This study was approved by the Institutional Review Board of Hallym University Dongtan Sacred Heart Hospital (IRB No. 2009-01-004-001). Because of the retrospective nature of this study, the need for informed consent from enrolled subjects or their surrogates was waived. Therefore, this study did not impact on the patients’ treatment.

Independent t-test and analysis of variance were used for parametric tests, and the Mann-Whitney test, Kruskal-Wallis test, and chi-square test were used for nonparametric tests. Univariate and multivariable logistic regression was performed to assess associations with overt DIC. A P-value of <0.05 (twotailed) was considered to be statistically significant. IBM SPSS ver. 24.0 (IBM Corp., Armonk, NY, USA) was used for all statistical analyses.

Among 226 patients included in the study, 157 were male and 69 were female. The median and interquartile range (IQR) of age was 54.0 years (IQR, 43.0 to 63.0 years). The median and IQR of body mass index was 23.9 kg/m² (IQR, 21.0 to 25.7 kg/m²). Seventy-two patients had underlying diseases such as hypertension, diabetes mellitus, dyslipidemia, chronic lung disease, chronic liver disease, and neurologic disorders. The median and IQR of hospital stay was 4.0 days (2.0 to 7.0 days). Five patients arrived at the hospital in a state of shock. One patient suffered hypotension and cardiac arrest within 30 minutes after the snakebite and died 2 days later because of multi-organ failure (Table 1). Only 44 patients (19.5%) were able to identify species of venomous snake.

We could assess the presence/severity of coagulopathy using the ISTH scoring system in 167 patients; coagulopathy could not be assessed in the other 59 patients (26.1%) due to incomplete examinations. Three subgroups were defined according to the ISTH scoring system: the normal group comprised 88 patients (52.7%), the simple coagulopathy group comprised 58 patients (34.7%), and the overt DIC group comprised 21 patients (12.6%). The one expired patient met the criteria of overt DIC. However, as we mentioned above, the overt DIC seemed to come from multiorgan failure after cardiac arrest, rather than direct effect from coagulopathic venom.

In statistical analysis of subgroups according to the ISTH scoring system, initial white blood cell (WBC) count, initial cholesterol level and hospital stay reached statistical significance (Table 2). Bleeding occurred in six patients, including major bleeding complications in two patients (diffuse alveolar hemorrhage and hemothorax) and minor bleeding in four patients (e.g., ecchymosis, petechiae, and hematoma in extremities). The two patients with major bleeding and two patients with minor bleeding met the criteria of overt DIC. The other two patients with minor bleeding met the criteria for simple coagulopathy. No definite thrombotic or embolic complications were identified in any patient.

We divided the patients into two groups, as patients without overt DIC (normal group plus simple coagulopathy group) and patients with overt DIC. Univariate analyses showed that initial higher WBC count, initial lower albumin level and initial lower cholesterol level were associated with the development of overt DIC. However, in multivariable logistic regression using the variables of which P-value less than 0.100 in the univariate analysis, only the initial lower cholesterol was independently associated with the development of overt DIC (Table 3).

The receiver operation characteristic curve of initial cholesterol level and overt DIC shows that the area under the curve is 0.785 (95% confidence interval, 0.703 to 0.853; P<0.001). The optimal cutoff value is 163 mg/dl by Youden index. Its sensitivity and specificity are 0.85 and 0.62, respectively at the cutoff value (Figure 1).

Among 21 patients with overt DIC, INR was immeasurably high in 19 patients (values exceeding the laboratory’s measurement limits of 8.93, 13, and 25 depending on machine used). For the other two patients, the INR did not increase above 1.5 in one patient and the other patient died before the INR exceeded the maximum measurable value. The median time from snakebite to an INR >1.5 was 2 days, ranging from 0 to 4 days. The median time from snakebite to immeasurably high INR was 3 days, ranging from 0 to 4 days. In 15 patients, an immeasurably high INR was recorded on the same day as INR >1.5. The maximum interval between INR >1.5 and immeasurably high INR was 2 days (Table 4).

In addition, there were three patients who reach immeasurably high INR in the simple coagulopathy group. They reached the immeasurably high INR in 3 days, 4 days and 5 days (but less than 120 hours) after the snakebite, respectively. Those days were same day of INR above 1.5. The highest INR was 1.43 in the patients who did not reach immeasurably high INR and who could assess ISTH scoring system except for the one dead patient.

Among the 21 patients with overt DIC, 16 patients did not experience bleeding (five patients were excluded; four patients had bleeding, the other one patient died before determining clinical result such as bleeding from overt DIC). All of these patients were administered antivenom (Agkistrodon halys antivenom injection; 6,000 units/vial; the only antivenom in South Korea). In 15 of the patients, the first dose was administered within 3 hours after the snakebite. The other one patient was administered the first dose of antivenom more than 9 hours after the snakebite.

INR was immeasurably high in 15 patients, and did not increase above 1.5 in the other one patient in the 16 overt DIC patients without bleeding. Among 15 patients with immeasurably high INR, the duration of INR above 3 and hospital stay were not significantly different between patients whose antivenom dose was ≥18,000 units and patients whose antivenom dose was <18,000 units (Table 5). Eleven patients had fresh frozen plasma (FFP) or cryoprecipitate transfusion. All the eleven patients had FFP transfusion (mean, 9.2 pints from 320 ml or 400 ml of whole blood) and two patients had cryoprecipitate transfusion (mean, 26.5 pints). Likewise, transfusion of FFP or cryoprecipitate was not associated with significant differences in the duration of INR above 3 or hospital stay in the 15 patients with immeasurably high INR (Table 6).

The World Health Organization estimated that there are more than 421,000 cases of venomous snakebites per year, with 20,000 deaths [12]. In a prior study in South Korea, it was reported that between 142 and 621 cases of snakebite occurred per year, with five deaths [3]. In previous studies in South Korea, there were more snakebites in men than in women, and most people were in their 40s or 50s [7,13]. In our study, all snakebites occurred between April and November because snakes in South Korea typically emerge in April and hibernate in November [13].

As described above, coagulopathy is one of the most common complications of snakebites. Kim et al. [11] reported that the cholesterol level was lower in patients with DIC. Winkler et al. [10] reported that the serum total cholesterol level on admission may indicate the severity of envenomation in patients bitten by snakes belonging to the Vipera family before the clinical syndrome fully develops. In our study, multivariable logistic regression showed that initial lower cholesterol level was associated with the development of overt DIC.

Snake venoms that cause coagulopathy can be classified as procoagulant toxins, anticoagulant toxins, platelet activity toxins, and vessel wall interactive toxins [2,14-16]. Coagulopathy manifested as prolonged INR, hypofibrinogenemia, thrombocytopenia, and increased fibrin degradation products in laboratory studies. In our study, 19 patients had immeasurably high INR among the 21 overt DIC patients, which could be explained as venom-induced consumption coagulopathy (VICC).

VICC describes any coagulopathy resulting from the consumption of clotting factors by procoagulant toxins present in snake venom [5]. It is caused by venom from various snake species, including members of the Viperid, Elapidae, and some Colubridae families [17]. DIC results from activation of clotting pathway and has a very high mortality. By contrast, VICC involves a different pathogenic process (specific enzyme activation) and needs to be treated differently, but has a more benign course than DIC [2]. Procoagulant toxins cause rapid clot formation in vitro, but, in vivo, they cause consumption of severe clotting factors therefore increase the risk of bleeding. Different snakes have different kinds of toxins, and the toxins activate various components in the clotting pathway [2]. The toxins in venom that cause VICC are classified by where they effect in the clotting pathway, with the important ones being thrombin-like enzymes (also known as fibrinogenase), prothrombin activators, factor V and X activators. [14]. Lee et al. [7] reported abrupt prolongation of prothrombin time (PT) and activated partial thromboplastin time (APTT) on days 2 and 4 of hospitalization. In our study, we found the peak INR occurred as late as 4 days after the snakebite in the overt DIC group, as late as 5 days (less than 120 hours) in the simple coagulopathy group as well. We also noted that INR prolongation showed an abrupt pattern rather than gradual pattern because INR increased rapidly from 1.5 to immeasurably high on the same day in 15 of 19 patients with an immeasurably high INR in the overt DIC group, three of three patients with an immeasurably high INR in the simple coagulopathy group as well. Therefore, we suggest that coagulopathy should be monitored for at least 4 to 5 days after a snakebite, even in patients without apparent symptoms. Clinicians should also consider the possibility of VICC if their patient’s INR rises above 1.5 within that period of time. The type of toxin and the administration of antivenom determine the duration of VICC. It took 24–48 hours to resolve VICC in Australian elapid envenoming regardless of antivenom treatment, and the rate of recovery was found to depend entirely on the production of new clotting factors [15]. On the other hand, Echis-induced VICC continued for days without antivenom treatment [18]. However, the patients recovered within 24–48 hours with antivenom treatment in Echis-induced VICC. This suggests that once the toxin is neutralized, the clotting factors recover normally. In the study by Lee et al. [7] it took a mean of 11.5±2.4 days for INR to fall below 1.2.

The aim of treating snakebite-induced coagulopathy is to prevent major bleeding and associated complications [2]. This is primarily achieved by neutralization of the procoagulant toxins followed by recovery of clotting factors up to a level where minor injury does not cause major bleeding.

Antivenom is the standard treatment for snake envenoming [19], and is highly effective for binding toxins and neutralizing their effects [20]. After the toxins have been neutralized, the patient’s clotting function should recover within 24−48 hours [21]. However, the clinical benefit of antivenom in VICC may differ for different snakes [22,23]. The antivenom available in South Korea neutralizes venom from three of four venomous snakes in the region [3]. However, there are no specific guidelines regarding its indications and optimal dosage [3].

Clotting factor replacement for VICC without bleeding is controversial because this strategy may worsen VICC by increasing the supply of clotting factors (substrates) to procoagulant toxins, and hence this may “add fuel to the fire” [24,25]. In an Australian study, ongoing reductions in fibrinogen was shown in patients with FFP given within 6 hours of a snakebite, despite prior neutralization with antivenom [25]. However, one study suggested that early use of FFP may allow quicker recovery from VICC [22]. In our study, transfusion of FFP or cryoprecipitate did not influence the duration of increased INR or hospital stay in patients without bleeding. However, patients have active bleeding, clotting factor replacement may expedite the recovery of VICC [2].

This study has several limitations to discuss. First, because this was a retrospective observational study, the admission, examination, treatment, and discharge of patients were at the attending doctor’s preference rather than a predetermined protocol. Second, this study focused on hospitalized patients, not outpatients. Therefore, the patients included in this study may have had more severe snakebites than those treated in outpatient settings. Third, total number of the overt DIC patients was small. Larger number of patients are required to get more definite conclusion.

Snakebites can cause a broad spectrum of symptoms with variable severity; coagulopathy is one of these symptoms. Initial lower cholesterol level could be a risk factor of developing overt DIC. The risk of coagulopathy should be assessed for at least 4 to 5 days following snakebite. Higher doses of antivenom and transfusion with FFP or cryoprecipitate may be unbeneficial for patients without bleeding.