Clinical Pharmacokinetics and Pharmacodynamics of Statins and Metformin in Obese Patients Pre- and Post-Gastric Bypass Surgery

Obesity is a major health concern, with 39.8% of adults in the US being obese in 2015-2016. Bariatric surgery is the recommended treatment strategy for obese patients with a body mass index (BMI) of ≥ 40 kg/m2 or ≥ 35 kg/m2 with comorbidities. Obese individuals are at a higher risk of diseases of diabetes, coronary heart diseases, hyperlipidemia, respiratory problems, sleep apnea, and others. As a result, obese individuals who undergo bariatric surgery are usually on multiple medications, and most of the medications may still be needed after the surgery. Roux-en-Y gastric bypass surgery (RYGB) is one type of bariatric surgery, that results in a reduction in the stomach pouch size and the bypass of the first section of small intestine. There is scarce research regarding to the effects of RYGB on the pharmacokinetics (PK) and pharmacodynamics (PD) of medications that are widely prescribed to severely obese patients. With the popularity of RYGB in the US, an understanding of its effects on the PK, PD and PK/PD correlations of the commonly prescribed medications in this patient population is timely and crucial for rational and proper dosing modification of the medications post-RYGB. Hyperlipidemia and type 2 diabetes are two of the most prevalent comorbidities in obese patients, with statins and metformin being the first medication choices for hyperlipidemia and type 2 diabetes, respectively. In this project, the PK and PD of 3 statins and metformin were studied longitudinally in each subject through the course of 12 months post-surgery. The objective of our study was to investigate the impacts of RYGB on the PK/PD of three of the most widely used statins which are simvastatin, atorvastatin, and their active metabolites as well as rosuvastatin, and metformin. Our study is innovative because it is the first longitudinal study for individual subjects with the same patient’s pre-surgery conditions as their own controls. In addition, our study monitored the active metabolites and parent drug of statins and investigated the effect of RYGB on statin metabolisms. We hypothesize that RYGB will decrease the absorption of simvastatin, atorvastatin, and rosuvastatin, as well as metformin and affect the PK/PD of these medications and their active metabolites, if any, in patients post-surgery. To achieve the objective of our study, three specific aims were formulated:

1. a) To develop and validate a simultaneous LC-MS/MS assay for statins, simvastatin, atorvastatin, and their active metabolites, as well as rosuvastatin (no metabolites) in lipemic plasma samples from obese patients to assess the concentrations of the drugs and metabolites before and after RYGB. b) To develop and validate an LC-MS/MS assay for metformin in lipemic plasma samples from obese patients to assess the concentrations of metformin before and after RYGB.
2. To monitor and characterize the effects of RYGB on the PK/PD of simvastatin, atorvastatin, and their active metabolites, as well as rosuvastatin for the time course of 12 months post RYGB.
3. To monitor and characterize the effects of RYGB on the PK/PD of metformin for the time course of 12 months post RYGB. First, we have successfully accomplished the objective of aim 1, where we have developed and validated two LC-MS/MS methods. The first method simultaneously quantifies simvastatin, atorvastatin, along with their active metabolites, and rosuvastatin. To our knowledge, this is the first report of simultaneous quantification of simvastatin, its active metabolite (simvastatin acid), atorvastatin, its two active metabolites (2-hydroxy atorvastatin and 4-hydroxy atorvastatin), and rosuvastatin in human plasma samples. The second LC-MS/MS method was utilized to quantify metformin concentrations in human plasma samples. Both methods have been validated in plasma with low (<300 mg/dl) and high (>300 mg/dl) triglyceride levels. The methods showed no interference with the quantification of all analytes, nor matrix effect from the triglycerides, and suitable for drug monitoring for statins and metformin, respectively, in lipemic plasma samples. Overall, two specific, accurate, robust and reliable methods were developed and validated for the quantification of three statins along with their active metabolites and metformin at the LLOQ of 0.25 ng/ml in plasma with either low or high triglyceride levels. Second, we have successively investigated the impacts of RYGB on statins and their active metabolites. Our study is the first report on the longitudinal effects of RYGB on the PK of simvastatin (n=9), atorvastatin (n=5), along with their active metabolites, and rosuvastatin (n=12) in the same subject pre- and post-surgery. The study showed a trend of significant decrease in the individual and mean plasma concentrations on the same dose per unit body weight [(nM)/(mg/kg)] of atorvastatin from 38.81 ± 2.36 at baseline to 16.30 ± 3.41 and 9.75 ± 2.52 (nM)/(mg/kg) at 3- and 6-months follow -up visits, respectively. The same trend of significant decrease was observed for the two hydroxy active metabolites of atorvastatin, 2-hydroxy and 4-hydroxy atorvastatin, with mean concentrations of 44.82 ± 6.11 and 18.75 ± 4.18 at baseline, respectively. The mean concentrations of 2-hydroxy and 4-hydroxy atorvastatin decreased to 14.62 ± 2.40 and 5.12 ± 1.20 (nM)/(mg/kg), respectively, by 3 months post-RYGB and stabilize at 10.26 ± 2.95 and 2.94 ± 0.80 (nM)/(mg/kg), respectively, at 6 months post-RYGB. Rosuvastatin individual and mean plasma concentrations on the same dose per unit body weight [(nM)/(mg/kg)] also showed the trend of decrease from 213.07 ± 22.87 at baseline to 122.56 ± 9.67 and 83.28 ± 6.39 (nM)/(mg/kg) at 3 and 6 months post-RYGB, respectively. For simvastatin and its active metabolite, simvastatin acid, the trend was opposite with an increase in their plasma concentrations on the same dose per unit body weight [(nM)/(mg/kg)] post-RYGB. The mean concentrations for simvastatin and simvastatin acid increased from 8.52 ± 3.04 and 9.96 ± 3.99 (nM)/(mg/kg) at baseline, respectively, to 11.29 ± 1.96 and 24.89 ± 6.90 (nM)/(mg/kg) at 3-month and stabilized at 28.45 ± 3.54 and 39.32 ± 7.95 (nM)/(mg/kg) at 6-month visits post-RYGB, respectively. The effect of RYGB on simvastatin and simvastatin acid concentrations seem to be normalized to the baseline levels at 12 months post-surgery with mean plasma concentrations of 7.09 ± 0.42 and 10.45 ± 2.39 (nM)/(mg/kg), respectively.
   These differential impacts on statin PK were consistent with the PD observations on LDL rebound with patients on atorvastatin and rosuvastatin, but the rebound was not apparent with patients on simvastatin when doses were reduced post-RYGB. A preliminary PK/PD correlation between the summation of the molar concentrations of atorvastatin, 2-hydroxy-atorvastatin and 4-hydroxy-atorvastatin and LDL values showed that the threshold of effective atorvastatin with active metabolites decreased from 40 nM pre-surgery to 20 nM at 3 and 6 months post-RYGB. The LDL concentrations were correlated with patients’ BMI and total atorvastatin (with metabolites) concentrations post-RYGB in a linear model. A preliminary PK/PD correlation between simvastatin acid molar concentrations and LDL values showed that a trend of decrease in LDL levels with increase in SMV-A concentrations at 3 and 6 months post-surgery. However, LDL reduction seems to level off at SMV-A concentrations higher than 5-10 nM. In addition, the LDL levels were correlated with the ratio of simvastatin acid/simvastatin concentrations and BMI in a linear model with no interaction between BMI and the simvastatin acid/simvastatin ratio post-RYGB. Moreover, both atorvastatin and simvastatin showed decreases in their metabolisms to the active metabolites post-RYGB. The ratios of 2-hydroxy atorvastatin/atorvastatin and 4-hydroxy atorvastatin/atorvastatin varied significantly at baseline among subjects, with 0.39-2.06 and 0.14-0.88, respectively. Among the subjects who completed the follow-up visits, the ratios decreased by 3 months post-RYGB by about 60%, then remained relatively constant at the lower levels. The ratio of simvastatin acid/simvastatin also varied significantly at baseline among subjects, with a range of 0.10-7.85. Among the 9 subjects on simvastatin, 5 had the PK data at the follow-up visits and showed a wide range of 3-60% decrease in simvastatin metabolism at 3 months or 6 months post-RYGB, as expressed by the ratios of simvastatin acid/simvastatin, then remained relatively constant at levels of lower ratios afterwards. The surgery itself influenced the lipid panel in patients by decreasing LDL in the non-statin group from 110 mg/dl at baseline (higher than the optimal level of <100 mg/dl) to 91 mg/dl at 1-year post-RYGB. Combined statin group had optimal LDL levels of 90, 77, 82, and 96 mg/dl at baseline, 3, 6, and 12 months post-RYGB, respectively. The significantly higher LDL in non-medicated patients was corrected by RYGB. The mean TG value pre-surgery in atorvastatin group (101 mg/dl) resembled those of non-statin group (118 mg/dl), but lower than those in simvastatin and rosuvastatin groups (165-183 mg/dl). At 1-year post-RYGB, TG levels decreased in all groups with 87 mg/dl in atorvastatin group, 103-116 mg/dl in simvastatin and rosuvastatin groups, and 91 mg/dl in non-statin group. The high pre-operative values of TG in simvastatin and rosuvastatin groups were corrected post-RYGB. Mean HDL levels in statin groups was 42-50 mg/dl at baseline and increased significantly at 12-month (58-64 mg/dl) follow-up visits. A similar time profile was observed in non-statin group with baseline mean HDL level of 47 mg/dl and increased to 58 mg/dl at 12-month visit. The weight loss was merely RYGB related, and statin treatments did not affect the weight loss outcomes post-RYGB comparing to non-statin group. Finally, we characterized the effect of RYGB on metformin PK/PD with 31 subjects in the metformin group. There was a trend of continuous decrease in metformin concentrations on the same dose per unit body weight [(ng/ml)/(mg/kg)] basis after surgery, with a range of 2.1-240.2 (ng/ml)/(mg/kg) at baseline that continuously reduced to 1.8-146.7, 0.6-110.6, and 0.1-12.9 (ng/ml)/(mg/kg) at 3, 6, and 12 months post-RYGB, respectively. The surgery itself proved to have a lowering effect on HbA1c. However, only 26.3-47.3% of the patients in our study reached complete remission state at any time point post-RYGB. Our results disagreed with the general notion that RYGB yields complete remission or resolution of type 2 diabetes in patients. The treatment failure in diabetic patients post-RYGB in our study could be correlated to the decreasing metformin concentrations post-surgery. Overall, we have achieved the goals of our specific aims. The complexity of the anatomical and physiological changes after RYGB makes the deduction of a conclusion challenging, especially with the small subject number as in our study. However, the observed longitudinal changes in the PK and PD of statins and metformin in the same individual subjects after the surgery suggest that the concentrations of statins and metformin should be monitored and the dosing regimen can be rationally adjusted after RYGB to ensure the therapeutic efficacy of the treatment with minimal adverse effects. In addition, patients who stop statins and/or metformin treatment post-RYGB should be followed-up closely to ensure they do not have recurrence of hyperlipidemia or diabetes. The merits of the study lie in the longitudinal monitoring and characterization of the impacts of RYGB on the PK and PD of three statins, atorvastatin, simvastatin, and rosuvastatin, along with their active metabolites, as well as the first line oral antidiabetic medication, metformin. The study utilizes the individual subjects as their own controls from pre-RYGB, to monitoring at 1 week, and at 1, 3, 6, and 12 months post-RYGB. In addition, the uniqueness of this study is being the first study to monitor the active metabolites of simvastatin and atorvastatin post-RYGB and provide PK/PD correlations of LDL with BMI and total active atorvastatin concentrations, as well as with BMI and simvastatin acid/simvastatin concentration ratios post-RYGB. The models with future validations offer the potential to rationally adjust the dose regimen of atorvastatin and simvastatin post-RYGB.