Healthy volunteers enrolled in three dexketoprofen bioequivalence clinical trials (EUDRA-CT: 2011-002728-42, 2011-003800-19 and 2019-002223-15. Ethics Committee approval numbers 1815, 1861 and 3818 respectively) conducted in the Clinical Trial Unit of Hospital Universitario de La Princesa (UECHUP), Madrid, (Spain) participated in the pharmacogenetic study. Hospital’s Research Ethics Committee and the Spanish Drugs Agency (AEMPS) revised and approved the study protocols; research was conducted complying with European and Spanish legislation and under Good Clinical Practice Guidelines; the Declaration of Helsinki (revised) was fully endorsed. All healthy volunteers (n = 96) signed their informed consent for their participation in the clinical trials, among which, 85 signed a specific informed consent for the pharmacogenetic study.

Inclusion criteria for the participation in the clinical trial were as follows: males or females aged 18 to 55, free from psychic conditions, with normal physical examination and medical records, without abnormalities in hematology, biochemistry, serology and urine tests. Exclusion criteria comprised any organic or psychic condition, having received prescribed pharmacological treatment in the last 15 days or any type of medication in the 48 h prior to study medication intake, body mass index (BMI) out of the 18.5–30 kg/m², history of drug sensitivity, suspected consumption of controlled substances, smoking, alcohol poisoning in the last week or daily alcohol consumption, having donated blood in the last month and pregnant or breastfeeding women.

The three clinical trials were phase I, oral single-dose, open-label, crossover and randomized, organized in 2 sequences and 2 periods evaluating bioequivalence between dexketoprofen 25 mg test formulations and Enantyum 25 mg (Menarini, Spain). Volunteers were hospitalized either the night before or 1 h before drug intake in the morning and remained hospitalized 12 h after dosing. Between drug intake (0 h), and 12 h after dosing, 17 or 18 blood samples were extracted for pharmacokinetic profiling, which was blinded. Formulations were administered by oral route with 240 ml of water under fasting conditions. Blood was collected in EDTA-K2 tubes; plasma was separated immediately after each blood extraction by centrifugation and samples were frozen until their shipment to the analytical laboratory. The method involved a liquid-liquid extraction procedure with tert-butyl methyl ether; dexketoprofen and internal standard were measured by reversed phase high performance liquid chromatography coupled to a tandem mass spectrometer (LC/MS/MS). Method complied with the European Medicines Agency (EMA) requirements for bioequivalence demonstration.

The CERTARA Phoenix WinNonlin software, most recent version at the moment of clinical trial performance (Certara United States, Princeton, NJ, United States) was used for the non-compartmental analysis of pharmacokinetic parameters. The area under the curve (AUC) from pre-dose to the last time-point (12 h) was calculated following the trapezoidal rule (AUC_t). By means of linear regression of the log-linear part of the concentration-time curve, the terminal rate constant (k_e) was calculated. The AUC from t = 12 to infinite was calculated as C_t/k_e (AUC_t-∞). AUC_∞ was calculated as: AUC_t + AUC_t-∞. By dividing the dose (D) by AUC_∞ and weight, drug clearance adjusted for bioavailability (Cl/F) was calculated. By dividing Cl/F by k_e, the volume of distribution adjusted for bioavailability (Vd/F) was calculated. Half-life (t_1/2) was estimated as–ln 2/k_e. The maximum concentration (C_max) and the time to reach the C_max (t_max) were directly obtained from concentration-time curves.

Abnormalities in analytical values, blood parameters, physical examinations or any other clinically relevant event were noted for safety assessment. A 12-lead electrocardiogram (ECG) was carried out at predose and 1 or 2 h after drug intake. Simultaneously, vital signs were monitored (systolic and diastolic blood pressure, heart rate and tympanic or armpit temperature). Volunteers were monitored for the occurrence of adverse events (AEs) by an open question, and, in addition, those reported spontaneously by themselves were noted. The Spanish Pharmacovigilance System (Aguirre and García, 2016) algorithm or Karch and Lasagna criteria (Karch and Lasagna, 1977) were used for causality determination. Only those AEs with a definite, probable or possible causality were defined as adverse drug reactions (ADRs) and therefore considered for analysis.

A MagNa Pure automatic DNA extractor (Roche Applied Science, United Status) was used for the extraction of DNA form peripheral blood collected in an EDTA-K2 tube. 46 SNPs in 14 genes were analyzed: CYP1A2*1C (rs2069514), *1F (rs762551), *1B (rs2470890), CYP2A6*9 (rs28399433), CYP2B6*9 (rs3745274), *5 (rs3211371), rs4803419, rs2279345, rs2279343, CYP2C8*2 (rs11572103), *3 (rs10509681), *4 (rs1058930), CYP2C9*2 (rs1799853), *3 (rs1057910), CYP2C19*2 (rs4244285), *3 (rs4986893), *4 (rs28399504), *17 (rs12248560), CYP2D6*3 (rs35742686), *4 (rs3892097), *6 (rs5030655), *7 (rs5030867), *8 (rs5030865), *9 (rs5030656), *10 (rs1065852), *14 (rs5030865), *17 (rs28371706), *41 (rs28371725), CYP3A4 *22 (rs35599367), rs55785340, rs4646438, CYP3A5*3 (rs776746), *6 (rs10264272), ABCB1 C3435T (rs1045642), G2677 T/A (rs2032582), C1236T (rs1128503), ABCC2 c.1247G>A (rs2273697), rs717620, SLCO1B1*1B (rs2306283), *5 (rs4149056), c.-910G > A (rs4149015), rs11045879, SLC22A1*2 (rs72552763), *3 (rs12208357), *5 (rs34059508), and UGT1A1*80 (rs887829). A CYP2D6 copy number variation assay (CNV) was performed following the methodology previously reported (Zubiaur et al., 2020). For the genotyping, a QuantStudio 12 k flex was used with two thermal blocks: the OpenArray for the genotyping of variants and the 96-well fast thermal-block for CNV determination (Zubiaur et al., 2020).

CYP2D6 (*3, *4, *5, *6, *7, *8, *9, *10, *14, *17, *41 and the gene copy number), CYP2C19 (*2, *3, *4, *17), SLCO1B1 (*1B, *5), CYP2B6 (*5 and *9) and CYP2C9 (*2, *3) variants were used to infer the enzymatic phenotype based on CPIC guidelines (Scott et al., 2013; Caudle et al., 2014; Ramsey et al., 2014; Caudle et al., 2019; Desta et al., 2019). For phenotype inference, those variants lacking a result were considered “not mutated.” Missing data due to genotyping errors (e.g., absence of amplification) were not analyzed. CYP1A2 (*1C, *1F and *1B) variants were used to infer the activity score and phenotype as described in previous publications (Saiz-Rodríguez et al., 2019; Zubiaur et al., 2019). SLC22A1 and ABCB1 variants were merged into haplotypes: the absence of any variant was assigned the wild-type haplotype, the presence of one variant was assigned the heterozygous haplotype and the presence of two or more variants was assigned the mutant haplotype.

The SPSS software was used for statistical analysis (version 19.0, SPSS Inc., Chicago, United States). AUC_∞ and C_max were divided by the dose/weight ratio (AUC/DW, C_max/DW) to correct the differences in weight between sexes or races which can produce pharmacokinetic variability. For distribution normalization, pharmacokinetic parameters were logarithmically transformed. Firstly, a univariate analysis was performed. The means of pharmacokinetic parameters or the incidence of adverse drug reactions (ADRs) were compared based on categorical variables (e.g., sex, race, haplotypes, phenotypes). For mean comparison, a T test (variables with two categories) or an ANOVA test followed by a Bonferroni post-hoc (variables with three or more categories) were used. For the comparison of the incidence of ADRs according to categorical variables, a Chi² test was used. Afterward, each pharmacokinetic parameter or ADR were individually analyzed with a multivariate analysis. By linear or logistic regression, pharmacokinetic parameters or ADRs were explored, respectively. As independent variables, any variable with p < 0.05 in the univariate analysis was explored (sex, race, ABCB1 C1236T, C3435T, G2677A/T, CYP1A2*1B, CYP2B6, and CYP2D6 phenotypes); in addition, pharmacokinetic parameters were introduced as independent variables in the logistic regression for the evaluation of ADR occurrence.

Study population was composed by 42 women (49.41%) and 43 men (50.59%). Men’s height, weight and body mass index (BMI) were significantly superior than that of women (p < 0.0001, p < 0.0001, and p = 0.030, respectively) (Table 1). Caucasian was the most prevalent race (82%) compared to Latin-Americans (15%), one Black and one Asian (both males). Demographics also differed significantly according to races (Table 1). Caucasians were younger than Latin-Americans (p < 0.003).

Dexketoprofen mean AUC_∞ was 4,031.88 ± 819.35 ng*h/mL (4,266.90 ± 832.59 ng*h/mL for females and 3,802.33 ± 746.18 ng*h/mL for males, p = 0.008) and mean C_max was 2,745.20 ± 690.81 ng/ml (3,013.30 ± 726.25 ng/ml for females and 2,483.34 ± 545.43 ng/ml for males, p < 0.001). After DW correction, women showed lower AUC/DW compared to men (ANOVA p = 0.007, unstandardized beta coefficient = −0.124, p = 0.011, model R ² = 0.151) and the differences in Cmax/DW disappeared (Table 2). The Black and Asian subjects showed higher AUC/DW, Cmax/DW and half-life and lower Cl/F and Vd/F, but statistical analysis was not possible. Latin-Americans showed lower AUC/DW and t_1/2 than Caucasians (Table 2); however, these associations disappeared after the multivariate analysis. Lastly, women exhibited higher Vd/F and Cl/F than men (unstandardized beta coefficients = 0.114 and 0.125; p = 0.012 and 0.010; model R ² = 0.097 and 0.155, respectively).

Some genetic polymorphisms were associated with dexketoprofen pharmacokinetic variability (Table 3). The CYP1A2*1B/*1B genotype was related to higher Cmax/DW (ANOVA p = 0.035) and to lower Tmax (ANOVA p = 0.022, unstandardized beta coefficient = −0.177, p = 0.015, R ² = 0.163) compared to the *1/*1 genotype (Table 3) and to higher Tmax compared to the *1/*1B genotype (p = 0.050). CYP2B6 intermediate (IM) and poor (PM) metabolizers were related to higher Cmax/DW compared to normal (NM) and rapid (RM) metabolizers (ANOVA p = 0.023 and 0.037, respectively). These associations were lost after multivariate analysis. Moreover, CYP2B6 intermediate and poor metabolizers were related to lower Tmax compared to RMs (ANOVA p = 0.024, unstandardized beta coefficient = −0.130, p = 0.043, R ² = 0.163). CYP2D6 IMs/PMs were related to higher AUC_∞/DW compared to NMs (p = 0.017). This association was not confirmed by multivariate analysis.

ABCB1 C1236T/T carriers showed lower AUC/DW (unstandardized beta coefficient = −0.072, p = 0.047, model R ² = 0.151), and Cmax (ANOVA p = 0.006 and 0.027, respectively) compared to C/T and C/C carriers; C/C carriers showed lower Cl/F compared to C/T + T/T (unstandardized beta coefficient = −0.072, p = 0.045, model R ² = 0.155). Moreover, ABCB1 C3435T T/T carriers showed lower Cmax/DW compared to C/C (ANOVA p = 0.021). This association was not confirmed by multivariate analysis. Finally, G2677T/A TT/AA carriers were related to lower Cmax/DW compared to G/G and GT/GA (ANOVA p = 0.008 and 0.009, respectively, unstandardized beta coefficient = −0.164, p < 0.001, R² = 0.188) (Table 3). The remaining genes (i.e., CYP2A6, CYP2C9, CYP2C8, CYP2C19, CYP3A4, CYP3A5, UGT1A1, SLC22A1, SLCO1B1, and ABCC2 had no effect on dexketoprofen pharmacokinetic variability, as shown in Table 4.

All the AEs reported during the three clinical trials were unrelated to dexketoprofen administration. No ADR was reported during any of the three clinical trials. No clinically relevant alteration of vital signs or ECG was observed.

First and foremost, the clinical trial study design did not allow to collect data on: 1) dexketoprofen effectiveness, 2) long-term pharmacokinetics and tolerability. It is possible that UGT2B7 nor UGT2B15 genes are involve in the elimination of dexketoprofen but they could not be evaluated because they were not included in the array used at our center. We were also unable to analyze the influence of Black and Asian race on dexketoprofen pharmacokinetics because there was only one subject of each race. Furthermore, it is important that these results are interpreted with caution given the small sample size. In contrast, bioequivalence clinical trials offer a controlled setting for the evaluation of pharmacokinetic variability based on genetic polymorphism or demographics as confounding factors are avoided. These results can be considered preliminary and further studies are needed to confirm the hypotheses that were raised in the study.