Skip to main content
Skip to main content

del Castillo JRE. 1998;5:189-202 Pharmacokinetic modeling of in-

   

Original research

September and October, 1998

Pharmacokinetic modeling of in-feed tetracyclines in pigs using a meta-analytic compartmental approach

Jerome R.E. del Castillo, DMV, IPSAV, MSc; Johanne Elsener, DMV; Guy P. Martineau, DMV, DESS

JREdC, GPM: Université de Montréal, Faculté de Médecine Vétérinaire. CP 5000 St-Hyacinthe Quebec, J2S-7C6 Canada; JE: Ayerst Laboratories Inc.

del Castillo JRE, Elsener J, Martineau GP. Pharmacokinetic modeling of in-feed tetracyclines in pigs using a meta-analytic compartmental approach. Swine Health Prod. 1998;6(5):189-202.

Also available in PDF format (1.6MB)

See also "Putting it into practice…" sidebar and Glossary.

Summary

Objective: To assess effect of pharmacokinetic differences between chlortetracycline (CTC) and oxytetracycline (OTC), the effect of day/night variations in ad libitum intake of medicated feeds on variation in plasma and tissue concentrations of both drugs, and the influence on the overall efficacy of metaphylactic strategies for bacterial respiratory disease in growing pigs.

Methods: A multidosage pharmacokinetic model was created for feed-administered CTC and OTC in pigs. Parameters obtained by meta-analysis of published pharmacokinetic articles and a computerized iteration procedure were used to find minimum CTC or OTC in-feed dosage to maintain a selected plasma drug concentration in pigs incorporating feeding behavior variables.

Results: The model accounted for daily feed consumption, ad libitum feeding behavior, time-kill activity of tetracyclines, and the span of their postantibiotic effect. The model successfully passed an external validation procedure. The bioavailability (P=.0001), apparent volume of distribution (P=.001), and elimination constant rate (P=.03) differed between CTC and OTC. When identical in-feed dosages are consumed, plasma concentrations of CTC were twice the concentrations of OTC. Ad libitum feeding behavior, which becomes increasingly diurnal as pigs grow from weaning to finishing, induced major daily variations in plasma concentrations of both drugs. We devised two equations that easily calculate any specific result generated by the computerized iteration procedure for CTC and OTC.

Implications: The use of either equation will allow swine practitioners to create more precise CTC or OTC metaphylactic strategies from weaning to finishing. Matching daily antibiotic dosages to what is needed can improve the efficacy of metaphylaxis.

Keywords: pharmacokinetics, antibiotics, feeding behavior, meta-analysis, chlortetracycline, oxytetracycline, in-feed

Received: August 18, 1997
Accepted: June 16, 1998

Controlling respiratory diseases is a major health concern for intensively raised weaner, grower, and finisher pigs. Often, the only realistic short-term option to prevent outbreaks of respiratory disease is to control infection pressure in colonized, subclinically infected animals (i.e., metaphylaxis–see glossary beginning page 200). Typically, metaphylaxis is accomplished through the large-scale use of antibiotics, which are delivered in feed and/or drinking water to minimize animal stress and labor and drug costs. However, while cost is an important factor in choosing a medication and administration route, oral bioavailability must also be considered. Tetracyclines, the most frequently used drugs for treating respiratory disease, inhibit bacterial protein synthesis (Figure 1).1 The number of bacterial cells that die when exposed to greater than the minimum inhibitory concentrations (MIC) of tetracyclines is related to the duration of the exposure, not to drug concentration (e.g., time-dependent kinetics of bactericidal activity) (Figure 2).2 With non-intestinal diseases, the antibiotic must be absorbed in the intestinal lumen before it can reach the targeted tissue. Thus, bioavailability is a factor in drug waste. The plasma concentration of a drug is influenced by three factors:

If any of these three factors is not addressed, then drug concentrations in plasma and tissue, when delivered in feed or water, are unlikely to meet the MIC for the targeted pathogens.3

A number of feed factors and characteristic pig behaviors affect the intake of medicated feeds. For example, when drugs are administered in moist or liquid feeds,4 top dressing, or drinking water, the drug residence in the stomach, bioavailability, and/or absorption rate of drugs can be different than when they are administered in dry feed.5 Supplemental citric acid in feeds medicated with tetracyclines is known to interfere with unavailable calcium-drug complexes that form in the intestines. The bioavailability of orally delivered tetracyclines, especially chlortetracycline (CTC), is thus enhanced by citric acid.6

Feed intake7 and ad libitum feeding behavior are also both influenced by pig age.8 Feed consumption in pigs is related to metabolic weight9 (W0.75) rather than crude weight (W); thus, ad libitum ingestion of a given medicated feed will result in higher drug dosages and higher plasma concentrations in 10-kg (22-lb) pigs than in 40-kg (88-lb) pigs. Moreover, Labroue, et al.,8 have observed that as pigs get older, they consume an increasing percentage of their overall feed intake during the day (Table 1). This suggests that feeding behavior might be less affected by day/night cycles in 10-kg (22-lb) piglets than in 40-kg (88-lb) pigs. For a given in-feed tetracycline dosage, the daily variations in plasma concentrations of the drug are positively associated with the age of the pigs.

Feeding behavior of pigs and the amount of feed ingested per meal are also affected by the size of groups placed within a pen,10 even when pen density is equal. Pigs housed in groups of <= 15 pigs per pen eat approximately twice as often as those penned in groups of >= 20.10 To maintain the same daily feed consumption, then, pigs housed in groups of >=20 must eat twice as much at each meal. These feed characteristics and feeding behaviors may have an effect on the drug concentrations in plasma and tissues and on the degree of protection against disease. No pharmacokinetic study has taken into account actual pig feeding behavior.

Pharmacokinetics is what the body does to the drug and pharmacodynamics is what the drug does to the body (both of the animal and of the bacterium).11 It is important to consider both pharmacokinetics and pharmacodynamics when designing metaphylactic strategies. Although both CTC and oxytetracycline (OTC) have been the subject of pharmacokinetic studies in pigs, the pigs were given a single oral treatment,1,4,6,12,13 or the drug plasma concentrations were modeled over time in a population of pigs fed according to a fixed-interval schedule.4 Little is known, therefore, about the pharmacokinetics of tetracyclines when medicated feeds are consumed ad libitum. It is risky to generalize from any one of these studies to a population of pigs that have ad libitum access to feed; however, by using meta-analysis we can synthesize the findings of many studies to obtain an average that can allow us to generalize more reliably to a wide variety of herd situations.

Meta-analysis is a process whereby a collection of research results from individual studies are statistically analyzed in a way that allows one to integrate their findings.14 Meta-analysis is currently used in epidemiological studies to obtain stronger, more reliable information that can be applied to a wider range of situations than can the findings of individual studies. It produces a set of parameters that represents average, not individual, findings across studies. This type of analysis can determine associations between disease and risk factors, risk estimates, or other parameter estimates when the original studies do not have adequate statistical power to report such findings.

The objectives of the present investigation were to:

Materials and methods

Meta-analysis and statistical analysis

Meta-analysis was performed with all available studies treating the pharmacokinetics of in-feed tetracyclines in pigs to obtain reliable parameters that could be further used in a multiple-dosage pharmacokinetic model, using the Mantel-Haenszel method.14 Briefly, Mantel-Haenszel meta-analysis consists of a series of steps:

Compartmental pharmacokinetic analysis15 was performed on all but one12 of the articles included in the meta-analysis; the statistical moment kinetic approach16 was used in the later study. A one-compartment kinetic analysis15 was thus performed with the individual sets of plasma CTC concentrations reported in this latter article.12 Experiments6,13 using feeds supplemented with 1.4% of calcium were included in the meta-analysis, because somewhat lower concentrations of calcium are recommended in Europe for weaned pigs,9 but their reported bioavailabilities were weighted at a 50% level.

To be included in our meta-analysis, a study had to meet several criteria:

Studies excluded from meta-analysis were reports of experiments performed on disease-challenged animals,1 those where moist or liquid medicated-feeds were used,4 those with citric acid-supplemented feed,6,13 and those that did not investigate the individual evolution of plasma concentrations of the drug over time.17,18

Five studies, two reporting on OTC1,13 and three reporting on CTC,4,6,12 met all the criteria and were included in the meta-analysis. To derive mean parameter values to use in calculations that describe the fluctuations in drug concentrations over time, two separate Mantel-Haenszel meta-analyses were then performed on the sets of studies (Tables 2–3). These pharmacokinetic parameters were compared statistically using the Mann-Whitney U tests19 at a .05 significance level.

The multiple-dosage pharmacokinetic model

The fluctuations of drug concentrations in plasma over time for one single dosage can be calculated with the following nonlinear equation:20

(Equation 1)

Where:

To calculate V, the equation is:

(Equation 2)

Where AUC is the area under the curve of plasma drug concentration over time measured with the linear trapezoidal rule.20

To conduct the meta-analysis, we used a crude average estimate of V(area) instead of a Mantel-Haenszel14 weighed average, because a majority of authors did not report the variance associated with their estimate (Tables 2–3).

Systemic clearance (CLb) and elimination half-life (t1/2) of CTC and OTC were calculated with the following equations:

(Equation 3)

(Equation 4)

We derived the estimates of CLb and t1/2 with parameters derived from the meta-analysis.

The multiple-dosage pharmacokinetic model was devised from equation 1 using the method of superposition;21 e.g., repeatedly adding Equation 1 (Figure 3) based on the proportion of daily meals taken during daytime (Table 1),8 and the amount of feed consumed ad libitum per meal.

Feeding behavior

Based on observations of the influence of the size of the group of pigs within a pen (not pig density) on feed consumption and feeding behavior,10 we modeled a scenario with <= 15 pigs per pen, and another scenario with >= 20 pigs per pen.

Computer iteration procedure

A computerized iteration procedure was devised to find the minimum in-feed dosage of either CTC or OTC (in ppm) that would maintain plasma MIC over a 20-hour interval (to allow for a 4-hour postantibiotic effect [PAE]). We focused on the minimum concentration of antibiotic in plasma to take into account the time-dependent antimicrobial effect of tetracyclines. A procedure was programmed in Turbo Basic (Corel Inc., Ottawa, Ontario). A Basic compiler is needed to run the program.

Variables we included in this program were the following:

Begin and end values for the in-feed dosage and the iteration step are also variables in this program.

Values of the parameters derived from meta-analysis were then input into Equation 1, which was superposed at even intervals according to the number of daytime meals during the 0- to 12-hour (daytime) period. Thus, if the pigs were assumed to eat 4 meals during the day, the equation was superposed 4 times at 4-hour intervals (e.g., 07:00, 11:00, 15:00, 19:00) (Figure 3). During the 12- to 24-hour (nighttime) period, Equation 1 was superposed at even intervals according to the number of nighttime meals. Daytime and nighttime superpositions were repeated for a continuous 120-hour period (Figure 4). Dose was determined by dividing respective feed intakes (i.e., daytime or nighttime) by the number of meals assumed to have been taken during daytime and nighttime.

A pharmacokinetic profile was then generated for 200 hours (to ensure that the data would include the period of equilibrium) with 0.2-hour intervals. The minimum in-feed dosage (in ppm) required to achieve TPL was found by backward iteration according to the following criterion: the resulting plasma concentration profile had to fall below TPL for exactly 4 hours (the span of post-antibiotic effect with tetracyclines) somewhere between 36 and 60 hours after the medicated feed is first offered, the time at which equilibrium of drug between plasma and tissue has been achieved. Finally, the program generated an ASCII file with the final pharmacokinetic profile, giving pairs of time and plasma drug concentration values in the 0- to 200-hour period in 0.2-hour intervals.

This iteration procedure was run twice, once for the OTC profile and once for the CTC profile.

Model validation

To assess the accuracy of the meta-analytic model, we used an external validation procedure to compare the antibiotic plasma concentrations the models predicted to actual plasma concentrations reported by Black and Gentry,22 Asanuma, et al.,17 Mevius, et al.,23 Reichert,24 Andrews, et al.,18 Hall, et al.,25 and Hunneman, et al.26 In some papers,18,25 the mean +/- SD of plasma concentrations was reported graphically; thus, we used approximate estimations of the mean and confidence intervals in our validity calculations. "Agreement" between the models and the literature was defined as >= 50% overlap between their respective 95% confidence intervals (CI).

The predicted confidence intervals of CTC and OTC plasma concentrations were obtained by using F and k values equal to the upper and lower 95% confidence limits in the models instead of average Mantel-Haenszel14 estimates. The upper limit represents a model with greater bioavailability (i.e., F at the upper limit) and slower elimination (i.e., k at the lower limit), whereas the lower limit represents a model with lesser bioavailability (i.e., F at the lower limit) and faster elimination (i.e., k at the upper limit) (Figure 5).

Results

Comparing CTC and OTC pharmacokinetics

The absolute bioavailability of CTC (13%) was significantly greater than that of OTC (4%) (P =.0001) (Table 4). Peak plasma concentration (Cmax) of OTC was half of the Cmax of CTC (Figure 6). The apparent distribution volumes (V(area)) of CTC and OTC were also significantly lower for OTC than for CTC (P =.0014) (Table 4). The speed of intestinal uptake of the drugs, measured by ka, was not significantly different (P =.9). On the other hand, CTC was eliminated more quickly than OTC, as shown by the statistically significant difference in their respective k (P =.033). The elimination half-life (t1/2) of CTC is thus 75% of OTC. Because CTC is absorbed more slowly and eliminated more quickly than OTC, the models predicted slightly greater daily variations in plasma concentrations of CTC compared to OTC (Figure 4).

Feeding behavior

Diurnal feeding patterns

Ad libitum feeding behavior of pigs induced a marked daily variation in both CTC and OTC plasma concentrations. In-feed dosages of both CTC and OTC differ between more-diurnal pigs and less-diurnal pigs when a targeted plasma drug concentration is required (Table 5). Because 40-kg (88-lb) pigs consume about 70% of their daily meals during the daytime,8 the model should predict that the drug accumulated in the body during the day and was eliminated during the night. The model did predict that drug plasma concentrations were lowest by the beginning of the diurnal period (approximately 7:00 a.m.), and highest at its end (7:00 p.m.) (Figure 4).

Pig numbers in pen

The number of pigs in the pen (holding pen density constant) did not prove to be a significant factor in determining plasma drug concentration over time (data not shown).

Model validation

In general, our model generated predictions that were consistent with the literature (Figures 7–8). The model for OTC was in slightly better agreement with reported plasma concentrations (Figure 7) than was the model for CTC (Figure 8).

Summarizing equation

The simulations were used to devise an equation where the drug dosage to be given to pigs is calculated by the antibiotic concentration in plasma, the daily feed intake of pigs, as well as their feeding behavior, in the nonlinear equation:

(Equation 5)

(Equation 6)

Where:

Discussion

Model validation

The plasma concentrations of OTC and CTC predicted by our models were in close agreement with those reported by Asanuma, et al.,17 Mevius, et al.,23 and Hunneman, et al.,26 strongly indicating that our model is valid. Moreover, the models made surprisingly accurate predictions for CTC and OTC plasma concentrations given by the 7-day use of medicated drinking water.24 However, the latter findings are not adequate to support the application of the models to administration of medicated water.

As expected, these models were not able to predict plasma concentrations of OTC in fasted piglets dosed with a medicated drench.22,23 Plasma OTC concentrations found with medicated drench administration to fasted22,23 or fed22 pigs were considerably higher than both concentrations predicted with the equations and those from their experiments using medicated feeds.25,26 Intestinal absorption of OTC is reduced by the presence of food.5

Our model for CTC was also in close agreement to observations reported by Andrews, et al.18 in pigs that were challenged with Pasteurella multocida and then received CTC-medicated feed. This agreement is somewhat surprising, because the disease challenge is known to modify drug disposition in animals.1

The effect of OTC and CTC dosage on consumption of medicated feed by pigs has not been considered in the models. We do not know whether higher in-feed dosages of CTC or OTC would significantly change the palatability of medicated feeds offered to pigs, but Pijpers, et al.,27 reported that pigs offered feed with 2400 ppm of OTC showed similar appetites to those of pigs fed with a much lower dosage of OTC-medicated feeds. Also, the rate of feed intake of 10-kg pigs offered a diet fortified with 3600 ppm of CTC was identical to that of a drug-free diet.28

The diurnal pattern of drug plasma concentrations that our model predicted is consistent with the observation of fluctuations in tetracycline plasma concentrations in pigs offered medicated drinking water for ad libitum consumption.29

Labroue, et al.,8 and Nielsen, et al.10 used an electronic monitoring device placed on each individual pig that allowed them to monitor feeding frequency and individual feed intake in a nonintrusive way. In their studies, then, normal feeding behaviors were not altered by human interaction. Unfortunately, in the studies used for model validation,17,18,22-26 feed intake of pigs was monitored by a stocksperson who entered the facility. This intrusive method of monitoring pig intake and plasma concentration is a potential confounder of pig feeding behavior. Factors related to feed characteristics (i.e., citric acid supplementation, calcium content, etc.) and pig feeding behaviors (i.e., diurnal feeding patterns, intrusive management) might be responsible for the differences between CTC and OTC plasma concentrations that were predicted by the models and those reported in some papers24,25 used for the validation procedure.

Pharmacokinetics

The pharmacokinetics of both CTC and OTC are linear. The time-evolution concentration parameters are thus assumed to remain constant throughout a wide range of in-feed dosages. The superposition method we used is valid and the drug plasma concentrations it predicts are directly related to the administered dosage, which is consistent with reports in literature.12,27 The better agreement between our model and the OTC literature than for the CTC literature may be due to the greater variation in bioavailability of CTC. The Mantel-Haenszel estimate of F for CTC may appear conservative, however, because two independent studies reported mean values near 17%. Large variance is associated with these F values.

Plasma concentrations are not necessarily equivalent to concentrations of drug in target tissue. Diffusion of antimicrobials to extra-vascular fluids is restricted to free molecules, whereas molecules transiently bound to proteins remain in plasma30 (Figure 9, 77k JPEG). Plasma-protein binding rate is negatively related to drug distribution to lung and other tissues in a nonlinear way; only drugs possessing plasma protein binding rates over 85%-90% have significantly reduced free drug-plasma concentrations.30 On the other hand, the rate of protein binding of tetracyclines is positively correlated to lipid solubility,30 which enhances tissue penetration. Consequently, the clinical significance of this balance between protein binding and lipid solubility is difficult to assess. As 20%-25% of less lipid-soluble OTC and 50%-70% of more lipid-soluble CTC are bound to plasma proteins,24 similar free drug concentrations should be found with identical plasma concentrations,31 and free CTC tissue penetration might be greater than that of OTC. Lung CTC and OTC concentrations reported by Asanuma, et al.,17 were higher than plasma concentrations and the mean of tissue:plasma concentration ratios of CTC is about 30% greater than that of OTC. Nonetheless, these findings are affected by last-meal-to-slaughter intervals, which were not specified. Concentrations of CTC in lung reported by Andrews, et al.,18 were also greater than plasma concentrations, but data represented in this article was from P. multocida-challenged pigs, and distribution kinetics could also be affected by this situation. The clinical significance of the comparison of CTC and OTC plasma and lung pharmacokinetic profiles at a given time is difficult to assess, because both functions are nonlinear and have different parameter values. In this case, however, it is less important than bioavailability; the dosage to be administered to pigs to obtain a given concentration of OTC in plasma is twice that of CTC (Figure 4). The greater lipid solubility of CTC is also likely to enhance diffusion to the site of infection.

Pharmacodynamics and antibiotic
sensitivity testing

Tetracyclines achieve full antimicrobial activity when supraMIC concentrations are reached at the site of potential infection: the prevalence of signs and lesions in pigs experimentally infected with pathogenic Actinobacillus pleuropneumoniae decreased as the steady-state OTC plasma concentrations approached the MIC.26 Andrews, et al.,18 performed a similar study with P. multocida, but the assessment of a prophylactic effect of subMIC concentrations of CTC was not included.

It is important, then, that the MICs of CTC and OTC serve as the basis for the design of strategies that ensure optimal prophylactic efficacy. Although this information is reported for several swine respiratory pathogens, only semi-quantitative information (sensitive, intermediate, or resistant) is typically given to veterinary practitioners. For human bacteria other than Neisseria spp., and Haemophilus spp.,32 the sensitivity breakpoint of tetracycline is equivalent to an MIC of 8.0 mg per mL. Hence, strains with MIC <= 4 µg per mL are sensitive, and strains with MIC >= 32 µg per mL are resistant. Many veterinary diagnostic laboratories use this breakpoint for veterinary pathogens. In the case of A. pleuropneumoniae, some veterinary laboratories33 use the breakpoint for human Haemophilus spp.,32 which is equivalent to 4 µg per mL. Strains with MIC <= 2 µg per mL are susceptible, while strains with MIC >= 8 µg per mL are resistant to tetracycline. Drug concentrations equivalent to these two breakpoints are easily exceeded with parenterally administered OTC, but rarely achieved with medicated feeds. The use of this breakpoint in diagnostic laboratories, consequently, creates misleading information for swine practitioners.34

Many papers report the MICs of CTC and OTC for Actinobacillus pleuropneumoniae,35-39 Bordetella bronchiseptica,35,36 Haemophilus parasuis,35,40 Mycoplasma hyopneumoniae35,37,41-43 and P. multocida.35-37,39 This list is restricted to articles that report the sample distribution of MICs. Differences noted among these reports may be associated with geographic variations, but may also be due to the MIC determination technique chosen.42

The 90th percentile of the MIC distribution (MIC90) is often considered as a target to reach. The use of such a parameter by itself is unfair, since it is often derived from bacterial isolates whose collection is based on taxonomy, with little consideration of their pathological and epidemiological importance as a cause of respiratory disease. Nonvirulent strains, or virulent ones that are seldom implicated with disease in the field, could be added to the test sample. This could lead to the inclusion of several subgroups of strains with different concentrations of both antimicrobial susceptibility and clinical interest. The MIC90 can thus be biased by strains possessing extreme MIC values but that are epidemiologically irrelevant, while much lower concentrations could inhibit a majority of more prevalent pathogenic strains.44

Precise evaluation of MIC sample distribution is necessary to discriminate between more- or less-sensitive subgroups of bacterial isolates and identify outlying values. The geometric dilution procedures that are used to determine MICs produce an increasing error to the right side of sample distributions, where the MIC90 is always located. Therefore, the bias created by outliers is exacerbated by the dramatic imprecision associated with its value.34,44,45

Another source of bias is when practitioners submit samples for bacteriological analysis to diagnostic laboratories only in cases of metaphylactic failure. The proportion of unusually resistant strains in the bacterial collection eventually increases, and inference of the MIC90 to the population of field strains will no longer be valid. The MIC target should be representative of field strains commonly implicated in disease outbreaks. Veterinary practitioners should routinely send animals or tissue specimens to diagnostic laboratories to maintain the relevance of the MIC target. Additionally, developing techniques to predict in vivo MIC would be of major clinical interest.

It is critically important to consider the type of activity provided by the antimicrobial to obtain effective metaphylaxis. So far, tetracyclines have been observed to exert a time-dependent bactericidal activity.2 In addition, subMIC concentrations of drug could reduce the growth rate of exposed bacteria, but without any decrease in bacterial population being induced. Nevertheless, this effect could help the immune system in controlling infectious pressure.

The post-antibiotic effect (PAE) of tetracyclines has only been studied on human enterobacteria, but the extrapolation to other Gram-negative rods, such as swine respiratory pathogens, seems to be acceptable.46 Escherichia coli cultures exposed for 1 hour to supraMIC concentrations of tetracycline did not show any noticeable in vitro growth for over 3 hours after drug had been washed off.47 This PAE has been reported with other protein synthesis inhibitors for E. coli and other Gram-negative bacilli as well.48 Theoretically, if in-feed medication regimens can deliver supraMIC concentrations of CTC and OTC to the site of infection for some period of time, their effect would be maintained for at least 3 hours even after an instantaneous and complete elimination.47 This instant elimination does not exist in animals; CTC and OTC elimination half-lives are > 4 hours in pigs (Table 4). SubMIC concentrations of drug are thus found for several hours before complete elimination, which may increase the length of PAE effect.48 However, the length of in vivo PAE effect caused by CTC and OTC is not yet known. Consideration of PAE is thought to be of major importance for the design of more effective treatment strategies. For example, schedules that give transient subMIC plasma concentrations of drug could be tolerated for antimicrobials possessing a PAE.49 The computerized iterative procedure we used in our study took into consideration a 4-hour-long postantibiotic phase, which allowed short-term subMIC plasma concentrations of CTC and OTC.

Post-antibiotic leukocyte enhancement effect (PALE) has not been taken into account in our model.46 After exposure to antibiotics, enhanced phagocytosis and intracellular killing of bacteria is noted during the antibiotic-free phase. Modifications in the bacterial cell surface caused by the drug may cause this phenomenon.48 The in vivo extent of PALE, which has been noted in vitro, is unknown but likely plays a major role in the outcome of metaphylaxis with antimicrobials. The addition of subMIC concentrations of some antibiotics has also been reported to induce a PALE effect.49 The PALE effect of the tetracyclines has not been studied to our knowledge, but is likely to exist as suggested by the sometimes successful results of metaphylaxis with low drug in-feed dosage concentrations.

Our observation that the number of pigs in the pen didn’t affect plasma concentrations of OTC and CTC could be predicted by Nielsen’s10 observation that pigs housed in larger groups (with pen density kept constant) ate less frequently, but ate more at each meal. Thus, the frequency of feeding would have an effect on dosage and dosage interval, but not an overall effect on drug concentrations in plasma. In our simulations, we kept pen density constant and altered only the number of pigs in the pen. Our result therefore cannot be safely generalized to situations in which pig density has been increased beyond standard commercial guidelines.

Implications

Acknowledgments

The authors are grateful to Dr. Denis Du Tremblay for his assistance in the computer programming of the multi-dosage pharmacokinetic simulations, to Dr. Michel Bigras-Poulin for his advice in meta-analysis, and to Dr. Jean-Guy Besner and Dr. Robert Higgins for the critical reviewing of the manuscript.

References

1. Pijpers A, Schoevers EJ, van Gogh H, van Leengoed LAMG, Visser IJR, van Miert ASJPAM, Verheijden JHM. The influence of disease on feed and water consumption and on pharmacokinetics of orally administered oxytetracycline in pigs. J Anim Sci. 1991;69:2947-2954.

2. Keck G, Borne PM. Nouvelles conceptions en antibiothérapie et leurs applications pratiques en médecine vétérinaire. Revue Med Vet. 1995;146:309-320.

3. Wages D. Volumetric versus mg/kg dosing in poultry. Proc 44th N Cent Avian Dis Conf Symp Poultry Ther. 1993; Columbus Ohio: 22-24.

4. Sutter HM, Wanner M. Futterzubereitung und pharmacokinetik von chlortetracyclin beim ferkel. Schweitz Arch Tierheilk., 1990;132:175-181.

5. Gibaldi M. Gastrointestinal absorption–Biologic considerations. In: Gibaldi M, ed. Biopharmaceutics and Clinical Pharmacokinetics. 4th ed. Malvern, Pennsylvania: Lea & Febiger. 1991; 24-39.

6. Wanner M, Walker W, Sutter HM, Riond JL, Broz J. Influence of dietary citric acid and calcium on the bioavailability of orally administered chlortetracyclin in piglets. J Vet Med. 1991;38:755-762.

7. National Research Council. Nutrient Requirements of Swine. 9th rev ed. Washington DC: National Academy Press. 1988;50.

8. Labroue F, Guéblez R, Marion M, Sellier P. Influence de la race sur le comportement alimentaire de porcs en croissance élevés en groupe. Journ Recherche Porcine en France. 1995;27:175-182.

9. Henry Y, Perez JM, Seve B. Alimentation des porcs en croissance. In: Blum JC, ed. L’alimentation des Animaux Monogastriques. Paris, France: I.N.R.A. 1984;49-66.

10. Nielsen BL, Lawrence AB. The effect of group size on the behaviour and performance of growing pigs using computerised single-space feeders. Pig News and Info. 1993; 14:127N-129N.

11. Holford NHG, Sheiner LB. Kinetics of pharmacologic response. Pharmacol Ther. 1982;16:143-166.

12. Kilroy CR, Hall WF, Bane DP, Bevill RF, Koritz GD. Chlortetracycline in swine – bioavailability and pharmacokinetics in fasted and fed pigs. J Vet Pharmacol Ther. 1990;13:49-58.

13. Wanner M, Nietlispach G, Sutter HM. Einfluß von citronensaüre une calcium auf die bioverfügbarkeit oral verabreichten oxytetracyclins beim ferkel. Dtsch Tierärztl Wschr. 1990;97:515-518.

14. Dickersin K, Berlin JA. Meta-analysis: State-of-the-science. Epidemiol Rev. 1992; 14:154-176.

15. Gibaldi M. Introduction to pharmacokinetics. In. Giabaldi M, ed. Biopharmaceutics and Clinical Pharmacokinetics. 4th ed. Philadelphia, Pennsylvania: Lea and Febiger. 1991.

16. Yamaoka K, Nakagawa T, Uno T. Statistical moments in pharmacokinetics. J Pharmacokinet Biopharm.1978;6:547-58.

17. Asanuma K, Shimazaki S, Hirata K. A study concerning the distribution of OTC and CTC in the lung and blood of pigs. J Anim Drugs – Getsukan Doyaku. 1984;6:1-14.

18. Andrews JJ, Lucas TE, Johnson DD. Prevention and control of experimentally induced Pasteurella multocida pneumonia in swine by the use of chlortetracycline administered in feed. Agri-Practice. 1988;9:33-38.

19. Daniel WW. Procedures that utilize data from two independent samples. In: Daniel WW, ed. Applied Nonparametric Statistic. 2nd ed. Boston, Massachusetts: PWS-Kent Publishing Co. 1990. 82-143.

20. Rowland M, Tozer TN. Clinical Pharmacokinetics: Concepts and Applications. Third ed; Media, Pennsylvania: Williams and Wilkins. 1994.

21. Gibaldi M. Method of Superposition. in Gibaldi M, ed. Biopharmaceutics and Clinical Pharmacokinetics. 4th ed. Malvern, Pennsylvania: Lea and Febiger. 1991; p. 379.

22. Black SD, Gentry RD. The distribution of oxytetracycline in the tissues of swine following a single oral dose. Can Vet J. 1984;25:158-161.

23. Mevius DJ, Vellenga L, Breukink HJ, Nouws JFM, Vree TB, Driessens F. Pharmacokinetics and renal clearance of oxytetracycline in piglets following intravenous and oral administration. Vet Quart. 1986; 8:274-284.

24. Reichert J. Blutspiegel von chlor- und oxytetracyclin beim ferkel nach futter- oder wasser-medication bei unterschiedlicher fütterungstechnik (MSc diss). Zurich, Switzerland:University of Zurich. 1988; 123 p.

25. Hall WF, Kniffen TS, Bane DP, Bevill RF, Koritz GD. Plasma concentrations of oxytetracycline in swine after administration of the drug intramuscularly and orally in feed. JAVMA.1989;194:1265-1268.

26. Hunneman WA, Pijpers A, Lommerse J, Crauwels APP, Verheijden JHM. Prophylaxis of pleuropneumonia in pigs by in-feed medication with oxytetracycline and the subsequent transmission of infection. Vet Rec. 1994; 134:215-218.

27. Pijpers A, Schoevers EJ, Haagsma N, Verheijden JHM. Plasma levels of OTC, DC, and MC in pigs after oral administration in-feed. J Anim Sci. 1991;69:4512-4522.

28. del Castillo J, 1997; unpublished observations.

29. Luthman J, Jacobsson SO, Bengtsson B, and Korpe C. Studies on the bioavailability of tetracycline chloride after oral administration to calves and pigs. J Vet Med. 1989. Series A, 36:261-268.

30. Craig WA, Suh B. Theory and practical impact of binding of antimicrobials to serum proteins and tissue. Scand J Infect Dis. Suppl. 1978;14:92-99.

31. Scheife RT. Protein binding: What does it mean? Drug Intell Clin Pharm. 1989;23:S27-S31.

32. National Committee for Clinical Laboratory Standards. Performance standards for antimicrobial susceptibility tests. 4th edition; approved. Villanova, Pennsylvania: NCCLS document M2-A4. 1990.

33. Messier S, Higgins R, Nadeau M. À propos des épreuves de sensibilité pour Actinobacillus pleuropneumoniae. Med Vet Québec. 1993;23:127-129.

34. Elsener J, del Castillo JRE, Martineau GP. MIC interpretation: Flaws behind the scene. SHAP. Submitted for publication.

35. Hannan PCT, O’Hanlon PJ, Rogers NH. In vitro evaluation of various quinolone antibacterial agents against veterinary mycoplasmas and porcine respiratory bacterial pathogens. Res Vet Sci. 1989;46:202-211.

36. Pijpers A, van Klingeren B, Schoevers EJ, Verheijden JHM, van Miert ASJPAM. In vitro activity of five tetracyclines and some other antimicrobial agents against four porcine respiratory tract pathogens. J Vet Pharmacol Therap. 1989;12:267-276.

37. Kuwano A, Yamamoto K, Takei M, Kato M, Tachi H. Evaluation of the effect of ofloxacin in experimentally induced mycoplasmal pneumonia in swine. Proc IPVS Cong. 1992; 12:320.

38. Kim BH, Jung BY. Serotypes and antimicrobial susceptibility of Actinobacillus pleuropneumoniae isolated from pneumonic lungs of Korean swine. Proc. IPVS Cong. 1994;13:126.

39. Flaus L, Tan ATSC. Synergy study between lincomycin and oxytetracyclin, and between lincomycin and chlortetracyclin against Actinobacillus pleuropneumoniae and Pasteurella multocida. Proc. IPVS Cong. 1994;13:186.

40. Barigazzi G, Candotti P, Raffo A. Determinazione della minima ccncentrazione inibente (MCI) di 23 farmaci antibatterici nei confronti di 97 ceppi di Haemophilus parasuis isolati dal suino. Atti del XXI Meet Ann della Societa Italiana di Patologia ed Allevamento dei Suini. Montichiari, Italia. 1994; 135-141.

41. Yamamoto K, Koshimizu K. In vitro susceptibility of Mycoplasma hyopneumoniae to antibiotics. Proc. IPVS Cong. 1984;8:116.

42. ter Laak EA, Pijpers A, Noordergraaf JH, Schoevers EC, Verheijden JHM. Comparison of methods for in vitro testing of susceptibility of porcine Mycoplasma species to antimicrobial agents. Antimicrob Agents Chemother. 1991;35:228-233.

43. Cooper AC, Fuller JR, Fuller MK, Whittlestone P, Wise DR. In vitro activity of danofloxacin, tylosin, and oxytetracycline against mycoplasmas of veterinary importance. Res Vet Sci. 1993;54:329-334.

44. del Castillo J. Biodisponibilités orale et parentérale des pénicillines naturelles utilisées en prophylaxie de la streptococcie chez le porcelet sevré (MSc thesis). St-Hyacinthe, Canada:Université de Montréal. 1995;145pp.

45. Turndidge JD. Prediction of antibiotic dosing intervals from in vitro susceptibility, pharmacokinetics and post-antibiotic effect: theoretical considerations. Scand J Inf Dis. 1991;Suppl.74:137-141.

46. Zhanel GG, Hoban DJ, Godfrey KM. The postantibiotic effect: A review of in vitro and in vivo data. Drug Intell Clin Pharm. 1991;25:153-163.

47. Gerber AU, Craig WA. Experimentelle studien zur frage des optimalen dosisintervalls in der antibiotikatherapie. Schweitz Med Wschr. 1982;112:42-45.

48. MacKenzie FM, Gould IM. The post-antibiotic effect. J Antimicrob Chemother. 1993;32:519-537.

49. Spivey JM. The postantibiotic effect. Clin Pharm. 1992;11:865-875.