Australian snake antivenom dosing: What is scientific and safe?

Abstract

Because the median dose of one vial ‘clears the blood of circulating venom’, the authors of the Australian Snakebite Project recommend restriction of antivenom to one vial for all envenomated victims. This is neither scientific nor safe. Methodological flaws in the case series include limited detection of venom toxins and misinterpretation of data. The recommendation fails to consider larger doses of venom than that neutralised by one vial of antivenom. Although one vial may be adequate for minor envenomation, the initial dose should be two vials with more on a clinical basis.

Introduction

The doses of Australian snake antivenoms, F(ab′)₂ equine antibodies, are uncertain and controversial. Results of the Australian Snakebite Project (ASP) have prompted investigators to recommend restriction of antivenom dose to one vial in all cases, ¹ with an implied benefit of cost-saving. ² However, rigid adherence to such a one-vial policy is unsafe based on scientific aspects, re-interpretation of the results of the ASP, the potential and actual amounts of venom injected by snakes, coronial comments and the uncertainty of antivenom neutralisation doses.

Scientific aspects of venom detection

Because a median dose of one vial ‘is sufficient to bind all circulating venom’, ¹ the authors of the ASP, a study of 835 envenomated patients using venom immunoassays, claim that one vial of relevant antivenom is a sufficient dose for all snakebites^1,2 and that additional antivenom is not beneficial.

The fundamental considerations in deriving clinical management from biochemical data are: what is being detected by the tests, what concentration is detected and at what points the tests are conducted in the course of venom–antivenom pharmacokinetics and pharmacodynamics. The pharmacokinetics of F(ab′)₂ antivenoms is best described as a two-compartment model, with zero-order input and linear elimination kinetics displaying a rapid half-life of distribution over hours and an extended half-life of elimination principally by the reticuloendothelial system over days. ³ Accordingly, antibodies could neutralise toxins in the blood, tissues and even those already bound to tissue targets.

Ideally, the pharmacokinetics of an antivenom should match the pharmacokinetics of all important toxins in a venom, ³ but this is unrealistic. The venom of each snake genus has multiple different toxins ⁴ whose pharmacokinetics have been studied inadequately in only a limited number of victims envenomated by Australian snakes.^5,6 The sparse data support a one-compartment model, with an elimination half-life of approximately ten hours, ⁵ but it is likely that venoms display multi-compartment pharmacokinetics. ⁵ In animals, the volume of distribution of venoms is large compared to blood volume, meaning that antivenoms would be much less effective once the venom leaves the circulation. ⁵

Antibodies should have a high affinity for toxins, have similar volumes of distribution to target toxins and remain in the blood for prolonged times to neutralise toxins reaching the blood late during envenomation and toxins dissociating from antivenom or returning to the blood from extravascular compartments (‘recurring envenomation’). Not all toxins, particularly those of brown snakes, are detected by immunological-based techniques, ⁷ and rabbit-derived enzyme-linked immunosorbent assay (ELISA) tests have inherent flaws in detecting snake toxins in the blood at concentrations which are low but clinically important (0.01–0.1 μg/mL). ⁸ In contrast, the ASP used a biotin-streptavidin amplified rabbit-derived venom enzyme immunoassay. ⁹ This assay is considerably more sensitive, with a limit of detection of 0.1–0.2 ng/mL, and better suited to deriving antivenom dosage, but it is unknown which of the multiple toxin(s) in different species were detected by the ASP and moreover which toxin(s) detected were bound to antivenom. ¹⁰

Essentially, the assay used in the ASP detects toxins, irrespective of whether they are bound to antivenom antibodies. Consequently, in the study with snapshot measurements of venom toxins in the blood after antivenom administration, it is uncertain whether such toxins are the result of inadequate antivenom or recurring envenomation. The latter problem was investigated by assaying the serum of victims treated with antivenom for Russell’s viper envenomation using antibodies raised against horse antibodies. ¹¹ The investigators found that the ELISAs for venom measured both free venom and venom complexed with antibodies, and concluded that venom detection after antivenom administration was largely due to venom bound to antibodies, but free venom could also be detected. The investigators concluded that measurement of venom and venom–antibody complexes was required to determine if there was sufficient binding of venom by antivenom. ¹¹

Although the pharmacokinetics of Australian snake toxins in humans are largely unknown, those targeting blood coagulation are undoubtedly different from those targeting extravascular nerve and muscle tissues and would be expected to be present only briefly in the blood and escape detection. In envenomated monkeys, neurotoxins entered the blood within 15 minutes, peaked within 60 minutes and then declined followed by appearance of neurotoxicty. ¹² Moreover, it cannot be assumed that venom toxins bound to antivenom have been neutralised—they may still exert pathological actions while bound to antibody or after dissociating from toxin–antibody complexes. ¹³

Reinterpretation of results of the ASP

It is difficult to reconcile recommendations of the ASP with the clinical outcomes of the victims. For example, the effects on victims of tiger snake (Notechis spp.) envenomation ¹⁴ suggest a more cautious approach than simply recommending one vial for every patient for all degrees of envenomation. Of 56 victims, 53 had procoagulant coagulopathy, 17 had neurotoxicity, 11 had myotoxicity and three had thrombotic microangiopathy with renal failure. Although one vial of antivenom ‘cleared blood of venom’, it is impossible to discern if this amount of antivenom, or more, neutralised venom in all tissue compartments and was responsible, or not responsible, for the outcomes.

It was not one vial of antivenom as claimed by the investigators that was responsible for the outcomes of all victims in the ASP. Moreover, the choice of the median dose as the required dose is not valid. Although less antivenom was used in 2014/2015 compared to 2005 with no differences in mortality, the upper-quartile ranges of antivenom doses used in 2014/2015 were actually considerably above median doses of one vial, and thus these doses were responsible for the outcomes. The upper-quartile dose for envenomation by brown snakes was approximately 1.5 vials, for tiger snakes (Notechis scutatus) it was 2.0 vials and for taipan snakes (Oxyuranus spp.) it was 1.5 vials. ² Envenomations by these species comprised 61% of all envenomations. These observations suggest that clinicians perceived that the neutralising doses of antivenom for these snakes are more than one vial, and it is thus misleading to claim that the use of antivenom has ‘decreased to one vial with no evidence of adverse consequences’. ² The upper-quartile range equated to the median dose only in bites by death adders (Acanthophis spp.), mulga snakes (Pseudechis australis) and red-bellied black snakes (Pseudechis porphyriacus), which together comprised only 23% of envenomations.

Arguably, the dose of antivenom should be based on the upper dose range or maximum effective dose. The claim that one vial resulted in good outcomes is simply fallacious and untenable.

Coronial and other comments

The one-vial policy has been criticised by a Victorian state coroner in the investigation of the death of a victim envenomated by a large tiger snake. Additional antivenom was denied to the victim after one vial had failed to improve his condition. Clinicians were repeatedly advised by a poison centre that additional antivenom would not be beneficial. ¹⁵ The coroner recommended a review of a guideline from the Department of Health and Human Services which restricted treatment to one vial. This case illustrates the imprudence of slavish adherence to a guideline based on poor evidence and which fails to consider all circumstances. ¹⁶ An expert at the investigation opined that the one-vial policy should be ‘disavowed’.

Other writers experienced in the management of snakebite have described the one-vial policy as a ‘precariously narrow clinical strategy’ and recommend two vials as initial treatment of serious envenomation. ¹⁷ Generally, clinical recommendations for medications encompass a range of dosage, according to the severity of illness. Antivenom dosage should be no different.

Amounts of venom injected

Antivenom potencies are standardised such that 1000 IU of antivenom neutralises 10 mg of venom by experimentally preventing death in rodent models. However, the possible yield of venom from snakes within species is extremely variable. For example, the average dry-weight yields of venom obtained manually from tiger snakes from South Australia and Victoria were 33 and 34 mg with maximums of 336 and 224 mg, respectively, but larger specimens from Kangaroo Island and Chappel Island yielded averages of 110 and 56 mg with maximums of 636 and 125 mg, respectively. ¹⁸ Similarly, the average from common brown snakes (Pseudonaja textilis) from South Australia and Queensland was 8–26 mg with maximums of 51–155 mg, while taipans yielded an average of 146 mg with maximum of 882 mg. Since one vial of tiger snake antivenom (3000 IU) theoretically neutralises 30 mg of venom, one vial of brown snake antivenom (1000 IU) theoretically neutralises 10 mg of venom and one vial of Taipan antivenom (12,000 IU) neutralises 120 mg, there is a high likelihood of under-treatment with one vial if treatment is based on the amounts of venom obtainable from snakes.

The amount of venom delivered during an actual bite is unpredictable and unknown in every human envenomation. Although smaller than amounts obtained manually, naturally delivered amounts are also quite variable, as determined in experiments using freshly killed mice offered to snakes.^19,20 The median quantity of venom delivered at the first bite of P. textilis was 3.8 mg (range 0.05–9.5 mg). Similarly, the median quantity injected by tiger snakes was 3.4 mg (range 0.003–36.7 mg), and the median quantity delivered by taipans was 19.8 mg (range 0.2–19 mg). Accordingly, one vial of antivenom would theoretically neutralise several times over the median amount of venom injected experimentally at a first bite by respective species, but the maximum quantity of venom injected by all species is very close to or exceeds the maximum experimentally determined neutralisation capability of antivenom.

Snakes may bite multiple times, and absorption of toxins from the bite site or sites of possible sequestration is probably a continual process, implying that initial provision of several vials or serial administration of antivenom may be required. Moreover, serial antivenom or infusion may be required to neutralise toxins from recurring envenomation.

With a one-vial policy, there is no margin for safety if larger quantities of venom are injected with multiple bites, bites by large specimens of snake or bites in which the snake must be forcibly detached from the victim. Unusual susceptibility to envenomation and pre-existing comorbidities may also determine the amount of antivenom required for the treatment of a particular envenomated patient.

Human neutralisation dose

The neutralising doses of antivenom for humans are unknown. It is an unjustifiable assumption that prevention of death in experimental animals equates to prevention of death and morbid effects of envenomation in humans. The neutralising dose in vivo for all effects of venom in humans cannot be extrapolated accurately from animal studies. Theoretically, the antivenom dose could be related to the amount of venom injected, which could be derived from the victim’s weight, the time since envenomation and the serum level at that time. However, since the pharmacokinetics of venom and of toxins are obscure and the measurement of venom in blood is presently impractical, ¹⁰ the neutralisation for each victim cannot be determined.

Arguably, the clinical neutralisation dose of antivenom for humans should be regarded as the maximum dose beyond which no further improvement is gained, but that is difficult to derive theoretically and likewise difficult to achieve during the practical clinical management of an individual victim, since the effects of antivenom are delayed. This applies to all effects of envenomation, including coagulopathy due to procoagulopathic toxins which cause consumption of coagulation factors. Even when adequate antivenom has been administered, coagulopathy will not resolve for at least six hours (while coagulation factors are remanufactured), while damaged tissues cannot be repaired by antivenom and will require many days for recovery.

In desperate circumstances of life-threatening envenomation, seemingly unresponsive to antivenom, it is difficult to differentiate reversible and irreversible organ dysfunction, leading to over-treatment. Clinicians managing an envenomated patient simply do not know whether additional doses of antivenom will prevent further damage.

Adverse effects and costs of antivenoms

Antivenom causes immediate-type hypersensitivity reactions in 25% of cases, but since 96% of these occur after one or the first vial, ²¹ this adverse effect alone should not preclude administration of more than one vial. Current costs of antivenoms are AUD$361 for brown snake, AUD$453 for tiger snake, AUD$1617 for death adder, AUD$1639 for black snake, AUD$1835 for taipan snake and AUD$2414 for polyvalent cases. Brown snake and tiger snake antivenoms are the most frequently used and, compared to their benefits, are not expensive.

Manufacturer’s advice

Seqirus Pty Ltd (formerly BioCSL Pty Ltd) does not support dosing with one vial and recommends higher doses. ²² Moreover, antivenoms may not be as potent as expected. A study of antivenom (some outdated) showed that their potencies as ratios of their specified neutralisation capacities were variable, with some considerably less potent and others a little more potent than specified: brown snake 0.6–1.3, tiger snake 0.6–1.2, taipan 0.9–1.0, death adder 0.3–1.0 and black snake 0.8–1.0. ²³

Conclusions

Valid conclusions on antivenom dosage, as ratios of their specified neutralisation capacities, cannot be drawn from the studies in the ASP, which, although of a high standard for case series, are uncontrolled and very low on scientific scales of evidence. Although one vial may ‘clear venom from blood’, it is flimsy evidence that one vial is enough and neutralises all effects of venom. For safety, two vials of appropriate antivenom should be the initial dose. Less antivenom may be sufficient for minor envenomation, but more may be clinically indicated according to the snake involved (species, size, location), number of bites and the severity of patient effects. Hospitals should stock at least two vials of each appropriate antivenom. The preparedness should always be patient-focused to treat life-threatening envenomation with a margin of safety. Under-treatment may be fatal or lead to morbidity with prolonged expensive hospital admission, while over-treatment is probably not harmful. The cost of antivenom is insignificant compared to other issues. Australia should adopt a two-vial policy for initial treatment of snake envenomation.