Getting hot and bothered…

How uneasy do we feel when our intensive care patients spike a high temperature or have a persistent high temperature for several days? During all of our training, we cannot recall anyone teaching us at how many degrees or at what duration of temperature we need to start worrying, or indeed if we should do anything specific about it at all. Raising our core temperature in sepsis is clearly physiological and is not confined to humans. Mammals, birds, reptiles and fish all do it. As a result, cytokine function improves, viruses and bacteria are better killed and it would appear that we may live longer as a result. But what about other causes of a raised temperature? Infection may only be the cause in the minority of cases. ¹ Why would hyperthermia after a marathon, or ecstasy, or propofol, confer any advantage? And there must be a temperature in sepsis beyond which the risks outweigh the benefits. We have learnt that ventilating people to a physiologically normal tidal volume is better than a higher tidal volume. ² We have also learnt that maintaining normal glucose levels is better than hyperglycaemia.^3,4 And more recently, that giving the right amount of oxygen is better than giving too much.^5,6 Why should we think that accepting more or less of anything is better than what we maintain for ourselves normally and have done for thousands of years?

In fact, what is the difference between ‘fever’, ‘pyrexia’ and ‘hyperthermia’ anyway? Depending on the reference source, the answer is somewhat different. With normal human temperature being as high as 38.2°C, the American College of Critical Care Medicine and the Infectious Diseases Society of America define fever as a core temperature above 38.2°C irrespective of the cause. ⁷ Fever has its etymological basis in Latin, where the original meaning is as broad as simply ‘heat’. Presumably Roman drug-taking was less prevalent than nowadays, negating the need for a more specific definition. Pyrexia, from the Greek ‘pyr' (fire, fever), should probably have the same meaning. The current International Statistical Classification of Diseases has a similarly broad definition of fever and pyrexia. Other sources reserve ‘fever’ for a raised temperature caused by the action of thermoregulatory pyrogens on the hypothalamus,^8,9 for instance, in sepsis. To confuse matters further, hyperthermia has variously been defined as a core temperature above 38.2°C irrespective of the cause; ⁸ other sources reserve hyperthermia for those conditions that increase the body’s temperature over the body’s thermoregulatory set point, due only to excessive heat production or insufficient heat loss, and therefore specifically excludes those conditions which cause fever. ⁹

Need we be more careful to differentiate pyrogenic and non-pyrogenic causes? Would hyperthermia do for the latter? But then again, are ‘pyrogenic’ and ‘non-pyrogenic’ temperatures really as different as we once thought? Certainly, the mortality and the effects of treatment are different, ¹⁰ but if so-called non-pyrogenic hyperthermia increases gut bacterial translocation, ¹¹ and the gut and brain are more permeable to these toxins, ¹² maybe not. What role do gut bacteria have in the subsequent multi-organ dysfunction? Certainly, giving antibiotics to dogs with heatstroke improves their survival. ¹³ Maybe there is less difference than we thought?

We are however learning that whether the cause is exercise, illicit street drugs, anaesthetic agents, prescribed medication or any of the other non-pyrogenic causes, even modest rises in temperature cause biochemical, structural and functional changes in many organs. We are beginning to learn that muscles and cells fall apart, enzymes stop working, our liver and kidneys stop working, bacteria get absorbed and our memory fails. With the exception of sepsis, are all raised temperatures as bad as each other? And how much of the effects are merely due to the hyperthermia per se, rather than the underlying cause? With increasing temperature, does the cause become less important than the temperature itself? Within the long list of conditions causing hyperthermia, the systemic effects are similar, and similar again to the effects of artificially induced hyperthermia, for example, in cancer treatment. ¹⁴

The Roman campaign in Arabia in 24BC gives us one of the earliest descriptions of heatstroke, when the army’s governor noticed that ‘the desert, the sun and the water, which had some peculiar nature,… caused his men great distress so that the larger part of the army perished.’ He was wrong though in his assertion that heatstroke ‘attacks the head and … descends to the legs, skipping all the intervening parts of the body’. ¹⁵ We know that ‘the intervening parts’ are affected in hyperthermia. Liver failure is often seen, on occasion sufficient to require transplant. Similarly, renal failure is much more common after exertional heatstroke (>40°C) than in endurance athletes without heat illness; military data suggest that one in six hospitalised exertional heat stroke victims will develop acute kidney injury, ¹⁶ in comparison with marathon runners generally; the Comrades marathon have reported an average of only one runner each year admitted with renal failure. ¹⁷ A reduction in liver and renal blood flow is at least part of the cause. ¹⁸ Cells, muscles and tissues break down, with corresponding elevated plasma CK, AST, ALT and lactate dehydrogenase levels. Again, a core temperature above 40°C is implicated. ¹⁹

Rat studies show us that hyperthermia damages blood vessels in specific organs, and blood flow, and probably contributes to the organ failure. Capillaries dilate, and vascular stasis is seen. Red blood cells leak into heart muscle and lung alveoli. The kidneys and adrenal glands are similarly affected. And again these changes may happen as rapidly as within 30 minutes, at temperatures as modest as 40.5°C. ²⁰

Brain dysfunction is common, not surprisingly given that hyperthermia increases cerebral metabolic rate, reduces cerebral blood flow and increases permeability of the blood–brain barrier. Excessive heat probably also causes direct neuronal damage and stimulates neuronal apoptosis. ²¹ The dysfunction may not be as reversible as we would hope – long-term neurological damage has been reported in exertional heat stroke, ²² neuroleptic malignant syndrome ²³ and serotonin syndrome. ²⁴ Even modest hyperthermia is bad – one study of induced hyperthermia in healthy volunteers showed that memory was impaired at a core temperature of only 38.8°C, compared with normothermia. ²⁵ Structural change and cerebral oedema have been reported in patients dying of heat-related illness with a core temperature at death of as low as 39°C. ²⁶ Rats, with a similar core temperature to humans, develop cerebral oedema, a permeable blood–brain barrier and neuronal cell changes at only 38.5°C–39°C. ²⁷

Coagulopathy is common and may well contribute to the multi-organ dysfunction in hyperthermia; thrombocytopenia, increased plasma fibrin degradation products, prolonged clotting times, and spontaneous bleeding are often seen. Liver dysfunction is probably the culprit; coagulopathy is rare without liver derangement and is temporally related to alterations in liver function. ²⁸ Hyperthermia inhibits platelet aggregation, which becomes increasingly marked at increasing temperature. But these changes begin to happen even at 38°C. ²⁹ Disseminated intravascular coagulation (DIC) may also be driven by release of pro-coagulant cellular components from damaged muscle. Are we seeing a combination of thermal damage, systemic inflammation, organ dysfunction and biochemical changes, and, in turn, affecting each other?

If non-pyrogenic hyperthermia is so bad, is cooling good? Probably, but it’s not clear. Even in sepsis, to which significantly more resources are devoted, it isn’t clear if normothermia or mild hyperthermia is better. ³⁰ There is little published work on the effects of cooling in non-septic patients. Observational data suggest that the mortality is 13-fold higher at a temperature of 39.5°C compared with normothermia; ¹⁰ whether artificially lowering the temperature has the same effect isn’t known. However, with the growing evidence that hyperthermia is bad, it probably should. If we cool, how rapidly should we be cooling? A core temperature above 40°C seems bad, whatever the pathology. In patients with classical heat stroke, cooling the patient to below 38.9°C within 60 min more than doubles the survival rate. ³¹ Most published guidelines and management strategies on these conditions advocate the need for urgent cooling, but few suggest a target temperature. The Faculty of Sports and Exercise Medicine, in their guidelines for treating exertional heat stroke, suggest a target temperature below 39°C, to minimise the effects of hyperthermia, balanced by minimising the risk of rebound hypothermia. Several national and international organisations concerned with limiting the risk of heat-related disorders during work in hot environments place an upper limit on body core temperature of 38°C. ²⁶ Given the effects of any temperature above normal, is normothermia therefore the only safe temperature?

And how to cool? 2000 years ago, our predecessors suffering with heat stroke were treated with ‘olive oil and wine, both taken as a drink and used as an ointment’. ¹⁵ Things were not much better 1500 years later: ‘A man suffering from fever should be bled immediately the moon passes through the middle of the sign of Gemini’. Times change. But antipyretics, paracetamol and NSAIDS, are probably as ineffective for these non-pyrogenic causes as bloodletting and olive oil. ¹⁰ Not surprisingly, given that it is a heat generation problem, not a hypothalamic-driven one. And with the risk of clotting abnormalities, renal dysfunction, and gastrointestinal dysfunction, ¹² NSAIDS feel like a bad idea. Then again, they may not be effective in sepsis, either. ³⁰ Medicine progresses. Ice packs in the groins and axillae, surface cooling and body cavity cooling became available but achieving a target temperature, and controlling the speed of change, are difficult. At last, technology now exists to effectively control temperature by a feedback system with rapid central cooling which is adjustable over a period of minutes. Perhaps, now, we are in the best position ever to be able to determine the effect of targeting a set temperature in a variety of conditions?

We have moved on from advocating olive oil and wine and moved on beyond venesection in a particular Zodiacal sign. But how far have we moved on? We don’t have an agreed definition yet. And that hyperthermia isn’t due to pyrogens may not actually be true. Could the intensive care community establish consensus on what is meant by fever and hyperthermia. And where should we now move to? Are we moving towards cooling all hyperthermic critically ill patients down to normothermia, just as we have moved to normoxia and normoglycaemia, or does the optimum temperature depend on the aetiology? Studies on the outcome at different temperatures and different aetiologies are needed now. We look forward to these studies with interest.