Determination of the biliary excretion of piperacillin in humans using a novel method

Abstract

Aim

To evaluate the applicability of a novel method to determine the biliary excretion of piperacillin.

Methods

Healthy volunteers were administered piperacillin i.v. Duodenal aspirates were collected via a custom-made oroenteric catheter; blood and urine also were collected. Gallbladder ejection fraction (EF) was determined by gamma scintigraphy and pharmacokinetic parameters were calculated using noncompartmental analysis.

Results

The fraction of the piperacillin dose excreted unchanged into bile was 1.1 ± 0.3% (biliary clearance corrected for EF was 0.032 ± 0.008 ml min⁻¹ kg⁻¹).

Conclusions

This methodology can be used to determine reliably the biliary clearance of drugs that are excreted only marginally into bile. Normalization of biliary clearance for EF significantly reduces intersubject variability of this parameter.

Introduction

Major challenges in developing an accurate method to quantify biliary clearance in healthy humans include the complex anatomy of the human hepatobiliary system, as well as the intermittent and incomplete expulsion of bile from the gallbladder. Pharmacological contraction of the gallbladder has been employed to collect as much bile stored within the gallbladder as possible [1]. However, gallbladder response can vary considerably among human subjects and is influenced by a variety of factors. The estimated extent of biliary excretion of the drug of interest may be inaccurately low and intersubject variability may be exaggerated unless the degree of gallbladder contraction is considered.

Recently, we developed a technique to quantify biliary clearance of compounds in humans that couples gamma scintigraphy with a specially designed oroenteric catheter [2]. This catheter is equipped with an inflatable occlusive balloon that facilitates periodic collection of duodenal fluids and bile. The gallbladder ejection fraction (EF) is determined by administering a hepatobiliary imaging agent (Tc-99m mebrofenin) that allows visualization of the gallbladder before, during and after pharmacological contraction. Administration of a gamma emitter that is excreted into bile also aids in the identification of bile leakage beyond the occlusive balloon in the duodenum. The versatility of this approach was investigated by applying it to piperacillin, a drug expected to exhibit low biliary clearance. Piperacillin is a third-generation, broad-spectrum, penicillin derivative used in the treatment of intra-abdominal infections and prophylactically for gastrointestinal and biliary tract procedures [3].

Additionally, the effect of prolonged intubation on gallbladder contractility and on the determination of biliary clearance of drugs was investigated by comparing the recovery of Tc-99m mebrofenin in the present and in our previous study [2].

Methods

Clinical protocol

Three male volunteers (23–40 years of age) within 20% of ideal body weight completed the study. The Clinical Research Advisory Committee and the Committee on the Protection of the Rights of Human Subjects at the University of North Carolina at Chapel Hill School of Medicine approved all procedures. Written informed consent was obtained from all subjects.

Since prolonged fasting can prevent gallbladder contraction in response to cholecystokinin-8 (CCK-8) administration, subjects were given a high-fat caffeine-free meal 12 h prior to the procedure and a high-fat snack 5 h later, in order to maintain gallbladder contractile response. After an overnight fast, a custom-made oroenteric tube was positioned in the upper duodenum as previously described [2]. The distal end of the tube was fitted with a polyethylene balloon, which was inflated to occlude the intestine during bile collection.

Subjects were administered 2 g of piperacillin (gift of American Pharmaceutical Partners Inc., Schaumburg, IL, USA) as a 15-min i.v. infusion and blood samples were collected at designated times through 600 min. To visualize the gallbladder, 2.5 mCi of Tc-99m mebrofenin (Choletec^®; Bracco Diagnostics, Princeton, NJ, USA) was administered as an i.v. bolus 2 h post piperacillin, subjects were positioned under a gamma camera (E-CAM Dual Head Gamma Camera; Siemens Inc., Hoffman Estates, IL, USA) and 1-min anterior planar gamma scintigraphic images of the abdomen were acquired for 4 h. To collect the majority of the bile containing drug accumulated in the gallbladder, contraction was induced with CCK-8 (Sincalide, Kinevac^®; 0.02 µg kg⁻¹; i.v. infusion over 30 min) 4 h after piperacillin administration. If CCK-8 failed to contract the gallbladder, 10–20 ml of half-milk/half-cream was delivered into the stomach ports of the catheter to stimulate endogenous CCK secretion by the presence of partially digested proteins and fats in the small intestine [4]. In one case, orange zest was used as an olfactory stimulant to promote gallbladder contraction upon failure of the previously described methods. Duodenal aspirates were collected for 6 h, urine and blood were collected for 10 h. Radioactivity was quantified in blood, bile and urine as previously described [2]. Piperacillin was quantified in plasma, bile and urine as described below.

Analytical methods

High-performance liquid chromatography (HPLC-UV) was used to quantify piperacillin in plasma and urine. An HP1050 system (Agilent Technologies Inc., Loveland, CO, USA) was equipped with a Phenomenex Luna column (C₁₈, particle size 5 µm, 250 × 4.60 mm), the flow rate was 1.5 ml min⁻¹, detection wavelength 254 nm and dynamic range 1–200 µg ml⁻¹. Samples were analysed with a 14-min gradient (sodium acetate 10 mm, pH 5.2/acetonitrile). For plasma, the gradient started at 18% acetonitrile and increased to 25% over 7 min; for urine, the gradient started at 13% acetonitrile and increased to 25% over 9 min. Prior to analysis, cefoxitin (internal standard, final concentration 20 µg ml⁻¹) was added to 300 µl of plasma; 100 µl of ammonium sulphate saturated solution was added to precipitate proteins. Upon addition of 600 µl of acetonitrile and centrifugation (15 000 g for 10 min), the supernatant was evaporated and reconstituted in sodium acetate. Urine samples were diluted with mobile phase containing cefoxitin. Accuracy ranged between 83 and 119%, and the interday and intraday CV% was ≤11% for both matrices.

HPLC-MS-MS was used to quantify piperacillin in bile. Samples were diluted 1 : 7 with methanol containing the internal standard cefazolin (4.5 µg ml⁻¹). An HP1100 system (Agilent) was coupled to an Applied Biosystems API4000 mass spectrometer with a TurboIonSpray source and controlled by Analyst software (Foster City, CA, USA). The analytes were eluted with a methanol/water (0.5% formic acid) gradient from an Aquasil column (C₁₈, particle size 5.0 µm, 50 × 2.1 mm) with a flow rate of 0.75 ml min⁻¹. The gradient was held at 95% water for 0.5 min, then decreased to 10% water over 2 min and held for 1 min and then returned to 100% water at 3.5 min, with a total run time of 4 min. The dynamic range was 1–100 µm. Accuracy ranged between 89 and 99%, and the interday and intraday CV% was ≤17.2%.

Pharmacokinetic analysis

A gallbladder EF was calculated from anterior planar scintigraphic images using ESOFT 2.5 software (Siemens) as described below. GB represents the counts over the gallbladder region and the time in all formulae refers to the minutes after piperacillin administration. The amount of piperacillin (X_bile) collected in bile during the time period used for the EF calculation was normalized by EF to account for the drug excreted into bile yet remaining in the gallbladder due to incomplete contraction (X_GB = Excreted dose).

EF = (GB^240min − GB^360min)/GB^240min

The amount of Tc-99m mebrofenin (X_GB) collected in bile over 4 h was calculated and corrected by EF in order to compare the recovery obtained in this study with previous results [2]:

Noncompartmental analysis was used to calculate pharmacokinetic parameters using WinNonlin Pro vs. 4.2 (Pharsight, Mountain View, CA, USA). The area under the plasma concentration–time curves (AUC) was calculated using the linear trapezoidal rule and, where appropriate, extrapolated to infinity using the slope obtained from linear regression of the last three to six time points. Total clearance (CL_total), biliary clearance (CL_biliary) and corrected for EF biliary clearance (in vivo CL_biliary) were calculated as follows, where X₀ was the administered dose and T_last was 180 min for Tc-99m mebrofenin and 360 min for piperacillin:

CL_total = X₀/AUC_0–∞

CL_biliary = X_bile/AUC_{0–T last}

in vivo CL_biliary = X_GB/AUC_{0–T last}

Results

The overall duration of intubation and fasting time between the administration of Tc-99m mebrofenin and CCK-8 in this study was 4 h longer than in the design adopted previously [2] to account for the longer elimination half-life of piperacillin. These modifications appeared to affect gallbladder contractility. All seven subjects enrolled in these two studies reported intestinal cramping in response to CCK-8. Three out of four subjects in our previous study had an EF ≥0.82, while only one of the volunteers enrolled here demonstrated normal gallbladder contraction by the end of the study [5]. Since CCK-8 did not completely contract the gallbladder of subjects A, B or C, these volunteers were administered 10–20 ml half-milk/half-cream at 160 min post Tc-99m mebrofenin administration in order to prompt endogenous CCK release. Subject A demonstrated satisfactory gallbladder contraction (EF =0.82) following fatty food (milk/cream) administration, while subjects B and C did not. Due to complete lack of response to CCK-8 and half-milk/half-cream, subject C was asked to smell orange zest 200 min post Tc-99m mebrofenin administration in an attempt to stimulate gallbladder discharge. As seen in Figure 1, the majority of piperacillin recovered in the duodenal aspirates corresponded to the successful contraction of the gallbladder and subsequent bile discharge into the duodenal lumen in all three subjects.

Figure 1.

Piperacillin mean plasma concentration (•) – time profile up to 360 min is plotted on the left axis (mean ± SD). Individual piperacillin cumulative biliary recoveries expressed as a percentage of the maximum recovery are plotted on the right axis for subjects A (——), B (–.–) and C (- - -). CCK-8, fatty food (milk) and orange zest were administered at 240, 280 and 320 min post piperacillin administration (arrows)

Piperacillin pharmacokinetics were similar to previously published data [6]. The mean plasma concentration–time profile and individual cumulative biliary excretion plots for subjects A, B and C are presented in Figure 1 and pharmacokinetic parameters are listed in Table 1. Piperacillin biliary recovery was minimal, with the majority of the dose excreted as unchanged drug in urine. Correction for incomplete contraction of the gallbladder decreased intersubject variability in the estimate of the amount of piperacillin excreted in bile and in the calculation of the corrected biliary clearance, by accounting for the amount of piperacillin that would have been collected if the EF had been complete (Table 1). Additionally, the calculation of Tc-99m mebrofenin recovery of excreted dose (mean ± SD) in this study was high (60.9 ± 14.9%) and comparable to our previous results (84.2 ± 9.3%[2]). The Tc-99m mebrofenin dose recovered in bile, corrected for EF, can be used to calculate the corrected biliary clearance of this probe for the four subjects enrolled in our previous study. When correcting biliary clearance for EF (in vivo CL_biliary = 16.1 ± 3.2 ml min⁻¹ kg⁻¹vs. CL_biliary 12.5 ± 3.6 ml min⁻¹ kg⁻¹), biliary elimination accounts for 93% of Tc-99m mebrofenin systemic clearance (17.3 ± 1.7 ml min⁻¹ kg⁻¹[2]).

Table 1.

Summary of individual and mean pharmacokinetic parameters, biliary recovery and urinary recovery of piperacillin and ejection fraction for subjects A, B and C

Parameter	Subject A	Subject B	Subject C	Mean	SD
Administered dose (mg)	1700	1764	1586	1683	90
Biliary recovery (% dose)	0.77	0.32	0.14	0.41	0.32
Urinary recovery (% dose)	72	85	60	72	13
Ejection fraction	0.82	0.28	0.07	0.39	0.39
Recovery of Excreted dose (% dose)	0.87	1.1	1.4	1.1	0.3
CLtotal (ml min−1 kg−1)	3.2	2.5	2.9	2.9	0.3
CLbiliary (ml min−1 kg−1)	0.025	0.008	0.004	0.012	0.011
in vivo CLbiliary (ml min−1 kg−1)	0.028	0.028	0.041	0.032	0.008

Discussion

Inconsistent contraction of the gallbladder in healthy volunteers after CCK-8 administration has not been reported extensively in the literature, but has been observed by our group during the development of the current methodology [7]. The lack of responsiveness to CCK-8 may be attributed to many factors, including prolonged fasting or the presence of the tube and inflated balloon in the duodenum for a prolonged period of time. It is well documented in animals that the duodenum is important in maintaining sphincter of Oddi response to CCK [8]. Thus, it is possible that the pressure exerted by the inflated balloon on the duodenal walls may have modified the response to CCK-8. Alternatively, it is possible that subjects in our study may have expressed rare polymorphisms or have had downregulation of the CCKA/CCK1 receptor on gallbladder smooth muscle, which would have caused hypomotility in response to CCK-8 [9]. It is also possible, but unlikely, that the infusion rate affected subject response to CCK-8; the CCK-8 infusion rate in this study was chosen to mimic closely physiological serum CCK profiles after a fatty meal and to minimize the adverse effects of CCK-8 (abdominal cramps, nausea) and variability in gallbladder contraction [5].

Piperacillin biliary excretion estimates in this study are consistent with published results [10, 11] and demonstrate that piperacillin is only marginally eliminated into bile in humans. The correction of the excreted dose of piperacillin for the gallbladder EF greatly diminished intersubject variability and yielded estimates for the fraction of dose excreted into bile similar to previously reported values in T-tube patients [11]. These results established that incorporation of the EF factor in the biliary clearance calculation can provide a method to account for some of the inherent variability associated with the determination of this parameter in healthy volunteers.

Some variability in determining the biliary clearance of drugs by collecting duodenal bile may be associated with the failure to account for the fraction of the dose excreted into bile by hepatocytes, but reabsorbed through the ductal and gallbladder epithelial cells. In addition, it is not possible to distinguish biliary excretion from duodenal secretion of drugs using this methodology. However, the bile-rich duodenal aspirates are collected as soon as they are expelled into the intestinal lumen in order to minimize intestinal reabsorption.

Investigating the mechanisms of biliary clearance of drugs in healthy volunteers presents a number of challenges, including the inability to collect liver bile readily or obtain complete bile collection utilizing non-invasive techniques. Nevertheless, this study demonstrates that a combination of tools can be used to obtain a reliable estimate of the biliary clearance of a drug in healthy humans.