Methods

Dissolution Media Preparation
Paediatric biorelevant media representative of infants were freshly prepared for each experiment, as described by Maharaj et al. (20). Infant fasted-state simulated gastric fluid (Pi-FaSSGF, pH 1.6) and fasted-state simulated intestinal fluid (FaSSIF-V2, pH 6.5) or infant fed-state simulated intestinal fluid (Pi-FeSSIF, pH 5.8) were used. Both fasted and fed intestinal state were simulated since the prandial state of the infant patients in the in vivo studies was not reported, and in order to investigate if medicine co-administration with a vehicle would induce a food effect in the infant. Double concentrated simulated intestinal fluids were prepared for the two-stage dissolution studies performed (“Biorelevant In Vitro Dissolution Studies”).

Sample Preparation
Formula was prepared as per manufacturer’s instructions: 1 scoop of powder (approximately 4.5 g) was added to 30 mL of boiled cooled water. Two formulations were tested: Singulair® granules (4 mg) and Actavis® chewable tablets (5 mg) which were crushed prior to mixing (following reported practices (7)). For the direct administration scenario, formulations were tested in the simulated GI fluids without prior mixing with a vehicle. For the mixing with vehicle scenario, each sample was prepared by addition of the formulation to milk (25 mL; as previously investigated (12)), applesauce (15 g) or formula (5 mL), followed by mixing with a stainless-steel spatula. Mixing with formula was performed only for the Singulair® granules to mimic the in vivo study dosing scenario (17). The preparation technique procedure was time-controlled (less than 2 min were spent between preparation and dosing of the mixture), and the mixing was performed in exactly 30 s.

Biorelevant In Vitro Dissolution Studies
Dissolution studies were performed with a mini-paddle apparatus (Agilent Technologies 708-DS apparatus configured with TruAlign 200 mL vessels and electropolished stainless-steel mini-paddles; Agilent, USA). Experiments were conducted at 37°C, and agitation rate was set to 50 revolutions per minute (rpm). A two-stage approach was followed: gastric conditions were simulated for 1 h (Pi-FaSSGF pH 1.6; total volume with sample: 100 mL), followed by intestinal simulated conditions (FaSSIF-V2 pH 6.5 or Pi-FeSSIF pH 5.8; final volume: 200 mL), for 3 h. Sample collection took place at 5, 15, 30, 45, 60, 75, 90, 120, 180 and 240 min. Samples of 2 mL were withdrawn (with volume replacement with the corresponding media), using a 2-mL glass syringe (Fortuna Optima® fitted with a stainless tubing) through a cannula fitted with a full flow filter (10 μm). All experiments were performed without direct light exposure to avoid photodegradation of montelukast (21). After collection, samples were filtered through a GF/D filter (2.7 μm), treated, placed into amber HPLC vials and injected into the HPLC. Treatment was as follows: 1000 μL of acetonitrile was added to 500 μL of the filtered sample, the mixture was vortexed (HTZ, UK) for 1 min and centrifuged (8000 rpm, 15 min, 4°C) (Beckman Coulter J2-MC centrifuge, UK) and the supernatant was filtered through a RC filter (0.45 μm). The pH of the media was measured at the end of each experiment to ensure the pH shift had been successful and that the vehicle did not alter the media pH.
The effect of different administration scenarios on drug dissolution was investigated by varying the mode of the introduction of the formulation in the simulated gastric fluid in the dissolution vessel: direct administration of the formulation or administration of the formulation after mixing with drinks (formula and milk) or soft food (applesauce). These vehicles were selected based on their impact on the dissolution of montelukast (12) and/or to mimic the in vivo studies performed in infants (12,16,17,19). The composition and physicochemical properties of these vehicles, including pH, buffer capacity and viscosity, have been recently published and discussed (13).
All experiments were performed in triplicate. Fresh calibration curves (concentration range: 0.5–60 μg/mL) were prepared in the corresponding media, by appropriate dilution of a 1000 μg/mL stock solution of montelukast analytical standard in methanol. Results were expressed as mean percentage (%) drug dissolved ± standard deviation (S.D.) at the given sampling time.

Chromatographic Conditions for Drug Analysis
The chromatographic method used for quantification of montelukast was a modification of the method by Raju et al. (22). Drug quantification was performed with HPLC with ultraviolet (UV) detection (Agilent HPLC system 1100/1200 series; Agilent, USA), using a C18 column (RP Agilent Eclipse XDB, 250 mm × 4.6 mm, 5 μm particle size), and ammonium acetate buffer pH 5.5 (A) and methanol (B) as mobile phase, delivered on a linear gradient. The selected gradient started with 10% of solvent B, which was increased to 50% within 2 min, and 90% within 4 min; at 11.30 min, the initial conditions of analysis were re-established. Injection volume was 100 μL, flow rate was 1 mL min−1, run time was 12.30 min, detection wavelength was 284 nm and column temperature was 20°C.

Data Analysis

In Vitro Data Analysis
The linear trapezoidal method was used to calculate the area under the curve of each in vitro % drug dissolved over 4-h profile (AUC0–4 h in vitro). One-way analysis of variance (ANOVA) with a post hoc Tukey honest significant difference (HSD) test was conducted to investigate differences in the AUC0–4 h in vitro calculated from the dissolution studies, after direct administration of formulation and after mixing the formulation with the different vehicles (p < 0.05 noting significance level). T test analysis was used to compare experimental results under fasted-state gastric conditions, followed by fasted- or fed-state intestinal conditions (represented as Pi-FaSSGF/FASSIF-V2 and Pi-FaSSGF/Pi-FeSSIF, respectively) (significance p < 0.05). Analyses were performed with GraphPad Prism® v.7 software (USA).

In Vivo Data Analysis
Published data of plasma concentration profiles of Singulair® granules (4 mg) co-administered with formula or applesauce to different infant subgroups (formula: 1 to 3 months; applesauce: 3 to 6 months and 6 to 24 months) were digitalised with WebPlotDigitizer® v4.1 software (USA) (16,17,19,23).
The corresponding in vivo drug absorption profiles were obtained after deconvolution of the oral data using the Wagner-Nelson equation (Eq. 1) (Excel, Microsoft®) (24).1 %absorbed=AtA∞×100=At+k∫τ=0tAτdτk∫τ=0∞Aτdτ×100
where A(t) is the amount of drug in the system at time t and k is the first-order elimination rate constant (24). The elimination rate constant was obtained from the slope of the terminal logarithmic concentrations of the in vivo montelukast oral data.
The linear trapezoidal method was used to calculate the area under the curve of each in vivo % drug absorbed over 4-h profile (AUC0–4 h in vivo).

In Vitro/In Vivo Relationship
An in vitro-in vivo relationship for Singulair® granules (4 mg) was investigated by comparing the in vitro dissolution (AUC0–4 h in vitro) and the in vivo absorption data (AUC0–4 h in vivo). Average differences between the obtained AUC0–4 h in vitro and the AUC0–4 h in vivo of the different subgroups were expressed as a ratio % (AUC0–4 h in vitro/AUC0–4 h in vivo × 100). For evaluation of the results, ratios lower than 100% indicate that AUC0–4 h in vitro was lower than the AUC0–4 h in vivo and higher values represent the opposite. To denote relevant discrepancies between the AUC0–4 h in vitro and AUC0–4 h in vivo, reference points corresponding to ratios of 80 and 125% were used. Mean ratios falling inside this reference range represent an in vitro-in vivo relationship, with in vitro dissolution results being predictive of the in vivo drug performance.