Society for Endocrinology - a world-leading authority on hormones

The Endocrinologist

Issue 130 Winter 2018

Endocrinologist > Winter 2018 > Features

Redefining the cortisol stress response

Sirazum Choudhury, Tricia Tan & Bernard Khoo | Features

The concept of stress derives from the pioneering work of Hans Selye. He took the term from physics to refer to a ‘non-specific response of the body to any demand’, where the body adapts initially to any condition which threatens to perturb homeostasis (‘the stressor’) via a co-ordinated and stereotyped response. This response depends on the acute release of neurotransmitters from the autonomic nervous systems as well as hormones from the adrenal cortex, adrenal medulla, pituitary and other endocrine glands.


A central part of this stress response is the secretion of cortisol, mediated by corticotrophin-releasing hormone/arginine vasopressin and adrenocorticotrophin (ACTH).1 From an endocrinologist’s point of view, patients with hypoadrenalism are unable to mount the stress response and are in danger of death from adrenal crises. Therefore, an understanding of the cortisol stress response is crucial to informing our treatment of hypoadrenal patients, based on the time-honoured principle of prescribing additional steroid replacement in patients who are undergoing situations of physiological stress, such as illness or surgery, to mimic the physiological levels observed in these situations.

The classical understanding of the cortisol stress response comes from the seminal work of Plumpton and Besser in 1969.2 Twenty euadrenal patients were recruited to undergo an insulin tolerance test (ITT), prior to elective surgery that would now retrospectively be considered as at least major surgery, under the Physiological and Operative Severity Score for the Enumeration of Mortality and Morbidity (POSSUM) criteria.

All 20 had plasma cortisol levels measured at fixed intervals during their ITT and subsequent surgery. As their surgery was uncomplicated, with cortisol levels responding appropriately to surgical stress, their ITT results formed the reference against which ITTs have been interpreted over the last nearly 50 years (peak cortisol >580nmol/l, with a maximal increment of >150nmol/l from baseline).

Through the years, these cut-off s have been adjusted down to reflect the differences in assay used, with guidelines now recommending 500nmol/l as a cut-off.3

Using surgical models as employed by Plumpton and Besser permits the simplest and most reproducible approach to understanding the clinical importance of the stress response.


The cortisol response to surgery was assessed in a recent meta-analysis of 71 studies between 1990 and 2016 by Prete et al.4 Their survey shows that the severity of surgery does influence the cortisol stress response, where minimally invasive procedures produce no peri-operative cortisol peak and increasingly invasive procedures cause correspondingly larger cortisol peaks.

Following major surgery, cortisol levels are elevated for up to 7 days, although this is based on more limited data. Open surgery and the use of general anaesthesia also increase the cortisol response. They found that older patients and female patients had higher cortisol responses.

However, the conclusions that can be drawn from these meta-analytic results are limited by the highly heterogeneous nature of the studies included, in terms of patient selection, peri- and post-operative care, anaesthesia and relatively few patients undergoing minimally invasive surgery. Importantly, the long time base of the meta-analysis means that the results are influenced by considerable advances in anaesthetic techniques, surgical approaches and quantification of cortisol through the last three decades.

To consider the last point in more detail, cortisol assays have graduated from being highly non-specific assays exhibiting large positive biases (for example the fluorometric assay used by Plumpton and Besser) to become increasingly more specific immunoassays which now align reasonably well to reference methods such as gas chromatography-mass spectrometry and liquid chromatography-mass spectrometry.5,6

Ideally, mass spectrometry would be used to ensure high specificity for cortisol and to reduce interference from related molecules such as 11-deoxycortisol, but this technique is not widespread and immunoassays on analyser platforms remain the mainstay in most clinical chemistry laboratories, due to their low cost and speed of analysis.

Another shortcoming of most studies on the surgical stress response is that the response of cortisol-binding globulin (CBG) is not studied. Some 80% of the total serum cortisol that is measured is bound to CBG (10% to albumin), and CBG is a key regulator of free (and hence bioavailable) cortisol levels. Levels of CBG fall during physiological stress, including (crucially) surgery,7 leading to elevations in free cortisol. This suggests that even if total cortisol is unchanged, a reduction in CBG may lead to an increase in bioavailable cortisol.

‘A true appreciation of the cortisol stress response to surgery will require a study which utilises validated methods for directly measuring free cortisol, for example equilibrium dialysis.’


Taking these factors into account, we carried out a study where we examined 93 euadrenal patients undergoing elective surgery in a single centre.8 We deliberately recruited patients with a wide range of surgical interventions, classified according to the POSSUM scale. Examples of procedures were as follows.

  • ‘Minor’ procedures included parathyroidectomy and diagnostic laparoscopy
  • ‘Moderate’ procedures included laparoscopic appendicectomy and cholecystectomy, open thyroidectomy and open hernia repair
  • ‘Major’/‘Major+’ procedures included cardiothoracic procedures, such as coronary artery bypass grafts as well as open major abdominal surgery.

We studied the cortisol response using a modern immunoassay aligned to reference methods (Abbott Architect), as well as measuring the CBG response to surgery in 83 patients. The magnitude of the cortisol stress response correlated positively with surgical severity, although the cortisol levels for ‘Minor’ procedures tended to fluctuate about patients’ respective morning baseline values, suggesting that, for these procedures, there is no marked cortisol stress response in most patients. We also found that total cortisol levels, even with ‘Major’ or ‘Major+’ surgery, often fell to baseline levels by post-operative day 1.


Our first take-home message is that a stratified approach to peri-operative adrenal replacement, as recommended by the Endocrine Society,9 may be a good approach to obviate adverse effects from over-treatment, especially in patients undergoing ‘Minor’ procedures. Based on the known pharmacokinetics of hydrocortisone, we believe that smaller induction doses of 25mg intramuscularly or orally for ‘Minor’/‘Moderate’ surgery and 50mg intramuscularly for ‘Major’/‘Major+’ surgery should be able to cater for any cortisol requirements with a safety margin.

Our second take-home message is based on our observation that CBG levels fall with surgery by up to 40% or so, and the magnitude of the fall is positively correlated with surgical severity. When bioavailable cortisol is estimated using the free cortisol index (FCI), we found little change in FCI with ‘Minor’ procedures, doubling of FCI with ‘Moderate’ surgery and tripling of FCI with ‘Major’ procedures. This suggests that the increase in bioavailable cortisol may be much larger than might be appreciated from looking at total serum cortisol. It may also explain the paradox of total serum cortisol levels falling to baseline levels in many patients on postoperative day 1: the free cortisol is likely to be higher than pre-operatively, given a fall in FCI and albumin levels after surgery. A true appreciation of the cortisol stress response to surgery will require a study which utilises validated methods for directly measuring free cortisol, for example equilibrium dialysis.

We found that the range of peak cortisol responses to ‘Major’/‘Major+’ surgery can vary widely, from 375 to 1452nmol/l in our dataset. Plumpton and Besser showed that in their cohort the cortisol range for their surgical procedures was higher by 30‒38% (equivalent to 607‒2070nmol/l). Whilst some of this large difference is likely to be due to the marked positive bias exhibited by the 1960s assay, some of it is also due to refinements in anaesthetic and surgical technique. As noted before, our cut-offs for judging adequacy of a cortisol response to ITT (and by extension to short synacthen test, SST) are based on this data, as the minimum adequate ITT cortisol response was 580nmol/l. The third take-home message, therefore, is that our criteria for judging the adequacy of ITT and SST response may need to be revised downwards. Indeed, the work of El-Farhan et al. suggests that the lower normal limit of the response to SST may be as low as 416‒430nmol/l for an Abbott Architect assay.5

‘Despite nearly 50 years of study, the cortisol stress response continues to be poorly understood in many ways. Further research is essential to develop the evidence base for the proper treatment of our hypoadrenal patients.’


Finally, one common question that is often asked of us by patients relates to the requirement for additional doses during times of psychological stress.

Emotional stress is cited as a precipitant of adrenal crisis in a large number of patients, for example 1 in 6 in a survey of hypoadrenal patients.10 Unlike the model of surgical stress that has been used to develop our current guidelines, psychological stressors are difficult to standardise and have been less well studied.

Modelling mental stress in healthy volunteers can be done by looking at individuals during a stressful period such as important exams. A study of 36 Spanish medical graduates undertaking career-defining exams measured salivary cortisol during the preparation, exam and post-exam periods.11 Anxiety during the study in each individual was tracked using a validated questionnaire. The measured salivary cortisol was noted to be significantly increased during the exam period, compared with the post-exam period when anxiety scores had normalised.

Other studies have, however, not shown similar findings, and the variability in responses may be due to some extent to methodological issues, such as the use of different stressors and the measurement of salivary cortisol. The use of measurements of serum cortisol or plasma ACTH is much less common in this research sphere, but may provide more consistent results.12

At present, there are no recommendations on the management of psychological stress. Patients are not routinely advised that they should change their regimen during such periods, although, anecdotally, patients do often take extra doses.

Despite nearly 50 years of study, the cortisol stress response continues to be poorly understood in many ways. The current guidance for management of surgery in hypoadrenal patients that is available from different expert groups varies, with some advocating far higher doses of hydrocortisone than others.3,9 Further research into the physiology of stress is essential to develop the evidence base for the proper treatment of our hypoadrenal patients.

Sirazum Choudhury & Tricia Tan, Department of Endocrinology, Imperial College Healthcare NHS Trust, and Department of Investigative Medicine, Imperial College London

Bernard Khoo, Division of Medicine, University College London


  1. Lim CT & Khoo B 2000 Endotext
  2. Plumpton FS & Besser GM 1969 British Journal of Surgery 56 216‒219.
  3. Husebye ES et al. 2014 Journal of Internal Medicine 275 104‒115.
  4. Prete A et al. 2018 Clinical Endocrinology 89 554‒567.
  5. El-Farhan N et al. 2013 Clinical Endocrinology 78 673‒680.
  6. Hawley JM et al. 2016 Clinical Chemistry 62 1220‒1229.
  7. le Roux CW et al. 2003 Journal of Clinical Endocrinology & Metabolism 88 2045‒2048.
  8. Khoo B et al. 2017 Clinical Endocrinology 87 451‒458.
  9. Bornstein SR et al. 2016 Journal of Clinical Endocrinology & Metabolism 101 364‒389.
  10. Hahner S et al. 2015 Journal of Clinical Endocrinology & Metabolism 100 407‒416.
  11. Gonzalez-Cabrera J et al. 2014 Stress 17 149‒156.
  12. Lopez-Duran NL et al. 2014 Stress 17 285‒295.

This Issue:

Winter 2018

Winter 2018