Amoxicillin capsules 250 mg (Towa, Osaka, Japan), Cefcapene pivoxil hydrochloride tablets 100 mg (Sawai, Osaka, Japan), and Magmitt tablets 330 mg (Nihon Shinyaku, Kyoto, Japan) as an MgO containing formulation were used. These formulations of amoxicillin, cefcapene pivoxil hydrochloride, and MgO are hereafter referred to as AMX, CFPN, and MG, respectively. Amoxicillin trihydrate (100%) and cefcapene pivoxil hydrochloride monohydrate (98.5%) were purchased from Tokyo Chemical Industry (Tokyo, Japan). Distilled water, acetonitrile, methanol, and trifluoroacetic acid (TFA) were purchased from Fujifilm Wako Chemicals (Osaka, Japan). These solvents were of high-performance liquid chromatography (HPLC) grade. Weakly acidic cation exchange resin AMBERLITE® (IRC76) was purchased from ORGANO (Tokyo, Japan). Membrane filters, hydrophobic DISMIC® − 03 JP (0.50 μm) and hydrophilic DISMIC® − 03 CP (0.45 μm), were purchased from Toyo Roshi Kaisha (Tokyo, Japan). Deuterated dimethyl sulfoxide (DMSO-d6) was purchased from Kanto Chemical (Tokyo, Japan).

The simple suspensions were basically prepared according to the previously described method [6, 17]. AMX (one capsule) or CFPN (one tablet) was soaked with or without MG (one tablet) in a 50-mL plastic centrifuge tube containing 20 mL of warm distilled water (55 °C). The centrifuge tube was allowed to stand at room temperature (ca. 25 °C) for 10 min or 5 h. The centrifuge tube was wrapped with aluminum foil for shading during standing. The mixture was then mixed with inversion 15 times and vortexed for one min to prepare a suspension, which was treated and analyzed as shown below. In separate experiments, the co-suspension of AMX and MG was prepared after 10 min soakage, and the aqueous supernatant after centrifugation was directly analyzed by HPLC and subjected to pH measurements using InLab® Routine Pro (Mettler Toledo, Ohio, USA). The suspension preparation was performed three times for each antibiotic in all of the experiments.

The suspension of AMX with or without MG in a centrifuge tube was completely transferred into a beaker using 180 mL of distilled water. After sonication for 2 min, the suspension was centrifuged (13000 × g) for 5 min. The aqueous supernatant obtained was passed through a column packed with weakly acidic cation exchange resin (ca. 20 g) for neutralization. The effluents from the column, which were confirmed to be neutral using a pH test paper, totaled 500 mL with distilled water. The solution was diluted by a factor of 10 with distilled water for the HPLC analyses. The suspension of CFPN with or without MG in a centrifuge tube was completely transferred into a beaker using 80 mL of methanol. The suspension was sonicated, centrifuged, and neutralized by the same procedures as for AMX. The effluents from the column totaled 200 mL with methanol. The solution was diluted by a factor of 50 with methanol for the HPLC analyses. These treatments for the suspensions were rapidly (within 30 min) completed.

A Shimadzu HPLC system with a photodiode array detector (Kyoto, Japan) was used [24]. A reversed-phase Atlantis T3 column (250 × 4.6 mm i.d., 5 μm; Waters, Milford, USA) fitted with a guard column (20 × 4.6 mm i.d., 5 μm) was used. The composition of the mobile phases was determined, modifying those in the literature [25]. The composition of the mobile phase was 0.1% TFA/acetonitrile (84/16) for AMX and 0.1% TFA/acetonitrile (62/38) for CEPN, respectively. The wavelengths used to draw the chromatograms of AMX and CFPN were 230 and 262 nm, respectively. The analyses were performed at a flow rate of 1.0 mL/min at 25 °C. The final solutions (10 µL) prepared above were injected after being filtered using the membrane filter. The concentrations of the samples were measured using the absolute calibration method based on the peak areas. The amounts of medicinal ingredients in the simple suspensions were determined from the concentrations and the percentages of the measured values to the labeled amounts of the formulations (recovery rates) were calculated. An Atlantis T3 column (250 × 10 mm i.d., 5 μm; Waters, Milford, USA) equipped with a guard column (10 × 10 mm i.d., 5 μm) was used as a semi-preparative column for the isolation of the degradation products, where the composition of the mobile phase was the same as mentioned above and the flow rate was maintained at 4.8 mL/min.

Standard stock solutions of amoxicillin (616 µg/mL) and cefcapene pivoxil hydrochloride (656 µg/mL) were prepared in distilled water and methanol, respectively. The solutions for the validation were prepared from stock solutions. The calibration curves were constructed by triplicate analysis of samples with five concentrations whose ranges were wide enough for the quantitation of the medicinal ingredients. The calibration curves exhibited good linearity with the coefficients of determination (r2) being greater than 0.999 (Supplemental Table 1). Quality control samples were prepared at three concentrations (middle, high, low) within the ranges of the calibration curves. The precision was determined by the intra- and inter-day variations through repeated analyses (N = 5) of the quality control samples on three different days. The relative standard deviations of the measured values were less than 2.0% for the intra- and inter-day variations (Supplemental Table 2). Limit of detection (LOD) and limit of quantitation (LOQ) were determined based on the following equations, where S and S.D. are the slope and residual standard deviation of regression line at low concentrations near to LOD and LOQ, respectively. The LOD and LOQ of amoxicillin were 25.8 ng/mL and 78.3 ng/mL, respectively, and those of cefcapene pivoxil were 3.3 ng/mL and 10.0 ng/mL, respectively (Supplemental Table 1). All of the validation parameters were within the acceptance ranges.

$$\:LOD\hspace=\hspace3.3\:\times\:\:S.D.\:/\:S$$

$$\:LOQ\hspace=\hspace10\:\times\:\:S.D.\:/\:S$$

The aqueous supernatant of the co-suspension of AMX and MG was concentrated using a rotary evaporator at room temperature. The residue was dissolved with distilled water (4 mL) and portions (300–500 µL) of the solution were injected into the HPLC apparatus with the semi-preparative column. The eluates corresponding to the peaks of the degradation products were collected and directly analyzed with a micro TOF-Q spectrometer (Bruker Daltonics, Kanagawa, Japan) in order to obtain the high-resolution electrospray ionization mass spectra (ESI-MS) and tandem mass spectra (MS/MS). The eluates were also evaporated to dryness using a rotary evaporator at room temperature. The residue was dissolved in DMSO-d6, and the solution was analyzed by nuclear magnetic resonance (NMR) spectroscopy with an AVANCE 500 NMR spectrometer (11.75 T, Bruker Japan, Kanagawa, Japan) using 5-mm sample tubes at 300 K.

All data are expressed as the mean ± standard deviation of three experiments. All statistical analyses were performed using one-way ANOVA followed by Tukey’s test to determine significance. The threshold for assessing significance was p < 0.05.