Physiological effects of corticosteroid therapy

This chapter is related to Section U2(i) and U2(v) from the 2017 CICM Primary Syllabus, which expects the exam candidate to "understand the pharmacology of glucocorticoids" and "understand the pharmacology of mineralocorticoids". After some agonising self-examination, the author concluded that it would serve the readers' psychological safety better to have this section separate from the rest, instead of grafting the entire corticosteroid pharmacology summary together into one 10,000-word abomination. Whereas the chapters on adrenal physiology deal with the physiological regulatory effects of endogenous cortisol and aldosterone, the following summary pertains to the effects of potent exogenous corticosteroid drugs, which are treated as a completely separate topic mainly because that is how it tends to appear in exams. Having said this, the CICM exam is notably without any questions on steroid side effects, which means most readers can safely stop reading at the end of this little grey box:

Links with specific articles are offered below under the relevant headings, which is the authors' cowardly method of outsourcing all the responsibility for explaining the detailed pathophysiological mechanisms to scientists and clinicians who actually understand them. The effects are obviously many, and it seems like each of them has had a group of dedicated investigators, which is something we can benefit from. The harried CICM trainee will of course have no time for any of that, and they can be referred to a single free article such as Fardet et al (2007) which covers all the adverse effects of steroids under a single structure. It is also only about twelve pages, making it manageable. 

Immunosuppressant effects of corticosteroids

It would have probably been criminal to leave this at the bottom of this chapter, as it is clearly the main reason people tend to use corticosteroids, and so it should occupy a position of prominence. And there is probably no better way to enhance that prominence than by namedropping a prominent author. Hence, it is probably worth mentioning with feigned nonchalance that the structure and content of the material below was mainly informed by an excellent article about this topic by none other than Anthony S. Fauci, written in 1978, many years before his war against the Sith. Obviously a few things have happened in the last fifty years, and more recent papers such as Baschant & Tuckermann (2010) are also available to shed light on this matter, but it never hurts to reference the classics, especially considering that the scientific progress into steroid immunology over those decades has mainly involved nightmarish contortions of molecular biology which the CICM exam candidate is not expected to remember anyway. The basic observations about how steroids affect the immune system have remained substantially the same since 1978.

Anyway, you could probably split the immune effects of corticosteroids into two broad groups, one of which is the effects on the innate immunity (which could be described as their "anti-inflammatory" activity) and the effects on specific immunity which relate more to lymphocyte function. 

• Effects of corticosteroids on innate immunity:
  • Prevention of neutrophil migration toward inflammatory loci
  • Impaired phagocytosis by neutrophils
  • Prevention of neutrophil degranulation
  • Suppression of the cytotoxic and phagocytic functions of monocytes, and induction of an anti-inflammatory phenotype which do not react to normal stimulatory cytokines and which perform largely janitorial duties (eg. consuming apoptotic neutrophils)
• Effects of corticosteroids on specific immunity:
  • Dendritic cell apoptosis, and induction of a "tolerogenic" phenotype, which is described using the excellent term "anergy", meaning that these cells fail to react to normally immunogenic stimuli
  • Lymphopenia, and specifically the selective depletion of T-helper cells from the circulation, which is mainly due to their sequestration in lymphatic tissue
  • Decreased B cell numbers and suppressed IgG production
  • Decreased function of lymphocytes on all levels, including cell activation, proliferation, differentiation, mediator production and release, response to mediators, antigen recognition, cytotoxic effector functions, antibody synthesis, and whatever else you could think of. 

What the hell does this mean, exactly:

• Positive or desirable effects:
  • Decreased tissue destruction, swelling, oedema and pain at the site of acute inflammation, mainly due to the reduced activity of neutrophils
  • Decreased fibrosis by reduced cytokine-mediated traffic of fibroblasts
  • Decreased production of destructive autoantibodies because of depressed B cell number and function
  • Increased tolerance for transplanted organs because of interrupted T-cell signalling and depressed T-cell function
• Negative or undesirable effects:
  • Increased risk of infection, which seems to be both dose-related and duration-related
  • Opportunistic infections, such as Aspergillus, Pneumocystis, Strongyloides, etc
  • Reactivation of suppressed latent infectious diseases, such as diseases that depended on macrophage function for suppression (tuberculosis) or those which required cytotoxic T cell function to remain under control (eg. CMV, hepatitis B and C)
  • Reduced immune response to vaccines, with the possibility that the vaccine will be entirely ineffective

Effects of steroids on airway oedema

After the shouting and blame-pointing is finished when a patient is emergently reintubated for post-extubation stridor, junior ICU doctors are often asked to prescribe some corticosteroids (usually dexamethasone) to reduce the vocal cord oedema and promote a more successful extubation attempt on the following day. The use of corticosteroids to reduce airway oedema is actually a very old-school approach that appears to have existed for many decades, and it has come to the ICU mainly from otolaryngology practice, where it was used for basically any nonspecific vaguely inflammatory swelling. Hawkins et al (1983) describe brutal ancient steroid rituals going back to the 1960s, for example for the management of epiglottitis and laryngotracheobronchitis, but also less clearly inflammatory processes like vocal cord trauma. This appears to be some sort of subacute effect, in the sense that most authorities would recommend waiting for at least 24 hours before trying something new with the swollen airway. This is on the basis of the observation that studies giving steroids 30-60 minutes prior to a planned extubation all failed to demonstrate any reduction in the incidence of stridor, whereas those that started steroids 12-24 hours earlier were almost uniformly successful. 

Steroids as bronchodilators

Putting aside their obvious immunomodulatory effects on the functions of mast cells and IgE-mediated hypersensitivity, steroids also seem to have direct effects on bronchial smooth muscle reactivity. Hirst & Lee (1998) outlined a lot of the proposed mechanisms, which range from direct effects on contractile intracellular proteins, effects on the availability of intracellular calcium, the effect of uncoupling smooth muscle vasoconstrictor receptors (such as histamine) and potentiation of relaxation by the increase of Na⁺/K⁺ ATPase activity and catecholamine sensitivity. It appears these effects are observed over the timeframe of hours, which suggests that this is a genomic effect, dependent on protein transcription. Long-term use tends to take over the processes of smooth muscle cell proliferation, which is where steroids help prevent the unhelpful chronic hyperproliferative response that makes chronic airflow limitation irreversible.

 Additionally, steroids reduce the blood flow to the bronchial mucosa, decreasing the hyperaemia of inflammation and infection. Various experiments have demonstrated that this is some kind of rapid, non-genomic effect, mediated probably by membrane-bound glucocorticoid receptors. The effect of inhaled fluticasone, for example, was very acute for Kumar et al (2000), with the nadir of mucosal blood flow reached around 30 minutes after the dose. 

Cardiovascular effects of corticosteroids

An ancient paper by Weil (1962) is still the best description (if not explanation) of what happens to haemodynamics and cardiovascular performance in response to systemic steroid therapy.  Broadly, the intensivist would describe these as "everything better", whereas a cardiologist would describe them as "everything worse". Clinically, steroids produce:μ

• Increased cardiac output, mainly due to an increase in contractility
• Increased peripheral vascular resistance
• Thus, increased blood pressure
• Thus, decreased heart rate (usually, i.e. by a reflex mechanism)
• Accelerated AV conduction
• Increased preload (due to mineralocorticoid effects)

A lot of the haemodynamic effects are owed to the increased sensitivity of various key components of the cardiovascular system to catecholamines. In the presence of steroids, noradrenaline works harder. Ullian (1999) outlines the experiments which have confirmed and defined this effect, and Boyer et al (2006) explain some of the cellular mechanisms (not that they are particularly well-understood). The bottom line is that vessel and cardiovascular reactivity to catecholamines is increased within minutes to hours, by some sort of non-genomic mechanism. In case you are interested, this effect seems to dissipate after about 36 hours following the cessation of steroids, according to some early human experiments by Kurland & Freedberg (1951).

Additionally, vessel tone seems to improve by some other non-catecholamine-related mechanism. Nobody seems to know the exact way this happens, but some tantalising information has come from experiments by d'Emmanuele di Villa Bianca et al (2006), who tested the effects of dexamethasone on vascular reactivity. From these, it appears that the steroid had immediately antagonised the action of ATP-sensitive potassium channels (normally responsible for regional blood flow autoregulation in arterioles). This means that steroids might be acting basically as anti-hydralazine, exerting something of a direct vasopressor (or at least counter-dilator) effect. 

Furthermore corticosteroids seem to defend the circulatory system from a lot of the nastiness of systemic inflammatory responses. Of particular interest is their effect on suppressing the endotoxin-mediated induction of nitric oxide synthase (within about an hour),  and their cardioprotective effects which defend the myocardial cells against the depressant effects of endotoxin. Multiple different mechanisms have been identified which seem to be responsible for these effects, and they are listed in the chapter on the use of steroids in septic shock, because that is where these effects are most important. The reader who is interested in a more professional take on this topic is redirected to some excellent corticosteroid propaganda by Djillale Annane (2011).

Lastly, it appears that mega-dose steroid pulses have other, poorly characterised and unexplained effects on the cardiovascular system. When Jain et al (2005) subjected thirty pemphigoid patients to pulse therapy, they observed all sorts of abnormalities, such as sinus bradycardias and an increase in the number of arrhythmias and ventricular ectopics. No explanation for this was offered, other than a series of theoretical notes on potential cellular mechanisms. 

Neuropsychiatric effects of corticosteroids

A veteran intensivist formerly in the authors' orbit had at one stage narrated to him an account of how he, following an injection of betamethasone into an aching shoulder, spent the whole night working feverishly on a solution to the political turmoil in the Middle East. In short, steroids can make a raving loon of any formerly normal person, given enough time and a high enough dose.  Ciriaco et al (2013) probably offers the best review of these central nervous system effects. To try to explain these has eluded the best of thinkers, which means the rest of us will have to satisfy ourselves with merely listing them:

• Euphoria
• Mania
• Sleep disturbance (well, insomnia)
• Poor concentration and memory (with chronic steroids)
• Psychosis with hallucinations

Good feels and abundant energy are not the uniform effect: some people actually report a worsening of their depression, or a decreased ability to regulate their mood. In case you are wondering what sort of steroid dose is needed to drive somebody insane, the answer is probably "a large dose"; a large retrospective audit from the 1970s found that the patients who received less than 40mg/day of prednisolone were highly unlikely to develop worsening features of psychosis (less than 1.3% of them did), whereas with doses in excess of 80mg about 18% of the patients became psychotic. This is clearly some sort of genomic effect, as it is not something seen abruptly - in most cases it takes about five days of treatment with normal doses, or about 24 hours with pulses.

Effects of steroids on muscle function

Though we might commonly associate steroids with proximal myopathy, it might seem logical that, when you're having a "stress response", some increased muscle strength and endurance would be a desirable effect from your stress hormones. This, reader, is exactly what happens from small and transient steroid courses- to the point where professional athletes can use these drugs to gain a competitive edge. As an example, when Casuso et al (2013) gave a mere 2mg of daily dexamethasone to a group of random males for five days, they found that with even this trivial dose they were able to perform a repetitive leg extensor exercise for 29% longer than without it. 

However, as everyone knows, this is a transient effect. Over time corticosteroids gradually produce myopathy, mainly as the result of their effects on protein metabolism and muscle RNA synthesis. Gupta & Gupta (2013) have probably the fastest most readable article on this subject, where the interested reader can turn for a little more detail; but without going in too deep, the mechanism involves the simultaneous depression of protein synthesis and the activation of structural protein catabolism, with the ubiquitin-proteasome and lysosome systems actively involved in degrading myofibrils. The result is muscle atrophy, which is something only seen over the course of months for patients in the community (presumably because they continue to use their muscles), or over the course of days and weeks for ICU patients, because they are immobile and critically ill. That sort of "acute" steroid myopathy is the one that is probably of the greatest interest to the intensivist. It probably has a slightly different underlying mechanism, as it progresses very rapidly (Haran et al report a time course of only three days for some patients), with some authors suggesting that this form is mainly non-genomic.

Electrolyte changes due to corticosteroid therapy

Virtually all corticosteroids (other than dexamethasone and betamethasone) have some sort of mineralocorticoid effect, which manifests itself as a trend to increase the total body sodium and decrease the total body potassium by the activity of the aldosterone-regulated ENaC channel in the distal nephron. Detailed discussions of renal handling of sodium and potassium are available elsewhere, and here it will suffice to note that this effect will be most pronounced when the normal water retention which is expected from these effects is somehow countered, like for example with the use of diuretics or with GI losses. The consequence is the concentration of sodium in the extracellular fluid, leading to an increase in the strong ion difference and therefore metabolic alkalosis.

So that's an undesirable excess of mineralocorticoid effect, generally seen with hydrocortisone and prednisolone more than with the other drugs. Dexamethasone and methylprednisolone tend not to have as much mineralocorticoid effect (dexamethasone in fact has none), which is one of the reasons that these drugs are preferred as the agents of choice for big steroid doses. However, over time, the use of these "pure" glucocorticoid agents can result in an undesirable deficit of mineralocorticoid activity. The main reason for this is the suppression of ACTH release. ACTH is suppressed by anything with a glucocorticoid effect, which means that in the sustained presence of something like dexamethasone or methylprednisolone ACTH secretion will decrease. This means aldosterone secretion will also decrease, as ACTH is an important aldosterone secretagogue. At the same time the "pure" glucocorticoid agent has no mineralocorticoid effect of its own. The result of this looks a little bit like primary hypoaldosteronism, or the consequences of spironolactone therapy:  hyponatremia hyperkalemia and a normal anion gap acidosis typically ensue. For this reason, the reader will probably be unable to recall ever seeing any patients on long-term oral dexamethasone; it is usually converted to an equivalent dose of prednisolone for chronic maintenance therapy, mainly for these sorts of reasons. 

Adrenal suppression by corticosteroids

Following from the above, it is worth mentioning the effect of corticosteroids on the hypothalamic-adrenal-pituitary axis that regulates the normal secretion of cortisol. One is reminded that this set of regulatory organs only has one sensor available to detect steroid concentrations, and this is the easily fooled glucocorticoid receptor. All steroids with a glucocorticoid effect are therefore likely to confuse the hypothalamus and pituitary. However, not all steroid courses will have the same effect. There are notable differences in the degree and duration of adrenal suppression, which depend on several factors, listed in detail by Helfer & Rose (1989):

• Duration of treatment: though the levels of endogenously secreted cortisol are already lower in the morning after the very first dose of steroids, they return to normal by the following day, i.e. short courses result in a short period of suppression. It looks like a five-day course of prednisolone resulted in a five-day period of suppression for Streck & Lockwood (1979), and a year-long period of steroid therapy resulted in more than nine months of suppression for Graber et al (1965). From this, it follows that a course of treatment in excess of 2 weeks duration should probably be followed by a weaning course which is as long as the initial course of treatment. 
• Timing of administration: provided the exogenous steroids are given well before the early morning cortisol surge, they will not suppress it. The steroid dose of yesterday will barely affect the morning cortisol of tomorrow. On the other hand, giving dexamethasone after midnight will definitely suppress the early morning cortisol of the same day.
• Duration of action: steroids with longer half-lives tend to cause more suppression than short-lived steroids. In this fashion, an equivalent dose of dexamethasone will produce more suppression than methylprednisolone, which is one of the reasons methylprednisolone is usually chosen for "pulse" mega-doses. 
• Route of administration: obviously you've got to get the steroid to its receptors in the CNS, which means systemic absorption is necessary, and therefore all the topical and inhaled methods of delivery are unlikely to produce adrenal suppression under normal circumstances.
• Dose: the normal amount of cortisol secreted by the adrenal glands over the course of a routine non-stressful day is about 30mg/day. Theoretically, any additional steroid that exceeds this total dose should cause a regulatory change in the secretion of endogenous cortisol. Practically, it seems to be very dose-related, with higher doses causing more suppression
• Dosing interval: it appears that having multiple doses over the course of 24 hours is much more adrenally suppressing than having a single dose, even where the total steroid dose remains the same. Adrenal suppression can be minimised even further by giving the same total dose every 48 hours (i.e. giving 100mg of prednisone every second day, instead of 50mg daily). Ackerman & Nolan (1968) found that alternate-day therapy with about 75mg prednisolone produced only a slightly depression in the plasma cortisol metabolites, very close to normal values, whereas daily therapy produced marked depression.

Helfer & Rose (1989) suggest that about 2-3 weeks of moderate-intensity steroid therapy should still allow an abrupt cessation of the steroid, without the fear of adrenal suppression. However, that person may still have sub-normal adrenal function (just not low enough to be significant for their daily function). They will still attend their day job, but any exposure to physiological stress (eg. surgery) may expose a hidden adrenal insufficiency. 

Effects of corticosteroids on carbohydrate metabolism

They don't call them glucocorticoids for nothing. If you considered yourself to be an open-minded glucose-tolerant person before, steroids will change that very quickly. If you had diabetes and considered it to be under good control, you can say goodbye to that as well. This effect seems to be dose-dependent, and has several mechanisms:

• Reduced insulin production by pancreatic β-cells
• Reduced insulin sensitivity of the liver, which means:
• Increased glucose production in the liver, because it is no longer listening to the pancreas, and
• Decreased glucose uptake by skeletal muscle and fat, by a mechanism which is thought to be related to a modification of post-receptor signalling (Rizza et al, 1982)

The net effect is hyperglycaemia.  According to Poetker & Reh (2010) and Radhakutty & Kurt (2018), the peak changes are seen within 8-12 hours of starting therapy, i.e. roughly at the same time that you should be seeing other genomic effects (such as the peak antiinflammatory activity).

The purpose of this, from a "why would the body do that" point of view, remains somewhat obscure. It appears that cortisol, like growth hormone and glucagon, are counterregulatory signals designed to balance the effects of insulin, which means that the hyperglycaemia we see with exogenous steroids is probably just an exaggeration of that effect, and not a part of a normal stress response. You can clearly see what would happen without these regulatory effects in the hypoglycaemia which presents along with Addison disease, where there is no cortisol to oppose the influence of insulin. Andrews & Walker (1999) offer a more rigorous breakdown, in case that is what is required, but the CICM First Part exam candidate is reminded that we are very far from core syllabuses knowledge here, and this is unlikely to ever appear in any of the papers or vivas.

Effects of corticosteroids on fat metabolism

The acute administration of corticosteroids results in an increase in the lipolysis taking place in adipose tissue. The intracellular molecular mechanisms of this are likely to be of little interest to the pragmatic intensivist, as there are no practical means of manipulating them, nor is there any reason to do so. Those who would like to know more are motioned towards Xu et al (2009), where the details are discussed. From the point of view of the ICU patient, this looks like an increase in the release of free fatty acids, and it contributes to the fatty liver disease which is often seen with corticosteroid therapy.  

At the same time, with chronic use, the user sees a redistribution of fat around the body, with a net increase in adipose tissue (some of which can be attributed to the increased appetite), but increased more in some places than in others. Clinically, this gives rise to the central obesity, typical "moon facies" and "buffalo hump" (dorsocervical fat pad) seen with chronic steroid use or Cushing's disease. This is obviously not something you would see over the period of a short ICU stay: it typically takes a long time to develop this sort of appearance, and the dose required is usually substantial. For example, rat studies by Rebuffe-Scrive et al (1992) suggest that, though some mesenteric fat deposition does occur with chronic stress, the native adrenal glands probably do not produce enough cortisol to push fat around on an industrial scale, i.e. exogenous corticosteroids are required. Poetker & Reh dig out some studies suggesting that modest doses (10-30mg of prednisolone) for modest periods (3 months) have a modest chance (15%) of producing this complication.

Effects of corticosteroids on protein metabolism

Corticosteroids reduce the rate of protein synthesis in skeletal muscle, and in fact favour protein catabolism and the generation of glucose from amino acids liberated by this process. Without going into the thick forest of molecular biology presented by Magomedova & Cummins (2015) it is possibly better to just summarise that steroids cause muscle cells to preferentially degrade functional myofibrillar components (eg. myosin heavy chain) and various essential-sounding regulatory proteins, with the ultimate outcome being atrophy and myopathy.

At the same time, the effects of glucocorticoids on the myocardium are completely the opposite: Ren et al (2012) found that they tend to increase the synthesis of bulky contractile proteins, and produce cardiac myocyte hypertrophy. Even more weirdly, this is not a mineralocorticoid effect, i.e. aldosterone definitely does stimulate cardiac hypertrophic remodelling but blocking the aldosterone receptor did not prevent myocyte hypertrophy for the investigators.

Still, the heart is small in terms of mass, and the skeletal muscle tissue mass is vastly greater. The net effect of systemic corticosteroid dosing is the mobilisation of amino acids. Wise et al (1973) found that the plasma alanine level increased by about 40% after three days of daily dexamethasone (2mg). Specifically, alanine is the amino acid that increases the most, which pushes fuel into the alanine cycle (otherwise known as the Cahill cycle), where it can act as the substrate for hepatic gluconeogenesis (i.e. the liver can create pyruvate from it). Thus, the net result of corticosteroids is the conversion of structural muscle protein into metabolic fuel.

But why, the reader might exclaim, looking at a mechanism that can't possibly be helpful in any way. Why would your own glucocorticoids want to do this. Well, reader, the stress response calls for desperate measures, and corticosteroids released in response to stress are analogous to a government commandeering the resources of the private economy during wartime. Braun & Marks (2015) explain this much better, and without lame military analogies. The long and short of it is that a systemic response to stress requires metabolic fuel in the form of glucose to be readily available and abundant, as well as free amino acids to act as substrates for the synthesis of acute phase reactants and the replication of immune cells. This sort of repurposing of skeletal muscle protein is probably a safe sacrifice to make when you are fighting off an episode of sepsis, as this will be shortlived and your robust young organism will probably not miss the 2% of the muscle mass that might disappear in this fashion every day.

Unfortunately, from an evolutionary standpoint, these sorts of quick acute stressors are all you were ever expected to handle before you had the opportunity to sow your reproductive oats. No illness more severe, or more prolonged, had yielded much of an evolutionary pressure, as our ancestors would have inevitably died, and so you lack the mechanisms to control this hypercatabolic process - there was simply no reason for them to ever evolve.  Fortunately, modern man has the benefit of some highly developed critical care services, and can remain profoundly ill for weeks and weeks, during which time this increasingly dysfunctional and misdirected stress response  can devastate their muscle tissue through catabolism. 

Association between corticosteroids and peptic ulcers

There is a widespread belief among the medical community that corticosteroid therapy produces an increased risk of peptic ulcers. This belief was originally based on some data from the 1980s and 1990s which, according to  Fardet et al (2007), were not entirely sound, at least in terms of drawing such firm conclusions. Looking at the (non-blinded, largely retrospective observational) studies, one fails to be impressed by the methodological rigor, and when later investigators pooled data from only respectable sources, the difference in peptic ulcer risk turned out to be 0.3% vs 0.4%. In the cold light of this modern enlightened era, Narum et al (2014) were ultimately able to identify some groups of increased risk (mainly hospitalised inpatients and the critically ill), in whom the rate of GI bleeding was increased by about 40%. However, the overall risk was very small - in the entire population (33,253 participants from 159 trials), only 2.5% of patients had this complication (2.9% in the steroid group). 

The mechanism underlying this increase in risk is not especially clear. In fact it appears that under normal physiological conditions glucocorticoids are actually gastroprotective. It is now believed that corticosteroids in vastly supraphysiological doses probably just make the gastric mucosa more vulnerable to other ulcer-generating stresses, and then prevent wound healing to exacerbate the injury. For example, Takeuchi et al (2007) were able to produce gastic ulceration in rats using prednisolone, but they had to give them 50mg/kg of the drug (a completely insane overdose for a human), and the rats had to have COPD for this to work (they were exposed to twelve weeks of cigarette smoke). Clearly, multiple factors must conspire to produce this steroid complication, and steroids alone are probably not to blame.

Corticosteroid-induced osteoporosis

Corticosteroids stimulate bone resorption by their direct activity on osteoclasts. Fardet et al (2007) list the molecular mechanisms, which basically consist of turning off the protective regulatory signals (OPG) while at the same time activating osteoclasts via a soluble cytokine pathway (RANKL). The net effect is osteoporosis, to which female patients are more vulnerable. The dose required to achieve bone loss is anything more than 5mg prednisolone, and the density bone loss is maximal over the first 6 months (10%), decreasing to 2-5% per year for the subsequent years. 

Avascular bone necrosis due to corticosteroids

This is a bizarre idiosyncratic complication of steroid therapy which probably needs to be mentioned as an aside, even though it dose not belong in a discussion of the "physiological effects of corticosteroid therapy". Nobody seems to have a clear explanation of how it happens, with fat hypertrophy, fat emboli or intravascular thrombosis all identified as possible mechanisms. It can occur at all skeletal sites, but seems to prefer the femoral head, and is surprisingly common (Chan & Mok, 2012, report that it is responsible for something like 10% of the total hip replacements in the US. It appears that the cumulative dose of steroids is the most important association, and the timeframe for the development of this complication is measured in months.

Leukocytosis due to corticosteroids

Following even the first dose of steroids, the white cell count is usually seen to increase, and this finding is dismissed by clinicians as something completely expected and unremarkable. The cells that appear in the circulation are mainly neutrophils, and they appear within the first 24 hours following a dose, remaining raised for at least two weeks during sustained treatment and then slowly decreasing (Shoenfeld et al, 1981).  That could get confusing if you also suspect the patient of having an infection. Fortunately, steroid-induced leukocytosis is usually without a left shift or toxic granulation, making it a bit easier to distinguish from a genuine inflammatory reaction. The mechanism behind this seems to be a combination of impaired tissue sequestration, increased release from the bone marrow, and decreased apoptosis (all laudable steps in a stress response where you expect an infection). There does not seem to be much of a dose-response relationship, with a similar magnitude of leukocytosis seen with 5mg and 80mg of prednisolone.

Impaired wound healing due to corticosteroids

Being something of a wet blanket for protein synthesis and cell proliferation, it is no surprise that steroids have a negative effect on tissue repair and wound healing. They interfere with it at every stage: the early macrophage infiltration, the angiogenesis and fibroplasia, the fibroblast proliferation, and the deposition and remodelling of scar collagen. 

What is the functional effect on wound integrity? Wang et al (2013) cite numerous horrific-sounding animal studies that have quantified this effect by inflicting wounds on steroid-treated animals and then putting quantifiable graded tension on the edges. The decrease in wound edge tensile strength is apparently about 30%. However, it is difficult to generalise this to the human population, as the doses of steroids required have been massive (15-40mg/kg of prednisolone) and there is clearly a lot of inter-species variability (for example, guinea pigs are completely unaffected). Human patients with Cushing syndrome seemed to have cutaneous wound tensile strength reduced by 40% (in one study from the 19050s with only five patients).

How does this poor wound healing affect the ICU population, from the perspective of clinical outcomes?  From the literature review by Wang et al, it would appear that in this sort of chronic setting the risk of wound breakdown (eg. anastomotic leak) is about 2-5 times greater than in the general population. For patients only recently commenced on steroids, there does not appear to be much of an effect over the first ten or so days of therapy. Moreover, some animal data suggests that after three days the main fibroproliferative steps have already been taken and the integrity of the wound will be maintained even in the face of steroid therapy. The surgeon who complains about the stress dose steroids being given to their septic patient should be reminded that uncontrolled septic shock with massive noradrenaline doses also has a negative effect on the anastomotic site.

Corticosteroid withdrawal syndrome

After a prolonged course of steroids, the abrupt cessation of treatment can produce a withdrawal syndrome which is only partially accounted for by the adrenal suppression.  Hochberg et al (2003), discussing endocrine withdrawal syndromes, have a great section on glucocorticoids which goes into some of the suspected mechanisms, which is a great read even though the majority of what is said there seems to be largely speculative. 

When steroids are abruptly ceased:

• Adrenal suppression will be revealed, and symptoms of hypoadrenalism may develop (hypotension, hypoglycaemia, hyponatremia, hyperkalemia, etc).
• Relapse of the illness that was being treated with steroids
• A nonspecific withdrawal syndrome may develop over the subsequent 7 days or so:
  • Depression, anorexia, nausea, emesis, weight loss (due to reduced central dopaminergic and noardrenergic activity)
  • Myalgia, fevers and headache (due to reduced secretion of POMC-related peptides)
  • Somnolence and lethargy (due to increased cytokine release)
• Psychological dependence may discolour the entire process

The most amazing version of this is the effect of pituitary resection on patients suffering with classical Cushing's disease, where a secretory ACTH-oma is removed, and the chronically maxed-out adrenal cortisol secretion suddenly stops. By post-operative day 4, Papanicolau et al (1996) noted all patients developed some degree of nausea, anorexia, somnolence and fatigue. Not everyone had fevers, but on average their temperature increased by about 0.6º C. With the immunosuppressant effects of cortisol now removed, the immune system awakens, showering systemic cytokines and producing a syndrome that resembles sepsis biochemically (for example, in the aforementioned case series, the IL-6 levels increased five-fold, with slightly lower rises in TNFα and IL-lβ). The 

Acute corticosteroid overdose

Remembering that one possible application of corticosteroids is a "pulsed" mega-course in doses of 1000mg per day for several days in a row, one can only but wonder: what even is an "overdose"? Historical examples of heroic steroid use are crowned by Woods et al (1973), who gave twenty-five grams of methylprednisolone to a renal transplant patient over the course of three weeks. That was a 50kg female patient who apparently did well, begging the question: how much is too much?

Well. Most "corticosteroid overdose" occurs subacutely, and represents either overzealous use or some kind of dose miscalculation, such as in this case report by Rottenstreich (2015) where a seven month old was accidentally given five times more dexamethasone than was planned, and developed Cushing syndrome. That's probably not what most people would be thinking of when they say "overdose", even though it is an unexpectedly high dose. When administering pulse steroids, some immediate adverse effects are also seen: Min et al, 2012, followed a group of patients and found abdominal pain and rash were the most commonly reported effects of high dose IV methylprednisolone, followed by hot flushes, a feeling of general weakness, constipation, insomnia, and facial swelling. Other authors also report arrhythmias, tachycardia and bradycardia both, as well as AV block, though nobody died and everything was very self-limiting.

Overall, even after discussing all these side effects and adverse reactions, the incredible harmlessness of these substances needs to be pointed out as a startling and remarkable feature. Consider corticosteroids in the context of other endocrine hormones (adrenaline, insulin, thyroid hormone, etc), and behold their enviable safety profile. Think about it: there are very few endogenous substances where, following the administration of an exogenous dose one hundred times the normal level, nothing serious happens.