DIAGNOSING AND TREATING CORTISOL INSUFFICIENCY IN ICU MOHD BASRI MAT NOR, IIUM, KUANTAN, MALAYSIA

DIAGNOSING AND TREATING CORTISOL INSUFFICIENCY IN ICU MOHD BASRI MAT NOR, IIUM, KUANTAN, MALAYSIA

Content Glucocorticoid physiology and effects of critical illness on HPA axis Assessment of tissue cortisol activity Controversies in adrenal function testing Lab diagnosis of adrenal insufficiency during critical illness Evidence based recommendations for the diagnosis and treatment of Relative adrenal insufficiency (RAI) Conclusion 2

Stress Response and HPA axis CORTISOL 3 PE Marik. Curr Opin Crit Care 2007

Total Plasma cortisol 1 ug/dl = 27.6 nmol/l Normal patients: 13.5 (5-24) ug/dl 95% bound to corticosteroid binding globulin (CBG) 5% as free plasma cortisol Free plasma cortisol is biologically active Routine assays measure only total cortisol levels Important Physiological Parameters 4

Cortisol and Glucocorticoid Receptor 5 NEJM 2005

Cortisol Cascade Plasma Interstitium Cytoplasm Nucleus 6 Best Pract Res Cin Endocrinol Metab 2011

Effects of Critical Illness Mechanisms known to affect tissue glucocorticoid activity 1. Change in Plasma free cortisol 2. Interstitial cortisol concentrations 3. 11β-hydroxysteroid dehydrogenase activity 4. Alterations in the glucocorticoid receptor 5. Glucocorticoid gene expression Cortisol cascade failure: dissociation between 1. Total vs. free plasma cortisol 2. Total and free cortisol vs. interstitial cortisol 3. Total plasma cortisol vs. pre-receptor changes 4. Plasma response vs. functional response of the tissues to stress 7

Assessing Adrenal activity Assessing adrenal insufficiency or relative adrenal insufficiency in critically ill patients is challenging Laboratory assays of plasma cortisol concentration and response to ACTH stimulation are unreliable in critically ill patients. 8

Mechanisms increasing and decreasing tissue glucocorticoid exposure 9 Best Pract Res Cin Endocrinol Metab 2011

Plasma free cortisol (PFC) Provides a better assessment of adrenal function than total cortisol PFC is biologically active, CBG and albumin levels decrease in critical illness, leading to increase in PFC PFC increments correspond to severity of illness in septic shock, total cortisol does NOT PFC concentrations cannot be predicted from total cortisol because of the non-linear relationship Total serum cortisol vary widely in critically ill patients Whether PFC level provides useful prognostic information in critically ill patients has not been formally investigated 10

Plasma Free Cortisol PFC levels may vary with changes in concentration or binding affinity of CBG. (Intensive Care Med 2001;27:1584) Hamrahian et al found that (NEJM 2004;350:1629) There was a greater relative increase in free compared with total plasma cortisol concentrations. Patients with hyporoteinemia had lower TPC concentration than patients without hypoproteinemia. Plasma free cortisol concentrations did not differ significantly between the two groups. Ho et al showed that in patients with sepsis or septic shock, plasma free cortisol concentrations better reflected illness severity than total plasma concentrations. (J Clin Endocrin Metab 2006). Cohen et al found in septic shock, there is dissociation between total and free cortisol both in response to corticotropin and over the course of illness. Shock 2012 11

Interstitial cortisol concentrations Neutrophil elastase, released from PMN leukocytes cleaves cortisol from CBG and increase interstitial levels Changes in interstitial fluid volume, increased capillary leakage and reduced peripheral tissue perfusion. 12

Critical Care 2009 Interstitial cortisol levels in patients with severe burns were measured by microdialysis Interstitial concentrations were elevated with respect to controls and poorly correlated with plasma free concentrations 13

Levels of interstitial cortisol obtained by microdialysis in patients with sepsis correlated only moderately with total plasma cortisol levels Suggesting plasma cortisol may not reflect tissue availability 14 Journal of Crit Care 2013; 28

11β-HSD system NEJM April 2013 11β-HSD system 11β-HSD 1 is reductase: converting inactive cortisone to cortisol. 11β-HSD 2 is dehydrogenase: converting cortisol to cortisone 15

Venkatesh et al published first data examining cortisol to cortisone ratios in critically ill. Intensive Care Med 2007 There is evidence of altered cortisol metabolism due to an increase in 11β-HSD activity as demonstrated by an elevation of plasma cortisol: cortisone ratios. This suggests an increase in intracellular cortisol generation independent of circulating plasma concentration 16

Total and free circulating cortisol levels were consistently higher in the patients than in controls, whereas corticotropin levels were lower Reduced cortisol breakdown, related to suppressed expression and activity of cortisol-metabolizing enzymes This finding has potential implications for the diagnosis of adrenal failure and its treatment in the ICU setting. NEJM April 2013 17

Adrenal Insufficiency Absolute AI among critically ill patients is rare, incidence is < 3% Suboptimal cortisol production during septic shock has been termed relative adrenal insufficiency or critical illness related corticosteroid insufficiency (CIRCI) There is no consensus about the diagnostic criteria or indication for treatment of RAI There is considerable disagreement over what level is appropriate in septic shock, what constitutes adequate response to ACTH and what dose of synthetic ACTH should be used 18

Relative Adrenal Insufficiency First proposed by Selye et al in 1940 Based on reduced plasma cortisol and a blunted cortisol response to corticotropin Changes in plasma cortisol reflect those in the interstitium, which in turn reflect the functional response at the cellular level? In critically ill, each of these criteria has significant limitations It is uncertain whether RAI is a true diagnostic entity A clear definition is lacking Cortisol assays that are available in most clinical laboratories are unreliable in the critically ill patients 19

Relative Adrenal Insufficiency This was supported by some important studies: 1. Jurney 1987 found that septic shock patients were either responders or non-responders to ACTH 2. The use of the short synacthen test became the cornerstone of RAI when Annane et al 2002 found that responders have better outcomes when treated with steroids than nonresponders e.g. patients with a cortisol level <9 μg/dl after 250 mcg synacthen do worst 20

Problems with RAI concept Leads to the prediction that patient s with lower levels of cortisol should have higher mortality in septic shock BUT the opposite is found! Short SynACTHen tests are nearly uninterpretable in septic shock patients what is an appropriate cut off? The RAI hypothesis is too simplistic free cortisol is biologically active, yet we only measure total cortisol (cortisol bind globulin levels change in sepsis) Response to cortisol may vary at any level: receptor levels and responsiveness, translocation into the nucleus, variations in the many hundreds of gene products that are transcribed 21

Suggested diagnostic criteria for RAI The diagnostic uncertainty has resulted in numerous published criteria for recognition of this entity Application of these criteria to the same patient population results in estimated incidence of 6.25-75% Surgical Intensive Care Med 2010 22

Variability of Total Plasma Cortisol Difficult to define a normal range for critically ill Wide range in stressed ICU patients: 400-1400 nmol/l Healthy volunteers: 200-450 nmol/l Trend toward mortality to be associated with higher cortisol concentrations. 23

Variability of TPC: Plasma cortisol rhythm The typical circadian rhythm of plasma cortisol level seen in healthy volunteers is absent Plasma cortisol had such marked variability that a random value had limited diagnostic utility 24 Anesthesia and Intensive Care 2005

Variability of TPC Time points of assessment Have been studied from as early as 8 h within onset of shock to 61 days after illness Cortisol concentrations vary over the time course of illness with peak levels in the early stage of sepsis and a nadir in the later phases 25 Best Pract Res Clin Endocrinol Metab 2011; 25(5)

Variability of TPC Sex-based differences in response Adrenocortical responses to stress differ between males and females, and between premenopausal and postmenopausal women In septic shock studies, male to female ratio varied from 0.8 to 3.4 26

Significant variability between cortisol assays Cohen et al first to report high degree of variability between assays (Intensive Care Med 2006) Potentially confound the diagnosis of RAI Erroneous labeling of an individual s adrenal status CORTICUS investigators reported variability in results when the same sample was assayed in different labs Resulting in 27% of patients being classified differently A patient may be potentially classified as hypo or euadrenal depending on the assay used Intensive Care Med 2009; 35(12) These inter-assay variations complicate the diagnosis of corticosteroid insufficiency. 27

Pharmacological Confounders Most data come were from Europe where the use of etomidate is common Etomidate inhibits adrenal steroidogenesis and may increase the incidence of adrenal insufficiency Data from countries where etomidate is not used clearly demonstrated a lower incidence of RAI in septic shock Jones et al. Anaesthesia and Intensive Care 2006;34(5) 28

Crit Care Med 2012;40(11) Meta-analysis Jan 1950 Feb 2012 Adrenal insufficiency determined by consyntropin stimulation test Seven studies addressed the development of adrenal suppression associated with the administration of etomidate; 1,303 subjects were included Etomidate administration increased the likelihood of developing adrenal insufficiency (pooled relative risk 1.33; 95% confidence interval 1.22 1.46) Conclusions: Administration of etomidate for rapid sequence intubation is associated with higher rates of adrenal insufficiency and mortality in patients with sepsis. 29 1.33(1.22,1.46)

30 Criticisms of the corticotropin stimulation test for diagnosing RAI: Lack of consensus on the appropriate change in cortisol after corticotropin simulation that indicates RAI: 250 nmol/l or 400 nmol/l Lack of consensus on the dose of corticotropin to be used: 1 μg or 250 μg 250 μg is supraphysiologic, Stimulates adrenal secretion of cortisol even when adrenal dysfunction exists. Use of low dose has been proposed In one study, Low dose test identified subgroup of patients with inadequate adrenal reserve that missed by high dose ACTH test

Corticosteroid Therapy Clinical evidence 1. French trial: Annane D, et al. Effect of treatment with low doses of hydrocortisone and fludrocortisone on mortality in patients with septic shock. JAMA. 2002 Aug MC DB RCT. n = 300 refractory septic shock patients in 19 French ICUs Low dose hydrocortisone (50 mg IV q6h) + fludrocortisone (50 mcg/d) for 7 days vs. placebo Outcomes: Non-responders to HD Synacthen test + given steroid, significant decrease in 28d mortality (P = 0.02): 53% vs 63% in favour of the intervention; but no difference at 1 year. Duration of vasopressor therapy shorter, more rapid reversal of shock, increased wound infections in the placebo group Commentary: High placebo group mortality Statistical methods used to describe outcomes Treatment commence within 8 hours of the onset of shock Use of etomidate Results were refuted by the CORTICUS study 31

Corticosteroid Therapy Clinical evidence 2. CORTICUS: Sprung CL, et al; CORTICUS Study Group. Hydrocortisone therapy for patients with septic shock. N Engl J Med. 2008 Jan 10;358(2) MC DB RCT 499 patients, Low dose hydrocortisone regardless of pressor dependency (50mg q6h tapered over 6 days) vs. placebo based on HD Synacthen test No survival benefit (32% vs 35%) regardless of patients responsiveness to stimulation test. No difference in whether or not shock was reversed (76% vs 70%) Shock reversed more quickly (2.8 vs 5.8 days) More superinfections, hyperglycaemia, hypernatremia Commentary and criticisms: 32 Stopped early due to slow recruitment Inadequately powered (target sample size was 800) Lower control death rate than expected No differences according to whether patients were responders or non-responders Enrolment was <72h compared with <8h in Annane 2002 Annane 2002 included sicker patients (based on SAPS scores) that had higher mortality

Comparison of French and CORTICUS Trials French: Enrolled patients within 8 hours, greater severity of patients that were vasopressor dependent. Septic shock was SBP < 90 mmhg despite adequate fluid resuscitation, or vasopressor administration for > 1 hour CORTICUS: Enrolled patients within 72hours of onset of shock, lower severity of illness. Septic shock was SBP < 90 mmhg despite adequate fluid resuscitation, or need for vasopressor for at least 1 hour Important methodological differences. 33

Corticosteroid Therapy Clinical evidence 3. COIITSS Study Investigators, et al. Corticosteroid treatment and intensive insulin therapy for septic shock in adults: a randomized controlled trial. JAMA. 2010 Jan MCRCT, n = 509 4 groups with septic shock: Intensive insulin therapy + hydrocortisone VS conventional insulin therapy + hydrocortisone vs. same treatments with fludrocortisone Inclusion criteria: adult, septic shock, MODS, hydrocortisone. Primary end points = hospital mortality and 90 day mortality Intensive Insulin Group double hypoglycaemic rate, no increase in mortality, no difference in secondary outcomes, no difference in synth-acth-en responders, no difference in fludrocortisone patients Fludrocoritisone Results: No increase in mortality, no difference in inotropes, excess superinfection rate Conclusions in septic shock patients treated with hydrocortisone there was: no benefit from intensive insulin therapy, it led to increased episodes of hypoglycemia no benefit from fludrocortisone, it led to increased rates of superinfection Criticisms: Did not reach required recruitment levels, not blinded, tested multiple 34 variables

Corticosteroid Therapy Clinical evidence 4. Low-dose steroids in adult septic shock: results of the Surviving Sepsis Campaign.Intensive. Care Med 2012 A large observational database of 17,847 patients who required vasopressor therapy despite fluid resuscitation Met eligibility criteria for low-dose systemic corticosteroids (hydrocortisone 50 mg intravenously every 6 hrs or 100 mg every 8 hrs) 50% received low-dose corticosteroids, and the hospital mortality was higher in those receiving corticosteroids than those without (41 versus 35 %) No significant difference in mortality was noted in those who received corticosteroids before 8 hours or between 8 and 24 hours. Criticism: The observational nature potentially limits the conclusions. 35

An approach to steroids in septic shock Pending results of adequately powered studies: Don t use in low risk patients Reserve for patients with severe septic shock (multi-organ failure), acknowledging that septic shock may reverse more quickly but will not change mortality Use low dose (hydrocortisone 50 mg q6h), no fludrocortisone Vigilance for super infection, hyperglycemia and myopathy. Discontinue steroids if patients fail to respond to treatment Don t use a short synacthen test Eagerly await the ADRENAL study! 36

The ADRENAL study protocol: ADjunctive corticosteroid treatment in critically ill patients with septic shock. Crit Care Resus June 2013 Background: There is considerable global uncertainty on the role of low-dose corticosteroids in septic shock, which translates into variations in prescribing practices. Objective: MC RCT in critically ill patients with septic shock, comparing the effects of hydrocortisone and placebo on 90-day mortality Methods: 3800 critically ill patients with septic shock treated in an ICU, to concealed, randomised, parallel assignment of hydrocortisone or placebo. The primary outcome will be all-cause mortality at 90 days postrandomisation. Secondary outcomes will include ICU and hospital mortality, length of ICU stay and quality of life at 6 months. Subgroup analyses will be conducted in two predefined subgroups. All analyses will be conducted on an intention-to-treat basis. Results and conclusions: The run-in phase has been completed and the main trial commenced in February 2013. The trial should generate results that will inform and influence prescribing of corticosteroids in septic shock. 37 ANZICS CTG INVESTIGATORS

Conclusion.. A functioning HPA axis is a prerequisite for survival from severe stress. In critical illness, HPA axis undergoes complex changes, and proportion of patients may have relative deficiency of cortisol or tissue cortisol resistance. Lab assays of plasma cortisol concentration and response to Synacthen test are unreliable in critically ill patients Accepted consensus on diagnosis of RAI or CIRCI is lacking Lack of a clearly defined plasma response in severe stress and the presence of adequate response at the cellular level suggest it is a sick euadrenal state, analogous to sick euthyroid state and not a sick adrenal indicating adrenal insufficiency The administration of low dose steroid is better reserved to severe shock whose BP is poorly responsive to fluid resuscitation and vasopressor therapy Awaiting for large adequately powered trial ADRENAL by the ANZICS group 38