Introduction

Top Abstract Introduction Methods Results Discussion References

Snake bites represent a major medical problem, especially in subtropical countries where they still cause high rates of morbidity and mortality. The specific treatment of envenoming is the administration of antivenoms, i.e., the injection of empirical amounts of antibodies, a part of which (about 20%) is directed against venom components. Although immunotherapy has proved its effectiveness in reducing mortality and morbidity of snake bites, it is also responsible for acute or delayed allergic reactions, the incidence of which depends mainly on the amount of heterologous antibodies injected and on the purity of the antivenom. To improve the safety and efficacy of antivenoms, two approaches can be followed: 1) improving the antivenom preparation and processing and 2) optimizing the use of antivenoms.

Acute side effects are mainly caused by immune sensitization to horse antibodies, which are widely used in human therapeutics. Some investigators have advocated antivenom produced in other animals, such as sheep (Sjostrom et al., 1994; Consroe et al., 1995) or hens (Carroll et al., 1992), although antibodies produced in this way were not shown to be less immunogenic than horse antibodies. Another approach to decreasing the side effects is to lower the amount of heterologous proteins injected by selecting those antibodies that specifically react with the venom, by use of immunopurification (Smith et al., 1992). The major drawback of this method is the high production cost. Other investigators have described the isolation of a fraction of immunoglobulins, called IgG_T, from the serum of hyperimmunized horses (Fernandes et al., 1991; Pépin et al., 1995). This fraction showed high protective ability but may be very immunogenic because of a high level of glycosylation of the Fc fragment (Sjostrom et al., 1994).

In parallel with the improvement of antivenom preparation, it would be useful to optimize their use during immunotherapy. At present, there is no definite protocol to guide clinicians in the choice of an appropriate treatment of envenomings. To assess the necessity of immunotherapy and to modify the dose of antivenom, the severity of the envenoming must be known as quickly as possible. Moreover, some parameters of antivenom administration, such as the delay after snake bites, the route of injection and the type of antibodies to be used (immunoglobulins or antigen-binding fragments) should be based on a clear understanding of the process of venom neutralization in the body. Such an understanding would lead to a precise definition of the antivenom quality and of the parameters of antivenom therapy applicable to different types of envenomings.

In France, the problem of envenoming is less acute than in subtropical countries, because only about 2,000 viper envenomings occur every year, with a mortality rate of about one to two deaths per year (Chippaux and Goyffon, 1989). An ELISA was set up to determine venom concentrations in biological samples of patients bitten by vipers to establish a correlation between venom plasma levels and clinical symptoms (Audebert et al., 1992). This assay was used to follow the disposition of Vipera aspis venom after intravenous and intramuscular injections to rabbits (Audebert et al., 1994b). After intravenous injection, V. aspis venom was shown to have a short distribution half-life (0.7 h) and a volume of distribution 20-fold larger than that of the central compartment. After intramuscular injection, the venom followed a complex resorption process, high concentrations being maintained in the blood for more than 3 days.

In the present work, we tested the effect of several doses of antivenom and of different routes of injection on the pharmacokinetics of venom. We also investigated the effect of early and delayed administration of antivenom, as well as the relative efficacy of different fragments of immunoglobulins.

	Methods

Top Abstract Introduction Methods Results Discussion References

Detection of Free Venom in Plasma

Purification of reagents. Immunoglobulin fragments used for the ELISA determination were purified from IPSER Europe antivenom (Pasteur Mérieux Sérums et Vaccins). IPSER Europe antivenom consists of equine F(ab)₂ fragments specifically directed against the venoms of V. aspis, Vipera berus and Vipera ammodytes. F(ab)₂ fragments were purified as described by Audebert et al. (1994a) and were labeled with peroxidase by glutaraldehyde coupling by a two-step procedure (Avraméas and Terninck, 1969, 1971).

ELISA. This test was based on the method described by Audebert et al. (1994b). Microplates (Nunc, Roskilde, Danemark) were coated with 100 µl per well of a solution of purified F(ab)₂ diluted in 0.1 M sodium bicarbonate buffer, pH 9.5. After 2 h at 37°C, the plates were washed six times with a solution of phosphate buffer saline, pH 7.4, containing 1 Tween 20 (PBS-Tween). The plates were blocked with PBS containing 3% of bovine serum albumin (PBS-BSA) during 1 h at 37°C. After washing, 100 µl samples were added in duplicate wells and the plates were incubated for 1 h at 37°C. The plates were washed six times with PBS-Tween and 100 µl of venom-specific antibodies conjugated with peroxidase diluted in PBS-BSA containing 1 Tween 20 were added to each well. The plates were incubated for 1 h at 37°C. After washing with PBS-Tween, 100 µl of substrate medium [0.01 M phosphate sodium buffer, pH 7.3, containing 2 mg·ml¹ of o-phenylenediamine dichloroamide (Sigma Chemical Co., St. Louis, MO) and 0.06% of H₂O₂] were distributed into each well. The colored reaction was stopped by adding to each well 50 µl a solution of 2 N H₂SO₄ containing 0.5% of sodium sulfate. Optical density was measured at 490 nm with a microplate reader MR 5000 (Dynatech Laboratories, Denkendorf, Germany). Venom antigen plasma concentrations (ng·ml¹) were determined by comparison with a standard curve established with crude V. aspis venom (Latoxan, Rosans, France) diluted in normal rabbit serum.

Iodination of Venom

Iodination was performed as described by Audebert et al., (1994b) with the iodogen method of Fraker and Speck (1978). Eppendorf tubes (1.5 ml) were coated with 0.3 mg of iodogen (Pierce, Rockford, IL) diluted in 750 µl of chloroform. Three milligrams of venom diluted in 300 µl of PBS were added to the radioactive iodine. After 6 min of incubation, free iodine was separated from radioactive proteins by Sephadex G-15 gel filtration. As described by Audebert et al. (1994b), venom was separated in two fractions: HMWP, which were radiolabeled, and LMWP, which were not. Venom was reconstituted by pooling HMWP and LMWP. Venom used in the pharmacokinetic experiments had a specific activity of about 100 µCi·mg¹.

Detection of Total Venom

Total venom (free or immunoconjugated venom) was detected by determination of the TCA-precipitable ¹²⁵I-radiolabeled proteins contained in rabbit plasma. TCA precipitations were performed as follows: 50 µl of plasma were loaded onto a GF/A filter (Whatman, Maidstone, England) and dried under a lamp. Two milliliters of 10% TCA were incubated with filters for 10 min to precipitate native proteins on the filter. Two milliliters of absolute ethanol were used to dry the filters, and the radioactivity of the filters was counted using a multigamma counter 1261 (Pharmacia, Uppsala, Sweden).

Determination of Plasma Concentrations of F(ab)₂ and Fab

Plasma concentrations of fragments of immunoglobulins [F(ab)₂ or Fab] were measured by ELISA. Unlabeled and peroxidase-labeled rabbit antihorse immunoglobulin antibodies were obtained from Biosys (Compiègne, France). Tests were performed as follows: microplates were coated with a solution of 2.5 µg·ml¹ of unlabeled antihorse immunoglobulin antibodies diluted in a 0.1 M sodium bicarbonate buffer, pH 9.5 (100 µl per well). After 2 h at 37°C, wells were saturated with 100 µl per well of PBS containing 0.5% of gelatin. Samples and standards diluted in PBS-Tween containing 0.5% of gelatin were added in duplicates after 1 h at 37°C. After incubation at 37°C during 1 h, 100 µl per well of peroxidase-labeled antihorse immunoglobulin antibodies diluted in PBS-Tween-gelatin were added to each well. A color reaction was performed as described above.

Lethality Assay

LD₅₀ values of cold and labeled venom were determined according to the WHO's method (1981). Male mice weighing 18 to 20 g (Charles River, St Aubin-lès-Elbeuf, France) were injected intravenously in the tail vein. Five doses of venom were injected (5, 7, 10, 14 and 20 µg per mouse). Three mice were used per dose.

Preparation of Fab Fragments

Fab was produced by cleaving the commercial F(ab)₂ preparation with papain, according to the method of Parham (1986) with minor modifications. F(ab)₂ preparation was dialyzed against 1 M Tris-HCl buffer, pH 7, containing 2 mM ethylenediaminetetraacetic acid. One milliliter of 1 M Tris-HCl buffer, pH 7.5, 1 ml of 20 mM ethylenediaminetetraacetic acid, 1 ml of 0.1 M cysteine, 1 ml of papain (10 mg, Boehringer, Mannheim, Germany) and 6 ml of water were added to the dialyzed preparation. This mixture was incubated for 24 h at 37°C. Then, 1 ml of papain was added and the mixture was again incubated for 24 h at 37°C. The reaction was stopped by addition of iodoacetamide at 5 mM (final concentration). Undigested and digested fragments were separated by gel filtration with a Sephadex G-100 column equilibrated with 5 mM sodium phosphate buffer, pH 8, containing 0.15 M NaCl.

Pharmacokinetic Experiments

New Zealand rabbits weighing 2.75 to 3 kg (CEGAV, St Mars-d'Egrenne, France) were placed in metabolism cages, which allowed the collection of urines and feces. Food and water were provided ad libitum. Experimental envenomings were performed by an intramuscular injection in the leg of 700 µg·kg¹ of radioactive venom in a final volume of 500 µl. Blood was serially collected in heparinized tubes from the central artery of the ears. Plasma was obtained by centrifugation at 1,500 × g for 15 min. Antivenom antibodies diluted in 5 ml of 0.15 M NaCl were infused into the marginal vein of the ear for an 8-min period by a syringe pump (Bioblock, Illkirch, France). Intramuscular immunotherapy was performed by injecting immunoglobulin fragments in a final volume of 500 µl into the contralateral leg used for the administration of the venom.

Pharmacokinetic parameters for the fragments of immunoglobulins were determined by the MKMODEL software (BioSoft, Cambridge, England). CL_T was determined as equal to D/AUC, Vd_ss as equal to (D·AUMC)/AUC² and MRT as equal to AUMC/AUC, where D was the dose injected, AUC was the area under the curve from injection to infinity and AUMC was the area under the mean curve; t_1/2 and t_1/2 represented the distribution and elimination half-lives.

The efficacy of immunotherapy was quantified by the following method: areas under the total and free venom curves (AUC_t and AUC_f, respectively) during 72 h after immunotherapy were calculated using the trapezoidal rule, the difference (AUC_d = AUC_t  AUC_f) was then used to allow comparisons.

Statistics

All results are given as mean ± standard error of the mean. Their significance was analyzed by the two-tailed unpaired or paired Student's t test; the level of significance was set at P < .05.

	Results

Top Abstract Introduction Methods Results Discussion References

Lethality assays. LD₅₀ values were determined with radioactive and native venoms; no difference in potency was observed between the preparations (LD₅₀ of 15 and 13.3 µg of venom per 18 to 20 g mouse, respectively), which showed that iodination of the venom did not modify its toxicity.

Pharmacokinetics of V. aspis venom after intramuscular injection. Six rabbits were injected intramuscularly with 700 µg·kg¹ of radiolabeled venom, and the time courses of venom concentrations in plasma were followed by use of ELISA and radioactivity (fig. 1A). The use of radiolabeled venom allowed the quantification of total venom components in plasma, whereas ELISA detected free venom that was not complexed by the antivenom antibodies. Levels of radioactivity indicated a slightly, but significantly different time course of the venom from that measured by ELISA: the AUCs were 7,200 and 4,500 ng·h·ml¹·kg¹ for radioactivity and ELISA, respectively (P < .05).

View larger version (22K): [in this window] [in a new window]   Fig. 1.   Effect of different doses of antivenom on the plasma pharmacokinetics of the venom of V. aspis. Rabbits were envenomed with 700 µg·kg1 of radiolabeled venom. Seven hours after envenomations, they received either none, 125, 25 or 5 mg (A, B, C and D respectively) of IPSER Europe antivenom (arrows). Concentrations of total venom () were measured with radioactivity, and free venom concentrations () were measured by ELISA. Values are the means ± S.E.M. of three independent experiments except for A (n = 6).

Influence of different doses of antivenom on the plasma pharmacokinetics of V. aspis venom. Four doses (125, 25, 5 and 2.5 mg of antivenom F(ab)₂ per rabbit) were tested in the case of an experimental envenoming with 700 µg·kg¹ of radiolabeled venom, a dose of venom which led to plasma concentrations corresponding to a severe envenoming in humans (Audebert et al., 1994a). Immunotherapy was performed 7 h after envenoming via the intravenous route.

1

1

1

1

View larger version (20K): [in this window] [in a new window]   Fig. 2.   Influence of different doses of antivenom injected 7 h after envenomation on the AUCd () and the AUCf between envenomation and 72 h later (). Values are the means ± S.E.M. of three independent experiments. AUCf and AUCd were calculated as described under "Material and Methods."

Route of administration of antivenom. The efficiency of an intramuscular injection of 25 mg of antivenom to rabbits was compared with the intravenous route. Intramuscular injection produced a delayed and partial neutralization of venom (fig. 3); moreover, the AUC_d and the maximal concentration of total venom were very low in comparison with the intravenous route (P < .05, table 1).

View larger version (10K): [in this window] [in a new window]   Fig. 3.   Effect of immunotherapy via the intramuscular route. Seven hours after experimental envenomation 25 mg of antivenom was injected intramuscularly in rabbits (arrow). The time courses of free () and total () venom concentrations were followed during 80 h. Values are the means ± S.E.M. of three independent experiments.

Effect of the delay between envenoming and immunotherapy. In a first experiment, 25 mg of antivenom were injected in rabbits 3 h after experimental envenoming. Immediately after infusion, a complete immunocomplexation was observed which lasted for 2 h (fig. 4A). Low concentrations of free venom (5-10 ng·ml¹) could be detected after that period. The AUC_d and the maximal concentration were not significantly different from those observed with a delay of 7 h.

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Effect of the delay between experimental envenomation and intravenous injection of antivenom.Twenty-five milligrams of antivenom were injected 24 h (A) or 3 h (B) after experimental envenomation (arrows). Free venom concentrations () and total venom concentrations () were observed until 100 h after envenomation. Values are the means ± S.E.M. of three independent experiments.

Effect of Fab fragments on the pharmacokinetics of the venom. Fab fragments were prepared from the antivenom by papain digestion, and 25 mg of Fab was injected via the intravenous route 7 h after envenoming. The Fab did not neutralize the venom antigens completely, and the redistribution induced by immunoglobulin fragments was very low, as shown by the measurement of AUC_d and maximal concentration of total venom (fig. 5 and table 1). Furthermore, 17 h after immunotherapy, the blood of treated rabbits contained toxic concentrations of uncomplexed venom (about 75 ng·ml¹).

View larger version (11K): [in this window] [in a new window]   Fig. 5.   Effect of an immunotherapy performed with Fab fragments. Twenty-five milligrams of Fab were injected intravenously (arrow) in rabbits 7 h after experimental envenomation with radiolabeled venom. Free () and total () venom concentrations were observed during 80 h. Values are the means ± S.E.M. of three independent experiments.

Pharmacokinetics of F(ab)₂ and Fab. We studied the pharmacokinetics of F(ab)₂ and Fab fragments to understand their different effects on venom neutralization and redistribution. F(ab)₂ or Fab (25 mg) were infused according to the same protocol used after experimental envenoming (fig. 6). Distribution and elimination half-lives of Fab were similar to those published in the literature (Ohning et al., 1994).

View larger version (12K): [in this window] [in a new window]   Fig. 6.   Pharmacokinetics of F(ab)2 and Fab fragments. Rabbits were injected intravenously with 25 mg of immunoglobulin fragments F(ab)2 or Fab (A and B, respectively). Curves have been drawn with pharmacokinetic parameters obtained after model-dependent analysis. Experimental values are the means ± S.E.M. of three independent experiments.

1

1

View this table: [in this window] [in a new window]   TABLE 2 Pharmacokinetic parameters of F(ab)2 and Fab determined after intravenous injection of 25 mg of fragments to 3 rabbits

1

1

1

	Discussion

Top Abstract Introduction Methods Results Discussion References

Although serotherapy was discovered a century ago, the mechanism by which antivenom antibodies neutralize venom proteins is still poorly understood. We therefore examined the effect of a specific antivenom on venom plasma pharmacokinetics after experimental envenoming by V. aspis venom in rabbits. The use of a ¹²⁵I-labeled venom allowed the quantification of total venom components in plasma, whereas only free venom antigens were determined by ELISA.

We first compared pharmacokinetic parameters obtained after intramuscular injection of V. aspis venom by the two methods of measuring plasma venom concentrations. Slightly different values of AUC and clearance were obtained for total and free venom. Nevertheless, the terminal half-lives (t_1/2) and the mean residence times were shown to be the same with the two methods (around 30 and 50 h for t_1/2 and mean residence time, respectively). A similar observation was described in the pharmacokinetic analysis of activated human protein C (Ishii et al., 1995) in which three methods were used to follow plasma levels of the protein: radioactivity, ELISA and amidolytic activity. In that case, the differences in kinetic parameters obtained with the three methods were explained by the formation of large complexes of protein C with plasma proteins, still detectable by radioactivity but less antigenic and biologically inactive. In V. aspis venom, three hypotheses can be considered: a binding of venom proteins to plasma proteins which can mask antigenic determinants, a degradation of some venom proteins still detectable by radioactivity but less antigenic or alternatively a differential detection of some proteins by radioactivity and ELISA. The two first hypotheses were tested by gel electrophoresis of radioactive venom samples under nondenaturing conditions. Radioactivity was shown to be associated with unchanged venom proteins and we did not observe any binding of venom proteins to plasma proteins (data not shown). Considering the third hypothesis, Audebert et al. (1994b) have shown that all the proteins of the venom are not radiolabeled with the same specific radioactivity and that they do not respond to ELISA with the same intensity. Therefore, the different values of clearance and AUC may be explained because venom proteins were not quantified exactly in the same way by ELISA and radioactivity. It is important, however, to recognize that these differences have no consequence on the further interpretation of the results obtained in this study, because radioactivity and ELISA data are considered independently to measure the extent of redistribution or immunoneutralization under different conditions.

After intravenous injection of a neutralizing dose of antivenom (125 mg of F(ab)₂, i.e., half a vial of IPSER Europe antivenom per rabbit) 7 h after venom injection, we observed a redistribution of venom antigens from the extravascular space into the central compartment as well as a complete immunocomplexation of venom antigens in the blood. The effect of antivenom therapy on venom levels in plasma has already been investigated by several authors in experimental models (Lwin et al., 1984; Labrousse et al., 1988; Theakston, 1989) who reported that the plasma levels of venom antigens determined by ELISA decreased after intravenous injection of the antivenom. However, these studies incompletely described the mechanism of action of antivenom, because ELISA does not quantify the antigens bound in vivo to the antibodies.

To minimize the risks caused by the injection of large quantities of heterologous proteins, we tested the effect of lower doses of antivenom to determine the minimal dose necessary to treat severe envenomings. Three doses (25, 5 and 2.5 mg) were injected intravenously to rabbits, 7 h after the injection of venom by the intramuscular route. The dose of 25 mg of antivenom was shown to induce a redistribution of venom from the extravascular to the vascular space, but this effect was not statistically different than that obtained with 125 mg. Furthermore, all the antigenic sites of venom proteins were complexed by antivenom during 2 h, and subsequent concentrations of free venom did not reach the toxic level defined by Audebert et al. (1994a), because concentrations did not exceed 20 ng·ml¹. The lower doses examined (5 and 2.5 mg of F(ab)₂ were inefficient, because toxic-free venom protein levels were detected by ELISA and only small quantities of venom were extracted from the extravascular space.

The amount of antivenom administered by clinicians is mainly determined by the clinical features, the size of the snake as well as the known efficacy of available antivenom (Chippaux and Goyffon, 1983). Clinical trials have been undertaken to reduce the dose of antivenom administered to envenomed patients (Jorge et al., 1995; Theakston et al., 1992). ELISA quantification of levels of venom antigens in the blood could allow clinicians to accurately adapt the dose of antivenom to the gravity of the envenoming and to follow the efficacy of antivenom treatment and thus to quickly evaluate the necessity of further antivenom injections.

The route of administration of antivenom is still controversial. The intravenous route has been considered by many authors to be the most efficient one, although it has not been systematically compared with the intramuscular route. Furthermore, in France, health regulations only allow intramuscular administration of antivenom. We, therefore, compared the effects of intravenous and intramuscular injections of 25 mg of antivenom on the pharmacokinetics of the venom. We observed a lower redistribution of the venom proteins in the intramuscular route. In addition, the immunoneutralization of antigenic sites was delayed and incomplete. This phenomenon can be explained in the light of the results published by Pépin et al. (1995), who determined the pharmacokinetic parameters of antivenom after an intramuscular injection to rabbits. The maximum blood concentration (C_max) of F(ab)₂ was reached only 48 h after the intramuscular injection of antivenom. Furthermore, the calculated bioavailability of IPSER Europe antivenom was about 42%. These two observations explain that the maximum immunocomplexation of venom proteins was delayed and that the redistribution was less efficient than with intravenous injection of antivenom, because less than half of the antibodies were able to reach the central compartment. The intravenous route is, therefore, the most efficient route of injection of antivenom, allowing a quick immunocomplexation, as well as an intense redistribution of the venom antigens. For these reasons, the intravenous route should be recommended. However, in the rare cases where antivenom cannot be administered by the intravenous route, the intramuscular route, although much less efficient, may still be used.

In the initial experiments, the administration of antivenom was performed after the absorption phase of venom proteins to determine the extent of redistribution induced by the injection of antibodies. However, the delays in arrival of envenomed patients at the hospital are highly variable, ranging from 1 h after envenoming to several days. We therefore tested the effect of an early and a delayed intravenous administration of antivenom on the pharmacokinetics of V. aspis venom. We observed a maximal redistribution when antivenom was injected 3 h after the envenoming. Furthermore, a complete immunocomplexation of venom antigens lasted for 2 h, and only very low concentrations of free venom were detected at later times. In conclusion, in V. aspis envenoming, when a dose of antivenom appropriate for the severity of envenoming was injected early, no reinjection of antivenom to rabbits was necessary. When antivenom was injected later (24 h) after envenoming, it still efficiently modified the pharmacokinetics of the venom. However the induced redistribution was lower and immunocomplexation lasted longer than in our standard conditions (25 mg of F(ab)₂ injected 7 h after experimental envenoming). This may be explained by the fact that some venom has already been cleared from the blood circulation after 24 h. Obviously, it is difficult to predict the clinical effect of an immunotherapy performed a long time after envenoming. Actually, the effect of a delayed administration of antivenom is still controversial in the literature. Karlson-Stiber and Persson (1994) have reported that immunotherapy was not efficacious for patients envenomed by V. berus when injected more than 18 h after envenoming, although one case of severe coagulation disturbances was shown to normalize when the antivenom was injected 30 h after the bite. On the other hand, two studies have reported successful treatment by immunotherapy 1 week after viper envenoming (Dwivedi Shubha Sheshadri and D'Souza, 1989; Tiwari and Johnston, 1986). In conclusion, antivenom is more efficient when injected early after the bite, but may still be effective even after a long delay.

Fab fragments have already been used for clinical therapy to treat intoxications by digitalis (Smith et al., 1976) and colchicine (Baud et al., 1995). Fab preparations have also been produced against desipramine (Brunn et al., 1992), phencyclidine (Valentine et al., 1994) and other compounds. Fab fragments are preferred over complete antibodies (IgG) or F(ab)₂ fragments because of the low incidence of adverse reactions which occur when they are infused. Only 0.8% of treated patients with specific antidigoxin Fab were shown to develop allergic reactions (Hickey et al., 1991), whereas 6 to 7% of envenomed patients showed adverse effects when treated with specific antivenom F(ab)₂ (Smith et al., 1992). These differences may be explained by the fact that Fab are not associated with type III hypersensitivity, do not bind complement or macrophages and are less immunogenic than intact IgG (Smith et al., 1979). Furthermore, they have a greater volume of distribution and a lower elimination half-life than IgG and F(ab)₂. However, the detoxification process is slightly different in digitalis or colchicine intoxication where the toxic agent penetrates into cells, from the case of envenoming by V. aspis, where, to our knowledge, no intracellular toxicity has been shown for the components of this venom. We therefore digested IPSER Europe F(ab)₂ with papain and injected 25 mg of Fab 7 h after experimental envenoming. In this case, we observed a very low redistribution of venom antigens, in comparison with the redistribution induced by F(ab)₂ in the same conditions. This phenomenon has already been described by Butler et al. (1977) in anti-digoxin Fab. Moreover, the concentration of free venom reached a toxic level 17 h after the injection of Fab fragments. In agreement with our observations, Ohning et al. (1994) reported that the effect of anti-gastrin Fab decreased 24 h after their injection. These results indicated that Fab have to be injected as an infusion rather than an intravenous injection.

The differences observed between the effects of F(ab)₂ and Fab on the pharmacokinetics of the venom were clearly linked to the pharmacokinetics of F(ab)₂ and Fab. Although F(ab)₂ remained mainly in the vascular compartment, Fab was largely distributed in the extravascular compartment. This is probably why the latter was less efficient in inducing a redistribution of venom antigens into the circulation. In the same way, although Fab and F(ab)₂ were shown to have similar affinity constants for antigens (approximately 10¹⁰ M, data not shown), Fab has a much shorter half-life than F(ab)₂ in the plasma and, therefore, loses its inhibitory effect rapidly, as evidenced by the reappearance of toxic venom concentrations in the plasma of rabbits.

Fab fragments therefore appeared less efficient in treating envenoming than in treating digitalis or colchicine intoxications. This is consistent with the dissociation kinetics of the different toxins from receptors. In digitalis intoxication, digoxin was shown to dissociate rapidly from its target with a dissociation half-life of 1 h, so that a rapid infusion of digoxin-specific Fab is able to reverse the intoxication. In the case of colchicine, however, the dissociation half-life from tubulin is 20 h and requires a longer infusion of Fab fragments. Similarly, in snake venom, Fab have to be frequently readministered or continuously infused to maintain sufficient levels of antibodies for a period of time at least equal to the half-life of the venom, which is 30 h by the intramuscular route. The elimination route of venom-antibody complexes has not yet been identified clearly. Reticuloendothelial tissues might be involved in the catabolism of circulating F(ab)₂-venom complexes. Fab-venom complexes are larger than the complexes formed with small molecules such as colchicine or phencyclidine: their molecular weight is higher than the glomerular filtration threshold so that they cannot be cleared via the renal route. Further investigations are in progress to clarify this important point.

In conclusion, this study of the in vivo neutralization of venom components by antivenom antibodies provides an experimental basis for the optimization of antivenom treatment in humans. It has to be completed by pharmacodynamic experiments to examine the effect of antivenom therapy on the symptomatology of the envenoming. Interestingly, the present study performed in the case of a European viper could be extended to envenomings by other poisonous animals and their immunotherapeutic treatment by antivenoms.