EDITORIAL

Should we question if something works just because we don’t

know how it works?
Michael C Reade

In 1687, Sir Isaac Newton observed planetary movement and benefit was observed only in the subgroup of 229 patients
(by his own account, inspired by a falling apple1) postulated who mounted a “subnormal” response to corticotropin. How-
the existence of a force between two objects that is propor- ever, in a multivariable model, the interaction of corticotropin
Crit
tional to Care Resusc ISSN:
the product 1441-2772
of their 7 December
masses and inversely proportional responsiveness and steroid efficacy was not significant.6 A
2009 11 4 235-236
to the square of the distance between them.2 Although this subsequent larger (499 patient) multicentre trial (CORTICUS)7
©Crit Care Resusc 2009
did not exactly accord with later observations (Einstein’s theory
www.jficm.anzca.edu.au/aaccm/journal/publica- found 46.7% of patients had “subnormal” responses to
tions.htmrelativity was required to explain inconsistencies
of general corticotropin, but this was not a predictor of mortality. Moreo-
Editorial
such as the movement of the planet Mercury), Newton’s ver, there was no mortality benefit associated with low-dose
theory has been used as the basis of much successful human hydrocortisone in corticotropin responders, non-responders, or
endeavour. Despite this, humanity does not yet really under- overall. Reconciling these two trials reportedly presented great
stand gravity, having little idea of how gravitational force is difficulty to the authors of the most recent Surviving Sepsis
transmitted. Similarly, in other branches of science, observa- Campaign guidelines,6 with the eventual recommendation
tions can be put to good use without an understanding of the being that low-dose hydrocortisone is indicated for patients
mechanism underlying the observed phenomena. An obvious who are “poorly responsive to fluid resuscitation and vasopres-
analogy in medicine is the use of inhaled general anaesthetics sor therapy” (ie, similar to patients in the 2002 Annane trial),
— the mechanism of which is yet to be fully explained.3 but that the corticotropin test should not be used to identify
Gravity, anaesthesia, and many other observed but poorly patients likely to derive the greatest benefit. More recent
understood phenomena are nonetheless the valid subject of consensus guidelines from the American College of Critical
mechanistic research. Deeper scientific understanding may Care Medicine8 agree that the corticotropin test should not be
lead to more technological applications. In this issue of the used, and even suggest that the term “relative adrenal insuffi-
Journal, Venkatesh and Cohen4 present a valuable summary of ciency” should be abandoned in favour of “critical illness-
the evidence against the concept of “relative adrenal insuffi- related corticosteroid insufficiency”, for many of the same
ciency” as the foundation for the use of corticosteroids in reasons articulated by Venkatesh and Cohen.
human septic shock (page 301). They note that conventional In contrast, a recent meta-analysis (incorporating the CORTI-
measurements of total plasma cortisol level and the response CUS results)9 disagrees with the second of the implications of
to corticotropin do not reflect disease severity, and that Venkatesh and Cohen’s argument, finding that in 12 ran-
changes in the free fraction of circulating cortisol and cellular domised placebo-controlled trials of low-dose steroids there is
alterations (for example, in intracellular cortisol metabolism overall evidence for benefit in septic shock. Current
and receptor binding) are probably significant confounding recommendations6,8 reflect this conclusion. On the best availa-
factors in the interpretation of these tests. Previous studies of ble trial evidence, steroids appear to be beneficial, even if
plasma cortisol levels in sepsis showing an “inadequate” “relative adrenal insufficiency” or the “sick adrenal state” do
response may thus be misleading, as the response may be not exist. This begs the question: do we need to understand
entirely appropriate for the, as yet, ill-defined needs of the cell how something works to use it? Physicists using gravity in their
(“sick euadrenal syndrome” to paraphrase the term applied to calculations, and anaesthetists using volatile anaesthetics in
the thyroid gland under similar conditions). their practice, would argue not.
The argument advanced by Venkatesh and Cohen has two Those who believe in the efficacy of steroid supplementa-
implications: first, that measurement of total plasma cortisol is tion in septic shock may be disappointed that Venkatesh and
unlikely to be helpful in patients with septic shock, and Cohen find the articulated biological rationale for this
second, that pharmacological replacement of “deficient” cor- approach is flawed. However, extrapolating knowledge of
tisol is not warranted. Evidence from clinical trials and consen- biological abnormalities to guide therapy has not been particu-
sus opinion support the first of these two contentions. The larly successful in other aspects of critical care endocrinology.
landmark major trial of low-dose corticosteroids for septic Levels of growth hormone and insulin-like growth factor-1 are
shock was that of Annane et al in 2002:5 of 299 patients both reduced in critical illness (reviewed by Taylor and
randomly allocated to receive placebo or 50 mg hydrocorti- Buchman10), but supplementing growth hormone in critical
sone every 6 hours plus 50μg fludrocortisone daily, mortality illness roughly doubled the risk of death.11 Thyroid hormone

Critical Care and Resuscitation • Volume 11 Number 4 • December 2009 235

 EDITORIAL

levels are also decreased in critical illness,12 predicting mortal- Author details
ity,13 but, despite promising evidence in an animal model,14 Michael C Reade, Associate Professor1,2
thyroxine supplementation has been unsuccessful in limited 1 Intensive Care Unit, Austin Hospital, Melbourne, VIC.
clinical trials.15,16 Patients with sepsis have insulin resistance,17 2 University of Melbourne, Melbourne, VIC.
and, although under certain conditions intensive insulin ther- Correspondence: [email protected]
apy was beneficial,18 when widely implemented this strategy
was harmful.19
Many parameters measured in critical illness are likely to be References
effects of, or even appropriate compensations for, the disease 1 White M. Isaac Newton: the last sorcerer. 1st ed. London: Fourth Estate,
process, rather than mediators of adverse outcome. Critical 1997.
2 Newton I. Philosophiæ naturalis principia mathematica. 1st ed. London:
care physicians increasingly understand that noting something
Jussu Societatis Regiae ac Typis Josephi Streater, 1687.
is abnormal does not mean that fixing it will help. When this is 3 Sear JW. What makes a molecule an anaesthetic? Studies on the
forgotten, much effort is expended in achieving “euboxia” (all mechanisms of anaesthesia using a physicochemical approach. Br J
the boxes on the pathology print-out in the normal range), but Anaesth 2009; 103: 50-60.
4 Venkatesh B, Cohen J. Sick adrenal or sick euadrenal? Crit Care Resusc
the patient still dies. A favourite example demonstrating the
2009; 11: 301-304.
irrelevance of the epiphenomena of critical illness is the number 5 Annane D, Sebille V, Charpentier C, et al. Effect of treatment with low
of relatives gathered around a patient’s bed. Having more than doses of hydrocortisone and fludrocortisone on mortality in patients
two relatives is generally associated with a poorer prognosis — with septic shock. JAMA 2002; 288: 862-71.
but asking relatives to leave rarely helps. (Of course, having no 6 Dellinger RP, Levy MM, Carlet JM, et al. Surviving Sepsis Campaign:
international guidelines for management of severe sepsis and septic
relatives by the bedside when critically ill is also often a bad shock: 2008. Crit Care Med 2008; 36: 296-327.
prognostic sign — when the number of relatives does not 7 Sprung CL, Annane D, Keh D, et al. Hydrocortisone therapy for patients
increase appropriately with illness severity, perhaps this could with septic shock. N Engl J Med 2008; 358: 111-24.
be called a “relative relative insufficiency”.) The point is that our 8 Marik PE, Pastores SM, Annane D, et al. Recommendations for the
diagnosis and management of corticosteroid insufficiency in critically ill
incomplete understanding of the pathogenesis of critical illness adult patients: consensus statements from an international task force by
has meant that attempts to normalise what we observe have the American College of Critical Care Medicine. Crit Care Med 2008; 36:
rarely been beneficial — particularly in critical care endocrinol- 1937-49.
ogy. The finding that measured indices of corticosteroid func- 9 Annane D, Bellissant E, Bollaert PE, et al. Corticosteroids in the treatment
of severe sepsis and septic shock in adults: a systematic review. JAMA
tion may not reflect the potential benefit of supplemental 2009; 301: 2362-75.
steroids should therefore not be too disappointing. 10 Taylor BE, Buchman TG. Is there a role for growth hormone therapy in
Current international guidelines6,8 (admittedly lacking Aus- refractory critical illness? Curr Opin Crit Care 2008; 14: 438-44.
tralasian endorsement20) recommend steroids for all patients 11 Takala J, Ruokonen E, Webster NR, et al. Increased mortality associated
with growth hormone treatment in critically ill adults. N Engl J Med
who do not respond satisfactorily to fluid and vasopressors.
1999; 341: 785-92.
However, probably there is a subset of patients who do derive 12 Baue AE, Gunther B, Hartl W, et al. Altered hormonal activity in severely ill
benefit, and another subset who do not. If we could measure patients after injury or sepsis. Arch Surg 1984; 119: 1125-32.
the factors listed by Venkatesh and Cohen in our patients, we 13 Slag MF, Morley JE, Elson MK, et al. Hypothyroxinemia in critically ill
patients as a predictor of high mortality. JAMA 1981; 245: 43-5.
might be able to distinguish these two groups. A likely major
14 Inan M, Koyuncu A, Aydin C, et al. Thyroid hormone supplementation in
advance in intensive care medicine in the next decade will be sepsis: an experimental study. Surg Today 2003; 33: 24-9.
an improved ability to select the appropriate patients for our 15 Brent GA, Hershman JM. Thyroxine therapy in patients with severe
existing treatments. This selection might be based on pharma- nonthyroidal illnesses and low serum thyroxine concentration. J Clin
cogenomic information (for example, genetic polymorphisms Endocrinol Metab 1986; 63: 1-8.
16 Klemperer JD, Klein I, Gomez M, et al. Thyroid hormone treatment after
that identify patients who will respond to recombinant human coronary-artery bypass surgery. N Engl J Med 1995; 333: 1522-7.
activated protein C21), but probably also on a better ability to 17 Shangraw RE, Jahoor F, Miyoshi H, et al. Differentiation between septic
characterise individual patients’ cellular function. and postburn insulin resistance. Metabolism 1989; 38: 983-9.
The article by Venkatesh and Cohen should therefore be 18 van den Berghe, Wouters P, Weekers F, et al. Intensive insulin therapy in
the critically ill patients. N Engl J Med 2001; 345: 1359-67.
seen as a call to better explore abnormalities of corticosteroid
19 Finfer S, Chittock DR, Su SY, et al. Intensive versus conventional glucose
function at a cellular level, as well as to develop technology control in critically ill patients. N Engl J Med 2009; 360: 1283-97.
that can rapidly characterise these abnormalities in individual 20 Hicks P, Cooper DJ; The Australian and New Zealand Intensive Care
patients. In the meantime, though, the clinical trial evidence Society (ANZICS) Board and Clinical Trials Group Executive Committee.
The Surviving Sepsis Campaign: International guidelines for manage-
for low-dose steroids in septic shock should be appraised on its
ment of severe sepsis and septic shock: 2008. Crit Care Resusc 2008; 10:
merits, not discounted simply because the mechanism of any 6.
effect has not yet been discovered. 21 Russell JA, Walley KR, inventors. Plasminogen activator inhibitor-1 (PAI-1)
haplotypes useful as indicators of patient outcome. Canada patent
See also Cohen and Venkatesh (page 287). C12Q 1/68. March 19, 2004. ❏

236 Critical Care and Resuscitation • Volume 11 Number 4 • December 2009