Journal of Acute Disease

: 2021  |  Volume : 10  |  Issue : 2  |  Page : 71--77

Early severity predictors of snakebite envenomation in the southern region of Tunisia: a multivariate analysis

Olfa Chakroun-Walha, Fadhila Issaoui, Abdennour Nasri, Haifa Bradai, Ayman Farroukh, Rim Karray, Mouna Jerbi, Noureddine Rekik 
 Emergency Department, University Hospital Habib Bourguiba Sfax, Tunisia

Correspondence Address:
Olfa Chakroun-Walha
Emergency Department, University Hospital Habib Bourguiba Sfax


Objective: To identify the early predictors of severe envenomation in the southern region of Tunisia. Methods: It was a retrospective monocentric study including consecutive patients admitted to the emergency department for snakebite envenomation. Snakebite envenomation was defined by a history of snakebite. Predictors of severe envenomation were determined by univariate and multivariate analyses. Results: Our sample consisted of 109 patients aged 30 (20-44) years with a 1.1:1 sex-ratio (56 males and 53 females). During the 24-hour surveillance period, 25 patients developed severe envenomation (22.9%). The in-hospital mortality rate was 4.6% (n=5). The independent predictors of severe snakebite envenomation were leucocyte count over 11 550/mm3 (OR: 18.7, 95% CI: 3.3-107.8), creatine kinase over 155 IU/L (OR: 6.16, 95% CI: 1.1-35.6), and/or tourniquet before arrival to the ED (OR: 32.14, 95% CI: 3.5-295.9). Conclusions: This study emphasizes the importance of early evaluation of snakebite envenomation. Further studies are required to approve a severity scale proper to snakebite envenomation in Tunisia.

How to cite this article:
Chakroun-Walha O, Issaoui F, Nasri A, Bradai H, Farroukh A, Karray R, Jerbi M, Rekik N. Early severity predictors of snakebite envenomation in the southern region of Tunisia: a multivariate analysis.J Acute Dis 2021;10:71-77

How to cite this URL:
Chakroun-Walha O, Issaoui F, Nasri A, Bradai H, Farroukh A, Karray R, Jerbi M, Rekik N. Early severity predictors of snakebite envenomation in the southern region of Tunisia: a multivariate analysis. J Acute Dis [serial online] 2021 [cited 2021 Apr 18 ];10:71-77
Available from:

Full Text

 1. Introduction

Snakebite is a serious threat to human health globally[1]. The mortality caused by snake envenomation is high with more than 100 000 deaths annually[2-5]. The most common mode of snake envenomation is biting[6]. Snakebites vary for an asymptomatic dry bite to the severe form. Life-threatening presentations are characterized by shock, and/or hemorrhage, limb necrosis, organ failure, and death[1],[3]. In Tunisia, a few experimental studies were conducted, however, epidemiological researches are still rare[5],[6],[7]. The historical features, clinical findings, and laboratory test results have different abilities to early predict the outcomes, in part because of the wide number of snake species. Besides, there is still no validated severity scale of snakebite envenomation in Tunisia. Therefore, we aim to identify the predicting factors of the severe form of snakebite envenomation in our region.

 2. Materials and methods

2.1. Clinical setting

This is a retrospective study conducted over five years (from January 1st, 2014 to December 31st, 2019). The records of 2016 were not analyzed (missing data). The study population consisted of consecutive patients admitted to the emergency department (ED) for snakebite.

2.2. Inclusion criteria and exclusion criteria

We reviewed all medical records and included consecutive patients who presented with snakebite at the ED. Patients with a history of coagulopathy or antiplatelet administration were excluded. Each time the snakebite envenomation was not clear, the case was excluded. Medical files with missing data were also excluded.

2.3. Management protocol in the ED

Following a snakebite, the patient is hospitalized in the emergency room, regardless of the clinical presentation on arrival to the ED. Patients presenting to a regional ED are also transferred to our institution.

On admission to the emergency room, blood tests are performed (prothrombin, platelets, serum creatine kinase level, serum creatine kinase level). These blood tests are repeated 12 and 24 h after admission.

The severe form was defined as a snake envenomation with at least one of the following complications: hematologic abnormalities, hemorrhage, organ failure (respiratory failure, shock, neurological abnormalities), limb necrosis, compartment syndrome or impending compartment syndrome requiring emergency fasciotomy, and death[3],[8],[9],[10],[11],[12].

We compared the group of patients who developed a severe form of snake envenomation, and those who had a moderate form within the 24-hour of follow-up.

2.4. Ethical approval

Given the nature of the study, the Habib Bourguiba Hospital Institutional Review Board considered this analysis to be exempt from ethical review. Patients’ data were anonymous in this study.

2.5. Used definitions

(1) Hematologic abnormalities (thrombocytopenia, low prothrombin rate, or disseminated intravascular coagulation). Thrombocytopenia was defined as a platelet count below the lower limit (150 000/mm3)[14]; Prothrombin rate was considered low when lower than 60%; Leukocytosis was defined as a leukocyte count over 10 000/mm3; Disseminated intravascular coagulation was defined by the presence of thrombocytopenia and low prothrombin rate. D-Dimer and fibrinogen measurements were not available during the study period in our laboratory.

(2) Organ failure (respiratory failure, shock, cardiac arrest, or neurological abnormalities).

Respiratory failure: pulse oximetry<90%, or needing for mechanical ventilation; Shock: systolic blood pressure<90 mmHg in adults, use of catecholamine (dobutamine, adrenaline or noradrenaline); Neurological abnormalities included in the severity signs were: seizures, confusion, paralysis, agitation, Glasgow coma scale<13; Rhabdomyolysis was defined as a serum creatine kinase level greater than 1 000 IU/L[13].

2.6. Statistical analysis

Statistical analysis was done using SPSS vision 20. Numeric variables were expressed as mean±standard deviation (SD), and dichotomous variables were expressed as percentages or ranges. To compare qualitative variables, we used the Pearson Chi-square test and the Fisher exact test. The Student’s t-test or Fisher exact test and the nonparametric Mann Whitney, test were used for between-group comparisons of discrete and continuous variables, respectively. Receiver-operating characteristic (ROC) curve was used to analyze the correlation between the levels of leucocytes, creatine kinase, prothrombin rate, and platelets and the risk of severe snake bite presentation. The area under the ROC curve was estimated by the method of Hanley and McNeill[15]. Snakebite circumstances (delay of arrival to the hospital, age, tourniquet carried at home), clinical findings on admission (dizziness, vomiting, diarrhea), and blood tests results (leucocytes, platelets, prothrombin rate, creatine kinase level) were evaluated by a multiple logistic stepwise regression procedure. Odds ratios were estimated from the b coefficients obtained, with respective 95% confidence intervals (95% CI). The significance level of the study was set at α=0.05.

 3. Results

We enrolled 109 out of 127 cases of suspected snakebite envenomation admitted in the ED [Figure 1]. The incidence was decreasing each year, and the monthly distribution of the cases showed a peak in the hottest and less rainy months in Tunisia (July, August, and September) [Figure 2]. Limbs were the most common injured parts (94.5% of the cases, n=103) [Table 1]. The snake species were identified in 84 patients (77.1%), encompassing Macrovipera lebetina (n=78; 71.6%), and Naja haje Vipera latastei (n=6; 5.5%). The species was not identified in 25 cases (22.9%).{Figure 1}{Figure 2}{Table 1}

The age of all patients was 30 (20-44) years. Thirteen patients were under 15 years old (11.9%) and 8 patients were above 65 years (7.0%) (range: 4 to 82 years). The sex-ratio was 1.1:1 (56 males and 53 females). Most patients (n=77; 70.6%) were from rural areas. Only 32 patients were from Sfax city (29.4%). In 26% of cases, patients were transferred from other regional hospitals. The arrival to the ED was within the first hour after the snakebite in 26% of cases (n=28) [Table 1].

On admission, blood tests showed leukocytosis in 29 patients (26.6%), thrombocytopenia in 10 patients (9.2%), low prothrombin rate in 11 patients (10.1%).

During the 24-hour follow-up, the observed complications were as follows: respiratory distress (n=16; 14.7%), cardiac arrest (n=5; 4.6%); shock (n=4; 3.7%); neurological trouble (n=6; 5.5%), deep vein thrombosis (n=2; 1.8%), disseminated intravascular coagulation (n=1; 0.9%), compartment syndrome (n=12; 11.0%) and necrosis (n=10; 9.2%).

Administered treatments were analgesics (n=84; 77.1%), oxygen (n=16; 14.7%), endotracheal intubation (n=7; 6.4%), catecholamine (n=5; 4.6%), transfusion (n=9; 8.3%), heparin (n=15; 13.8%), antivenom therapy (n=18; 16.5%), fasciotomy (n=9; 8.3%), tetanus prophylaxis (n=31; 28.4%), and/or antibiotics (n=37; 33.9%).

Forty patients were not discharged within 24 h; the mean inhospital length of stay was (2.9±0.5) d. Eleven patients were transferred to the intensive care unit. The in-hospital mortality rate was 4.6% (n=5).

The severe form was observed in 25 patients (22.9%). [Table 2] shows the predictors of developing severe snakebite envenomation. The sex proportions were comparable in the two groups of patients (P=0.5). On admission, leucocytes count higher than 11 550/mm3, creatine kinase level higher than 155 IU/L, prothrombin rate lower than 76.5%, and/or platelets count lower than 190 000 /mm3 were statistically associated with the severe form [Figure 3].{Figure 3}{Table 2}

Multivariate analysis showed that leukocyte count over 11 550/mm3, creatine kinase over 155 IU/L, and/or carried tourniquet before arrival to the ED were early independent predictors of severe snake bite presentation [Table 3].{Table 3}

 4. Discussion

Snakebites account for one of the serious injuries worldwide[16]. Accurate statistics on snake envenomation are still lacking, and the number of injured patients is expected to be much higher than reported[2],[16]. In Tunisia, during the '90s, the annual incidence of snake envenomation is about 250, with two deaths per year[17]. Nevertheless, snakes are still among the least known reptile species in Tunisia. Compared to scorpions, snakes’ diversity and distribution are rarely studied as well[18],[19]. The most-reported species in Tunisia are Daboia deserti, Daboia mauritanica, Echis leucogaster, Macrovipera lebetina, Cerastes cerastes, and Naja haje Vipera latastei[20]. The three latter ones are the most species found in our region. In our study, the snake species were not systematically recorded in the medical file. However, our data are coinciding with the reported species in our region[20].

Snakebites vary from a dry, also called asymptomatic, bite to severe envenomation[8],[9]. Clinical presentations of snake envenomation may include these three events: tissue injury, hematologic abnormalities, and/or systemic dysfunctions[8],[9],[10],[21]. These different clinical presentations are explained by the complex effects of the venoms. In Tunisia, the venoms of Macrovipera lebetina, Cerastes cerastes, and Naja haje Vipera latastei have hemostatic, hemodynamic, and/or myotoxic effects[5].

The prevalence of severe envenomation is estimated between 14% and 27% globally[1],[21]. In our study, we found severe snake envenomation in 22.9% of the cases. As reported in previous studies, historical factors (age, sex, site of snakebite, medical condition) did not discriminate severe snakebite envenomation[1]. Oppositely, the time between the snakebite and the first medical contact was associated with the severity of the envenomation as was reported in several studies[1]. One of the most concerning historical characteristics of snakebites is the size of the snake. Several studies have shown that patients bitten by large snakes have a threshold risk to develop systemic envenomation[1],[22],[23],[24].

In-vitro researches have demonstrated that the pharmacological effects of snakes’ poison are dose-dependent. The studied species were namely Macrovipera lebetina and Cerastes cerastes, which are major species in our region[7],[25],[26]. The poison dose is frequently correlated with snake size and/or bites number. Snake size description is commonly based on patients’ or relatives’ descriptions. In our study, this information had been missing in almost one-quarter of cases. The number of snakebites had no association with the envenomation severity. Among the historical features, faulty tourniquet application was the unique early independent predictor of severe snakebite envenomation. This finding is coinciding with previous studies[24].

Due to the enormous number of snake species, the physical examination findings are varying in the literature[1]. Their association with the outcomes is varying with the species and the regions[1],[23],[24],[27]. In our sample, we found no clinical early predictors of severity in multivariate analysis.

Concerning the accuracy of laboratory tests in predicting the severity of snakebite envenomation, the data are rare[1],[11],[23],[28]. In a recent literature review, laboratory values were not discriminatory in determining severe envenomation, and only one study found a correlation between thrombocytopenia four days after the snakebite and the severe form[1],[23]. None of these studies has been conducted in the North African region. In our study, we aim to determine early biological predictors of severity. In multivariate analysis, leukocyte count over 11 550/mm3 and/or creatine kinase over 155 IU/L were associated with a severe snakebite form (OR: 18.7 and 6.16; 95% CI: 3.3-107.8 and 1.1-35.6, respectively). This finding can be explained by the pharmacological mechanisms of snake venoms. Recently, Baîram et al. characterized the biochemical effects of snake venom phospholipases A2 isolated from Cerastes cerastes and Macrovipera lebetina transmediterranea[7]. This in-vitro research has demonstrated a higher activity of these phospholipases compared to a thermoactive enzyme activity in these species[7]. Phospholipases A2 plays an important role in diet lipid catabolism and the metabolism of lipid membranes. Furthermore, their hydrolysis products, namely the arachidonic acid, are the precursor of inflammatory activities, via prostaglandins, prostacyclins, thromboxanes, and leukotrienes[29-31]. Thanks to their antitumoral effect, snake venom inoculated in animals or humans induces an increase of polymorphonuclear leukocytes in the early stages of tumor growth[29],[32]. Due to the enormous number of snake species, the physical examination findings are varying in the literature[1].

The management of snake envenomation is based on early symptomatic treatment of organ failures and specific antivenom. In Tunisia, the antivenom is provided by the Pasteur Institute of Tunis (Gamma-VIP). It was generated by hyper immunization of horses with a mixture of Cerastes cerastes and Macrovipera lebetina venoms. It consists of purified F(ab’)2 fragments generated by digestion with pepsin of ammonium sulfate-precipitated IgG molecules[33]. The antivenom is indicated in patients of at least a Grade Π of severity. In our study, it was used in only 16.5% of cases. The in-hospital mortality rate in our study was 4.6%; which is quite higher than other studies[34]. The rate of compartment syndromes is also higher than in some studies[35],[36],[37],[38]. These differences may be explained by the delays between the bite and the arrival to the ED. Besides, these rates are overestimated; several patients did probably not present to the hospital. Most snakebites are in rural areas where no health facility is available. At-home carried first-aids may also explain these outcomes; one-third of the patients with severe snakebite envenomation have had a tourniquet. Awareness about the risk of tourniquet has to be improved.

The limitations of this study are as follows. First, this study included the results of only one emergency department, and it was based on a retrospective medical record review. The sample size was small, and there were insufficient factors to identify confounding variables to determine the historical or clinical features and outcomes of patients. Second, these results are among the species reported in Tunisia. Larger and multi-institutional studies are necessary to generalize these findings to other snake types.

Despite these limitations, we think that our work will be interesting to ED physicians. To the best of our knowledge, this is the one of the first studies evaluating early predictors of snakebite severity in Tunisia. It can be a new base for future studies to approve a severity scale.

In summary, this study shows that the snakebite envenomation is still a serious accidental pathology in Tunisia. Leukocytosis and rhabdomyolysis are predictors of severity. Further studies are required to confirm these findings.

Conflict of interest statement

The authors report no conflict of interest.

Authors’ contributions

Study concept and design (O.C.W., F.I.); acquisition of the data (F.I., H.B., A.F.); analysis of the data (I.F., O.C.W.), drafting of the manuscript (A.N., R.K., M.J.); critical revision of the manuscript (O.C.W., N.R.); approval of final manuscript (O.C.W., N.R.).


1Gerardo CJ, Vissoci JRN, Evan S, Simel DL, Lavonas EJ. Does this patient have a severe snake envenomation? The rational clinical examination systematic review. JAMA Surg 2019; 154(4): 346-354.
2World Health Organization. Guidelines for the management of snakebites, 2nd edition. [Online] Available from: http: // handle/10665/249547. [Accessed on 16th March 2020].
3Kyung Hoon Park, Hyungoo Shin, Hyunggoo Kang, Changsun Kim, Hyuk Joong Choi, Kyunghun Yoo, et al. Effectiveness of repeated antivenom therapy for snakebite-related systemic complications. J Int Med Res 2019; 47: 4808-4814.
4Warrell DA. Envenoming and their treatments. In: Bon C, Goyffon M, (Eds.). Lyon: Foundation Marcel Merieux; 1996, p. 63-76.
5Bazaa A, Marrakchi N, El Ayeb M, Sanz L, Calvete JJ. Snake venomics: Comparative analysis of the venom proteomes of the Tunisian snakes Cerastes cerastes, Cerastes vipera and Macrovipera lebetina. Proteomics 2005; 5(16): 4223-4235.
6Chaari AN, Bahloul M, Tlijani M, Dammak H, Bouaziz M. An unusual mode of snake poisoning in Tunisia. Med Trop 2010; 70(2): 201-202.
7Baîram D, Aissa I, Louati H, Othman H, Abdelkafi-Koubaa Z, Krayem N, et al. Biochemical and monolayer characterization of Tunisian snake venom phospholipases. Int J Biol Macromol 2016; 89: 640-646.
8Dart RC, Hurlbut KM, Garcia R, Boren J. Validation of a severity score for the assessment of crotalid snakebite. Ann Emerg Med 1996; 27(3): 321-326.
9Gold BS, Dart RC, Barish RA. Bites of venomous snakes. N Engl J Med 2002; 347(5): 347-356.
10Seifert SA, Kirschner RI, Martin N. Recurrent, persistent, or late, new-onset hematologic abnormalities in crotaline snakebite. Clin Toxicol 2011; 49(4): 324-329.
11Bucaretchi F, Herrera SR, Hyslop S, Baracat EC, Vieira RJ. Snakebites by Crotalus durissus ssp in children in Campinas, São Paulo, Brazil. Rev Inst Med Trop São Paulo 2002; 44(3): 133-138.
12Hifumi T, Sakai A, Kondo Y, Hifumi T, Sakai A, Kondo Y, et al. Venomous snake bites: clinical diagnosis and treatment. J Intensive Care 2015; 3(1): 16.
13Chavez LO, Leon M, Einav S, Varon J. Beyond muscle destruction: a systematic review of rhabdomyolysis for clinical practice. Critical Care 2016; 20(1): 135.
14Isbister GK. Snakebite doesn’t cause disseminated intravascular coagulation: coagulopathy and thrombotic microangiopathy in snake envenoming. Semin Thromb Hemost 2010; 36(4): 444-451.
15Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 1982; 143(1): 29-36.
16Park KH, Shin H, Kang H, Kim Ch, Choi HJ, Yoo K, et al. Effectiveness of repeated antivenom therapy for snakebite-related systemic complications. J Int Med Res 2019; 47(10): 4808-4814.
17Hamouda C. Envenimations par les serpents en Tunisie. Tawarek 2009; 4: 33-40.
18Kalboussi M, Achour H. Modelling the spatial distribution of snake species in northwestern Tunisia using maximum entropy (Maxent) and Geographic Information System (GIS). J For Res 2018; 29: 233-245.
19Nouira S, Blanc CH, Ktari H. Biodiversité de l’Herpétofaune tunisienne. I. Les Ophidiens. Bull Soc Sci Nat 1995; 24: 76-94.
20World Health Organization. Venomous snakes distribution and species risk categories. 2017. [Online] Available from: bloodproducts/snakeantivenoms/database/default.html. [Accessed on 16th March 2020].
21Long B, April M. What history, examination, and laboratory findings suggest severe snake envenomation? Ann Emerg Med 2019; 74(3): 450-452.
22Janes DN Jr, Bush SP, Kolluru GR. Large snake size suggests increased snakebite severity in patients bitten by rattlesnakes in Southern California. Wilderness Environ Med 2010; 21(2): 120-126.
23Ribeiro LA, Jorge MT, Lebrão ML. Prognostic factors for local necrosis in Bothrops jararaca (Brazilian pit viper) bites. Trans R Soc Trop Med Hyg 2001; 95(6): 630-634.
24Bandyopadhyay SK, Ghosh S, Bandyopadhyay R, Dutta A. Prognostic factors in haemotoxic viper bite: analysis of data from a referral hospital. J Indian Med Assoc 2009; 107(1): 12-13.
25Zouari-Kessentini R, Luis J, Karray A, Kallech-Ziri O, Srairi-Abid N, Bazaa A, et al. Two purified and characterized phospholipases A2 from Cerastes cerastes venom, that inhibit cancerous cell adhesion and migration. Toxicon 2009; 53(4): 444-453.
26Bazaa A, Pasquier E, Defilles C, Limam I, Kessentini-Zouari R, Kallech-Ziri O, et al. MVL-PLA2, a snake venom phospholipase A2, inhibits angiogenesis through an increase in microtubule dynamics and disorganization of focal adhesions. PLoS One 2010; 5(4): e10124.
27Jorge MT, Ribeiro LA, O’Connell JL. Prognostic factors for amputation in the case of envenoming by snakes of the Bothrops genus (Viperidae). Ann Trop Med Parasitol 1999; 93(4): 401-408.
28Barraviera B, Bonjorno Júnior JC, Arkaki D, Domingues MA, Pereira PC, Mendes RP, et al. A retrospective study of 40 victims of Crotalus snake bites: analysis of the hepatic necrosis observed in one patient. Rev Soc Bras Med Trop 1989; 22(1): 5-12.
29Calderon LA, Sobrinho JC, Zaqueo KD, de Moura AA, Grabner AN, Mazzi MV, et al. Antitumoral activity of snake venom proteins: new trends in cancer therapy. Biomed Res Int 2014; 2014: 203639.
30Gutierrez J, Lomonte B. Phospholipase A2 myotoxins from Bothrops snake venoms. Toxicon 1995; 33(11): 1405-1424.
31Teixeira CFP, Landucci ECT, Antunes E, Chacur M, Cury Y. Inflammatory effects of snake venom myotoxic phospholipases A2. Toxicon 2003; 42(8): 947-962.
32da Silva RJ, da Silva MG, Vilela LC, Fecchio D. Cytokine profile of Ehrlich ascites tumor treated with Bothrops jararaca venom. Mediators Inflamm 2002; 11(4): 197-201,
33Fahmi L, Makran B, Pla D, Sanz L, Oukkache N, Lkhider M, et al. Venomics and antivenomics profiles of north African Cerastes cerastes and C. vipera populations reveals a potentially important therapeutic weakness. J Proteomics 2012; 75(8): 2442-2453.
34Senek MZF, Kong SY, Shin SD, Sun KM, Kim J, Ro YS. Epidemiological profile and outcomes of snakebite injuries treated in emergency departments in South Korea, 2011-2016: a descriptive study. Trans R Soc Trop Med Hyg 2019; 113(10): 590-598.
35Hsu CP, Chuang JF, Hsu YP, Wang SY, Fu CY, Yuan KC, et al. Predictors of the development of post-snakebite compartment syndrome. Scand J Trauma Resusc Emerg Med 2015; 23: 97.
36Juckett G, Hancox JG. Venomous snakebites in the United States: management review and update. Am Fam Physician 2002; 65(7): 1367-1374.
37Frangides CY, Koulouras V, Kouni SN, Tzortzatos GV, Nikolaou A, Pneumaticos J, et al. Snake venom poisoning in Greece. Experiences with 147 cases. Eur J Intern Med 2006; 17(1): 24-27.
38Hall EL. Role of surgical intervention in the management of crotaline snake envenomation. Ann Emerg Med 2001; 37(2): 175-180.