• anaphylaxis
• anaphylactic shock
• adrenaline
• epinephrine
• neurocardiogenic syncope

Anaphylaxis today still generates as much excitement, fear, rhetoric, and ripostes as it must have done for Charles Richet and Paul Portier at the turn of the 19th century. While they were guests on board Prince Albert of Monaco’s yacht in the Mediterranean, they reported on their experiments on dogs rechallenged with Physalia extracts, and first coined the phrase “anaphylaxis”, literally meaning “against protection”, when some dogs unexpectedly died. Since then, anaphylaxis has come to symbolise one of medicine’s great clinical bedside challenges, demanding rapid recognition without the benefit of an immediate laboratory test, and urgent management to avert a potentially fatal outcome usually in an otherwise healthy, young patient. Its evanescent nature has mitigated against the development of a solid scientific database to guide clinicians, and has generated as spiritedly polarised views on management as any therapeutic topic. No more so than when the use, dose, and delivery of adrenaline (epinephrine) is being argued.

Brown et al in this issue contribute reliable clinical evidence supporting the use of carefully titrated intravenous adrenaline with volume resuscitation for treating significant anaphylaxis.¹ In their case this followed jack jumper ant (Myrmecia pilosula) sting challenge on 68 healthy volunteers in Tasmania known to have a history of hypersensitivity to this ant. Their original paper in the Lancet attested to the efficacy of the ant venom immunotherapy they had developed,² while this paper reporting on the same group of patients describes in detail their management. All received a sting challenge, and in a randomised, double blind protocol they received either venom immunotherapy or placebo, followed by resuscitation in a supervised resuscitation area. The ensuing sometimes dramatic reactions were suffered almost exclusively by the placebo group, and were to an extent ameliorated by the study exclusion criteria that had eliminated volunteers with hypertension, heart disease, poorly controlled lung disease, ACE inhibitor or β blocker therapy, and age less than 17 or over 65 years. None the less, of the 21 systemic reactions from 29 patients in the placebo group, eight were grade IV or severe according to the grading system developed by Műller, with features of hypotension, collapse, loss of consciousness, incontinence of urine or faeces, or cyanosis. Three were grade III reactions manifesting dyspnoea, wheeze or stridor and two or more of dysphagia, dysarthria, hoarseness, weakness, confusion, or a feeling of impending disaster, and the remaining 10 had grade II (three patients) or grade I (seven patients) reactions. There is no question, this really was significant anaphylaxis being studied, enough to make even the most torporous ED physician coming off a long nightshift galvanise back into action!

Brown’s results highlight some important features about anaphylaxis of relevance to us all, some of which are already recognised, some merely assumed to be true and some would be less well known, even surprising. Thus, all his patients developed cutaneous features albeit often subtle, such as erythema, itch or urticaria, which highlight an essential diagnostic spectrum that must always be looked for as corroborating evidence of the possibility of anaphylaxis, whether by properly undressing the patient in the resuscitation room; or by the anaesthetist peering under the surgical drapes in theatre, as the patient suddenly loses their blood pressure or develops raised airway pressures.

The hypotensive anaphylactic reactions Brown measured in the most severe group were characterised by an initial fall in diastolic blood pressure from systemic vasodilatation, followed by a drop in systolic pressure too, with mean arterial pressures ranging from zero (unrecordable) to 55 mm Hg (with a median at 45 mm Hg). All these patients were initially tachycardic, but of interest all then developed a relative bradycardia as hypotension ensued, with heart rates from 15–65 per minute (median 32). The mechanism of this bradycardia is consistent with neurocardiogenic syncope as suggested by Brown, although it is not possible to exclude direct ant venom mediator effects on the heart.

Neurocardiogenic syncope, vasodepressor or vasovagal syncope, is notable for its variable afferent limb and threshold, with triggers ranging from posture, pain, fear, and psychological stress to the use of vasodilators, inferior myocardial ischaemia (Bezold-Jarisch reflex), and severe haemorrhage. However, the efferent response is uniform with a vagally mediated bradycardia, and paradoxical vasodilatation with the concurrent interruption of sympathetic vasoconstriction.³ Thus sensory input from arterial baroreceptors as well as cardiac mechanoreceptors appear to modulate the balance of the parasympathetic and sympathetic nervous systems. A neurocardiogenic mechanism was further supported by a dramatic response to intravenous atropine 600 μg in two of Brown’s most parlous patients in virtual cardiac arrest. Rather than ascribing this to near terminal hypoxia, the successful use of atropine in these situations to counter vagal tone deserves further study.

Neurocardiogenic syncope may also explain Pumphrey’s personal observations on 214 anaphylactic fatalities, where some deaths clearly followed a change in the victim’s posture to a more upright position, either sitting or standing.⁴ He exhorts us to keep victims of anaphylaxis lying down, even those self administering adrenaline, and to support or raise their legs to maintain vena caval filling at all costs.

What Brown did show unequivocally was the efficacy of intravenous adrenaline and fluid to successfully treat all his subjects with anaphylactic shock, without the use of any antihistamines or corticosteroids in the acute phase. The adrenaline was given cautiously at a dilution of 1:100 000 (that is,10 μg/ml), starting at 5–15 μg/min to a total dose of around 5–20 μg/kg, with more adrenaline being required the more severe the hypotension. This approach gratifyingly supports the hereto intuitive (non-evidence based) recommendations of several authors including this editorial’s writer.⁵ It also supports the perennial warning of others to always administer adrenaline for significant anaphylaxis with care, suitably diluted, given slowly and titrated against response in an adequately monitored patient.⁶

The limitations of Brown’s paper are to be aware that certain patients were excluded from the study as mentioned previously, and that as there were few life threatening respiratory complications, extrapolating the universal success of intravenous adrenaline to all forms of severe anaphylaxis, although highly likely to be efficacious, still effectively remains an eminence based recommendation. Likewise, there seemed to be no clear answer as to how much fluid should be being given at the same time for hypotension, with Brown’s dose comparatively small at one litre of normal saline (with only one patient receiving a total of three litres). Also, as Brown mentions, the early use of intramuscular adrenaline, particularly prehospital or in the unmonitored patient, still warrants direct comparison with intravenous adrenaline to quantify their relative efficacy compared with complication rates.

The other paper on suspected allergy in this issue is a far more sanguine reminder that “all is not gold that glitters”. Goodyear et al describe a case of orbital cellulitis in a 14 year old boy that was initially diagnosed and treated as an allergic reaction on the strength of unilateral periorbital oedema and a history of known dog hair allergy.⁷ Two days later, despite a high temperature of 39.3°C the condition was still considered to be allergic, delaying the final referral to an ophthalmology specialist until day 4 when the serious, true infective diagnosis was rapidly made. As the authors pointed out, a delay in diagnosing orbital cellulitis contributes to a higher rate of serious complications such as blindness or even death from cavernous venous thrombosis. A useful table highlighting the differences expected between an infective or allergic presentation of periorbital swelling is given, and the authors emphasise the pivotal role of the CT scan in investigating suspected orbital cellulitis.

Anaphylaxis naturally has a potentially vast differential diagnosis, although the rapid onset, accompanying cutaneous features, and direct relation to a potential trigger or particular iatrogenic precipitant suggest the diagnosis in most cases. In this instance, the authors did emphasise that unilateral orbital swelling is more likely to be attributable to infection than allergy. I agree with this, and unfortunately on this occasion I am unable to blame a dog for this unfortunate boy’s illness!