Effects of a negative pressure venom extraction device

Transcription

1 Wilderness and Environmental Medicine, 11, (2000) ORGNAL RESEARCH Effects of a negative pressure venom extraction device (Extractor) on local tissue injury after artificial rattlesnake envenomation in a porcine model SEAN P. BUSH, MD, FACEP; KEVN G. HEGEWALD, MD; STEVEN M. GREEN, MD, FACEP; MCHAEL D. CARDWELL; WLLAM K. HAYES, PhD From the Lorna Linda University School of Medicine, Lorna Linda. CA (Drs Bush, Hegewald, and Green), the San Bernardino County Sheriffs Department, San Bernardino, CA (Mr Cardwell), and Lorna Linda University, Lorna Linda, CA (Dr Hayes). Objectives.-To determine if a commercially available negative-pressure venom extraction device (Extractor) reduces local tissue injury after artificial rattlesnake envenomation in a porcine model. Methods.-We prospectively studied 10 pigs using a crossover design. After the pigs were anesthetized, 25 mg Crotalus atrox venom was injected obliquely with a 22-gauge needle 7 mm deep into subcutaneous tissues proximal to the ventral hind hoof. Pigs were randomized to receive either the Extractor (applied 3 minutes following envenomation and left in place for 30 minutes) or no Extractor. The protocol was repeated 14 days later by using the alternate treatment group and opposite hind leg for each animal. We measured leg circumference at standardized locations on the hoof, foreleg, and thigh at baseline and then 1, 2, 3, 4, 5, 6, 24, 48, 72, and 96 hours following venom injection. Maximal changes in circumference at 6 hours were compared using the paired t test. Minimum residual swelling at up to 96 hours was similarly compared. Results.-Maximal 6-hour swelling was similar with and without the Extractor: the hoof difference with the Extractor was -0.1% (95% C1 = -3.4% to 3.2%, P =.95), foreleg difference was 0.3% (95% C = -4.1% to 4.7%, P =.88), and thigh difference was -2.8% (95% C = -10.0% to 4.4%, P =.40). Minimum residual swelling at up to 96 hours was also similar with and without the Extractor: hoof difference with the Extractor was 1.2% (95% C = -5.6% to 8.0%, P =.70), foreleg difference was 0.6% (95% C = -3.7% to 4.9%, P =.76), and thigh difference was 0.3% (95% C = -2.4% to 3.0%, P =.81). A circular lesion identical in size and shape to the Extractor suction cup, which later necrosed and resulted in tissue loss, developed where the device had been applied in 2 animals. No such lesions occurred in legs not treated with the Extractor. Conclusion.-No benefit was demonstrated from Extractor use for artificial rattlesnake envenomation in our animal study. The skin necrosis noted in 2 Extractor-treated extremities suggests that an injury pattern may be associated with the device. Key words: Extractor, first aid, rattlesnake, snakebite, envenomation, animal model ntroduction A negative-pressure venom extraction device (Extractor; Sawyer Products, Safety Harbor, FL; Figure 1) has Abstract presented at the Society for Academic Emergency Medicine National Meeting in San Francisco, CA, May 23, Corresponding author: Sean P Bush, MD, FACEp, Department of Emergency Medicine, Lorna Linda University Medical Center, Room AOS, Anderson St, Lorna Linda, CA ( famensean@aol.com). been advocated by many experts, including the Wilderness Medical Society, for use as a temporizing or adjuvant first aid measure in the field immediately after a snakebite. 1-4 The device is applied directly over fang marks, requires no cutting, and delivers approximately 1 atmosphere of negative pressure. For optimal benefit, experts, in accordance with manufacturer's instructions, recommend that the device be applied within 3 minutes of snakebite and left in place for 30 to 60 minutes. ts use has been considered in the prehospital setting as a

2 Extractor 181 Figure 1. The Extractor. Photo by S.P.B. low-cost, easily portable, widely available, and rapidly applicable means of first aid prior to antivenin availability. t has also been suggested that the device may be useful in austere or remote environments, such as resource-challenged countries, where antivenin administration may be exceedingly delayed or not available at all. s For the past 15 years, most authorities have agreed that use of the Extractor for snakebite under the circumstances described could be beneficial and probably would cause no harm. - 4 However, other experts have suggested that the device may be dangerous and may exacerbate tissue damage, adding insult to injury after serious pit viper envenomation. 6-9 Study of the device has resulted only in 2 abstracts, which have not yet led to full-length publication, 10, and 2 reports of cases. 12,13 No study has ever shown that the Extractor influences outcome after any type of envenomation. We present the first randomized controlled trial to determine whether the Extractor reduces local tissue injury after rattlesnake envenomation. Methods All procedures involving animals were approved by the Animal Research Committee at our institution. The protocol outlined complies with the regulations and guidelines of the Animal Welfare Act, the NH guidelines for the care and use of laboratory animals, and American Association for Accreditation of Laboratory Animal Care. We chose a porcine model because pig skin and subcutaneous tissue are anatomically similar to their human counterparts l4-18 and because pigs have been used successfully in prior envenomation research The similarity of subcutaneous tissue is important because most human envenomations occur subcutaneously.14 Ten female domestic pigs (13-27 kg) were studied in the setting of an animal care facility. The pigs were housed individually and fed standard pig rations. They were kept on a 14-hour light: lo-hour dark cycle, and experiments were performed in the light phase. They were allowed to recover from transport at least 1 day before the experiment. Animals were anesthetized with

3 182 Bush et al 25% Cl :a 20%!.5 15% i u E.. 10% 5% o Extractor (n=10) fj, Control (n=10) 0% o Hours After njection Figure 2. Mean comparative increases in swelling between Extractor extremities and controls in the hoof. ketamine 20 mglkg 1M and xylazine 5 mg/kg 1M and given buprenorphine 10 j.lglkg 1M for analgesia before venom injection. soflurane was administered by mask at the time of injection. After induction, each pig was weighed, numbered, and marked for measurements. We performed the trial in a crossover design, with each animal serving as its own control. Half of the animals were randomized in the first phase to the experimental group, and then after a l4-day venom washout period, they were entered in the control group; the opposite hind leg was used. The other half were first in the control group, then in the experimental group. n the experimental phase, an Extractor was applied to the envenomation site 3 minutes after venom injection and left in place for 30 minutes. n the control phase, no Extractor was applied. A new Extractor was used for each application. The device was closely monitored and immediately reapplied if the vacuum was lost. All experiments utilized lyophilized Crotalus atrox venom from a single lot (Sigma Chemical, St Louis, MO), which was reconstituted by mixing 1000 mg in 10 ml normal saline for a final concentration of 100 mg! ml. To ensure uniformity, a single investigator injected 0.25 ml (25 mg) of venom in all animals obliquely with a 22-gauge needle 7 mm deep into subcutaneous tissues on the nonwalking surface proximal to the ventral hoof. A spacer was used on the needle to ensure uniform injection depth. The venom dose for our study was based on 2 prior reports,14,15 on preliminary findings from venom expenditure studies,19 and on doses used in previous research by one of us (W.K.H.).2o n a pilot trial, 3 pigs died after receiving 50 mg of venom. The dose was reduced to 25 mg for the experiment, with no fatalities. Needle size and depth were chosen based on a prior report,14 comparison of previously published measurements of fang length,21 and the authors' visual comparison of 27 actual rattlesnake fangs. We measured leg circumference at standardized locations around the hoof, foreleg, and thigh prior to venom injection and then 1, 2, 3, 4, 5, 6, 24, 48, 72, and 96 hours following injection. Standardized locations were the hoof (at the injection site at the base of the proximal phalanges), foreleg (2 cm proximal to the metatarsophalangeal joint), and thigh (3 cm proximal to the ankle joint). The distal hind leg was used in the experimental procedure because of its shape and accessibility. We recorded leg circumference, times, medication dosages, and vital signs on a standardized data form. All measurements were taken in the same manner by two of us (S.P.B. and K.G.H.), who were blinded to the intervention at the time of the measurements. Each extremity was photographed by one of us (M.D.C.) at several time intervals. The investigators' techniques and materials used were also photographed. Our main outcome measures were the maximum relative increase in swelling compared to baseline during the first 6 hours and the minimum residual swelling compared to baseline during the 24 to 96-hour period. The first measure was chosen to quantify the degree of immediate toxicity during the acute phase, with greater swelling signifying increased.toxicity. The latter measure was chosen to quantify the degree of resolution dur-

4 Extractor % Cl.= 20%!.= 15% ell i o Exlraclor (n=10) 10% i t:., Control (n=10) &! 5% 0% '-..,-----r--,----,---,---,-----,-----f f--t------rl----,--1---""1- o Hours After njection Figure 3. Mean comparative increases in swelling between Extractor extremities and controls in the foreleg. ing the nonacute phase, with lesser swelling signifying improved resolution. Clinically these were felt to correlate with the emergency care phase and hospitalization phase, respectively. We compared outcome measures between study and control legs for each pig with the paired t test, using Stata (version 6.0) statistical software (Stata Corporation, College Station, TX). Statistical significance was accepted at P < Results Mean swelling at serial time points is compared in Figures 2 through 4 for the hoof, foreleg, and thigh, respectively. Maximal swelling in the first 6 hours and minimum residual swelling at up to 96 hours were similar between Extractor and control groups (Table). Swelling occurred rapidly at the hoof, especially in the first 25% tll.; 20% 1) 0 Extractor (n=10).= 15% 6. Control (n=10) ell.. ca l!? u.5 ell 10%. Oi Gi a:: 5% 0% o / Hours After njection Figure 4. Mean comparative increases in swelling between Extractor extremities and controls in the thigh.

5 184 Bush et al o... <"!<"') NO'<:!" C'i '-' We report the first randomized controlled trial to determine whether the Extractor reduces local tissue injury after rattlesnake envenomation and found no benefit from this device. We observed no statistical reduction in swelling in the Extractor group and no trends towards efficacy in any measurement. The maximum potential 6 hour swelling reduction in this sample (ie, upper limit of 95% C) would be 3.4%, 4.1%, and 10.0% for the hoof, foreleg, and thigh, respectively. Conversely, it is equally possible that the Extractor could have worsened swelling at 6 hours by 3.2%, 4.7%, and 4.4% at the hoof, foreleg, and thigh, respectively. Similarly, the 95% Cis demonstrate that the largest beneficial effect that the Extractor could have had at 96 hours would have been as much as 5.6%, 3.7%, and 4.9% in terms of minimal residual swelling at the hoof, foreleg, and thigh, respectively. However, the Extractor could have just as easily worsened minimal residual swelling at 96 hours by 8.0%, 4.9%, and 3.0% at the hoof, foreleg, and thigh, respectively. We believe that even the largest potential differences (as quantified by our 95% Cis) are so small as to lack clinical importance, and we thus believe that our data provide strong evidence of the Extractor's inefficacy. We also noted 2 occurrences of skin necrosis with the Extractor and none in the control group. t appears poshour. Swelling was progressively less rapid at sites more distal to the bite. Although swelling did not always completely return to baseline within 96 hours, the change in percentage of swelling from 24 to 96 hours decreased over time. This was particularly obvious at the thigh measurement, where swelling frequently returned to baseline. However, it was less obvious at the hoof, where swelling did not return to baseline within 96 hours. A circular lesion identical in size and shape to the Extractor suction cup developed where the device had been applied in 2 animals. The lesions initially appeared ecchymotic and subsequently formed an eschar, which later necrosed, sloughed, and resulted in tissue loss (Figure 5). No such lesions occurred in legs not treated with the Extractor. Whereas swelling resolved in most animals within 4 to 7 days, these lesions required more than 3 weeks to heal. The volume of serosanguinous fluid retrieved from any Extractor application was no more than 0.1 ml. A wound infection developed in 1 animal initially randomized to the Extractor group, and this animal was excluded from the remainder of the experiment. Thus, the study group was composed of 10 pigs. There were no fatalities in the experimental trial N <'"iv) '<:!" <'"i 0. N "" Discussion

6 Extractor 185 Figure 5. A lesion developed in 2 pigs after application of an Extractor following artificial rattlesnake envenomation in our experimental model. Photo taken 13 days after Extractor application. Photo by M.D.C. sible that there may be a pattern of injury associated with the device, which could prolong, prevent, or complicate healing. Venom sequestration at the site of envenomation, suction-induced tissue ischemia, or both, are possible explanations. Our finding that the Extractor may cause or contribute to a specific pattern of injury without lessening envenomation severity is supported by previous literature. Use of the Extractor clinically in humans has been documented in a few case reports in nonmedical literature. 2,13 n one series, although no obvious benefit from the Extractor was seen, "a dark bleb formed over the fang puncture site conforming to the size of the suction cup used" in at least 2 out of 5 cases. 12 n this same series, severe envenomation and necrosis developed in some cases, despite use of the Extractor. One patient developed necrosis around the puncture site, which required 3 weeks to heal. Whether the Extractor increased the severity of the necrosis or prolonged healing was not established by this previous study. n our study, healing around the puncture site was similarly delayed in pigs with Extractor lesions, which took at least 3 weeks to heal. n another separate account, 2 Extractors were placed (lover each fang puncture) between and 2 minutes after a timber rattlesnake (C horridus horridus) bite to the base of the thumb. 3 Although no tissue damage resulted, uncertainty regarding the quantity of venom delivered makes it unclear what role the Extractor may have played in this positive outcome. Other experts have argued that suction and "aspiration devices" should be avoided and may be potentially harmful.8.9 First aid treatment (including suction) had no relationship to envenomation severity, according to one large retrospective series. 22 Two small studies suggest that some degree of venom retrieval may be achieved with the device, although the amount removed varied extremely. 10, n an uncontrolled series of 2 humans bitten by C atrox with an Extractor applied within minute, "an average of 27.5 j..lg of venom per ml of serosanguinous fluid was removed during the initial suction period, with this amount falling to 4.4 j..lg/ml by the end of 5 repeated periods." 11 Although the concentration of rattlesnake venom varies considerably,23 dividing the concentration of venom in the material removed in the initial Extractor application by the approximate concentration of C atrox venom reveals that the concentration of venom in the material removed was less than 1/10000 the concentration of the venom. Authors do not state what volume was extracted but indicate that the device was left in place until the cup filled with serosanguinous fluid 5 times. The largest cup in the Extractor kit holds about 5 ml. Even if a large amount of fluid was extracted (25 ml, for example), this equates to less than a single milligram of total venom retrieved. n our model, no more than 0.1 ml was removed from any Extractor application. This volume was consistent with the volume obtained by the Extractor in actual clinical cases in humans personally observed by the authors. n comparison with

7 186 typical venom expenditures after defensive rattlesnake bites,19 the amounts of venom removed by the Extractor are miniscule. Venom expenditure after a defensive rattlesnake bite often exceed 25 to 50 mg and may be greater than 200 mg. 19 Unfortunately, tissue t<ffects were not described in these 2 patients, and their outcome is not revealed in the abstract. However, removal of such a' trivial amount of venom would probably be clinically insignificant and is very unlikely to have any positive impact on outcome. n a small, unpublished series of 16 patients with pit viper envenomations, venom antigen was detected in fluid aspirated by the Extractor as late as 9 hours after the snakebite. 5 However, no quantitative determination of the percentage of venom removed could by made. Whether the amount removed led to a change in clinical outcome is not known. Only 1 other animal study has evaluated the efficacy of the Extractor in terms of venom removal after artificial envenomation, and this abstract has never led to a full-length publication. lo n their preliminary study, the authors claim that as much as 34% of 125odine radiolabeled C viridis helleri venom was removed by application of an Extractor for 30 minutes after simulated envenomations in a rabbit model, as measured in a gamma counter. t is not clear how this percentage was obtained, because the paper was published only in abstract form and methodology is not described. This study did not address clinical efficacy at all. Tissue effects and outcome are not described. Because there was no control in this study, it is impossible to know whether the material retrieved was even toxic. t is possible that the may have adhered to an inactive venom component that was aspirated, leaving only toxic portions. There is no description of statistical significance. The extreme variation in the amount of venom removed between studies suggests these findings are not conclusive. n addition, there is no description of how the envenomations were "simulated." Therefore it is not clear that the depth or gauge of needle used realistically approximated envenomation by a rattlesnake fang. The model used and the measurements taken are difficult to extrapolate to a clinical scenario. Rabbit skin differs from that of humans in its thick coat of fur, thin epidermis and dermis, and sparse subcutaneous layer with little elastic tissue. 24 Therefore, the external validity of rabbit studies is limited because of dissimilarities in skin structure. The porcine model was used in our study because pig skin and subcutaneous tissue are anatomically more similar to those of humans The findings from these sparsely documented studies remain unconvincing. Our study is limited by the inherent precautions in extrapolating results from animal studies to humans. An- Bush et at ecdotal experience with the device is consistent with our finding that a pattern of injury may develop. However, without a control group, it will be difficult to establish whether the Extractor reduces tissue effects in humans. Prospective human studies are unlikely to be approved by institutional review boards in light of the possibility of injury associated with the device. Mortality was not addressed in this study, as the animal model is not appropriate. We chose to deliver sublethal doses of venom to our subjects. Some will argue that an increase in local injury is a reasonable trade-off for decreased mortality. However, mortality is extremely rare in the United States after pit viper envenomation, and local injury is very common. Of nearly reptile bites reported to American Association of Poison Control Centers between 1983 and 1998, only 10 deaths were associated with rattlesnake bites. 25-4o n some of these 10 cases, death was most likely due to complicating factors or unusual circumstances. For example, 2 of these fatalities were likely due to anaphylaxis to antivenin. One fatal case involved a bite to the face. One case was suspicious for an intravascular envenomation. The Extractor is unlikely to influence fatal outcomes after these types of envenomations. Although the manufacturers of the Extractor recommend seeking medical help immediately, some critics of the device caution that using the device may give false reassurance, leading to delays in seeking medical care. 7 Delays in seeking medical care are one of th leading contributing factors to fatal rattlesnake envenomation, according to one report. 41 Mortality after snakebite in the United States is difficult to study. Because envenomation is a fairly uncommon occurrence and death is exceedingly rare, studies would have to be extremely large to demonstrate a statistically significant improvement in mortality due to any particular intervention. Such a study would probably not even be possible. n contrast to the extreme rarity of fatal envenomation, some degree of local injury occurs after almost every serious pit viper envenomation. Swelling was used as the outcome measure in our study because it is by far the most common effect seen ater any serious pit viper envenomation. Long-term morbidity is almost exclusively limited to local injury.42 Permanent local sequelae occur in as many as 10% of all bites. Venom sequestration at the site of envenomation may worsen local injury, leading to increased tissue loss. Although we feel the site of envenomation in our model mot closely represented an actual rattlesnake bite to a human extremity, a different site could have dispersed or absorbed venom differently, affecting its availability to suction. Results may have varied with higher

8 Extractor or lower doses of venom, which may be evaluated in future studies. Our study is also limited by constraints of the data collected. Minimal swelling, as a measure of recovery, may have continued to improve over the next few days. Perhaps a future study would address recovery in terms of time to complete resolution of swelling. Similarly, although swelling is clearly increasing at all 3 measurements from 0 to 6 hours, peak swelling may not occur until after 6 hours. This is true clinically, as reports of delayed progression of envenomation severity are known to occur. 43,44 A future study may follow hourly measurements for a longer period, such as 12 or 24 hours. Extrapolation of our results to snakebite in parts of the world other than the United States should be done with caution. Extractor efficacy after bites by other types of snakes (such as elapids) was not addressed by our study. Future studies might evaluate the device after bites by snakes possessing nonnecrotizing venom, such as coral snakes (Micrurus sp) or spiders, such as widow spiders (Latrodectus sp). n our controlled study evaluating the Extractor, there was no difference in swelling, which is the most common outcome following pit viper envenomation, after artificial envenomations in a porcine model. However, our study and others suggest that there may be a specific injury pattern associated with the device. Although most authorities have previously agreed that the Extractor could be beneficial and would probably cause no harm, our study suggests the opposite may be true. Acknowledgments This study was supported by the Wilderness Medical Society Research Training Award We thank David Cadogan, Korbin Haycock, James Kim, Barbara Anderson, A'me Saverino, and Abigail Hegewald, without whom we could not have completed this study. References 1. Forgey WW, ed. Wilderness Medical Society Practice Guidelines for Wilderness Emergency Care. Merrillville, N: CS Books; Forgey WW. More on snake-venom and insect-venom extractors [letter]. N Engl J Med. 1993;328: Gold BS. Snake venom extractors: a valuable first aid tool [letter]. Vet Hum Toxicol. 1993;35: Forgey W, Norris RL, Blackman, Walter FG, Hultgren HN. Viewpoints: response. J Wild Med. 1994;5: Callahan MV, Hamer D, Pitts RM. Epidemiology and management of viperidae and elapidae snake envenomation. Proceedings of the Annual Scientific Meeting ofthe 187 Wilderness Medical Society. August 2-8, 1997; Sun Valley, Gellert GA. Snake-venom and insect-venom extractors: an unproved therapy [letter]. N Engl J Med. 1992;327: Gellert GA. More on snake-venom and insect-venom extractors [letter reply]. N Engl J Med. 1993;328: Winkel KD, Hawdon GM, Levick N. Pressure immobilization for neurotoxic snake bites. Ann Emerg Med. 1999; 34: Warrell DA. Snake bite and snake venoms. QJ Med. 1993; 86: Bronstein AC, Russell FE, Sullivan B, et al. Negative pressure suction in the field treatment of rattlesnake bite [abstract]. Vet Hum Toxicol. 1985;28: Bronstein AC, Russell FE, Sullivan JB. Negative pressure suction in the field treatment of rattlesnake bite victims [abstract]. Vet Hum Toxicol. 1986;28: Hardy DL Sr. A review of first aid measures for pitviper bite in North America with an appraisal of Extractor suction and stun gun electroshock. n: Campbell la, Brodie ED Jr, eds. Biology of the Pitvipers. Tyler, TX: Selva; 1992: Levell JP. The Vivarium mailbag. Vivarium. 1994;5: 31, Burgess JL, Dart RC, Egen NB, Mayesohn M. Effects of constriction bands on rattlesnake venom absorption: a pharmacokinetic study. Ann Emerg Med. 1992;21: Bania Te, Bernstein SL, Baron B, et al. ntraarterial vs intravenous administration of antivenin for the treatment of Crotalidae atrox envenomation: a pilot study. Acad Emerg Med. 1998;5: Hobbs GD, Anderson AR, Greene T, et al. Comparison of hyperbaric oxygen and dapsone therapy for Loxosceles envenomation. Acad Emerg Med. 1996;3: Hobbs GD, Yealy DM. Development of an acute cutaneous swine model for Loxosceles envenomation. Vet Hum Toxicol. 1994;36: Forbes PD. Vascular supply of the skin and hair in swine. n: Montagna W, Dobson RL, eds. Advances in Biology of the Skin. 9th ed. New York, NY: Pergamon; 1969: Herbert SS. Factors nfluencing Venom Expenditure During Defensive Bites by Cottonmouths (Agkistrodon piscivorus) and Rattlesnakes (Crotalus viridis, Crotalus atrox) [Master's thesis]. Loma Linda, CA: Loma Linda University, Hayes WK. Factors associated with the mass of venom expended by prairie rattlesnakes (Crotalus v. viridis) feeding on mice. Toxicon. 1992;30: Klauber LM. Rattlesnakes: Their Habits, Life Histories, and nfluence on Mankind. 2nd ed. Berkeley, CA: University of California Press; Wingert WA, Chan L. Rattlesnake bites in southern California rationale for recommended treatment. West J Med. 1988;148: Glenn L, Straight RC. The rattlesnakes and their venom yield and lethal toxicity. n: Th AT, ed. Rattlesnake Ven-

9 188 oms: Their Actions and Treatments. New York, NY: Marcel Dekker nc; 1982: Hobbs GD. Brown recluse spider envenomation: is hyperbaric oxygen the answer? Acad Emerg Med. 1997;4: Litovitz TL, Klein-Schwartz W, Cararati EM, et al Annual report of the American Association of Poison Control Centers Toxic Exposure Surveillance System. Am J Emerg Med. 1999;17: Litovitz TL, Klein-Schwartz W, Dyer KS, et al Annual report of the American Association of Poison Control Centers Toxic Exposure Surveillance System. Am J Emerg Med. 1998;16: Litovitz TL, Smi1kstein M, Feldberg L, et al Annual Toxic Exposure Surveillance System. Am J Emerg Med. 1997;15: Litovitz TL, Feldberg L, Soloway RA, et a Annual Toxic Exposure Surveillance System. Am J Emerg Med. 1996;14: Litovitz TL, et a Annual report of the American Association of Poison Control Centers Toxic Exposure Surveillance System. Am J Emerg Med. 1995;13: Litovitz TL, Clark LR, Soloway RA, et a Annual Toxic Exposure Surveillance System. Am J Emerg Med. 1994;12: Litovitz TL, Holm KC, Clancy C, et al Annual report of the American Association of Poison Control Centers Toxic Exposure Surveillance System. Am J Emerg Med. 1993;11: Litovitz TL, Holm KC, Bailey KM, et al Annual National Data Collection System. Am J Emerg Med. 1992;10: Litovitz TL, Bailey KM; Schmitz BF, et a Annual report of the American Association of Poison Control Cen- Bush et a1 ters National Data Collection System. Am J Emerg Med. 1991;9: Litovitz TL, Schmitz B, Bailey KM Annual report of the American Association of Poison Control Centers National Data Collection System. Am J Emerg Med. 1988; 8: Litovitz TL, Schmitz B, Holm Ke Annual report of the American Association of Poison Control Centers National Data Collection System. Am J Emerg Med. 1989;7: Litovitz TL, Schmitz BF, Matyunas N, et a Annual National Data Collection System. Am J Emerg Med. 1988;6: Litovitz TL, Martin TG, Schmitz B Annual report of the American Association of Poison Control Centers National Data Collection System. Am J Emerg Med. 1987; 5: Litovitz TL, Nonnann SA, Veltri Je Annual report of the American Association of Poison Control Centers National Data Collection System. Am J Emerg Med. 1986; 4: Litovitz TL, Veltri JC Annual report of the American Association of Poison Control Centers National Data Collection System. Am J Emerg Med. 1985;3: Veltri JC, Litovitz TL Annual report of the American Association of Poison Control Centers National Data Collection System. Am J Emerg Med. 1984;2: Hardy DL. Fatal rattlesnake envenomation in Arizona: Clin. Toxicol. 1986;24: Dart C, McNally JT, Spaite DW, Gustafson R. The sequelae of pitviper poisoning in the United States. n: Campbell JA, Brodie ED Jr, eds. Biology ofthe Pitvipers. Tyler, TX: Selva; 1992: Guisto JA. Severe toxicity from crotalid envenomation after early resolution of symptoms. Ann Emerg Med. 1995; 26: Hurlbut KM, Dart RC, Spaite D, et al. Reliability of clinical presentation for predicting significant pit viper envenomation. Ann Emerg Med. 1988;17: