Single‐Dose and Multiple‐Dose Pharmacokinetics and Dose Proportionality of Intravenous and Intramuscular HPβCD‐Diclofenac (Dyloject) Compared with Other Diclofenac Formulations

To evaluate single‐ and repeated‐dose pharmacokinetics (PK) and dose proportionality of hydroxypropyl‐β‐cyclodextrin (HPβCD)‐diclofenac compared with Voltarol after intravenous (IV) and intramuscular (IM) administration.

Balanced inhibition of both cyclooxygenase (COX)-1 and COX-2 by the nonsteroidal antiinflammatory drug (NSAID) diclofenac allows it to be highly effective in decreasing pain and inflammation. 1,2 Recent studies have further demonstrated that diclofenac opens KCNQ2/3 potassium channels and inhibits sensory neuronal depolarization, resulting in analgesia. 3 Diclofenac has long been approved as safe and effective for both acute and chronic pain through a variety of routes. However, no injectable diclofenac formulation is currently approved in the United States, in part due to the poor aqueous solubility of diclofenac. Despite its proven efficacy, 4-6 use of the current injectable diclofenac formulation (Voltarol) marketed in Europe, Latin America, and other regions is limited by its cumbersome preparation and administration requirements. These include dilution, buffering with sodium bicarbonate, instability with consequent need for immediate administration following preparation, and slow administration rate (30 minutes to 2 hours). 7 In addition, this formulation uses two organic solvents, propylene glycol and benzyl alcohol, each of which is a known vascular irritant causing pain on injection [8][9][10] or, occasionally, local necrosis following intramuscular (IM) injection (Nicolau syndrome). 11 To overcome the poor aqueous solubility of diclofenac, we have used hydroxypropyl-b-cyclodextrin (HPbCD), a solubilizing agent currently used in several other pharmaceutical products. 12,13 HPbCD-diclofenac (diclofenac sodium solubilized with HPbCD, 37.5 mg/ml) was approved for marketing in October 2007 in the United Kingdom for the treatment or prevention of postoperative pain via intravenous (IV) or IM routes. 14 This ready-to-use HPbCD-diclofenac solution is stable at room temperature and designed to minimize the complexity and risks resulting from multistep preparation and administration of parenteral drugs. 15 Because it lacks irritating organic solvents and may be administered as a single rapid IV bolus injection, the HPbCD-diclofenac formulation optimizes the analgesic efficacy and safety profile of parenteral diclofenac for acute pain.
To evaluate the pharmacokinetics and safety of the HPbCD-diclofenac formulation, we conducted two separate studies in healthy volunteers. Study 1 used different routes of administration in comparison with Voltarol. Study 2 evaluated single and repeated administration of HPbCD-diclofenac in comparison with oral immediate-release diclofenac (Cataflam). The primary objective of these studies was to characterize the pharmacokinetic profile of HPbCD-diclofenac via different routes of administration compared with parenteral and oral formulations of diclofenac during single and repeated dosing. The secondary objective was to assess the safety of HPbCD-diclofenac following IV and IM administration. cology Research Unit, Baltimore, Maryland. The protocol and informed consent were reviewed and approved by the Chesapeake Research Review, Inc., Columbia, Maryland. The studies were conducted in accordance with the ethical principles of the Declaration of Helsinki.

Subjects
Healthy male and female nonsmoker volunteers 18 years of age or older were enrolled, with similar inclusion/exclusion criteria in studies 1 and 2. Females were required to have a negative pregnancy test, be nonlactating, and practice contraception. Subjects were excluded if they had any significant medical history or clinically relevant laboratory test results; were serologically positive for the human immunodeficiency virus, hepatitis B, or hepatitis C; had hypersensitivity to NSAIDs; or were substance abusers.
Subjects were screened 7 to 14 days prior to randomization. Subjects were evaluated prior to each visit to confirm compliance with inclusion and exclusion criteria.

Study 1
Subjects were randomized to each of four treatments using a computer-generated random sequence: A = one dose of HPbCD-diclofenac 75 mg/2 ml, IV bolus, delivered as previously described; B = one dose of HPbCD-diclofenac 75 mg/2 ml, deep intragluteal IM injection; C = one dose of Voltarol 75 mg/3 ml, 30-minute IV infusion; and D = one dose of Voltarol 75 mg/3 ml, deep intragluteal IM injection. Four treatment sequences were utilized to randomize six subjects per sequence. There was a minimum 5-day washout between treatments.

Study 2
Subjects were randomized to receive each of three treatments in a sequence determined by a computer-generated randomization list: A = HPbCD-diclofenac 18.75 mg IV bolus; B = 37.5 mg IV bolus injection, delivered as previously described; and C = Cataflam 50 mg orally. Each subject received four doses of each of these three diclofenac formulations at 6-hour intervals separated by a 48-hour washout. Study drug administration occurred on days 1, 4, and 7.
Blood samples from both studies were collected in 5-ml heparinized Vacutainers (BD, Franklin Lakes, NJ, USA) and centrifuged immediately after collection. The plasma was decanted into polypropylene tubes and stored at À20°C until assay.

Assay Methodology for Study 1 and Study 2
Diclofenac plasma concentrations were measured by liquid chromatography with tandem mass spectrometry detection, a method previously validated for the detection of diclofenac in human plasma and urine. 16 The assay used atmospheric pressure ionization with turbo ion spray followed by multiple reaction monitoring of the characteristic deprotonated molecular ion to product-ion transitions for diclofenac and internal standard. This assay was linear from 25 to 30,000 ng/ml. Its

Pharmacokinetic Analyses for Study 1 and Study 2
Pharmacokinetic parameters were calculated using noncompartmental analysis. Only those plasma concentrations equal to or greater than the lower limit of quantitation (LOQ; 25.0 ng/ ml) were used in the analyses. Actual sampling times were used in all pharmacokinetic analyses. Per protocol times were used to calculate mean plasma concentrations for graphical displays.
C max and T max were taken directly from the data. The elimination rate constant, kz, was calculated as the negative of the slope of the terminal log-linear segment of the plasma concentration-time curve. The data used for each subject and treatment were determined by visual inspection of a semilogarithmic plot of concentration versus time. Elimination half-life (t ½ ) was calculated as Area under the curve (AUC 0-t ) from zero to the final sample with a concentration ≥LOQ (AUC 0-t ) was calculated using the linear trapezoidal method and extrapolated to infinity using where C tf is the final concentration ≥LOQ. For each treatment in study 2, the following pharmacokinetic parameters were calculated using noncompartmental analysis: • First dose: C max , T max , AUC (0-t) , and t ½ • Fourth dose: C max , T max , AUC (0-6) , and t ½ Safety Safety assessments included adverse event (AE) monitoring, physical examinations, clinical laboratory tests, vital signs, and electrocardiograms (ECGs). Thrombophlebitis assessment by the clinical staff of the subject's IV site at 4 and 8 hours following IV administration was included after IV bolus administration. A 6-point thrombophlebitis grading scale 17 was used (0 = no reaction; 1 = tenderness along the vein; 2 = continuous tenderness or pain with redness; 3 = palpable swelling or thrombosis within the length of the cannula; 4 = palpable swelling or thrombosis beyond the length of the cannula; 5 = as for grade 4, with overt infection).

Study 1
Statistical sample size calculations were conducted using data from previous clinical studies. Within-subject coefficients of variation for natural log-transformed AUC 0-t and AUC 0-∞ were both approximately 13%, indicating that sample sizes of 16-24 subjects should provide 80% power to obtain 90% confidence intervals within 80% ? 125% range if the true difference between observations for the different formulations was 5% or less.
Comparison of AUC 0-t and AUC 0-∞ between the HPbCD-diclofenac (test) and Voltarol (reference) treatments was conducted on natural logarithms of the primary data using an analysis of variance (ANOVA) model with sequence, subject within sequence, treatment, and period as classification variables. 18 Sequence was tested using subject within sequence as the error term; all other terms were tested using mean squared error. Confidence intervals (90%) were constructed for the ratios (test to reference) of the two parameters using the log-transformed data and the one-sided t-test procedures. The point estimates and confidence limits were exponentiated back to the original scale.

Study 2
The analysis used descriptive statistics (mean and standard deviation). Pharmacokinetic parameters were compared for different doses (18.75 and 37.5 mg) and multiple-dose administrations. C max and AUC (inf) for the first dose and C max and AUC (0-6) for the fourth dose were compared across treatments using ANOVA with subject and dose number as classification variables, using the natural logarithm of the data. 18 Pharmacokinetic data from both studies were analyzed using WinNonlin v.3.3 (Pharsight, Mountain View, CA, USA).

Study 1
Twenty-four healthy subjects were enrolled. were included in the pharmacokinetic analysis and in the statistical comparisons for which they had data for all treatments.

Study 2
Thirty-six healthy subjects were enrolled. All 36 subjects completed the pharmacokinetic component of the study and were included in the pharmacokinetic analysis.

Study 1
Consistent with the distinct administration protocols of rapid IV bolus versus 30-minute infusions, mean plasma concentration-time profiles after IV administration differed between HPbCD-diclofenac and Voltarol (Table 1, Figure 1). HPbCD-diclofenac was administered IV over 15 seconds and had a mean C max approximately 4-fold higher (21,524 ng/ml) than the C max for Voltarol (5668 ng/ml) administered IV over 30 minutes. Following IV administration of HPbCD-diclofenac, median C max occurred at 0.05 hours versus 0.5 hours for Voltarol. The mean half-life (t ½ ) of diclofenac was equivalent for both formulations, averaging 1.17 AE 0.32 hours for IV HPbCD-diclofenac and 1.23 AE 0.31 hours for IV Voltarol.
The mean C max (2569 ng/ml) after IM HPbCD-diclofenac was higher compared with Voltarol IM (C max 1541 ng/ml; Table 1). The higher C max with HPbCD-diclofenac compared with Voltarol after IM administration indicates a more rapid early exposure profile. The median T max following IM administration was 0.642 hours for HPbCD-diclofenac and 0.792 hours for Voltarol. The mean values for HPbCD-diclofenac and Voltarol for AUC 0-∞ were 4304 AE 908 ngÁhour/ml and 3932 AE 627 ngÁhour/ml, respectively, with a point estimate of 107.91% (90% CI 101.05-115.24), demonstrating equivalent bioavailability in terms of AUC after IM administration (Table 2, Figure 2).
The mean C max for Voltarol infused IV over 30 minutes was 5668 ng/ml compared with a C max of 2569 ng/ml for IM HPbCD-diclofenac. The T max for IV Voltarol was 0.5 hours versus 0.642 hours for IM HPbCD-diclofenac. Based on Value is higher than expected because one subject had a plasma concentration at the 3-minute blood draw time point that was 10-fold higher than expected. The clinical site deems it possible that the same cannula was used for drug administration and for the 3-minute blood draw for that subject. this pharmacokinetic data, the IM administration of HPbCD-diclofenac offers comparable delivery.
The mean t 1/2 following administration of IV HPbCD-diclofenac was 1.17 AE 0.32 hours versus 1.17 AE 0.31 hours after IM HPbCD-diclofenac, suggesting similar disposition profiles with this formulation by both IV and IM routes.

Study 2
Study 2 compared the pharmacokinetics of HPbCD-diclofenac following IV administration of single and multiple doses of 18.75 and 37.5 mg versus oral Cataflam ( Figure 3). As expected, mean plasma concentrations were maximal immediately after IV dosing. Oral Cataflam exhibited slower absorption, with a median T max of 1.5 hours ( Table 4). The C max following the first IV dose of HPbCD-diclofenac approximately doubled from a mean of 2904-6031 ng/ml for 18.75 versus 37.5 mg. The AUC 0-∞ also approximately doubled from a mean of 898-1859 ngÁhour/ml across IV doses of 18.75-37.5 mg of HPbCD-diclofenac, demonstrating dose-proportional pharmacokinetics for IV HPbCD-diclofenac. The C max following the    fourth IV dose was 3090 g/ml and 5617 ng/ml, for 18.75 and 37.5 mg, respectively. AUC (0-6) following the fourth dose of 18.75 and 37.5 mg were 935 and 1839 ngÁhour/ml, respectively. The C max and AUC following the first and fourth doses were similar, suggesting no accumulation of IV HPbCD-diclofenac. The calculated accumulation factor based on half-life was close to 1, indicating a lack of accumulation during successive doses given every 6 hours, approximately 3.5 9 t½. Plasma exposures to IV HPbCD-diclofenac 18.75 mg (866 ngÁhour/ml) and 37.5 mg (1843 ngÁhour/ml), as measured by AUC uncorrected for dose, bracketed that after administration of the first oral Cataflam 50 mg dose (1473 ngÁhr/ ml). This is consistent with the bioavailability of oral Cataflam, which was 64.1% after the first dose and 54.6% after the fourth dose.
The C max after the fourth dose of oral Cataflam 50 mg was 851 ng/ml compared with a C max of 1246 ng/ml following the first dose. The AUC following the first dose was 1473 ngÁhour/ ml compared with 1350 ngÁhour/ml following the fourth dose of Cataflam, suggesting variability in diclofenac's oral absorption. The IV profiles of HPbCD-diclofenac were similar and consistent following the first and fourth dose compared with a substantial decrease in C max from the first to the fourth dose of Cataflam.

Study 1
No serious AEs occurred during the study. There were no clinically significant changes in clinical laboratory tests, ECGs, or vital signs reported. A total of eight treatment-emergent AEs were reported by five subjects (HPbCD-diclofenac: two subjects; Voltarol: three subjects).
The specific AEs for Voltarol included anemia, dizziness, headache, sweating, and vasovagal attack; for HPbCD-diclofenac they included dysgeusia, postural dizziness, and headache. Of the reported adverse events, only one event (dysgeusia following IV administration) was considered "related" to the study drug. All other AEs were considered "not related" or "unlikely to be related" to the study drug and were mild or moderate in severity.
The thrombophlebitis assessment revealed that one subject (1 of 23, 4.3%) given HPbCD-diclofenac had mild irritation (1 = tenderness along the vein) at the 4-and 8-hour time points. Similarly, one subject (1 of 24, 4.2%) had mild irritation (1 = tenderness along the vein) only at the 4-hour time point following Voltarol administration.

Study 2
There were no serious AEs, and none of the subjects were discontinued from the study due to an AE. A total of 14 treatment-emergent AEs were reported by seven subjects; all were mild, resolved spontaneously, and most were unrelated to study treatments. Treatment-related AEs included three mild gastrointestinal events and one instance of mild injection site pain.
None of the chemistry or hematology changes were considered clinically significant. There were no clinically significant changes in vital signs. There were no clinically significant findings or observable differences between treatment sequences for quantitative or qualitative ECG parameters.

Discussion
HPbCD, a cyclic glucose-derived oligomer consisting of linked a-1,4-glucose units, was used to enhance the solubility of diclofenac for injection. 19 Compared with the previous formulation (Voltarol), this approach allows a reduction in dosing volume and lessened irritation provoked by the nonphysiologic pH and organic solvents. 20 HPbCD has likewise been used to enhance the solubility of poorly soluble drugs such as the marketed antifungal itraconazole 8 and a novel formulation of propofol under development. 9 When diclofenac is solubilized with HPbCD, a therapeutic dose of diclofenac is available in a smaller volume, 75 mg/2 ml versus 75 mg/3 ml as in Voltarol. Furthermore, when Voltarol is to be administered IV, it must first be diluted to 50 to 100 ml. The effects of route of administration on pharmacokinetics of this novel formulation were examined by comparing IV and IM administration, and by evaluating IV versus a dose-adjusted oral comparison. The first comparison was of the new formulation versus the preexisting product when both were administered IV. We found that AUC 0-t and AUC 0-∞ were equivalent between HPbCD-diclofenac and Voltarol after IV administration. The 90% CIs for the geometric mean ratios of HPbCD-diclofenac to Voltarol with respect to AUC 0-t and AUC 0-∞ were well within the accepted 80-125% equivalence window for bioavailability. Thus HPbCD-diclofenac and Voltarol have equivalent bioavailability following IV administration. However, C max was not used in this study to assess equivalence, as the two products differed with respect to rate of administration. Current labeling for Voltarol indicates that the dose be administered over a 30-minute infusion, whereas HPbCD-diclofenac is administered as a rapid IV bolus. As would be expected, C max was higher and T max earlier after HPbCD-diclofenac administration compared with Voltarol. This difference may contribute to the clinical observation of a more rapid onset of analgesia for HPbCD-diclofenac than Voltarol. 21 Despite the higher C max , there was no increased safety risk based on AEs, laboratory tests, and vital signs. These data corroborate a safety metaanalysis of seven single-dose clinical trials 22 that found the incidence of thrombophlebitis observed following IV HPbCD-diclofenac treatment was 1.2% versus 6.5% following IV Voltarol.
When compared for a similar route of administration (IM), AUC 0-t and AUC 0-∞ were equivalent for HPbCD-diclofenac and Voltarol. The 90% CIs for the geometric mean ratios of HPbCD-diclofenac to Voltarol with respect to AUC 0-t and AUC 0-∞ were within the 80-125% equivalence window for bioavailability. Thus HPbCD-diclofenac and Voltarol have equivalent bioavailability following IM administration.
Finally, the route and process of administration can be examined for IM HPbCD-diclofenac and IV Voltarol. Of note is that the pharmacokinetics for the different formulations and routes were similar. The longer t 1/2 following IM administration could be potentially attributed to the flip-flop phenomenon. 23 The flip-flop phenomenon occurs when the process of absorption is the rate-limiting factor in the overall disposition of the drug. The terminal t 1/2 under flipflop condition reflects the rate and extent of absorption and is not a true t 1/2 ; for Voltarol it can be potentially attributed to the formulation characteristics that result in erratic absorption. The lack of flip-flop phenomenon after IM administration of HPbCD-diclofenac compared with Voltarol could be attributed to superior solubilization using HPbCD versus organic solvents. Due to the extended infusion time for the IV Voltarol, C max was lower compared with IM HPbCD-diclofenac. Furthermore, the AUCs were equivalent.
When developing dosing guidelines, it is important to establish proportionality using the pharmacokinetic profile of a product. Two doses of HPbCD-diclofenac (18.75 and 37.5 mg) within the therapeutic range were examined. The pharmacokinetics were dose proportional after IV administration of both doses, indicating that C max and AUC were dose proportional.
The plasma exposures to 18.75 and 37.5 mg IV HPbCD-diclofenac, as measured by AUC uncorrected for dose, bracketed that after administration of oral Cataflam 50 mg, thus offering the option for transition of therapy from IV for inpatients to oral following their release home from the hospital.
Most products to treat acute postoperative pain require multiple doses, raising concern that repeated dosing many lead to increasing exposure secondary to frequent dosing or a reduced metabolic clearance. However, we found that overall, pharmacokinetic parameters after single and multiple doses did not differ; nor was there evidence for accumulation of diclofenac after IV administration of HPbCD-diclofenac every 6 hours for four doses. Diclofenac is predominantly eliminated via hepatic biotransformation to 4-hydroxy-diclofenac as the major metabolite, a reaction catalyzed by the cytochrome P450 enzyme CYP2C9. The lack of accumulation of diclofenac following IV and IM administration reduces the potential for clinical drug-drug interactions with substrates and inhibitors of CYP2C9, a genetically polymorphic enzyme.
IV bolus administration of HPbCD-diclofenac did not raise safety concerns. HPbCD-diclofenac administered as 18.75 and 37.5 mg IV boluses every 6 hours over 24 hours was safe and well tolerated.
Recent clinical trials evaluating the use of rapid bolus injections of HPbCD-diclofenac have indicated a faster onset of analgesia compared with other NSAIDs, which is consistent with the pharmacokinetic profile of this new formulation demonstrated in the current study. In two separate double-blind placebo-controlled trials in patients undergoing third-molar extraction, HPbCD-diclofenac had a faster onset of pain relief than either Voltarol 21 or ketorolac. 24 More recently, HPbCDdiclofenac showed a faster onset of action than IV ketorolac in a population of patients having undergone orthopedic surgery. 25 The lower incidence of thrombophlebitis with HPbCD-diclofenac compared with Voltarol and subsequent reduced need for treatment of adverse events, as well as HPbCD-diclofenac's lack of need for reconstitution, dilution, and buffering prior to each dose compared with Voltarol provide support for potential cost savings with HPbCD-diclofenac compared with Voltarol. 26

Conclusion
HPbCD-diclofenac, a novel formulation of diclofenac with convenience in both IV and IM dosing, demonstrated a higher peak plasma concentration (C max ) and earlier time to peak plasma concentration (T max ), as compared with Voltarol. Overall plasma exposures for HPbCDdiclofenac and Voltarol were equivalent. HPbCDdiclofenac was safe and well tolerated, with few AEs reported. Lack of accumulation and linear pharmacokinetics of HPbCD-diclofenac were also demonstrated, which could potentially provide added benefits in patients with complex analgesic regimens or receiving multimodal analgesia.