Chronic kidney disease (CKD) is a significant public health problem worldwide, impacting national health expenditures and patient morbidity and mortality (United Nations, 2020). The overall prevalence of CKD in the Thai population is remarkably high (17.5% of the population), and the prevalence increases with advancing age (Ingsathit et al., 2009). Thailand is currently in a significant demographic shift, with a rapid increase in the ageing population and the elderly, defined in this study as those aged >60 years (World Health Organization, 2015). Due to physiological changes in the elderly, such as a reduction in renal mass, renal blood flow, glomerular filtration rate, and tubular function, the ability of renal drug elimination can be declined (Hutchison and O'Brien, 2007). For renally excreted drugs, ageing causes changes in the pharmacokinetics of drugs leading to unexpected drug efficacy and safety (Tieu et al., 2016). Age-related alterations in liver function may also change drug metabolism and disposition through alteration of the cytochrome P450 (CYP) enzymes and drug transporter activity (Eldesoky, 2007; Nolin et al., 2008; Klotz, 2009; Yeung et al., 2014). Because of multiple comorbid conditions, polypharmacy is also more common in ageing patients. Physiologic changes in drug metabolism and clearance in the elderly increase the risk of adverse drug reactions and drug-drug interactions in this patient group (Dagli and Sharma, 2014; American Geriatrics Society Beers Criteria®Update Expert Panel, 2019).

Cytochrome P450 3A (CYP3A), the most abundant isozyme in the liver and intestine (Blech et al., 2008), plays an important role in metabolizing several drugs and is commonly implicated in pharmacokinetic drug-drug interactions (Lynch and Price, 2007). For the past decade, increasing numbers of clinically important drug-drug interactions mediated by drug transporters, including intestinal P-glycoprotein (intestinal P-gp), organic anion-transporting polypeptide 1B (OATP1B), and breast cancer resistance protein (BCRP), have been reported (Giacomini et al., 2010).

Clinical studies using probe substrate cocktails are recognized by both the United States Food and Drug Administration (United States FDA) and the European Medicines Agency (EMA) to evaluate the activity of multiple CYP enzymes and drug transporters simultaneously (European Medicines Agency, 2012; Center for Drug Evaluation and Research (CDER) U.S. Department of Health and Human Sciences Food and Drug Administration, 2020). A cocktail comprising a sub-therapeutic dose (microdose) of 5 drugs was recently validated in healthy subjects (Prueksaritanont et al., 2017). The cocktail drugs include 1) midazolam (MDZ), a specific and selective substrate for CYP3A; 2) dabigatran etexilate (DABE), a selective substrate for intestinal P-gp; 3) pitavastatin (PTV), a relatively selective substrate for OATP1B; 4) rosuvastatin (RSV), a substrate of BCRP and OATP1B; and 5) atorvastatin (ATV), a substrate of CYP3A, OATP1B, BCRP, and P-gp. This cocktail was subsequently studied in patients with CKD, and the study results suggested that CKD reduces intestinal P-gp and BCRP activity (Tatosian et al., 2020). However, whether the observed changes were due solely to CKD or as a consequence of advancing age, remains inconclusive.

This study aimed to evaluate age- versus CKD-related changes in the activity of CYP3A and the drug transporters, P-gp, OATP, and BCRP, using a microdose cocktail in the elderly with and without CKD compared to young healthy participants.

This study was a clinical pharmacokinetic study (Clinical trial registration number: TCTR 20180312002 registered on March 07, 2018). The study was approved by the Institutional Review Board of the Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand, and was conducted in accordance with the Declaration of Helsinki and Good Clinical Practice Guidelines. Written informed consent was obtained from all participants before the start of the study. The nomenclature of enzymes and drug transporters conformed to the International Union of Basic and Clinical Pharmacology/British Pharmacological Society (IUPHAR/BPS) Guide to Pharmacology Nomenclature Classification (Alexander et al., 2017).

Our study investigated the activity of CYP3A and drug transporters in three participant groups using a validated microdose cocktail containing five probe substrates, with no inhibitor activity of any enzymes and/or transporters (U.S. Food and Drug Administration, 2020). We found that CYP3A and intestinal P-gp activities are reduced in ageing and CKD, respectively. There is a trend of changes in OATP1B and BCRP activity measured by microdose cocktail probe drugs.

The probe substrates in the drug cocktail should be specific and devoid of any drug-drug interactions to allow precise quantification of enzymes or activity of drug transporters (Fuhr et al., 2019). Our microdose cocktail had specificity (MDZ) and relative selectivity (DABE and PTV). The microdose cocktail has been validated in Caucasians, showing a potential of simultaneous assessment for enzyme and drug transporter activity (Prueksaritanont et al., 2017). Here, we verified the utilization of the cocktail in the Thai population. Our microdose cocktail-RIF drug-drug interaction study in the Thai population showed similar findings as reported in the earlier study (Prueksaritanont et al., 2017).

Reductions in CYP3A activity were observed with similar magnitude in both elderly with or without CKD, shown by the increased AUC and C_max and prolonged T_1/2 of MDZ. It should be noted that oral MDZ as a probe drug represents CYP3A activity in both intestine and liver and that the change in MDZ pharmacokinetics found in this study could be attributed to changes in intestinal and/or hepatic CYP3A activity. Additionally, as MDZ is an intermediate hepatic extraction drug, a physiological reduction of hepatic blood flow anticipated in the elderly might also contribute to the observed reduction in hepatic clearance of MDZ (Dundee et al., 1986; Amrein and Hetzel, 1990). Our finding that CKD was not associated with changes in pharmacokinetic profiles of MDZ was consistent with a previous study where neither AUC nor C_max of MDZ was affected by renal impairment (Tatosian et al., 2020).

Genetic polymorphisms of the CYP3A gene were reported in the Thai population (Mauleekoonphairoj et al., 2020). CYP3A4 polymorphisms are rarely found in Asians, whereas 65–85% of Asians have CYP3A5*3 (Lamba et al., 2002; Hutchison and O'Brien, 2007; Zhou et al., 2017; Mauleekoonphairoj et al., 2020). Although some studies indicated that polymorphisms of CYP3A4/5 might affect the clearance of their substrate such as tacrolimus, the genetic variation in CYP3A enzymes has no significant correlation with in vivo MDZ metabolism and disposition (He et al., 2005; Miao et al., 2009; Singh et al., 2009). Here, CYP3A5*3 genotyping and pharmacokinetic parameters analysis were also assessed (Supplementary Table S3). The results showed that elderly patients with CKD who were with wild-type CYP3A5 (n = 5) had increased AUC and C_max of MDZ (GMR∼2.0–3.0) when compared to healthy young participants with wild-type CYP3A5 (n = 5, Supplementary Table S6). This magnitude of increases in AUC and C_max of MDZ in participants with wild-type CYP3A5 was similar to that in all other participants studied (Table 2). The comparison between wild-type and variant genotypes of CYP3A5 within the group of healthy young participants was also analyzed, and no significant difference was observed (Supplementary Table S7). Therefore, the genetic variations of the CYP3A5 enzyme did not affect the pharmacokinetics of MDZ in our study.

A widely used probe substrate for an intestinal efflux transporter P-gp is DABE. DABE is rapidly converted to DABI after absorption (Chu et al., 2018). The reduction of P-gp activity may decrease the efflux of DABE in the intestine, increasing plasma DABI levels (Chu et al., 2018). In this study, the AUC of DABI was increased in the elderly both with and without CKD compared to healthy young adults. Considering that DABI is 80% excreted into the urine in an unchanged form (Muñoz-Corcuera et al., 2016), CL_R of DABI was measured. The magnitude of DABI’s AUC augmentation was prominent in elderly patients with CKD, corresponding with a more significant CL_R reduction (Table 2). Compared to healthy young participants, the theoretical AUCRs for DABI in healthy elderly and elderly patients with CKD were increased by 1.47 and 2.84 fold, respectively. These data suggested that ageing may not affect intestinal P-gp function since the observed AUCR for DABI of healthy elderly in our study (1.46 fold, Table 2) was closed to the theoretical AUCR (1.47 fold). However, intestinal P-gp is possibly affected by CKD as the observed AUCR for DABI in elderly patients with CKD (4.26 fold, Table 2) is much higher than the theoretical AUCR (2.84 fold). By this, intestinal P-gp function might be impaired, leading to an approximately 1.42 fold increase in AUCR for DABI in elderly patients with CKD. Likewise, a previous study also reported no significant AUC elevation of digoxin, another probe substrate of P-gp at intestinal and systemic levels in the elderly, compared to young adults (Larsen et al., 2007). Reports in a rat model also suggested a downregulation of intestinal P-gp activity in CKD associated with increased drug absorption (Veau et al., 2001; Naud et al., 2007). The latest clinical study in CKD (Tatosian et al., 2020) observed a progressive increase in AUC_0-inf and C_max of DABI with increasing severity of renal impairment up to 4.9 and 1.7 fold, respectively, which is in keeping with our data. Their findings were confirmed with the co-administration of RIF. Here, we also showed that elderly patients with CKD had increased AUC of DABI and the change in AUCRs supports a reduction of intestinal P-gp activity in renal failure. However, unlike the AUC, an increase in C_max of DABI in elderly patients with CKD compared to healthy elderly did not reach statistical significance. Hence, the intestinal P-gp activity in CKD might not be meaningfully affected.

OATP1B is an uptake drug transporter in the liver, where PTV is a probe substrate for this transporter (Prueksaritanont et al., 2014). OATP1B transfers PTV from the portal vein into hepatocytes. PTV is not a significant substrate of human CYP450 but mainly undergoes rapid glucuronidation by UDP-glycosyltransferase 1A1 (UGT1A1), UGT1A2, and UGT2B7 to PTV-lactone. Unlike PTV, PTV lactone is not a substrate for OATP1B (Alagona, 2010). The reduction of OATP1B activity can increase plasma drug levels of an OATP1B substrate (Shitara, 2011). In our study, only OATP1B activity in the elderly patients with CKD declined compared to healthy young participants and healthy elderly, as demonstrated by increased AUC and C_max of PTV (Table 3). This finding is consistent with a previous pharmacokinetic study in patients with non-hemodialysis moderate and severe renal impairment (Morgan et al., 2012). A reduction in OATP1B transporter expression and activity was reported in CKD rat models (Sun et al., 2006; Naud et al., 2008). Additionally, a physiologically-based pharmacokinetic (PBPK) study showed a 60% reduction in hepatic OATP1B transporter activity in patients with severe CKD (Tan et al., 2019). Our findings were different from Tatosian, et al. (Tatosian et al., 2020) as we also observed an increase in AUC and C_max of PTV-lactone, and no differences were found with the AUC_0-last of PTV-lactone/PTV ratio in the elderly patients with CKD (Table 3). The underlying reasons for the discrepancy between the two studies remain unclear.

BCRP is an intestinal efflux transporter capable of reducing systemic exposure of many oral drugs and substrates, including RSV (Huang et al., 2006). Consistent with other studies (Zhang et al., 2006; Keskitalo et al., 2009; Lee et al., 2013), we found that polymorphisms of ABCG2 affected pharmacokinetic profiles of RSV as evidenced by increased AUC and C_max of RSV in participants with ABCG2 variant (Supplementary Table S5). After excluding the participants with this genetic confounder, we observed a significant increase in AUC of RSV in healthy elderly and a more prominent increase of AUC in elderly patients with CKD (Table 4). In addition, the observed AUCRs of RSV in healthy elderly and elderly patients with CKD when compared to healthy young adults were 2.02 and 2.09 fold, respectively (Table 4), which were higher than the theoretical AUCRs of RSV (1.19 fold in healthy elderly and 1.32 fold in elderly patients with CKD). This finding suggested that BCRP transporter activity might be affected by both ageing and CKD. However, a more significant effect on C_max of RSV in healthy elderly and elderly patients with CKD should have occurred if intestinal BCRP transporter activity was reduced (Table 4). Hence, an association of ageing and CKD with BCRP represented by changes in pharmacokinetic profiles of RSV still cannot be entirely determined in our study. Other studies reported no association between ageing and RSV exposure (Martin et al., 2002; Davidson, 2007; Kim et al., 2019). Our data is consistent with Tatosian, et al. (Tatosian et al., 2021) where RSV pharmacokinetics alteration was not associated with CKD. In addition, RSV is another substrate of hepatic OATP1B transporter but is less sensitive and less selective than PTV (Prueksaritanont et al., 2014). Therefore, when considering the pharmacokinetic changes of RSV with PTV results as discussed above, a definitive effect of CKD on hepatic OATP1B transporter activity is still yet to be concluded.

ATV is mainly metabolized by CYP3A and transported by hepatic OATP1B (major), intestinal BCRP, and possibly, to a minor extent, intestinal P-gp (Kellick et al., 2014; Kellick, 2017). Here, we found a significant increase in AUC and C_max. of ATV associated with ageing and CKD. The pharmacokinetic changes of ATV were in support of the pharmacokinetic changes seen with MDZ. CYP3A4 converts ATV into its major metabolite, 4-OH-ATV (Hoffart et al., 2012). A reduction in the AUC_0-last ratio for 4-OH-ATV/ATV and the association of ageing and changes in ATV plasma levels observed in this study (Table 4) was in line with findings observed with MDZ on CYP3A. The effect of ageing on pharmacokinetic profiles of ATV (Table 4) is similar to a previous report of reduced CYP3A enzyme activity in ageing (Gibson et al., 1996).

On the other hand, the effects of CKD on pharmacokinetic profiles of ATV and its metabolite in our study were inconclusive. In the recent study by Tatosian, et, al., increases in AUC and C_max of ATV were associated with the severity of renal impairment (Tatosian et al., 2020). Therefore, P-gp and BCRP may be additional factors involved in alterations of ATV pharmacokinetics in elderly patients with CKD. Unfortunately, BCRP activity was inconclusive in our study due to the results of RSV as discussed above.

It is important to note that although age- and CKD-associated changes in OATP1B and BCRP drug transporter activity could not be definitively concluded, our results suggest a trend in their changes. Furthermore, these findings are potentially beneficial to clinicians as most older people are generally taking multiple drug treatments and are therefore at risk of drug-drug interactions.

It is also worth noting that our study did not directly measure unbound plasma concentrations due to the assay sensitivity limitation. However, plasma protein binding was indirectly determined by ex-vivo experiments using pooled predose plasma samples from each study group, spiked with relevant plasma concentrations of substrates (5–30 ng/ml). No significant difference of unbound fraction amongst the three studied groups was found for all drugs used in this cocktail (data not shown). These data suggested a negligible effect of plasma protein binding in our study, and consequently, similar conclusions would be expected with unbound drug pharmacokinetics. Currently, efforts are underway to develop a PBPK model using data from this study to delineate the effect of ageing and CKD on the underlying mechanisms related to drug metabolizing enzymes and drug transporters in elderly patients with CKD.

This study has limitations. In general, gender could affect the pharmacokinetics of drugs because of the differences in their physiology. For instance, women usually have lower body weight, lower muscle mass, and less intestinal enzymatic activity than men (Whitley and Lindsey, 2009). Although participants in this study were not intentionally matched by gender, which is one of our limitations, the significant difference in pharmacokinetic parameters was not found after gender adjustment in the multivariable analysis. Total body weight determines body composition (lean weight: adipose weight), affecting the volume distribution of drugs (Muñoz-Corcuera et al., 2016). A significant difference in body weight was seen at baseline between healthy young participants and elderly patients with CKD, but no effect of body weight on pharmacokinetic of drugs was found in the multivariate analysis. We have also carried out a subgroup analysis stratified by severity of CKD, and there was no significant difference shown between moderate and severe renal impairment (data not shown). No potential pharmacokinetic drug-drug interaction was observed between concomitant drugs and probe drugs. However, the screening for pharmacodynamics drug-drug interaction has not been performed.

Additionally, further studies with larger cohort sizes that stratify based on CKD severity, together with patients on dialysis and RIF co-administration in all study populations, may be warranted. Population pharmacokinetic studies would be helpful to elucidate if and how other factors affect the pharmacokinetics of the probe drugs profiles. Finally, PBPK studies using data from this study would provide more information to inform predictions of changes in drug metabolizing enzymes and drug transporter activity in elderly patients with CKD.

In conclusion, the microdose cocktail approach provides a valuable and safe screening tool to determine pharmacokinetic parameter alterations showing reduced CYP3A activity in the elderly. Intestinal P-gp activity might be reduced in CKD, but as the results in AUCR and C_max were conflicting, this has to be confirmed in further studies. Health care providers should be aware of the potential consequences of administering CYP3A and intestinal P-gp substrates or inhibitors to these target populations.