediatrics, Seth G.S. Medical College and KEM Hospital, Mumbai, Maharashtra, India
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0022-3859.113820
Source of Support: None, Conflict of Interest: None
Objective: To study the clinical and biochemical parameters that can predict cortisol insufficiency in children with septic shock. Design: prospective, observational study. Setting: tertiary health-care center. Patients/Subjects: Fifty children admitted with the catecholamine resistant septic shock to a tertiary health-care center. Materials and Methods: At the time of hospitalization all patients underwent detailed clinical evaluation including, history and physical examination, evaluation with the complete blood count, serum cortisol, renal function tests, liver function tests, prothrombin time activated partial thromboplastin time, arterial blood gas analysis, urine analysis, chest roentgenogram, ultrasonography of the abdomen and chest, urine, and blood culture for bacteria and fungi. Results: Out of 50 children with the catecholamine resistant septic shock, seven had adrenal insufficiency (serum cortisol <18 μg/dl). Of all parameters studied, only arterial bicarbonate at the time of admission to intensive care predicted adrenal insufficiency. On Receptor operative characteristic curve analysis, a bicarbonate level of 10.9 mEq/L had the best accuracy to predict adrenal insufficiency. Conclusion: Arterial bicarbonate may be used as a rapid test for provisional identification of adrenal insufficiency among children with the catecholamine resistant septic shock.
Keywords: Adrenal insufficiency, children, septic shock
Adrenal insufficiency is associated with an increased mortality in children with septic shock.  Adrenal insufficiency is a common and underdiagnosed disorder that develops in the critically ill patients.
There is consensus on the use of intravenous hydrocortisone in adult septic shock patients after blood pressure is identified to be poorly responsive to fluid resuscitation and vasopressor therapy;  however, use of hydrocortisone in all pediatric catecholamine resistant septic shock patients is controversial. A retrospective study from a large administrative database recently reported that the use of any corticosteroids in children with the severe sepsis was associated with an increased mortality.  The surviving sepsis guidelines recommend reserving the use of hydrocortisone in children with catecholamine resistance and suspected or proven adrenal insufficiency.  Patients at risk for adrenal insufficiency include children with severe septic shock and purpura, children who have previously received steroid therapies for chronic illness, and children with pituitary or adrenal abnormalities. Identifying patients with an inadequate cortisol response among the catecholamine resistant septic shock children without features suggestive of risk for adrenal insufficiency may be crucial for appropriate use of hydrocortisone.
In most centers from India, random serum cortisol cannot be obtained on an emergency basis to identify patients with adrenal insufficiency. Evaluation for relative adrenal insufficiency is further difficult due to limited availability of cosyntropin in India. Hence, it would be interesting to identify factors that can predict adrenal insufficiency in catecholamine resistant septic shock children. Hence, we have studied the clinical and biochemical parameters that can predict cortisol insufficiency in this group of children.
The study was conducted in the Department of Pediatrics of a Tertiary Care Hospital. The study was approved by the Institutional Ethics Committee. Informed consent was taken from the guardians of all participants and assent was obtained from all participants aged ≥7 years. It was a prospective study including 50 consecutive children who presented with catecholamine resistant septic shock without any risk factors for absolute adrenal insufficiency. Systemic inflammatory response syndrome, sepsis, and septic shock were defined according to international pediatric sepsis consensus conference criteria.  Fluid refractory shock (shock not responding to isotonic saline or colloid boluses up to and over 60 ml/kg) was treated with dopamine or dobutamine. When shock did not improve with these measures, it was called as Fluid refractory, dopamine/dobutamine resistant shock. Subsequently shock was treated with titrated doses of epinephrine for cold shock, norepinephrine for warm shock if shock persisted despite these measures; it was labeled as Catecholamine-resistant shock. Children at risk for adrenal insufficiency like children with severe septic shock and purpura, children who have previously received steroid therapies for chronic illness, and children with pituitary or adrenal abnormalities were excluded from the study.
A detailed clinical evaluation including history, general examination, and systemic examination was carried out. At the time of hospitalization, all patients underwent laboratory evaluation including, the complete blood count, renal function tests, liver function tests, coagulation parameters (prothrombin time, activated partial thromboplastin time), arterial blood gas analysis, urine analysis, chest roentgenogram, ultrasonography of the abdomen, and chest, urine, and blood culture for bacteria and fungi. In addition to the above routine investigations all the children were tested for a random serum cortisol by using a solid phase, competitive chemiluscent enzyme immunoassay (Immulite 1000, Seimens, LosAngelis; assay sensitivity: 0.5 μg/dl; intra assay coefficient of variation: 5.3-11.5%, inter assay coefficient of variation: 8.9-11.5%).
All patients were managed according to the surviving sepsis guidelines.  Data were analyzed using the SPSS version 16.1. Independent t test is used for the comparison of the group with adrenal insufficiency with that of adrenal sufficiency. Pearson's correlation coefficient is used to find out correlation between various parameters with the serum cortisol levels. Receptor operative characteristic (ROC) curve analysis is done to identify the arterial bicarbonate level that can best distinguish children with adrenal insufficiency from those with adrenal sufficiency. P<0.05 was considered significant.
The mean age of patients was 2.76±3.6 years (range: 1 month-12 years). Infants constituted 42% of total cases while 84% of children were less than 5 years of age. There were 26 girls and 24 boys. Fever was the most common symptom (82%) while the respiratory and gastrointestinal system related complaints were the other common symptoms.
Out of 50 patients, eleven had pneumonia, six had meningitis, six had a urinary tract infection, five had gastroenteritis, two had pyoderma, two had cellulitis, two had congenital heart disease with cerebral abscess, eight had proven septicemia without focus and eight had suspected sepsis.
Subjects were divided into two groups based on serum random cortisol using cut offs of 18 μg/dl [Table 1]. There was the trend towards significance with respect to age, those with adrenal insufficiency being younger. There were no significant differences between the groups with respect to grade of malnutrition, pediatric risk mortality (PRISM) score, Glasgow coma scale score, number of involved systems, and duration of intensive care units (ICU) stay duration of mechanical ventilation, dyselectrolytemia or derangement of renal/liver functions. The only parameter, which differed significantly between the groups was arterial bicarbonate value at the time of admission. Overall mortality was 24% and did not differ between the groups.
There were no significant correlation between any of the parameters and random serum cortisol levels except with arterial bicarbonate (Pearson's correlation coefficient = 0.295; P: 0.03) though all parameters were more severely affected in patients with adrenal insufficiency. On stepwise multiple regression analysis, arterial bicarbonate at the time of admission accounted for 9.8% of variation and was the only significant predictor of random serum cortisol levels.
In ROC curve analysis [Figure 1] to identify random serum cortisol of less than 18 μg/dl, arterial bicarbonate of 10.9 mEq/l provided the best sensitivity (85.7%) and specificity (85.7%).
In this study, arterial bicarbonate significantly predicted inadequate cortisol response in children with catecholamine resistant septic shock. Prediction of adrenal insufficiency using the arterial bicarbonate, which is a rapid and commonly available test in most of the pediatric ICU, may help in the appropriate use of corticosteroids in septic shock children.
A retrospective study from a large administrative database which reported increased mortality with the use of any corticosteroids in children with severe sepsis alerted pediatricians against its routine use in this situation.  This study suggested that use of corticosteroids may be limited to severe sepsis. Surviving sepsis guidelines recommended that due to lack of data in children and potential risk, steroids should not be used in those children who do not meet minimal criteria for adrenal insufficiency. 
Adrenal insufficiency may be absolute or relative. There are no strict definitions of adrenal insufficiency. Surviving sepsis guidelines suggest a random total cortisol concentration <18 μg/dL as absolute adrenal insufficiency in the case of catecholamine-resistant septic shock. An increase in serum cortisol of ≤9 μg/dL after 30- or 60-min of adrenocorticotrophic hormone (ACTH) stimulation has been used to define relative adrenal insufficiency. 
Most of studies have used an ACTH stimulation test as a standard test to assess adrenal insufficiency during septic shock. ,,] Septic shock (stress) itself is a strong enough stimulus for the full-fledged expression of normally functioning hypothalamic pituitary adrenal axis. Hence, an individual with septic shock and a normally functioning hypothalamic pituitary adrenal axis should produce an adequate amount of ACTH and cortisol as a stress response. Defect of hypothalamic pituitary adrenal axis at any level will in turn result in low random serum cortisol and whatever the level of the defect the management remains the same. In countries like India, the availability of cosyntropin is limited. Hence, use of random cortisol alone may have to be relied upon to identify adrenal insufficiency. Moreover, the treatment of relative adrenal insufficiency in children with septic shock is controversial.
One study has shown an association of hypercholesterolemia with hypocortisolism in children with meningococcal sepsis. In that study, total cholesterol, high density lipoprotein (HDL) cholesterol, and Low density lipoprotein (LDL) cholesterol levels on admission were inversely associated with disease severity. However, we have not studied lipid profile of our patients. 
Another study has shown that the most severely ill children with the septic shock had signs of adrenal insufficiency. In that study, age and mechanical ventilation were independent predictors of cortisol to ACTH ratio, but on multivariate regression analysis Interleukin IL-6 and intubation with etomidate were the only predictors of cortisol to ACTH ratio, but none of the clinical or routine biochemical investigations did. However, that study did not identify a single clinical or routine biochemical parameter, which can predict adrenal insufficiency. 
Adrenal insufficiency is usually associated with hyponatremia, hyperkalemia, hypoglycemia, and metabolic acidosis. Unlike, arterial bicarbonate, serum electrolytes, and blood glucose levels did not predict adrenal insufficiency in our cohort. Since, our center is a tertiary care center, most of children with septic shock had received glucose containing intravenous fluids before reaching us. This might be responsible for the lack of correlation between blood glucose at ICU admission and adrenal insufficiency.
Our study has few limitations. First, we have not carried out ACTH stimulated cortisol increase to diagnose relative adrenal insufficiency. Second, since all children were managed according to the survival sepsis guidelines, all children with adrenal insufficiency received 50 mg/m 2 /24 h of intravenous hydrocortisone. Hence, the effect of hydrocortisone treatment on the outcome of children with adrenal insufficiency could not be studied.
In conclusion, arterial bicarbonate could be a potential surrogate marker for adrenal insufficiency in septic children with catecholamine resistant shock. Its utility for this purpose should be investigated in further larger prospective studies. We suggest arterial bicarbonate of <10.9 meq/L at admission to be used as a cut-off value to identify adrenal insufficiency in this population.