Venom-induced consumption coagulopathy (VICC) is a major effect of snake envenoming.
Venom-induced consumption coagulopathy (VICC) is a major effect of snake envenoming.
To investigate whether fresh frozen plasma (FFP) given after antivenom resulted in more rapid correction of coagulation.
This was a multicenter open-label randomized controlled trial in patients with VICC of FFP vs. no FFP within 4 h of antivenom administration. Patients (> 2 years) recruited to the Australian snakebite project with VICC (International Normalized Ratio [INR] > 3) were eligible. Patients were randomized 2 : 1 to receive FFP or no FFP. The primary outcome was the proportion with an INR of < 2 at 6 h after antivenom administration. Secondary outcomes included time from antivenom administration to discharge, adverse effects, major hemorrhage, and death.
Of 70 eligible patients, 65 consented to be randomized: 41 to FFP, and 24 to no FFP. Six hours after antivenom administration, more patients randomized to FFP had an INR of < 2 (30/41 [73%] vs. 6/24 [25%]; absolute difference, 48%; 95% confidence interval 23–73%; P = 0.0002). The median time from antivenom administration to discharge was similar (34 h, range 14–230 h vs. 39 h, range 14–321 h; P = 0.44). Seven patients developed systemic hypersensitivity reactions after antivenom administration – two mild and one severe (FFP arm), and three mild and one severe (no FFP). One serious adverse event (intracranial hemorrhage and death) occurred in an FFP patient with pre-existing hypertension, who was hypertensive on admission, and developed a headache 6 h after FFP administration. Post hoc analysis showed that the median time from bite to FFP administration was significantly shorter for non-responders to FFP than for responders (4.7 h, interquartile range [IQR] 4.2–6.7 h vs. 7.3 h, IQR 6.1–8 h; P = 0.002).
FFP administration after antivenom administration results in more rapid restoration of clotting function in most patients, but no decrease in discharge time. Early FFP administration (< 6–8 h) post-bite is less likely to be effective.
Snake envenoming is responsible for at least 20 000 deaths annually, making it one of the world's most neglected tropical diseases according to the World Health Organization . There are > 440 000 snake envenoming cases annually, with the main burden falling on tropical and subtropical regions of Africa, Asia, Latin America, and Oceania [1, 2]. Venom-induced consumption coagulopathy (VICC) is a major clinical envenoming syndrome that can result in major hemorrhage and death [2, 3]. Antivenom is the main treatment for VICC, but, although it appears to neutralize procoagulant toxins, clotting factor resynthesis and full recovery of clotting function takes 24–48 h . Therefore, envenomed patients remain at risk of major hemorrhagic complications such as intracranial hemorrhage for a significant time period after antivenom treatment. This has prompted the use of clotting factor replacement in VICC [4-8], despite limited evidence for its safety and effectiveness.
VICC is the commonest manifestation of severe snake envenoming in Australia, where ~ 75% of cases with VICC were given antivenom . A number of important snakes in Australia, including brown snakes (Pseudonaja spp.), tiger snakes (Notechis spp.), and the coastal taipan (Oxyuranus spp.), cause VICC, owing to the presence of prothrombin activators in their venoms [3, 10, 11]. These prothrombin activators activate the clotting pathway and cause complete consumption of fibrinogen, factor V, and FVIII, resulting in unrecordable times for clotting tests (International Normalised Ratio [INR] and activated partial thromboplastin time), as well as very high D-dimer levels .
Factor replacement has been suggested for the treatment of snakebite, to rapidly restore clotting factor levels and consequently reduce the risk of major hemorrhage [6, 7]. However, there is a concern that the provision of clotting factors will worsen the coagulopathy, as more substrate will be available for the procoagulant toxins present in the venom to activate . Furthermore, there are also risks associated with the use of blood products, and there is an increasing recognition of adverse effects with fresh frozen plasma (FFP) . There have been only a few studies of factor replacement for VICC in humans [4, 5, 7, 8], with two observational studies in Australia suggesting that the use of clotting factor replacement speeds the recovery from VICC [7, 8].
We aimed to investigate whether FFP administered after antivenom would lead to earlier correction of clotting function in envenomed patients with VICC.
This was a multicenter open-label randomized controlled trial in Australia of FFP vs. no factor replacement within 4 h of antivenom administration for the treatment of snake envenoming in patients with VICC. The study was approved by 14 Human Research Ethics Committees covering all hospitals involved in the study. The trial was registered with the Australian Clinical Trials Registry (ACTRN12607000620426).
The Australian Snakebite Project (ASP) involves clinical, laboratory and academic investigators located in hospitals throughout Australia, as well as the national Poisons Information Centre Network, which prospectively recruits all snakebite presentations in Australia. Recruitment is performed via a national free-call telephone number, whereby patients notified to the ASP were assessed for suitability for the FFP trial, with those who met the inclusion criteria being invited to participate.
Patients (aged ≥ 2 years) were eligible for inclusion if they had VICC and an INR of > 3.0, antivenom had been administered for the appropriate snake (brown snake, tiger snake, or taipan), there were no known allergies or objections to blood products, and four units of FFP were available and could be commenced within 4 h of antivenom administration. Exclusion criteria were severe cardiac failure or significant renal failure where there was a risk of fluid overload. Patients were recruited from 28 hospitals from four States of Australia, between 1 March 2008 and 30 June 2012.
Written informed consent was obtained from the patient or the patient's parent/guardian by the treating doctor or local site investigator with the assistance of the chief investigators. The patient was then randomly allocated to receive either FFP or nothing in a 2 : 1 ratio of active treatment to no treatment. The imbalance in allocation was to ensure that sufficient numbers of patients were given FFP, in case patients randomized to active treatment did not receive it within 4 h because of logistic reasons and remote location.
The study used an adaptive biased coin randomization schedule, with an online data entry and randomization system. Randomization and allocation to FFP or not to receive FFP was performed by the chief investigators (G.I., S.B., C.P.) or a research assistant (L.C.). This was done by accessing a password-protected single web page, where the inclusion criteria, exclusion criteria and consent were confirmed, before randomization to either ‘give FFP’ or ‘no FFP’. Randomizing the patient locked the system, and the allocation was recorded on an online database that could not be changed. The online database was monitored by one chief investigator (N.B.) who was not involved in patient treatment, random allocation, or data collection. If the allocation became unbalanced (i.e. > 40% or < 25% in the non-treatment arm or > 75% or < 60% in the FFP arm), then the probability was biased to correct the imbalance (i.e. 0.75/0.25 instead of 0.67/0.33). This was done after every 20 patients had been recruited.
Antivenom was administered to all patients in the study. For patients randomized to receive FFP, the treating hospital cross-matched and ordered the FFP. These patients received 10–15 mL kg–1 up to a maximum of 4 units of FFP (~ 1000 mL) over 30–60 min within 4 h of antivenom administration. Patients in the no-FFP arm were not to receive FFP until after the 6-h primary outcome unless there was evidence of major hemorrhage.
The randomized trial utilized the existing ASP infrastructure [9, 14]. The datasheets used for ASP already contained the information required for the study outcomes. All patients recruited to the ASP have datasheets filled out by the local investigators or treating clinicians, which are then faxed back to the chief investigator and entered into a purpose-built relational database in Microsoft Access. This includes information about the patient, bite, time of administration and dose of antivenom, administration of factor replacement, time and results of laboratory tests, including coagulation studies, adverse effects from antivenom or factor replacement, and any complications (hemorrhage or death). Systemic hypersensitivity reactions to antivenom or FFP were defined as anaphylaxis if they met NIAID-FAAN consensus criteria , and defined as severe according to the Brown grading system .
Serum and citrate samples were collected from all recruited patients and processed according to the ASP protocol. Briefly, serum was centrifuged at 2000 × g for 10 min, and citrated plasma was double centrifuged at 2000 × g for 10 min, before being stored at − 80 °C for the later measurement of venom concentrations and clotting factors, respectively. Snake venom concentrations were measured in serum with a previously described venom-specific enzyme immunoassay (EIA) . Venom-specific EIA was used for identification of the snake involved, including testing for brown snake, tiger snake, rough-scaled snake, taipan and Stephen's banded snake venoms. The citrated samples were used in clotting assays to measure the levels of FI (fibrinogen), FII (prothrombin), FV, and FVIII, as well as for the measurement of D-dimer levels. All assays were performed on either a Behring Coagulation System or Sysmex CA-1500 analyzer (Dade Behring, Marburg, Germany), with standard coagulometric or immunoturbidimetric methods as provided by the manufacturer.
The primary outcome was the proportion of patients with a significant return of coagulation function, as defined by an INR of < 2.0, at 6 h after antivenom administration was commenced. Secondary outcomes included time from antivenom administration to discharge, adverse effects from FFP (transfusion-related acute lung injury or systemic hypersensitivity reactions), major hemorrhage as defined by the ISTH , and death prior to hospital discharge. Clotting factor levels after antivenom administration constituted a predefined secondary outcome, but citrate samples were available for fewer than half of the patients, owing to problems with immediate processing and freezing in some regional laboratories.
The sample size was based on previous studies from the ASP suggesting that ~ 20% of patients not given FFP recovered to an INR of < 2 after 6 h [7, 8]. We considered an absolute increase of 30% of patients recovering to an INR of < 2 in 6 h to be clinically important. To detect this absolute increase with a significance level (alpha) of 5% and a power of 80%, a minimum of 34 and 68 patients would need to be recruited to each arm of the trial, respectively (i.e. a total of 102 patients). To allow for protocol non-compliance (failure to deliver FFP within 4 h), we initially planned to recruit 120 patients over 3 years for the duration of funding. There was no plan for interim analyses. Compliance with FFP allocation was higher than expected, but enrollment rates were lower than expected, and, after 4 years, the study had to be stopped after randomization of only 65 patients, because the grant funding for the study had finished.
At the completion of the study, one chief investigator (G.I.), who was blinded to the treatment allocation, extracted the data from the ASP outcomes database. Cases were only identifiable by study numbers. The classification of primary and secondary outcomes, including the presence or absence of adverse reactions or complications, was then determined (G.I.) and reviewed by two investigators (C.P. and S.B.). Finally, the fourth investigator, who was not involved with the day-to-day conduct of the study (N.B.), checked for missing or inconsistent data and then unblinded treatment allocations by using information from the online randomization database.
All continuous variables were summarized as medians and interquartile ranges (IQRs), and proportions were presented with 95% confidence intervals (CIs) to make the interpretation easier. The dichotomous primary outcome was analyzed by intention-to-treat with Fisher's exact test. A ‘per protocol analysis’ was also performed, comparing the group randomized to FFP who actually received a full dose of FFP within 4 h with those randomized to no FFP who did not receive it prior to the 6-h INR. For secondary outcomes, appropriate statistical tests based on data distribution were used for continuous outcomes, and proportions with 95% CIs were calculated for dichotomous outcomes. All analyses were performed, and all graphs were made, with graphpad prism version 5.03 for Windows (GraphPad Software, San Diego, CA, USA; www.graphpad.com).
There were 322 patients recruited to the ASP cohort study during the period March 2008 to June 2012. Seventy patients with VICC and an INR of > 3.0 were eligible for the FFP study, and 65 patients from 28 hospitals consented and were randomized (Fig. 1). Forty-one patients were randomized to receive FFP and 24 to not receive FFP. The two study arms had similar baseline characteristics, including patient demographics and type of snake (Table 1).
|FFP treatment (n = 41)||%||No FFP (n = 24)||%|
|Age (years), median (IQR)||39 (18–53)||45 (29–59)|
|Venom concentration (ng mL–1), median (IQR)||1.0 (0.2–5.4)||1.2 (0.3–3.7)|
|Tiger snake, rough-scaled snake, Hoplocephalus†||11||27||6||25|
|New South Wale||9||22||7||29|
|Time to antivenom administration (h), median (IQR)||4.5 (2.9–5.8)||3.5 (1.9–4.9)|
Six hours after antivenom administration, 30 of 41 (73%) patients randomized to receive FFP had an INR of < 2, as compared with only six of 24 (25%) patients not given FFP (absolute difference of 48%; 95% CI 23–73%; P = 0.0002) (Fig. 2). Figure S1 shows the time course of the INR for patients in each arm. The median time to recovery of the INR to < 2 after antivenom administration for patients randomized to FFP was 6 h (IQR 3.8–9.1 h), as compared with 14 h (IQR 7.9–21.9 h; P = 0.0005) for those patients who did not receive FFP (Fig. 3). Only one patient did not receive the assigned treatment (randomized to the FFP arm but not given FFP) and the ‘per-protocol analysis’ results were not different from the intention-to-treat analysis results (data not shown). Of the 47 patients with brown snake envenoming, 21 of 30 (70%) randomized to FFP had an INR of < 2 at 6 h, as compared with five of 17 (29%) not given FFP (P = 0.0136).
Of the 11 patients who received FFP but did not respond to the treatment, one was given FFP at > 6 h after antivenom administration, two received reduced doses of FFP (2 and 3 units), and two had an early recovery at 3 h after antivenom administration (INRs of 1.8 and 2.0) and then a rebound to higher INRs, although not to the levels seen prior to treatment (Fig. S1). Table 2 provides a comparison of FFP patients who did and did not respond to treatment. Non-responders were given antivenom, and therefore FFP, much earlier than the responders, and a poor response to FFP was not associated with age, sex, or snake type. There was evidence of further clotting factor consumption in patients who did not respond to FFP, with a more prolonged decrease in fibrinogen being associated with the administration of FFP (Fig. 4). This is supported by a much higher ratio of D-dimer to fibrinogen in non-responders than in responders (Fig. 5). A receiver operating characteristic curve was created to explore the overall effect of the time to FFP administration post-bite in predicting non-response to FFP treatment (Fig. 6; area under the curve = 0.84; P = 0.001) and also the threshold value that best predicted non-response. Administration of FFP earlier than 5.8 h post-bite was the optimal cut-point based on Youden's index, and the most sensitive cut-point was 8 h (suggesting that the risk of non-response after 8 h is very low).
|Response to FFP treatment (n = 30)||%||No response to FFP (n = 11)||%||P-value|
|Age (years), median (IQR)||43 (18–60)||38 (10–41)||0.49|
|Tiger snake, rough-scaled snake, Hoplocephalus||8||27||2||18||NA|
|Time to antivenom administration (h), median (IQR)||5 (4.1–6)||2.4 (1.5–4.4)||0.001|
|Time to FFP administration (h), median (IQR)||7.3 (6.1–8)||4.7 (4.2–6.7)||0.002|
|Time from antivenom administration to FFP administration (h), median (IQR)||2 (1.3–2.9)||2.7 (1.5–3)||0.32|
|Number of units of FFP|
There were no significant differences between patients receiving FFP and those not receiving FFP in any of the secondary outcomes, which are summarized in Table 3. In 49 patients who were discharged with a normal INR (< 1.3), the time to reach this complete recovery was significantly shorter in the FFP patients (Table 3). Sixteen patients were discharged with the coagulopathy largely recovered, but without a normal INR being recorded (last INR ranged from 1.3 to 1.9).
|FFP treatment (n = 41)||No FFP (n = 24)||P-value|
|Time from antivenom administration to discharge (h), median (range)||34 (14–230)||39 (14–321)||0.44|
|Reactions after antivenom administration, no. (%)|
|Systemic hypersensitivity reaction||3 (7)||4 (17)||0.41|
|Severe anaphylaxis||1 (2)||1 (4)||1.0|
|Time to INR of < 1.3 (normal) (h), median (IQR)||16.5 (4–58) n = 33||31 (12–69) n = 16||0.0051|
Seven patients developed systemic hypersensitivity reactions after antivenom administration but before FFP administration – two mild and one severe in the FFP arm, and three mild and one severe in the no FFP arm. One patient in the FFP arm had a severe reaction, with urticaria, angioedema and hypoxia, to a second FFP treatment given 24 h after antivenom administration. One serious adverse event occurred in the form of an intracranial hemorrhage in a patient with brown snake envenoming in the FFP arm. She had pre-existing hypertension and was hypertensive on admission, but did not develop a headache until 6 h after FFP administration, when the INR was still > 12. She died 1 day later.
This study shows that the administration of FFP within 4 h of antivenom administration results in more rapid restoration of clotting function in the majority of patients with VICC. Almost three-quarters of patients receiving FFP had an INR of < 2 at 6 h after antivenom administration, and this was also associated with more rapid complete recovery to a normal INR. However, this was not associated with more rapid hospital discharge of patients, and the numbers were too small to determine whether the administration of FFP could reduce the risk of major hemorrhage. An interesting finding was that non-responders to FFP treatment were given FFP earlier after the bite than those who responded well to FFP.
There are a number of limitations of this trial, including the small size and the unblinded allocation of FFP. Although the study was stopped before it reached the intended sample size, the observed effect of FFP on the INR was much greater than expected from the power calculation, and major clinical outcomes (death and major hemorrhage) were very uncommon, so little further information was likely to be gained from continuing the study. The trial was of a sufficient size to demonstrate that the use of FFP rapidly corrects the INR, and it therefore provides the basis for larger studies to investigate clinical outcomes that are uncommon and investigate the safety of FFP. The trial could not be blinded, because of the impossibility of providing a convincing placebo for factor replacement. However, the primary outcome was an objectively measured laboratory parameter, and there was concealed allocation to FFP, so the potential for bias was low .
The dose of FFP in this study was 10–15 mL kg–1 up to a maximum of 4 units. A retrospective review of FFP use suggested that 2 or 3 units of FFP was insufficient, with recovery being seen in fewer than half of the patients in that study receiving 2 or 3 units, as compared with all patients receiving 4 units . Insufficient FFP (two patients) or late administration (one patient) may explain the lack of response in three of the 11 FFP patients who did not have an INR of < 2 at 6 h (Table 2). Two further patients had a rebound in the INR (Table 2; Fig. S1, thick line). The remaining six patients with treatment failure appeared to recover at the same rate as the non-FFP patients. This suggests that, although the majority of patients improve rapidly with FFP, there is a small subgroup of patients who do not respond. This subgroup of patients received FFP significantly earlier after the time of bite.
This raises the possibility that FFP can be consumed in the first few hours of envenoming, despite antivenom treatment. Figure 4 shows that, in the FFP non-responders, the fibrinogen remains low as compared with patients not receiving FFP, and that responders recover more rapidly. In addition, the D-dimer to fibrinogen ratio in Fig. 5 shows that this slow recovery is likely to be attributable to further consumption. It is unlikely that this consumption of FFP administered early is attributable to active venom, as antivenom was administered to all patients, and previous in vitro data have confirmed the efficacy of antivenom for neutralizing the procoagulant effect of venom . Ongoing consumption may be attributable to the presence of active clotting factors in the initial period of resolution of VICC. Until this issue is resolved, FFP administration within 6–8 h of a bite should only be given if there are compelling reasons for administration, such as active major bleeding. The use of alternative therapies, such as specific factor concentrates, requires further investigation in this setting.
FV and FVIII in FFP are affected by the timing of storage, thawing, and subsequent delays in administration . Previous studies have demonstrated that deficiencies of FV and FVIII are the most important in Australian elapid VICC, and the recovery of these factors correlates best with the recovery of the INR . It is therefore reasonable to consider that the patients who did not respond to FFP may have been given FFP that had low activity of FV and FVIII, because of either poor storage or premature thawing (> 24 h) prior to administration. Unfortunately, this information was not collected during our study, and should be recorded in future studies. The measurement of factor activity in FFP prior to administration would also assist in determining whether this is the cause of non-response in some patients.
VICC is caused by a number of important groups of snakes around the world , including Echis spp. (saw-scaled and carpet vipers) , Daboia russelli (Russell's viper) , Calloselasma rhodostoma (Malaysian pit viper)  and Atrox spp. and Bothrops spp. (rattlesnakes from the Americas) . Although this study of Australian elapids cannot be immediately translated to other snakes, the study provides support for the use of FFP after sufficient time and antivenom has been given. In addition, FFP is cheap and readily available throughout the world, and, in many resource-poor countries, may be more easily accessible than antivenom. Further and larger studies are required with snakes in other parts of the world.
G. Isbister, S. G. Brown, and N. A. Buckley: designed the study in consultation with M. Seldon; G. Isbister, S. G. Brown and C. B. Page: coordinated recruitment; G. Isbister and S. G. Brown: undertook data collection; F. E. Scorgie: undertook laboratory studies under the supervision of L. F. Lincz; G. Isbister, S. G. Brown, and N. A. Buckley: undertook the analysis; G. Isbister: wrote the paper; N. A. Buckley, C. B. Page, F. E. Scorgie, L. F. Lincz, M. Seldon, and S. G. Brown: reviewed drafts. G. Isbister takes responsibility for the study.
Patients were recruited by: S. Brown (Albany Hospital [two], Carnarvon Hospital [two], Geraldton Hospital [three], and Karratha Hospital [four]); C. Macrokanis (Broome Hospital [two]); A. Coulson (Bunbury Hospital [one]); G. Isbister (Calvary Mater Newcastle [three] and Port Macquarie [two]); A. Holdgate (Campbelltown Hospital [two] and Liverpool Hospital [one]); A. Tankel (Coffs Harbour Hospital [two]), R. Greenberg (Dubbo Base Hospital [one]); R. Ellis (Fremantle Hospital [six]); D. Spain (Gold Coast Hospital [four]); D. Bitmead and K. Tay (Ipswich Hospital [two]); P. Bailey and I. Vladd (Joondalup Hospital [three]); T. Fraser (Mackay Base Hospital [four]); J. Willis (Manning Base Hospital [one]); P. Garrett (Nambour Hospital [three]); P. Thompson (Rockhampton Hospital [four]); R. Ellis (Rockingham Hospital [one]); D. McCoubrie (Royal Perth Hospital [two]); S. Alfred and J. White (Royal Adelaide Hospital [one]); O. Pascui (Sir Charles Gairdner Hospital [four]); A. Stafford (Swan Districts Hospital [one]); N. Ryan (Tamworth Hospital [one]); B. Close (Townsville Hospital [one]); and N. Gunja (Westmead Hospital [three]). We acknowledge the many referrals from the National Poison Centre Network and clinical toxicologists, and the help of the many other nurses, doctors and laboratory staff in recruiting patients and collecting samples. We thank L. Calver for her help in recruiting and randomizing patients, R. Kearney for data entry and follow-up of patient information and medical records, and E. MacDonald for assistance with ethics applications and data collection. We thank M. O'Leary for undertaking enzyme immunoassays for venom concentrations. We thank J. Walker for development of the online randomization and secure database.
The study was supported by NHMRC Project Grant ID490305. G. K. Isbister is supported by NHMRC Clinical Career Development Award ID605817. S. G. A. Brown is supported by NHMRC Career Development Fellowship Award ID1023265.
The authors state that they have no conflict of interest.