https://orcid.org/0000-0001-8006-4917
Correction
4 Jan 2023: Lai CS, Liu PY, Lee CH, Ho CH, Chen WL, et al. (2023) Correction: The development of surgical risk score and evaluation of necrotizing soft tissue infection in 161 Naja atra envenomed patients. PLOS Neglected Tropical Diseases 17(1): e0011035. https://doi.org/10.1371/journal.pntd.0011035 View correction
Figures
Abstract
Background
Naja atra bites cause wound necrosis, secondary infection, and necrotizing soft tissue infection (NSTI) requiring repetitive surgeries. Little information is known about the predictors for surgery after these bites.
Materials and methods
We retrospectively evaluated 161 patients envenomed by N. atra, 80 of whom underwent surgery because of wound necrosis and infection. We compared the patients’ variables between surgical and non-surgical groups. To construct a surgical risk score, we converted the regression coefficients of the significant factors in the multivariate logistic regression into integers. We also examined the deep tissue cultures and pathological findings of the debrided tissue.
Results
A lower limb as the bite site, a ≥3 swelling grade, bullae or blister formation, gastrointestinal (GI) effects, and fever were significantly associated with surgery in the multivariate logistic regression analysis. The surgical risk scores for these variables were 1, 1, 2, 1, and 2, respectively. At a ≥3-point cutoff value, the model has 71.8% sensitivity and 88.5% specificity for predicting surgery, with an area under the receiver operating characteristic curve of 0.88. The histopathological examinations of the debrided tissues supported the diagnosis of snakebite-induced NSTI. Twelve bacterial species were isolated during the initial surgery and eleven during subsequent surgeries.
Discussion and conclusions
From the clinical perspective, swelling, bullae or blister formation, GI effects, and fever appeared quickly after the bite and before surgery. The predictive value of these factors for surgery was acceptable, with a ≥3-point risk score. The common laboratory parameters did not always predict the outcomes of N. atra bites without proper wound examination. Our study supported the diagnosis of NSTI and demonstrated the changes in bacteriology during the surgeries, which can have therapeutic implications for N. atra bites.
Author summary
N. atra bites cause extensive wound necrosis and/or NSTI instead of neuroparalytic effect. We retrospectively evaluated 161 patients envenomed by N. atra, 80 of whom underwent surgeries because of wound necrosis and secondary infection. We found five variables—lower limb as the bite site, a ≥3 swelling grade, bullae or blister formation, GI effects, and fever that occurred rapidly after the bites—that were significant associated with surgery in the multivariate logistic regression analysis. The risk scores for these variables when integer the regression coefficients were 1, 1, 2, 1, and 2, respectively. The predictive value of these factors for surgery is acceptable, with a ≥3-point score. The common laboratory parameters did not always predict the outcomes of N. atra bites in the absence of proper wound evaluation. Deep tissue cultures and histopathological examinations supported the diagnosis of snakebite-induced NSTI. The bacterial species obtained during initial surgery differs from those obtained in subsequent surgeries, suggesting secondary colonization or nosocomial infections. These findings may be helpful for clinicians in better management and disposition of patients with N. atra bites.
Citation: Lai C-S, Liu P-Y, Lee C-H, Ho C-H, Chen W-L, Lai K-L, et al. (2022) The development of surgical risk score and evaluation of necrotizing soft tissue infection in 161 Naja atra envenomed patients. PLoS Negl Trop Dis 16(2): e0010066. https://doi.org/10.1371/journal.pntd.0010066
Editor: Timothy Craig Hardcastle, Inkosi Albert Luthuli Central Hospital, SOUTH AFRICA
Received: July 28, 2021; Accepted: December 6, 2021; Published: February 10, 2022
Copyright: © 2022 Lai et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: Because this is a human research, Data cannot be shared publicly because of legal regulation in Taiwan. However, data can be available from the Taichung Veterans General Hospital after agreement by the Ethics Committee for researchers who meet the criteria for access to confidential data. The contact information to TCVGH Institutional Review Board is: irbtc@vghtc.gov.tw (email); 886-4-23592525#4006 (TEL); 886-4-23592525#4408 (FAX). The data could be available after application assessment by the TCVGH-IRB.
Funding: The author(s) received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Six medically important venomous snake species are found in Taiwan: Trimeresurus stejnegeri stejnegeri, Protobothrops mucrosquamatus, and Deinagkistrodon acutus in the Viperidae family and Crotalinae subfamily; Daboia siamensis in the Viperinae subfamily; and Naja atra and Bungarus multicinctus multicinctus in the Elapidae family [1]. N. atra also is widely found in southeastern China, North Laos, and North Vietnam [2]. In Taiwan, N. atra is the only cobra species, and its bite is uncommon (i.e., 6% [3]), except in the Central Plain region [1].
N. atra envenomation causes serious wound complications, including extensive wound necrosis; necrotizing soft tissue infection (NSTI); finger or toe wet gangrene necessitating distal extremity amputation; and systemic manifestations, such as gastrointestinal (GI) effects and fever [3,4]. Neuroparalytic effects were trivial in adults despite the presence of a short-chained alpha-neurotoxin in the venom [4,5]. Treatment consists of specific antivenom and antibiotics administration and surgical debridement when the necrosis margin is well demarcated [3,4,6].
Although the N. atra bite has been recognized as a risk factor for surgery among all venomous snakebites in Taiwan [7,8] and nearly half of the patients with bites underwent various surgeries because of wound necrosis and secondary infection [4], the risk factors for surgery remain less understood [9]. Therefore, to help first-line medical staff and clinicians better understand the risk factors for surgery and patients’ disposition, we retrospectively analyzed the data for 161 envenomed patients from the Taichung Veterans General Hospital (TCVGH) in central Taiwan.
Materials and methods
Ethics statement
This retrospective observational study was approved by the TCVGH Institutional Review Board (CE16225A and CE17219A). Informed consents were waived for such a retrospective study because of unidentifiable private information after obtaining approval from the Institutional Review Board.
Study population
The patients were admitted to TCVGH between January 2001 and September 2016.
First we identified all possible cases of N. atra bites by searching the computerized database at TCVGH. Two authors independently reviewed the medical records of all patients with possible N. atra envenoming. Only definite and suspected cases were included for analysis [4]. We diagnosed a definite case by examining the culprit snake or laboratory testing of the venom [10–12] and a suspected case by having the patient identify the snake in a standard picture exhibited in the emergency department (ED). Some study participants overlapped with previous studies with a different reference period [3,4,13].
We extracted the following data: sex; age; body part bitten; first aid administered; comorbidities (e.g., diabetes mellitus, liver diseases [i.e., abnormal liver biochemistry, hepatic virus infection, or liver cirrhosis], vascular diseases (i.e., coronary artery disease [CAD] or cerebrovascular accident [CVA]), and malignancies); clinical manifestations; laboratory findings; and management details, including time elapsed between bite and antivenom administration, antivenom dosage, antibiotics administered, indications for surgery, time to initial and last surgery, surgery type, hospital stay length, and follow-up period [4,14]. We used consensus to resolve any disagreement between the authors during the data review and abstraction process.
Definition of the variables
We extracted the variables from the patients’ medical charts. If no anomalies were mentioned in the case notes, we assumed that none were present.
The degree of swelling, which was modified from Blaylock’s classification [15], was categorized as Grade 1 (local swelling at the bite site), Grade 2 (swelling involving a whole hand or foot), Grade 3 (swelling from the hand to the forearm or from the foot to the leg), or Grade 4 (swelling extending to the whole arm, thigh, or the area above). Acute compartment syndrome (ACS) was diagnosed on the basis of typical signs and symptoms and operative notes of intracompartmental pressure >30 mmHg [16]. We classified the case with only clinical signs and symptoms suggestive of ACS but without documentation as suspected cases.
Local numbness was described as an effect that did not extend beyond the affected limb. Lymphangitis was clinically identified as a red line originating from the wound, whereas swollen tender lymph glands draining the affected area denoted lymphadenitis [4,14,17]. NSTI, including necrotizing changes in the layers of the soft tissue compartment and finger or toe wet gangrene were documented in the operations’ notes [4,18]. Fever was defined as a body temperature of ≥38°C measured using a tympanic thermometer. If a patient underwent surgery, only preoperative fever was included in analysis. GI effects included the presence of vomiting or diarrhea [4,11]. Ptosis was defined as partial and complete [19]; motor deficits were defined according to the modified Medical Research Council Scale for Muscle Strength [20].
All patients received blood tests on arrival at the hospital, and these data were repeatedly measured at the clinicians’ request thereafter. We analyzed the blood tests, including white blood cell (WBC) count, neutrophil-to-lymphocyte ratio (NLR), hemoglobin (Hgb), C-reactive protein (CRP), serum sodium, serum creatinine, blood glucose, and creatine kinase (CK), taken within 24 h post-bite and before surgery for those patients who underwent surgery [21,22]. Deep tissue culture was performed during surgical debridement, bullae fluid culture was performed with needle fluid aspiration, and blood culture was performed during febrile episodes. The culture sampling technique has been described previously [23]. Antibiotics administration was defined as any antibiotics used within 48 h after the bite.
Statistical analysis
We compared the data between the definite and suspected cases as well as the surgical and non-surgical cases using the Mann–Whitney U test for the continuous variables and the chi-square test or Fisher’s exact test for the categorical variables. We developed a logistic regression model to predict surgery [21].
First, we included the factors including patients’ characters, clinical variables and laboratory findings, that had showed a significant association with surgery in the univariate logistic regression analysis in the multivariate backward elimination logistic regression analysis, with laboratory data obtained within 24 h post-bite and before surgery. We reported odds ratios (ORs) with corresponding 95% confidence intervals (CIs).
Second, to construct a diagnostic scoring system, we converted the predictors’ regression coefficients in the final model into integers. We expressed the score’s predictive accuracy as an area under the receiver operating characteristic (AUROC) curve. We used the Hosmer–Lemeshow analysis to test the model’s performance. We further examined the Laboratory Risk Indicator for Necrotizing Fasciitis (LRINEC) score [21] in those patients with NSTI [4].
We analyzed all data with SPSS version 26.0 (released 2019; IBM Corp., Armonk, NY, USA). A two-tailed p-value of <0.05 was considered statistically significant. If the Mann–Whitney U test was used, the statistical significance indicated a difference in groups and not association.
Results
Patient characteristics and clinical variables
A total of 170 patients with N. atra bites met the inclusion criteria. Nine were bites without noticeable signs or symptoms of envenoming; therefore, we analyzed data for 161 envenomed cases (Table 1). A total of 57 patients (35.4%) were definite cases, and 104 were suspected cases (Table 1, left columns); 80 patients (49.7%) had surgery, and 81 did not (Table 1, right columns). No significant variation was observed in the symptomatology between the definite and suspected cases.
Local swelling and pain were observed in all envenomed patients. Suspected ACS was found in eight patients (5%); seven of these received surgery because of concomitant wound necrosis and suspected secondary infection, and one patient did not have wound necrosis and refused surgery, ultimately recovering well after antivenom therapy. Wound necrosis; bullae or blister formation; local numbness; lymphadenitis or lymphangitis; fever; GI effects including vomiting or diarrhea; and neurological effects such as partial ptosis or equivocal, barely detectable weakness (Grade 5–) occurred in 68.3%, 20.5%, 23.6%, 6.8%, 34.2%, 32.3%, and 5% of the patients, respectively. NSTI was identified in 44.7%.
The median time elapsed between bite and GI effects and that between bite and fever was 2.3 h [interquartile range (IQR), 1–4.2 h) and 22 h (IQR, 13–37.6 h), respectively. Bullae or blisters in patients appeared before surgery, whereas swelling in patients reached its highest grade but not necessarily most swelling before surgery if patients underwent surgery. No deaths happened during the study period.
Patient laboratory findings
The blood WBC, NLR, Hgb, CRP, serum sodium, creatinine, blood glucose, and CK levels could be obtained in 142, 138, 142, 83, 139, 142, 103, and 129 patients, respectively. No snakebite-related systemic bleeding or coagulopathy occurred.
Aerobic or anaerobic deep tissue cultures were performed in 33 of the 72 patients with NSTI, including a positive result in 30 of the 33 patients who received an aerobic culture and in 1 of the 19 patients who received an anaerobic culture (Table 1). Thirteen patients had polymicrobial infections. Blood culture was performed in 43 patients, and 3 had a positive result, including Shewanella algae, Bacteroides fragilis, and an unidentified gram-positive bacillus. The antibiotics administered in the three patients before obtaining the positive blood culture were cefazolin, gentamicin, and piperacillin/tazobactam in Patient 1; cefazolin, gentamicin, and ampicillin/sulbactam in Patient 2; and ampicillin/sulbactam in Patient 3. Bullae fluid culture was available in one patient in whom a positive blood culture (S. algae) was obtained at the same time; the isolated pathogens were S. algae and Morganella morganii.
Eighteen bacterial species were isolated in the deep tissue cultures (Table 2). These cultures were further grouped based on their acquisition during the initial or subsequent surgeries. During the initial surgery, 22 patients and 1 patient had a positive report in 24 aerobic and 16 anaerobic culture samplings, respectively. The most common isolated pathogens during the initial surgery were Enterococcus sp. and M. morganii, which also contributed to the polymicrobial infection. In addition, Group A Streptococcus sp., S. putrefaciens, Aeromonas hydrophila, Citrobacter freundii, Proteus mirabilis, Yokenella regensburgei, and B. fragilis were seen only in the initial surgery. Coagulase-negative Staphylococcus sp., Staphylococcus aureus, Enterobacter cloacae, Providencia rettgeri, Pseudomonas aeruginosa, and Stenotrophomonas maltophilia were seen only in the subsequent surgeries. The antibiotics administered before the initial surgery to obtain the culture are summarized in Table 2.
Patient management and outcomes
A total of 137 patients (85.1%) received the specific antivenom within 6 h after the bite (Table 3). The exact time of first antivenom administration was available in 88 cases; and the median time elapsing between the bite and administration of initial antivenom dose was 2.6 h (IQR, 1.4–5.6 h), and the median total dose was 10 vials (IQR, 6–14 vials). A total of 150 patients (93.2%) received antibiotics.
The surgical indications were wound necrosis and secondary infection in all 80 patients. The initial surgery was performed at a median of 3 days (IQR, 1.9–5.4 days), whereas the last surgery was performed at a median of 25.7 days (IQR, 18.0–41.6 days) after the bite. The median hospital stay was 6.8 days (IQR, 2.0–23.8 days) for all patients; the stay was 2.1 days (IQR, 0.9–4.9 days) for the non-surgical patients and 23.4 days (IQR, 9.7–30.8 days) for the surgical patients. The median follow-up period was 37 days (IQR, 14–70 days) (Table 2). Surgical pathology was obtained in 14 patients, and acute necrotizing inflammation was commonly mentioned (Table 4).
Statistical findings
From the univariate logistic regression analysis, the significant clinical variables for surgery were lower limb as the bite site, swelling grade, bullae or blister formation, GI effects, and fever, and the significant laboratory variables were WBC and NLR (Tables 5 and S1). Because no surgery was performed in patients with Grade 1 swelling, we grouped the swelling grade into a dichotomous variable for better comparison (i.e., Grade ≥3 and <3). Wound necrosis is the indication for surgery but not predictor; therefore, it was not entered into the logistic regression model. As a result, we found that lower limb as the bite site, a ≥3 swelling grade, bullae or blister formation, GI effects, and fever were significant in the multivariate backward elimination logistic regression analysis. The WBC and NLR were potential confounders. We then fit these significant variables into a final multivariate logistic regression model containing the entire study population. The regression coefficients of the predictors in the final model were 1.4, 1.4, 2.1, 1.1, and 1.7, respectively. The scores were 1, 1, 2, 1, and 2, respectively, when converting the regression coefficients. No interaction was observed between any two of the five variables.
The AUROC curve of the score for surgery was 0.88 (95% CI of 0.83–0.94, p-value <0.001). At a ≥3-point cutoff value, the model had 71.8% sensitivity and 88.5% specificity, whereas the positive likelihood ratio was 6.22 (95% CI, 3.31–11.68) and the negative likelihood ratio was 0.32 (95% CI, 0.22–0.46) in the prediction of surgery. The goodness-of-fit for the final logistic regression model was assessed using the Hosmer–Lemeshow test. Thus, we found that p-value was 0.191 (>0.05), indicating the proper model fitting. Additionally, we observed a positive association between bullae or blister formation, GI effects, or fever and tissue swelling grade and the direction of association was not changed in the final multivariate logistic regression model (Table 5) suggesting that the model assumption was not violated.
All the required laboratory indicators to construct the LRINEC score (i.e., WBC, Hgb, CRP, sodium, creatinine, and glucose; see S2 Table) were available together in 64 patients. The AUROC curve of the LRINEC score for NSTI was 0.55 (95% CI, 0.4–0.69), with an insignificant p-value of 0.52.
Discussion
N. atra has distinct biological features that makes visual identification relatively easy [6,24]. The clinical manifestations caused by N. atra envenoming, however, overlap with those of the common crotaline snakebites [11,14,25]. Therefore, in this study we only included patients with a definite diagnosis and those who saw the culprit snake to construct the surgical risk score and excluded those with snakebites for which the culprit snake could not be identified (i.e., clinical diagnosis [4]).
Only a few studies on the risk factors for surgery in snakebites from a single species or genus exist. These factors were finger as the bite site and bullae or blister formation in P. mucrosquamatus bites analyzed with multivariate logistic regression [14] and finger as the bite site, blister and abscess formation, and venom-induced coagulopathy in Bothrops spp. bites analyzed with chi-square analysis [26]. The risk factors for wound necrosis and, therefore, possible surgical intervention were finger or toe as the bite site, bullae or blister formation, and venom-induced coagulopathy in green pit viper (Trimeresurus spp.) bites in Thailand [27] and cold packing as the first aid, bullae and blister formation, and suspected wound infection in T. s. stejnegeri bites; both studies were analyzed using multivariate logistic regression [25]. Another study has found risk factors of ecchymosis and cyanosis in rattlesnake bites in North America, analyzed as relative risk assessment [28]. Our study demonstrated that lower limb as the bite site, a ≥3 swelling grade, bullae or blister formation, GI effects, and fever were significantly associated with surgery in N. atra bites.
From the clinical perspective, these effects appeared quickly after the bite and before surgery. The predictive value for surgery is acceptable, with a ≥3-point risk score. In addition, to avoid determining surgical risk mainly for ACS, which had been reported as extremely high (i.e., 6.6%–39.0%) in Taiwan patients with snakebites [8,29–31], we comprehensively reviewed the medical records of eight patients with suspected ACS, finding that seven had wound necrosis and suspected wound infection before fasciotomy or fasciectomy; hence, we included these in our analysis. We did not incorporate wound necrosis and secondary infection as a predictor in the model, as these already were surgical indications [4,14].
We also found a positive association between bullae or blister formation, GI effects, or fever and swelling grade. Swelling is generally accepted as one measurement of the severity of snakebites [15,32,33]; in Vipera spp. bites, the swelling grade is correlated with the serum venom level [34,35]. N. atra venom consists of a complex mixture of proteins and peptides, mainly cardiotoxins (CTXs, 52.9%–59.4%), neurotoxins (NTXs, 20.5%–23.5%), phospholipase A2s (PLA2s, 14%–16.8%), cysteine-rich secretory proteins (CRISPs, 2.2%–2.4%), snake venom metalloproteinases (SVMPs, 1.3%–1.7%), and other trace constituents [36,37]. Wound swelling, pain, and necrosis are probably induced by the synergistic actions of CTXs, PLA2s [38–42], and other cytotoxic enzymes (e.g., CRISPs) [36,43]; however, the mechanisms of bullae or blister formation are poorly understood, which may be partly induced by the proteolytic effects of SVMPs in the skin [44].
Mao et al. have suggested removing the bullae or blisters to reduce the venom load in the bite wound [14] because venom has been detected in the bullae fluids after the P. mucrosquamatus bite [45]. Coincidentally, the venom proteins of the Naja species (i.e., CTXs and PLA2s) have been detected in the bullae fluids after the bite [46]. Even though the current snakebite management guideline recommends that these bullae or blisters should be kept intact and aspirated only if they threaten to rupture [17], the bullae or blisters in N. atra bites are almost always developed in patients with wound necrosis, so early removal likely will be helpful for better wound inspection and to reduce the toxin load, similar to the management of chemical burns [47].
Notably, a 57-year-old male patient who was bitten by an N. atra on his left foot had repeated vomiting and diarrhea within 1 h post-bite, developing bullae and blisters, wound necrosis, lymphangitis and lymphadenitis, fever, and necrotizing fasciitis requiring multiple surgeries. His blood and bullae fluid cultures before initial surgery yielded S. algae and S. algae and M. morganii, respectively. Bacteria may have a role in the formation of bullae or blisters [48]. Furthermore, we have provided evidence that Shewanella soft tissue infection is a route of bacteremia transmission [49,50].
GI effects were reported only in N. atra bites, in contrast to the other common crotaline snakebites (i.e., T. s. stejnegeri and P. mucrosquamatus) in Taiwan [4,14,25]. Typically, patients presented with repeated vomiting and diarrhea in the first few hours post-bite before treatment and promptly responded to the administration of specific antivenom and/or anti-emetics [4]. Muscarinic PLA2s and/or CTXs could be involved [51,52].
However, fever, which also correlates with swelling grade, is more likely to be caused by multiple factors, including snakebite-related tissue inflammation and/or infection [3,15]. Pyrogenic response to the administration of antivenom is less favorable, as that has not been reported in the past 20 years in Taiwan [53]. Moreover, this response has not been observed in B. m. multicinctus bites that cause minimal tissue reaction, and patients receive the same bivalent specific antivenom for N. atra and B. m. multicinctus bites [54]. Therefore, we believe that fever is an important feature of N. atra bites and suggest that the presence of GI effects and fever are caused by high venom inoculation and, hence, more tissue swelling and/or wound complications (i.e., wound necrosis, NSTI, and surgery).
The reasons for increased risk for surgery in lower limbs as the bite sites for N. atra bite remain unknown. We suspect a higher likelihood of wound contamination (e.g., water and soils) and a secondary infection in lower limb than in other anatomic regions.
We observed a high incidence of NSTI following N. atra bites. The diagnosis of NSTI requires prompt surgical exploration in the cases of high clinical suspicion, especially when a patient’s pain is out of proportion [18] and the patient is unresponsive to the administration of antivenom [55,56]. NSTI’s histopathology findings include necrosis of the superficial fascia; polymorphonuclear WBCs in the deep dermis and fascia; edema of the reticular dermis, subcutaneous fat, and superficial fascia; fascial artery and vein fibrinous involvement; and visible bacteria on the gram stain of the fascia and dermis [57,58].
The patients in our study supported the diagnosis of snakebite-induced NSTI. However, a median time of 8.6 h between NSTI diagnosis and surgery has been reported [59], contradicting our study’s findings. Our patients’ surgeries were delayed, likely because of the aggressiveness of the pathogens and the relatively healthy status of the patients with N. atra bites [13,60]. NSTI has been described as three microbiological classes based on the intraoperative bacterial culture findings [61]: Type I (polymicrobial), Type II (monomicrobial), and Type III (clostridial myonecrosis) [61,62].
We found that the most common pathogens (i.e., M. morganii and Enterococcus spp.) isolated during the initial surgery differed from those in Type I NSTI [63,64]. These pathogens were also identified in the oropharynx of N. atra [13,65,66], which suggested that they may come from the snake’s mouth during envenoming. In addition, a few pathogens were identified only in the later surgeries, suggesting secondary colonization or nosocomial infection. This information may have therapeutic implications for clinicians tailoring antibiotic administration at different periods post-bite.
We further examined the association between the laboratory data (i.e., WBC, NLR [22], and LRINEC score [21]) and surgery or NSTI [4]. However, these parameters obtained on arrival in the ED failed to predict surgery after being adjusted with the clinical risk factors, probably because the predictive role of these non-specific inflammatory markers was mediated through these factors. In addition, the progress of bacterial inoculation to infection is dynamic [67]. Serial laboratory data could be better correlated with the infection continuum than only with a single data set obtained on patients’ arrival.
The Taiwan government produces four types of antivenom against six venomous snakebites: a bivalent for N. atra and B. m. multicinctus, a bivalent for T. s. stejnegeri and P. mucrosquamatus, a monovalent for D. acutus, and a monovalent for D. siamensis. These formulas have not changed since the 1990s, and they all are ammonium sulfate-precipitated F(ab′)2 fragments in a lyophilized form [6,68].
According to the position statement from the Taiwan Poison Control Center (PCC), 6–10 vials are recommended to treat an envenomed patient, a figure based on the average amount of venom milked from a cobra (mean, 48 or 217 mg for the western and eastern types of N. atra, respectively) and the neutralizing activity of the antivenom (8.58 or 17.42 mg per vial, respectively) [6]. Our findings are consistent with those of a previous study and the PCC’s suggestion that the median antivenom dose of 10 vials be administered [4]. However, a recent rodent study has demonstrated that the bivalent antivenom was insufficient to neutralize the cytotoxicity produced by CTXs [69]. Prospective evaluations of the optimal dosages of antivenom to counteract the synergistic cytotoxic effects of the enzymes that participate in the induction of wound necrosis and a better design of antibodies and medications against CTXs in N. atra or other cytotoxic Naja bites are urgently required [70].
Study limitations
This study has several limitations. First, because N. atra bites are uncommon in Taiwan, a major limitation is the lack of a second population to validate the surgical risk score. The inclusion of suspected cases also raises the possibility of some of them not having been bitten by cobras since it is widely accepted that the identification of the culprit snake by the patient is usually imprecise. Nonetheless, the comparison between definite and suspected cases (Table 1) did not show significant differences, which is in support of a correct diagnosis.
Second, the wound bacteriology might have been underestimated because we uncovered only 18 bacterial species compared to 23 in Mao et al.’s study [3]. Possible reasons include the lower number of bacterial species in the deep tissue cultures than in the wound swab sampling, the fastidious or uncultivable bacteria, and the several treatment forms (e.g., wound cleansing and topical herb application) initiated before the sampling. In addition, all patients received antibiotics before obtaining the surgical samples, which might have changed the bacterial loads and complexity. Therefore, we suggest a next-generation sequencing technique to better discriminate wound bacteriology [13]. Nevertheless, we reported only deep tissue cultures, which generally are accepted as the gold standard to diagnose wound infection and as guidance for further antibiotics administration [71].
Third, the histopathological examination of the debrided tissue in the N. atra bites was not routinely performed at our hospital. The debrided tissues did not always contain all layer parts for examination, and the specific changes in tissue layers due to various pathogens could not be demonstrated clearly [58,72]. Moreover, the biological mechanisms of the venom-induced tissue injury remains undetermined [42].
Finally, the data of the study were extracted from medical charts that cover a long period of time, and this introduces the possibility of variations in the way the information has been introduced into the charts; that is, different physicians may introduce a different degree of detail in the charts. Additionally, it is likely that different physicians were involved in the therapeutic and surgical decisions in these patients, with the likelihood of having differences in the trends to do or not to do surgery at different times. Because this study is retrospective, which has certain inherent limitations, the results should be interpreted cautiously. Nevertheless, ours is the first study to investigate the risk factors for surgery for N. atra bites, which can cause serious wound complications in Taiwan.
Conclusions
Although N. atra bites are relatively uncommon in Taiwan, they can cause serious local wound complications in contrast to the common crotaline snakebites (i.e., T. s. stejnegeri and P. mucrosquamatus), thus, requiring repetitive debridement and reconstructive surgeries to restore limb function. Our study suggested that lower limb as the bite site, ≥3 swelling grade, bullae or blister formation, GI effects, and fever are risk factors for surgery. We further analyzed wound bacteriology during the initial surgery and provided histopathological evidence for NSTI. These findings can be helpful for patient management and disposition following an N. atra bite. When used alone, common laboratory parameters, although helpful for measuring inflammation and infection, may not accurately predict the outcomes of N. atra bites.
Supporting information
S1 Table. Univariate logistic regression analysis of risk factors associated with surgery in 161 Naja atra envenomed patients.
https://doi.org/10.1371/journal.pntd.0010066.s001
(DOCX)
S2 Table. Laboratory findings in NSTI and non-NSTI groups.
https://doi.org/10.1371/journal.pntd.0010066.s002
(DOCX)
References
- 1.
Mao YC, Hung DZ. Epidemiology of Snake Envenomation in Taiwan. Gopalakrishnakone P, Faiz MA, Fernando R, Gnanathasan CA, Habib AG, Yang CC, eds. Clinical Toxinology in Asia Pacific and Africa. Singapore: Springer. 2015:3–22.
- 2. Uetz P, Freed P, Hošek J, et al. The Reptile Database. Retrieved from http://wwwreptile-databaseorg/, accessed May 20, 2021.
- 3. Mao YC, Liu PY, Hung DZ, Lai WC, Huang ST, Hung YM, et al. Bacteriology of Naja atra Snakebite Wound and Its Implications for Antibiotic Therapy. Am J Trop Med Hyg. 2016;94(5):1129–35. pmid:26976881
- 4. Mao YC, Liu PY, Chiang LC, Lai CS, Lai KL, Ho CH, et al. Naja atra snakebite in Taiwan. Clin Toxicol (Phila). 2018;56(4):273–80. pmid:28830248
- 5. Tan CH, Wong KY, Chong HP, Tan NH, Tan KY. Proteomic insights into short neurotoxin-driven, highly neurotoxic venom of Philippine cobra (Naja philippinensis) and toxicity correlation of cobra envenomation in Asia. J Proteomics. 2019;206:103418. pmid:31201947
- 6.
Mao YC, Hung DZ. Management of Snake Envenomation in Taiwan. Gopalakrishnakone P, Faiz MA, Fernando R, Gnanathasan CA, Habib AG, Yang , eds. Clinical Toxinology in Asia Pacific and Africa. Singapore: Springer. 2015:23–52.
- 7. Shih YC, Ma H, Yeh FL, Lin JT, Hwang CH, Wang MS, et al. Risk factors of surgical intervention in the management of venomous snakebite in Taiwan. J Plast Surg Assoc ROC. 2006;15(4):367–76.
- 8. Hsieh YH, Hsueh JH, Liu WC, Yang KC, Hsu KC, Lin CT, et al. Contributing Factors for Complications and Outcomes in Patients With Snakebite: Experience in a Medical Center in Southern Taiwan. Ann Plast Surg. 2017;78(3 Suppl 2):S32–6.
- 9. Su HY, Wang MJ, Li YH, Tang CN, Tsai MJ. Can surgical need in patients with Naja atra (Taiwan or Chinese cobra) envenomation be predicted in the emergency department? Hong Kong Med J. 2016;22(5):435–44. pmid:27516567
- 10. Huang YP, Yu YJ, Hung DZ. Sandwich enzyme-linked immunosorbent assay for Taiwan cobra venom. Vet Hum Toxicol. 2002;44(4):200–4. pmid:12136964
- 11. Hung DZ, Liau MY, Lin-Shiau SY. The clinical significance of venom detection in patients of cobra snakebite. Toxicon. 2003;41(4):409–15. pmid:12657310
- 12. Hung DZ, Lin JH, Mo JF, Huang CF, Liau MY. Rapid diagnosis of Naja atra snakebites. Clin Toxicol (Phila). 2014;52(3):187–91. pmid:24580058
- 13. Mao YC, Chuang HN, Shih CH, Hsieh HH, Jiang YH, Chiang LC, et al. An investigation of conventional microbial culture for the Naja atra bite wound, and the comparison between culture-based 16S Sanger sequencing and 16S metagenomics of the snake oropharyngeal bacterial microbiota. PLOS Negl Trop Dis. 2021;15(4):e0009331. pmid:33857127
- 14. Mao YC, Liu PY, Chiang LC, Lee CH, Lai CS, Lai KL, et al. Clinical manifestations and treatments of Protobothrops mucrosquamatus bite and associated factors for wound necrosis and subsequent debridement and finger or toe amputation surgery. Clin Toxicol (Phila). 2021;59(1):28–37. pmid:32400229
- 15. Blaylock R. Antibiotic use and infection in snakebite victims. S Afr Med J. 1999;89(8):874–6. pmid:10488365
- 16. Tiwari A, Haq AI, Myint F, Hamilton G. Acute compartment syndromes. Br J Surg. 2002;89(4):397–412. pmid:11952578
- 17.
Guidelines for the management of snakebites. World Health Organization: Regional Office for South-East Asia, New Delhi, India. 2016.
- 18. Hakkarainen TW, Kopari NM, Pham TN, Evans HL. Necrotizing soft tissue infections: review and current concepts in treatment, systems of care, and outcomes. Curr Probl Surg. 2014;51(8):344–62. pmid:25069713
- 19. Silva A, Maduwage K, Sedgwick M, Pilapitiya S, Weerawansa P, Dahanayaka NJ, et al. Neuromuscular Effects of Common Krait (Bungarus caeruleus) Envenoming in Sri Lanka. PLOS Negl Trop Dis. 2016;10(2):e0004368. pmid:26829229
- 20. Jackson CE. A clinical approach to muscle diseases. Semin Neurol. 2008;28:228–40. pmid:18351524
- 21. Wong CH, Khin LW, Heng KS, Tan KC, Low CO. The LRINEC (Laboratory Risk Indicator for Necrotizing Fasciitis) score: a tool for distinguishing necrotizing fasciitis from other soft tissue infections. Crit Care Med. 2004;32(7):1535–41. pmid:15241098
- 22. Elbey B, Baykal B, Yazgan ÜC, Zengin Y. The prognostic value of the neutrophil/lymphocyte ratio in patients with snake bites for clinical outcomes and complications. Saudi J Biol Sci. 2017;24(2):362–6. pmid:28149174
- 23. Bowler PG, Duerden BI, Armstrong DG. Wound microbiology and associated approaches to wound management. Clin Microbiol Rev. 2001;14(2):244–69. pmid:11292638
- 24.
Mao YC, Liu PY, Chiang LC, Yang CC. Bacteriology of Naja atra snakebite wound and its implications for antibiotic therapy. Vascular Surgery, Neurosurgery, Lower Extremity Ulcers, Antimicrobials, Wound Assessment, Care, Measurement and Repair. Switzerland: Springer. 2018:95–107.
- 25. Chiang LC, Tsai WJ, Liu PY, Ho CH, Su HY, Lai CS, et al. Envenomation by Trimeresurus stejnegeri stejnegeri: clinical manifestations, treatment and associated factors for wound necrosis. J Venom Anim Toxins Incl Trop Dis. 2020;26. pmid:32983233
- 26. Jorge M, Ribeiro L, O’connell J. Prognostic factors for amputation in the case of envenoming by snakes of the Bothrops genus (Viperidae). Ann Trop Med Parasitol. 1999;93(4):401–8. pmid:10656041
- 27. Rojnuckarin P, Mahasandana S, Intragumthornchai T, Sutcharitchan P, Swasdikul D. Prognostic factors of green pit viper bites. Am J Trop Med Hyg. 1998;58(1):22–5. pmid:9452286
- 28. Heise CW, Ruha A-M, Padilla-Jones A, Truitt Hayek C, Gerkin RD. Clinical predictors of tissue necrosis following rattlesnake envenomation. Clin Toxicol (Phila). 2018;56(4):281–4. pmid:28885114
- 29. Su HY, Li YH, Tang CN, Su CL, Tsai MJ. Can surgery in patient with Protobothrops mucrosquamatus envenomation be predicted in emergency department? Hong Kong J Emerg Med. 2016;23(4):210–9.
- 30. Hsu 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. pmid:26561300
- 31. Ho CH, Mao YC, Tsai YD, Lin CS, Liu SH, Chiang LC, et al. Descriptive study of snakebite patients in Northern Taiwan: 2009 to 2016. J Med Sci. 2019;39(3):114.
- 32. Resiere D, Mégarbane B, Valentino R, Mehdaoui H, Thomas L. Bothrops lanceolatus bites: guidelines for severity assessment and emergent management. Toxins. 2010;2(1):163–73. pmid:22069552
- 33. Gerardo CJ, Vissoci JR, Evans CS, Simel DL, Lavonas EJ. Does this patient have a severe snake envenomation?: the rational clinical examination systematic review. JAMA Surg. 2019;154(4):346–54. pmid:30758508
- 34. Audebert F, Sorkine M, Bon C. Envenoming by viper bites in France: clinical gradation and biological quantification by ELISA. Toxicon. 1992;30(5–6):599–609. pmid:1519251
- 35. Audebert F, Sorkine M, Robbe-Vincent A, Bon C. Viper bites in France: clinical and biological evaluation; kinetics of envenomations. Hum Exp Toxicol. 1994;13(10):683–8. pmid:7826686
- 36. Huang HW, Liu BS, Chien KY, Chiang LC, Huang SY, Sung WC, et al. Cobra venom proteome and glycome determined from individual snakes of Naja atra reveal medically important dynamic range and systematic geographic variation. J Proteomics. 2015;128:92–104. pmid:26196238
- 37. Shan LL, Gao JF, Zhang YX, Shen SS, He Y, Wang J, et al. Proteomic characterization and comparison of venoms from two elapid snakes (Bungarus multicinctus and Naja atra) from China. J Proteomics. 2016;138:83–94. pmid:26924299
- 38. Wang JP, Teng CM. Roles of mast cells and PMN leukocytes in cardiotoxin-induced rat paw edema. Eur J pharmacol. 1989;161(1):9–18. pmid:2721549
- 39. Liu YB, Xu K. Lack of the blocking effect of cobrotoxin from Naja naja atra venom on neuromuscular transmission in isolated nerve muscle preparations from poisonous and non-poisonous snakes. Toxicon. 1990;28(9):1071–6. pmid:2175458
- 40. Wang JP, Teng CM. Effects of anti-inflammatory drugs on rat hind-paw swelling caused by phospholipase A2 from Naja naja atra venom. Naunyn-Schmiedeb Arch Pharmacol. 1991;344(3):377–81. pmid:1683689
- 41. Lai MK, Wen CY, Lee CY. Local lesions caused by cardiotoxin isolated from Formosan cobra venom. J Formos Med Assoc. 1972;71(6):328–32. pmid:4512908
- 42. Rivel M, Solano D, Herrera M, Vargas M, Villalta M, Segura Á, et al. Pathogenesis of dermonecrosis induced by venom of the spitting cobra, Naja nigricollis: An experimental study in mice. Toxicon. 2016;119:171–9. pmid:27288896
- 43. Tadokoro T, Modahl CM, Maenaka K, Aoki-Shioi N. Cysteine-Rich Secretory Proteins (CRISPs) from Venomous Snakes: An Overview of the Functional Diversity in a Large and Underappreciated Superfamily. Toxins. 2020;12(3):175. pmid:32178374
- 44. Jiménez N, Escalante T, Gutiérrez JM, Rucavado A. Skin pathology induced by snake venom metalloproteinase: acute damage, revascularization, and re-epithelization in a mouse ear model. J Invest Dermatol. 2008;128(10):2421–8. pmid:18449209
- 45. Lin CC, Wang PJ, Liu CC. Venom concentrations in blisters and hemorrhagic bullae in a patient bitten by a Taiwan habu (Protobothrops mucrosquamatus). Rev Soc Bras Med Trop. 2019;52.
- 46. Macêdo JK, Joseph JK, Menon J, Escalante T, Rucavado A, Gutiérrez JM, et al. Proteomic analysis of human blister fluids following envenomation by three snake species in India: Differential markers for venom mechanisms of action. Toxins. 2019;11(5):246. pmid:31052189
- 47. Gnaneswaran N, Perera E, Perera M, Sawhney R. Cutaneous chemical burns: assessment and early management. Aust Fam Physician. 2015;44(3):135. pmid:25770579
- 48. Nishifuji K, Sugai M, Amagai M. Staphylococcal exfoliative toxins:“molecular scissors” of bacteria that attack the cutaneous defense barrier in mammals. J Dermatol Sci. 2008;49(1):21–31. pmid:17582744
- 49. Huang YT, Tang YY, Cheng JF, Wu ZY, Mao YC, Liu PY. Genome analysis of multidrug-resistant Shewanella algae isolated from human soft tissue sample. Front Pharmacol. 2018;9:419. pmid:29755354
- 50. Liu PY, Lin CF, Tung KC, Shyu CL, Wu MJ, Liu JW, et al. Clinical and microbiological features of shewanella bacteremia in patients with hepatobiliary disease. Intern Med. 2013;52(4):431–8. pmid:23411697
- 51. Lin-Shiau S, Huang H, Lee C. A comparison of the actions of cobra cardiotoxin and scorpion toxin II on the guinea-pig taenia coli. Toxicon. 1986;24(2):131–9. pmid:3705092
- 52. Huang LF, Zheng JB, Xu Y, Song HT, Yu CX. A snake venom phospholipase A2 with high affinity for muscarinic acetylcholine receptors acts on guinea pig ileum. Toxicon. 2008;51(6):1008–16. pmid:18281071
- 53. Chen JC, Bullard MJ, Chiu TF, Ng CJ, Liaw SJ. Risk of immediate effects from F(ab)2 bivalent antivenin in Taiwan. Wilderness Environ Med. 2000;11(3):163–7. pmid:11055561
- 54. Mao YC, Liu PY, Chiang LC, Liao SC, Su HY, Hsieh SY, et al. Bungarus multicinctus multicinctus Snakebite in Taiwan. Am J Trop Med Hyg. 2017;96(6):1497–1504. pmid:28719273
- 55. Wu CH, Hu WH, Hung DZ, Peng YC, Yang DY. Snakebite complicated with Vibrio vulnificus infection. Vet Human Toxicol. 2001;43(5):283–5. pmid:11577933
- 56. Gonzalez M, Sanchez R, Carrillo EH. Septic shock syndrome resulting from snakebite. J Trauma. 2010;68(4):1015. pmid:20386292
- 57. Stamenkovic I, Lew PD. Early recognition of potentially fatal necrotizing fasciitis: the use of frozen-section biopsy. N Engl J Med. 1984;310(26):1689–93. pmid:6727947
- 58. Keller N, Andreoni F, Reiber C, Luethi-Schaller H, Schuepbach RA, Moch H, et al. Human streptococcal necrotizing fasciitis histopathology mirrored in a murine model. Am J Pathol. 2018;188(7):1517–23. pmid:29684366
- 59. Gunter OL Jr, Guillamondegui OD, May AK, Diaz JJ. Outcome of necrotizing skin and soft tissue infections. Surg Infect. 2008;9(4):443–50. pmid:18759681
- 60. Liu H, Zhu J, Hu Q, Rao X. Morganella morganii, a non-negligent opportunistic pathogen. Int J Infect Dis. 2016;50:10–17. pmid:27421818
- 61. Giuliano A, Lewis F Jr, Hadley K, Blaisdell FW. Bacteriology of necrotizing fasciitis. Am J Surg. 1977;134(1):52–57. pmid:327844
- 62. Ustin JS, Malangoni MA. Necrotizing soft-tissue infections. Crit Care Med. 2011;39(9):2156–62. pmid:21532474
- 63. Smith AJ, Daniels T, Bohnen JM. Soft tissue infections and the diabetic foot. Am J Surg. 1996;172(6):7s–12s. pmid:9003684
- 64. Tessier JM, Sanders J, Sartelli M, Ulrych J, De Simone B, Grabowski J, et al. Necrotizing soft tissue infections: a focused review of pathophysiology, diagnosis, operative management, antimicrobial therapy, and Pediatrics. Surg Infect. 2020;21(2):81–93.
- 65. Shek KC, Tsui KL, Lam KK, Crow P, Ng KH, Ades G, et al. Oral bacterial flora of the Chinese cobra (Naja atra) and bamboo pit viper (Trimeresurus albolabris) in Hong Kong SAR, China. Hong Kong Med J. 2009;15(3):183–90. pmid:19494373
- 66. Lam KK, Crow P, Ng KH, et al. A cross-sectional survey of snake oral bacterial flora from Hong Kong, SAR, China. Emerg Med J. 2011 Feb;28(2):107–14. pmid:20511640
- 67. Cutting KF, White R. Defined and refined: criteria for identifying wound infection revisited. Br J Community Nurs. 2004;9(3):S6–S15. pmid:15029002
- 68. Liau MY, Huang RJ. Toxoids and antivenoms of venomous snakes in Taiwan. Toxin Rev. 1997;16(3):163–75.
- 69. Liu CC, Chou YS, Chen CY, Liu KL, Huang GJ, Yu JS, et al. Pathogenesis of local necrosis induced by Naja atra venom: Assessment of the neutralization ability of Taiwanese freeze-dried neurotoxic antivenom in animal models. PLOS Negl Trop Dis. 2020;14(2):e0008054. pmid:32032357
- 70. Lee CH, Leu SJ, Lee YC, Liu CI, Lin LT, Mwale P, et al. Characterization of chicken-derived single chain antibody fragments against venom of Naja naja atra. Toxins. 2018;10(10):383. pmid:30248928
- 71. Esposito S, De Simone G, Gioia R, Noviello S, Pagliara D, Campitiello N, et al. Deep tissue biopsy vs. superficial swab culture, including microbial loading determination, in the microbiological assessment of Skin and Soft Tissue Infections (SSTIs). J Chemother. 2017;29(3):154–8. pmid:27376439
- 72. Stevens DL, Bryant AE. Necrotizing soft-tissue infections. N Engl J Med. 2017;377(23):2253–65. pmid:29211672