Effect of the timing of food intake on the absorption and bioavailability of carbamazepine immediate-release tablets in beagle dogs

ABSTRACT

The effect of dosing time on the bioavailability of carbamazepine immediate-release (IR) tablets was investigated when administered to beagle dogs who were fasting, with co-administration of food (Co-food), and 0.5 h before food and 2 h after food. The study was conducted using a single dose of 200 mg (tablets/solution) with a 2 week washout period in a crossover design. Food intake significantly increased the rate and extent of tablet absorption. The C_max (µg·ml⁻¹, 8.13/3.65) and t_max (h, 1.83/0.92) were increased more than two-fold and the AUC_0-24 (µg·h·ml⁻¹, 20.09/8.19) was 2.5 times that of the values obtained under fasting conditions. The bioavailability of the tablets under fasting conditions was 91.2%, but increased to 223.5%, 182.8% and 148.4% in the Co-food, 0.5 h before food and 2 h after food groups, respectively (p < 0.05). Although there was no significant difference in the C_max or AUC_0-24 between the treatments with food, the absorption appeared to be reduced to some extent when the tablets were given 2 h after food. The oral bioavailability of CBZ IR tablets was significantly affected by the timing of the food intake. This is maybe favored by the fluctuations in the level of bile salts with the timing of food intake. To obtain acute therapy for a drug with narrow therapeutic window, attention should be given to the dosing time and food intake interactions. Copyright © 2012 John Wiley & Sons, Ltd.

Introduction

Food–drug interactions may influence the gastrointestinal (GI) absorption of orally administered drugs, usually as a result of food-induced changes in bioavailability [1, 2]. Since the bioavailability and clinical effect of most drugs are related, the bioavailability is an important parameter as far as pharmacokinetic effects are concerned. Often, drug absorption is inadvertently reduced [3, 4, 5] or increased [6, 7] by co-administration of an oral drug with meals, resulting in therapeutic failure or increased toxicity. Such interactions are frequently caused by the physicochemical properties of drugs, GI physiology as well as food factors, such as food components [8], food volume [9] and the timing of drug administration in relation to meals [4, 10]. In the fed state, the most important concern is a high risk of treatment failure arising from a significant variability in bioavailability. In particular, for drugs with a narrow therapeutic index and/or steep dose–response curve, this variability may have clinically significant implications. In susceptible patients, even moderate food–drug interactions may cause serious side effects, and drug administration with a consistent relationship to food intake should be considered.

Carbamazepine (CBZ, antiepileptic drug) was chosen as a model drug because of its low aqueous solubility (about 170 µg·ml⁻¹ at 25 °C) and narrow therapeutic window (4–12 µg·ml⁻¹) [11, 12, 13]. Being poorly water soluble, its absorption is limited by its dissolution rate, which often results in irregular and delayed absorption [14, 15]. The interactions between food and CBZ are largely unknown. The bioavailability of CBZ in sustained release forms is not significantly affected by concurrent food intake [16, 17] which is mainly due to the role of formulation factors. In contrast, a standard food diet or grapefruit juice has been reported markedly to enhance the oral bioavailability of CBZ immediate-release (IR) tablets [18, 19, 20, 21]. However, the effect of the timing of food intake on the absorption and bioavailability of CBZ IR tablets and the underlying mechanisms of the interaction are not completely understood and further studies are still necessary.

Despite increasing reports of considerable variability in CBZ drug levels in the blood [15, 22], CBZ IR tablets are widely used in clinical situations in China partly due to the price of IR tablets, which is 10–15 times lower than that of sustained release forms. Because frequent (three or four times daily) dosing of CBZ IR tablets is required and the effects of food on their gastrointestinal absorption, the timing of food intake may affect the fluctuations in CBZ plasma concentrations. It is of the utmost importance to identify the optimal timing of drug administration in relation to food intake.

To reduce the risks to human subjects and to minimize the research costs, a suitable animal model for testing is usually a good idea. In spite of the higher intestinal pH and the shorter intestinal transit time, the gross physiology of the stomach in humans is very similar to that in dogs in the fasted state [23]. Moreover, beagle dogs have been reported to be a good animal model for studying food effects on the absorption of drugs [24, 25, 26]. Beagle dogs were used to see if they were a good in vivo model for predicting the potential food intake effect of CBZ IR tablets in the present work.

No pioneering work has been carried out to demonstrate the impact of the timing of food intake on the bioavailability of CBZ IR tablets. Therefore, the objectives of this study were three-fold: (1) to determine the effect of food on the bioavailability of CBZ IR tablets; (2) to investigate the optimal timing of CBZ IR tablet intake in relation to food intake; and (3) to explore whether dogs could serve as useful surrogates for evaluating food effects on the absorption of CBZ.

Materials and Methods

Materials

Standards and bulk powder CBZ were purchased from Chinese Pharmaceuticals Ltd (99.5%, Beijing, China) and Suzhou Wanqing Pharmaceutical Co. Ltd (CBZ, 99%, Suzhou, China), respectively. Carbamazepine immediate-release (IR) tablets (100 mg) were obtained from Beijing ShuGuang Pharmaceutical Co., Ltd, Beijing, China (lot no. 070933). Phenacetin was supplied by the National Institute for the Control of Pharmaceutical and Biological Products (Beijing, China). Polyethyleneglycol 400 (PEG 400) was purchased from Tianjin Kemiou Pharmaceutical Co. Ltd (Tianjin, China). Methanol and ethyl acetate were HPLC grade. All other chemicals and solvents were of analytical reagent grade, and deionized double-distilled water was used throughout the study. All tests were completed before the expiry dates on the materials.

Animals and dosing

All the animal studies were performed according to the Guidelines for the Care and Use of Laboratory Animals approved by the Committee of Ethics of Animal Experimentation of Shenyang Pharmaceutical University. Ten beagle dogs of both sexes (age range 10–12 months; body weight 10 ± 0.5 kg) were provided by the Animal Center of Shenyang Pharmaceutical University (Shenyang, China). Animals were housed in a room with a controlled temperature and humidity and allowed free access to food (standard chow) and water. All dogs were given 7 days to acclimatize to the facility before the experiments began and were divided randomly into five groups (each n = 2 animals).

The study was investigated using a 2 week washout period with a five-way crossover design. Carbamazepine solution was prepared by dissolving CBZ in PEG 400 to give a final concentration of 20 mg·ml⁻¹ for oral administration. Dogs received 200 mg CBZ (2 tablets) or CBZ solution (10 ml) in the following treatments: solution under fasting conditions, tablets under fasting conditions, co-administration with food (Co-food), 0.5 h before food and 2 h after food. The details of the dosing protocol were as follows: under fasting conditions, tablets (Fasted) and CBZ solution were orally given to dogs at least 12 h after the last meal and 6 h before the next meal. For the fed state, the dogs were fasted for 12 h until 5 min (Co-food), before 2 h (2 h after food) and after 0.5 h (0.5 h before food) prior to dosing, and were then given 200 g solid food (dog diet containing 27% protein and 16% fat, Longjin, Kim Oil Co., Ltd, China) mixed with 350 ml water and 250 ml milk.

Blood samples (3 ml) were collected from foreleg vein with heparinized catheters at 0 (predose), 0.25, 0.5, 0.75, 1, 1.5, 2, 3, 4, 5, 6, 8, 10, 12 and 24 h after administration. Then, the plasma was obtained from the whole blood by centrifugation at 6000 rpm for 5 min and stored at −20 °C until analysis.

Sample analysis

The concentration of CBZ in plasma was determined by HPLC using a Jasco PU1580 pump equipped with a UV1575 ultraviolet detector (Tokyo, Japan). The analytical column was a Diamonsil® C₁₈ column (250 mm × 4.6 mm i.d., 5 µm), the mobile phase was a mixture of methanol and distilled water (70:30, v/v), and the flow rate was 0.8 ml/min. UV-detection was performed at 284 nm. Phenacetin was employed as the internal standard (IS). Samples were analysed by a published method with minor modification [27]. Briefly, to aliquots of 0.5 ml plasma, 15 µl of IS solution (400 µg·ml⁻¹) was added followed by homogenization and extraction with 3.0 ml ethyl acetate on a vortex mixer (XW-80A, Shanghai, China) for 3 min. After centrifugation at 4000 rpm for 5 min, 2.5 ml of the organic layer was transferred to another clean tube and evaporated to dryness under nitrogen gas at 40 °C. Then, the residue obtained was reconstituted in 100 µl mobile phase and 20 µl of the sample was injected into the HPLC for analysis.

Pharmacokinetic and statistical analysis

The pharmacokinetic parameters of CBZ were derived from plasma concentration–time curves using a non-compartmental model with Data Analysis System Version 2.0 software (Center for Drug Clinical Evaluation, China). The areas under the plasma concentration–time curve (AUC_0–24) were calculated from the observed plasma concentrations from 0 to 24 h. Each AUC_0–∞ was calculated as the AUC_0–24 and the ratio of C_t/k_e, where C_t was the last measured concentration and k_e was the slope of the linear regression of the log-transformed plasma concentration time in the terminal phase. The C_max and t_max values were the observed maximal drug concentration and its time, respectively. The lag-time of the CBZ absorption (t_lag) was used as an estimate of the gastric residence time. The bioavailability was calculated based on the following equation:

• 

OriginPro 8.1 SR3 (OriginLab Corp., MA, USA) was used for plotting the arithmetic means of the plasma concentration–time curves.

Statistical analyses were based on untransformed values for C_max, t_max and AUC_0-24 (mean ± SD). Statistical significance was assessed by a two-way factorial analysis of variance test followed by Duncan's multiple comparison between groups with the statistical programs SPSS Statistics 17.0 software (SPSS Institute, Inc., Shanghai, China) [28]. Differences were only considered significant at p ≤ 0.05.

Results

Pharmacokinetic studies

The tolerability of both CBZ tablets and solution was good. Drug-related side effects were mild and transient in all ten beagle dogs. No laboratory abnormalities were noted.

The CBZ analytical method was well controlled and stable. Calibration curves constructed daily throughout the studies showed an average recovery of 93.9 ± 1.80% over the range 0.04–10.0 µg·ml⁻¹. The limit of sensitivity for the analysis of CBZ was 0.025 µg·ml⁻¹. The inter-day and intra-day precision over the range of the standard curve were characterized by CVs of 2.0–7.30% and 0.50–8.50%, respectively.

The mean plasma concentration versus time profiles (C-t) of CBZ IR tablets and solution in the fasted state are shown in Figure 1, and the profiles in relation to the timing of food intake and their comparisons with the fasted control are shown in Figure 2. Relevant pharmacokinetic parameters corresponding to the C–t curves are summarized in Table 1. The detailed statistical analysis of the pharmacokinetic data between dosing times is presented and compared in Table 2.

Figure 1. Pharmacokinetic profiles following oral administration of carbamazepine tablets and solution (200 mg) in the fasted state (mean ± SD, n = 10)

Figure 2. Pharmacokinetic profiles following oral administration of carbamazepine tablets (200 mg) in the fasted and fed state (mean ± SD, n = 10)

Table 1. Pharmacokinetic parameters following oral administration of CBZ tablets and solution (200 mg) to dogs under the five treatments (mean ± (SD), n = 10)

Dosing time (Preparation)	Cmax (µg·ml^−1)	tmax (h)	AUC0-24 (µg·h·ml^−1)	tlag (h)	Bioavailability (%)
Fasted (Solution)	6.24 (1.61)	0.31 (0.17)	8.99 (1.77)	0.07 (0.01)	100.0
Fasted (Tablets)	3.65 (0.89)	0.92 (0.14)	8.19 (1.59)	0.20 (0.03)	91.2
Co-administration with food (Tablets)	8.13 (2.62)	1.83 (0.42)	20.09 (6.20)	0.80 (0.21)	223.5
0.5 h before food (Tablets)	8.34 (1.30)	1.20 (0.27)	16.43 (1.82)	0.23 (0.11)	182.8
2 h after food (Tablets)	5.93 (1.05)	1.30 (0.74)	13.34 (0.74)	0.24 (0.15)	148.4

Table 2. Statistical analysis of the main pharmacokinetic parameters of CBZ in the fasted/fed state by Duncan's multiple comparisons

Comparison between the treatments	Cmax	tmax	AUC0-24	MRT	tlag
** Indicates significant differences among the means (p ≤ 0.05) of CBZ and no superscripts indicate no significant difference.
CBZ solution vs Fasted					**
CBZ solution vs Co-food	**	**	**		**
CBZ solution vs 0.5 h before food	**	**	**		**
CBZ solution vs 2 h after food		**	**		**
Fasted vs Co-food	**	**	**	**	**
Fasted vs 0.5 h before food	**		**	**
Fasted vs 2 h after food			**	**
Co-food vs 0.5 h before food				**	**
Co-food vs 2 h after food					**
0.5 h before food vs 2 h after food				**

The use of food as a co-component significantly increased the oral absorption of CBZ compared with that in the fasted state. The timing of food intake also had a complicated effect on the bioavailability of CBZ IR tablets. As shown in Figure 1, no marked differences in the shape of the C–t curves between the solution and tablets treatments were seen, expressed by higher C_max (6.24/3.65), prolonged t_max (0.31/0.92) and delayed t_lag (0.07/0.20) values for the CBZ solution. The bioavailability of CBZ tablets in the fasted state was 91.2% with CBZ solution as the reference. Significant differences in the shape of the C–t curves between the Co-food and fasted state were observed. As illustrated in Figure 2 and Table 1, the mean C_max estimate from tablets ingested with food (8.13 µg·ml⁻¹) was found to be 223% higher and statistically significantly different (p = 0.003) compared with the mean from the fasted treatment (3.65 µg·ml⁻¹). However, the C_max values were not significantly different among the treatments with food (p > 0.05), which were 8.13 ± 2.62, 8.34 ± 1.30 and 5.93 ± 1.05 µg·ml⁻¹ at the dosing time of 0 h (Co-food), 0.5 h before food and 2 h after food, respectively. Likewise, the two primary comparisons, which were 0.5 h before food vs Co-food and 2 h after food vs Co-food, showed similar differences in the AUC_0-24 data (p > 0.05). The bioavailability of CBZ tablets increased significantly to 223.5% when the tablets were administered together with food compared with the fasted control, showing an increase of 132.4%.

When the tablets were given 0.5 h before food, the absorption of CBZ was rapid and complete (bioavailability = 182.8%, p = 0.000). Similar to the Co-food, the values of the AUC_0-24 and C_max were more than twice that in the fasted state. However, the time parameters (t_max and t_lag) were not significantly different (p > 0.05) compared with the fasted control. Again, marked differences in the C–t courses between the dosing time of 2 h after food and the other treatments were found, reflected in the relatively complex and lower curves. Although there was no statistically significant difference (p > 0.05), the values of AUC_0-24 (13.34 µg·h·ml⁻¹) and C_max (5.93 µg·ml⁻¹) were lower than those from the other two fed states. Also, CBZ levels when given 2 h after food were much higher than those obtained in the fasted state (bioavailability =148.4%, p = 0.000).

The fed/fasted C_max and AUC in dogs and humans following administration of a single oral dose are compared in Table 3. Data for humans were adapted from those reported in the literature [19, 20]. The absorption of CBZ in the two species showed a clear increase with the food intake, and the extent of absorption in humans was a little lower than that observed in dogs. A 1.22-fold (or 1.41-fold) increase in AUC and a 1.26-fold (or 1.40-fold) increase in C_max were observed in the fed state in humans, compared with a more than 2.2-fold increase in dogs. Moreover, the ratios of fed/fasted for the two pharmacokinetic parameters in dogs were different under the different dosing time with the food intake. The AUC_fed/AUC_fasted and C_max-fed/C_max-fasted of 2 h after food group in dogs were about 1.60, which were closer to the results of humans.

Table 3. Comparison of fed/fasted pharmacokinetic parameters in the dog vs human for carbamazepine tablets with various propensities for food effect

Parameter	Human fed/fast		Dog fed/fast
	Co-administration with food		Co-administration with food	0.5 h before food	2 h after food
Data for humans from literature ^a Levy et al. (1975) [19] and ^b Malhotra et al. (2002) [20].
Cmax	1.26a	1.39b	2.23	2.28	1.62
AUC	1.22a	1.38b	2.45	2.00	1.63

Discussion

Several authors have reported that the plasma levels of carbamazepine-10,11-epoxide (CBZ-E), which is the main metabolite of CBZ, were very low and could not be measured accurately until 72 h after a single dose of the parent drug [20, 29, 30]. No plasma CBZ-E was detected during the single dose administration in the present study. The present study only examined the plasma concentrations of parent CBZ in dogs. Also, the results showed that the method described has good reproducibility and overall accuracy for dog plasma samples over the concentration range studied.

The aim of the comparison between CBZ solution and tablets in fasted dogs was to evaluate the effect of the formulation on drug absorption. The gastric residence time (t_lag) of the drug itself (0.07 h) was obtained by investigating the oral solution. The disintegration time of CBZ IR tablets in the stomach was less than 8 min (t_{lag tablets} − t_{lag solution} = 0.13 h). The absorption of CBZ from IR tablets was not affected by any drug–formulation interaction compared with that of the solution. In addition, the highest oral absorption of poorly water-soluble drugs could usually be seen when drugs were administered as solution. However, the oral absorption of CBZ from solution was significantly lower than that from tablet with food, 0.5 h before food and 2 h after food. The observed phenomena were not due to inappropriate experimental design or unreliable results, but to the faster absorption and metabolism of the CBZ solution, on the contrary, the delayed absorption, the higher levels of bile salts and the slower metabolism of the CBZ tablets with food. Moreover, the bioavailability of CBZ tablets in fasted dogs was 91.2%, which indicates that the dosing formulation of solution used in the study design was very appropriate.

Comparing the ‘Co-food’ and ‘Fasted’ states, it is clear that food consumption produced a significant increase in CBZ plasma concentrations from 0.8 h onwards. This could be explained by the prolonged time parameters (t_max and t_lag), especially in the Co-food state (the ratio of Co-food /fasted for t_max and t_lag, was 1.83/0.92 and 0.80/0.20, respectively). This showed that food delayed the gastric emptying of CBZ and prolonged the gastrointestinal absorption time. Although the bioavailability of CBZ tablets increased when taken with food, there was a pronounced inter-individual variation in this interaction, which was reflected by the SD of the C_max and AUC_0-24 from the treatment with food which were higher than those of the other treatments. Taking CBZ without a consistent relationship to meals may result in undesirable fluctuations in drug concentrations, and the significantly increased and larger fluctuations in bioavailability may pose some clinical risk. The co-administration of CBZ IR tablets with food is not a good idea from the observed results in dogs.

The dosing time at 0.5 h before food, and the bioavailability of CBZ IR tablets were markedly increased. The most plausible explanation for this is that CBZ tablets completely disintegrate and begin to be absorbed (t_lag under fasting conditions was 0.20 h) when food was ingested. Meanwhile, food intake may result in an increase in drug solubility because of stimulation of biliary secretion and the solubility of CBZ is increased by bile micelles [31]. The absorption of CBZ is limited by its solubility [14, 15], and such increased results will certainly promote the bioavailability of CBZ. Under this dosing treatment, the absorption of CBZ was markedly increased, which may result in serious toxicity given the narrow therapeutic index of CBZ.

The dosing time at 2 h after food was chosen because the cyclic pattern of gastric motility in dogs is about 2 h (ranging between 1 and 3 h) [23], which means that gastric contraction can drive most of the gastric contents into the small intestine. This is supported by the similar values of t_lag between 2 h after food (0.24 h) and the fasted state (0.20 h) (p > 0.05). When most of food enters the small intestine (SI), biliary secretion is not markedly stimulated. Thus, the solubility of CBZ was not significantly increased, which did not markedly affect its bioavailability. This may be why the C–t curve in the 2 h after food group was significantly lower than the other two treatments of food. Compared with the state of Co-food or 0.5 h before food, the kinetic curve in the 2 h after food group showed appropriate absorption behaviors (drug concentration was not so high). Also, the SDs of the C_max (1.05) and AUC_0-24 (0.74) were significantly lower than the other two treatments of food. Compared with the fasted control, the pharmacokinetics of the 2 h after food group produced higher drug concentrations and a longer retention time, which were satisfactory results. Compared with the fed/fasted of AUC and C_max , the results of the 2 h after food group in dogs showed a closer prediction for those reported for co-administration of food in humans. Moreover, the ingested food will have a protective effect on the gastric mucosa and reduce the drug stimulation and side effects. In summary, a dosing time of 2 h after food may be the optimal time to administer CBZ IR tablets.

The multi-peak in the 2 h after food group seen in the kinetic curves was due to the combined effects of several factors. The 0.25–1 h peak may have been due to rapid absorption from the upper gastrointestinal (GI) tract [32]. The 1–5 h peak may be attributed to the multi-site and inhomogeneous absorption in the GI tract [33]. The 5–8 h peak may be due to the food intake during the experiment. Another important contribution comes from the interrupted character of CBZ absorption that is caused by the very poor solubility of CBZ [34].

The most important pharmacokinetic food–drug interactions are caused by changes in the absorption of a drug because of chemical reactions between the drug and food (e.g. chelation) or the physiological response to food intake (food may change the physiology of the SI, which includes changes in SI pH, an increase in splanchnic blood flow and a change in the luminal metabolism of a drug) [2, 35, 36]. These factors may also produce changes in drug absorption. However, compared with the state of Co-food or 0.5 h before food, it can be seen that the effects of these factors on the bioavailability of CBZ were not major. Carbamazepine is a neutral compound, and its solubility and dissolution rate are almost unaffected by the medium pH [12]. The saturation solubility of CBZ in pH 1.2–7.4 buffer solution was 316 ± 30 µg·ml⁻¹ in this study (n = 3, data not shown). The variation of pH in the gastrointestinal tract may not influence the solubility of CBZ. It is well known that the absorption of CBZ is limited by its solubility. So, the main reason for the effect on CBZ absorption may be the increased biliary secretion, which is supported by the pharmacokinetic results for the treatment at 0.5 h before food and 2 h after food.

The fed/fasted C_max and AUC of CBZ in dogs showed the food effect to be greater than those reported in humans. Although failure of the dog quantitatively to predict the human food effect, the trends of the food intake effect of CBZ tablets can still be predicted accurately using dogs as suitable surrogates. Further investigation is needed fully to understand this species difference in food on the absorption of CBZ. The food effects studies in dogs and humans were conducted under different conditions that could have contributed to the differences in the magnitude of the absorption response to food. The healthy subjects received a 420 mg dose (6 mg/kg for about 70 kg body weight) in the literature [19] and the dogs received a 200 mg dose in the present study. More importantly, the total volume of test food administered to the dogs was similar to a full human meal [19], and also, the composition of the fed state test food varies substantially between dogs and humans [19, 20]. These factors may partly contribute to the over-prediction. In addition to the consideration of differences in presystemic drug metabolism and the higher levels of bile salts in fed dogs resulting in the marked increased in the extent of absorption of CBZ [23, 36, 37], may explain the disconnect to observed human data. It generally appears that control of dog gastric pH and the amount and/or composition of the fed state are important parameters to improving the predictability of the dog overall as a human food effect model.

In summary, the timing of the food intake has a marked effect on the absorption of CBZ IR tablets. The shorter the interval between the dosing time and food intake, the greater the variation in the bioavailability of the CBZ tablets. The main reason may be contribute to the variation of the level of bile salts with the timing of food intake. The fluctuations in drug plasma concentrations for a drug with a narrow therapeutic window may result in serious toxicity. To obtain optimum oral bioavailability and tolerability, attention should be given to the dosing time and food intake interactions.

Acknowledgements

The authors thank all members of the laboratory for their help. Dr Jack David is gratefully thanked for revising the manuscript.