Effect of bariatric surgery on the pharmacokinetics of drugs used for attention‐deficit hyperactivity disorder—A case series

Abstract

Background

Changes in gastrointestinal physiology following bariatric surgery may affect the pharmacokinetics of drugs. Data on the impact of bariatric surgery on drugs used for attention‐deficit/hyperactivity disorder (ADHD) are limited.

Methods

In patients treated with ADHD medication and undergoing bariatric surgery, serial drug concentrations were measured for 24 h preoperatively and one, six and 12 months postoperatively. Primary outcome was change in area under the concentration‐time curve from 0 to 24 h (AUC_0–24), with other pharmacokinetic variables as secondary outcomes.

Results

Eight patients treated with lisdexamphetamine (n = 4), dexamphetamine (n = 1), methylphenidate (n = 1) and atomoxetine (n = 2) were included. In total, 409 samples were analysed. Patients underwent sleeve gastrectomy (n = 5) and Roux‐en‐Y gastric bypass (n = 3). AUC_0–24 and C_max of dexamphetamine increased after surgery in those using the prodrug lisdexamphetamine. There was no clear‐cut reduction in t_max postoperatively. For ritalinic acid and atomoxetine, no changes in AUC_0–24 were observed, but for atomoxetine, a higher C_max and a shorter t_max were observed postoperatively.

Conclusion

Bariatric surgery may increase the systemic exposure of dexamphetamine after intake of lisdexamphetamine. Patients using lisdexamphetamine should be followed with regard to adverse drug reactions after bariatric surgery, and, if available, therapeutic drug monitoring should be considered.

Plain English Summary.

After surgery for obesity the systemic exposure of drugs may be affected, which may lead to changes in the efficiency of medications. The aim of this study was to evaluate the effects of surgery on the pharmacokinetics of dexamphetamine, lisdexamphetamine, methylphenidate and atomoxetine. We obtained serum concentrations from eight patients before surgery and at different times after surgery. We found increased concentrations of dexamphetamine after surgery in patients using lisdexamphetamine but did not find such changes after intake of dexamphetamine, atomoxetine and methylphenidate. Clinicians should be aware of adverse drug reactions in patients using lisdexamphetamine after surgery for obesity.

1. INTRODUCTION AND BACKGROUND

Obesity is widely recognized as a major health risk, and bariatric surgery is an effective and sustainable method to reduce body weight. ¹ The prevalence of obesity has increased dramatically, and the World Obesity Federation estimates that 1 billion adults are living with obesity in 2024. ² New evidence indicates that attention‐deficit/hyperactivity disorder (ADHD) is associated with obesity, and in a recent meta‐analysis, ADHD was found to be present in 21% of patients undergoing bariatric surgery. ³

ADHD is a neuropsychiatric disorder characterized by developmentally inappropriate levels of hyperactive, impulsive and inattentive behaviours. Although the pathophysiology of ADHD remains incompletely understood, converging evidence supports an underlying dysregulation of noradrenergic and dopaminergic pathways associated with higher cortical function, including attention and executive functioning. Drugs used in the treatment of ADHD, such as amphetamine and its derivatives methylphenidate and atomoxetine, exert their effects by enhancing the levels of noradrenaline and dopamine in the synaptic cleft.

The two most commonly performed weight loss operations are sleeve gastrectomy (SG) and Roux‐en‐Y gastric bypass (RYGB). Both procedures reduce the volume of the stomach, whereas RYGB also reduces the absorptive surface of the gut by bypassing parts of the intestine. Oral drug pharmacokinetics can be significantly altered after bariatric surgery. Post‐operative changes in gastric mixing, gastric pH, gastric emptying and drug solubility in the intestine may underlie the alterations in the pharmacokinetics of drugs. ⁴ In addition, changes may also be caused by a reduction in the absorptive area in the gut and in the pre‐systemic metabolism by enzymes located in the intestinal mucosa, including cytochrome P450 (CYP) enzymes. Pharmacokinetic changes, including alterations in the expression of hepatic CYP enzyme activities, are also likely to occur as a consequence of weight loss as such. Given the diversity of physiological changes taking place, theoretical considerations alone cannot be used to anticipate the impact of bariatric surgery on the pharmacokinetics of specific drugs. Most drugs studied have shown unchanged or decreased post‐operative drug bioavailability after bariatric surgery, but also increased bioavailability is reported. ⁴ , ⁵

To our knowledge, the only publications so far on ADHD medication and bariatric surgery are two case reports ⁶ , ⁷ and a controlled experimental study. ⁸ The first case report describes a patient who experienced toxic symptoms of methylphenidate on the same dose 2 weeks after RYGB. ⁷ In contrast, in the second case report the patient, who also used methylphenidate, experienced lack of effect 2 weeks after RYGB. ⁶ Unfortunately, drug concentrations in serum or plasma were not available in these cases.

In the experimental study, ⁸ 10 subjects who had undergone RYGB 9–24 months previously and 10 nonsurgical control subjects were given a single dose of 50 mg lisdexamphetamine. Thereafter, plasma samples were obtained during a period of 24 h, and concentrations of lisdexamphetamine and the active metabolite dexamphetamine were measured. No significant differences between groups were observed in any of the pharmacokinetic variables studied, including area under the concentration‐time curve (AUC_0–24), maximum plasma concentration (C_max) and time to maximum plasma concentration (t_max).

As no longitudinal concentration data are available for any ADHD medication in subjects undergoing bariatric surgery, the aim of the present study was to assess the impact of RYGB and SG on lisdexamphetamine, dexamphetamine, methylphenidate and atomoxetine pharmacokinetics by measuring drug concentrations prospectively before surgery as well as one, six and 12 months postoperatively.

2. MATERIALS AND METHODS

2.1. Study design

The study is part of the pharmacokinetic platform study Changes in oral health and pharmacokinetics of drugs after bariatric surgery (BAR‐MEDS), which investigates serum concentrations of various medications before and after bariatric surgery. The study has been approved by the Regional Committee for Medical and Health Research Ethics in Mid Norway (ref. 2016/1145) and is registered at www.clinicaltrials.gov. It was performed in accordance with the ethical standards as laid down in the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards. Patients treated with lisdexamphetamine, dexamphetamine, methylphenidate or atomoxetine for ADHD and who were referred for bariatric surgery at St. Olav University Hospital, Trondheim, Norway, were invited to participate in this sub‐study. All eligible patients were included and gave their written consent before inclusion.

Serial blood samples were obtained preoperatively and one, six and 12 months after surgery. Blood samples were obtained at 0, 0.5, 1, 1.5, 2, 2.5, 3, 3.5, 4, 6, 12 and 24 h after drug intake. In patients taking the drug twice daily, additional samples were obtained the first hours after intake of the second dose. After centrifugation and pipetting, serum was stored at −80°C until analysis.

Body composition (body weight, body mass index [BMI], body fat, muscle mass and visceral fat area) was measured on each of the four study days using a multifrequency impedance analyser (InBody 720, Seoul, Korea).

2.2. Drug quantification

For lisdexamphetamine, which is inactive but is rapidly and completely metabolized to its active metabolite dexamphetamine, ⁹ , ¹⁰ dexamphetamine was chosen for analysis. As methylphenidate may undergo spontaneous hydrolysis in samples while stored, ¹¹ its primary, and inactive, metabolite ritalinic acid was chosen for analysis.

Dexamphetamine and ritalinic acid were analysed with an ultrahigh performance liquid chromatography–tandem mass spectrometry (UPLC‐MS/MS) method developed in our laboratory, as described previously. ¹²

Atomoxetine was analysed with another UPLC‐MS/MS method developed in our laboratory. In brief, sample preparation prior to analysis consisted of protein precipitation with acetonitrile and filtration through a phospholipid removal plate. Atomoxetine‐d₃ was used as internal standard. The method employed an Acquity UPLC BEH C18 1.7 μm, 2.1 × 50 mm column (Waters, Milford, MA, USA) and a mobile phase consisting of 5 mmol/L ammonium formate in methanol. MS/MS detection was performed on an Acquity UPLC I‐Class instrument with a Xevo TQ‐S detector (Waters). Positive electrospray ionization and multiple reaction monitoring were applied (m/z 256 > 44 for quantification and m/z 256 > 148 for qualification). The calibrated range was 40–4000 nmol/L (aprox. 160–1600 ng/mL). The between‐assay relative standard deviations were in the range of 3.4%–4.6%.

2.3. Pharmacokinetic and genetic analyses

In patients not using the same daily dose on the four study days, concentrations were adjusted based on the main dose for that subject, assuming linear pharmacokinetics, before any calculations were performed. After this procedure, trough concentrations (C₀; measured immediately before intake of the daily dose—or the first dose that day in subjects taking the drug twice daily), maximum serum concentrations (C_max) and the times to achieve C_max (t_max) were obtained directly from the measured values. Other pharmacokinetic variables were calculated using the pharmacokinetic programme package Kinetica, version 5.0 (ThermoFisher Scientific, Waltham, MA, USA).

AUC from 0 to 24 h (AUC_0–24) was calculated using a mixed log‐linear model. Apparent clearance (Cl/F) was calculated as dose/AUC_0–24. By applying a non‐compartment model, the parameter estimate describing the decrease of the log‐concentration (λ_z) was calculated using the best fit log‐linear regression line of the samples representing the elimination phase. The elimination half‐life (t_1/2) was calculated as ln2/λ_z. Apparent volume of distribution (V_d/F) was calculated as (Cl/F)/λ_z.

Patients using atomoxetine were genotyped for CYP2D6, which is the main enzyme involved in the metabolism of this drug. The inactivating variants *3, *4, *5, *6 as well as the partially inactivating variants *9, *10 and *41 were tested using allele‐specific polymerase chain reaction (PCR). In addition, duplication/multiplication of the whole gene, causing increased enzyme activity, was investigated.

2.4. Outcomes

The primary outcome was changes in systemic drug exposure after surgery, measured as AUC_0–24. Other pharmacokinetic variables were considered secondary outcomes. As few subjects were included for each drug, only descriptive statistics are presented.

The study was conducted in accordance with the Basic & Clinical Pharmacology and Toxicology policy for experimental and clinical studies. ¹³

3. RESULTS

Eight patients treated with lisdexamphetamine (n = 4), dexamphetamine (n = 1), methylphenidate (n = 1) and atomoxetine (n = 2) were included (Table 1). Five patients underwent SG whereas three underwent RYGB. There were seven females and one male with a mean age of 39.8 ± 9.2 years. All patients were Caucasians. Mean BMI preoperatively was 39.0 ± 5.2 kg/m². Individual demographic and clinical characteristics as well as details regarding medication are summarized in Table 1. Results from clinical biochemistry tests were, with a few exceptions, within normal limits (Table S1).

TABLE 1.

Patient characteristics in eight subjects undergoing bariatric surgery and treated with lisdexamphetamine (n = 4), amphetamine (n = 1), methylphenidate (n = 1) and atomoxetine (n = 2). Drugs taken twice daily (patient number 2 and 5) were ingested in the morning and at lunchtime.

Patient number	Age	Sex	Drug (brand name)	Type of surgery	Daily dose before surgery	Daily dose 1 month postop	Daily dose 6 months postop	Daily dose 12 months postop	Other medications g
1	56	M	Lisdexamphetamine (Elvanse)	SG	100 mg × 1	100 mg × 1	100 mg × 1	100 mg × 1	Amlodipine, codeine/paracetamol, hydrochlorothiazide, irbesartan, paracetamol h
2	45	F	Lisdexamphetamine (Aduvanz)	SG	30 mg × 2	30 mg × 2	30 mg × 2	30 mg × 2	Paracetamol
3	37	F	Lisdexamphetamine (Elvanse)	SG	50 mg × 1	50 mg × 1	50 mg × 1	60 mg × 1 c	None
4	34	F	Lisdexamphetamine (Aduvanz)	RYGB a	30 mg × 1	30 mg × 1	30 mg × 1	30 mg × 1	Oxycodone h , tramadol i
5	47	F	Dexamphetamine (Attentin)	SG	15 mg × 2	15 mg × 2	15 mg × 2	10 mg × 2 d	None
6	28	F	Methylphenidate (Ritalin LA capsule) b	RYGB	30 mg × 1	30 mg × 1	30 mg × 1	20 mg × 1 e	Metoprolol, desogestrel, esomeprazole, ursodeoxycholic acid
7	30	F	Atomoxetine (Strattera)	SG	40 mg ×  f	40 mg × 1 f	60 mg × 1	80 mg × 1 f	None
8	41	F	Atomoxetine (Strattera)	RYGB	60 mg × 1	60 mg × 1	60 mg × 1	60 mg × 1	Amlodipine h , metoprolol i

Abbreviations: RYGB, Roux‐en‐Y gastric bypass; SG, sleeve gastrectomy.

^a Sleeve gastrectomy 5 years previously, now Roux‐en‐Y gastric bypass.

^b “Long acting” (extended release).

^c Adjusted to 50 mg × 1 in the pharmacokinetic calculations.

^d Adjusted to 15 mg × 2 in the pharmacokinetic calculations.

^e Adjusted to 30 mg × 1 in the pharmacokinetic calculations.

^f Adjusted to 60 mg × 1 in the pharmacokinetic calculations.

^g Recorded preoperatively; see separate footnotes for drugs discontinued during the study period.

^h Discontinued before 1 month postoperatively.

ⁱ Discontinued before 6 months postoperatively.

Concentration‐time curves for each individual at baseline (preoperatively) and one, six and 12 months postoperatively are displayed in Figure 1. Corresponding pharmacokinetic data are presented in Table S2. Individual changes in BMI, body fat, AUC_0–24, C_max, Cl/F, t_½, t_max and Vd/F are displayed in Figure 2 for those using lisdexamphetamine and dexamphetamine and in Figure 3 for those using methylphenidate and atomoxetine.

FIGURE 1.

Individual time‐concentration plots at baseline (preoperatively; red squares) and 1 month (yellow diamonds), 6 months (green circles) and 12 months (brown triangles) postoperatively. In a patient not using the same daily dose during the study, concentrations were adjusted to the main dose for that subject (Table 1). Note that the scale on the y‐axis varies between subjects. RYGB, Roux‐en‐Y gastric bypass; SG, sleeve gastrectomy.

FIGURE 2.

Body mass index (BMI), body fat, AUC_0–24, C_max, Cl/F, t_1/2, t_max and V_d/F at baseline (preoperatively) and 1, 6 and 12 months after bariatric surgery in four patients treated with lisdexamphetamine undergoing sleeve gastrectomy (patients 1–3) or Roux‐en‐Y gastric bypass (patient 4), and one patient treated with dexamphetamine undergoing sleeve gastrectomy (patient 5). Values for t_max are not shown for patient 2 and 5 since these patients took their drugs twice daily and the time intervals between intakes varied between study days.

FIGURE 3.

Body mass index (BMI), body fat, AUC_0–24, C_max, Cl/F, t_1/2, t_max and V_d/F at baseline (preoperatively) and 1, 6 and 12 month after bariatric surgery in one patient treated with methylphenidate and undergoing sleeve gastrectomy (patient 6), one patient treated with atomoxetine and undergoing sleeve gastrectomy (patient 7) and one patient treated with atomoxetine and undergoing Roux‐en‐Y gastric bypass (patient 8). Values for Cl/F and Vd/F are not shown for patient 6 as the pharmacokinetic variables are presented for the metabolite ritalinic acid.

In patients 1–4, who used lisdexamphetamine, AUC_0–24 for dexamphetamine increased postoperatively; for two of the patients there was an increase of about 50%–100%, whereas for the two others the increases were smaller, roughly about 30%. In contrast, in patient 5, who used dexamphetamine, no such increase was observed. For all these patients, C_max increased postoperatively. There was no clear‐cut reduction in t_max after surgery, but reliable data were available only for three of the five patients as patient 2 and patient 5 took their drugs twice daily and the time interval between these intakes varied between study days.

Patient 6, who used methylphenidate, and patient 7–8, who used atomoxetine, had fairly stable AUC_0–24 values, although it should be noted that for patient 6 the inactive metabolite ritalinic acid, and not methylphenidate as such, was analysed. In the patients using atomoxetine, there was a tendency towards increased C_max and decreased t_max postoperatively. Patient 7 had the CYP2D6 genotype *1/*1, indicating normal metabolic capacity, whereas patient 8 had the genotype *4/*41, indicating decreased metabolic capacity. This difference is also reflected in the atomoxetine concentrations, with patient 8 having three–fjve times higher AUC_0–24 values than patient 7 despite using the same dose.

Changes in body fat and weight, BMI, muscle mass and visceral fat after surgery are presented in Table S3. As two of the four patients using lisdexamphetamine had large increases in AUC_0–24 of dexamphetamine after surgery, whereas the two others had modest increases, we compared these two pairs with regard to body composition. It then turned out that those with large increases in AUC_0–24 had lost more weight than those with modest increases (relative weight loss after 12 months 35.2% vs. 25.2%). We did not find any discernible relationship between alterations in body composition and changes in pharmacokinetic variables for any of the other drugs.

4. DISCUSSION

The main finding in our prospective cohort study of patients using ADHD medication and undergoing SG or RYGB was an increased AUC_0–24 and C_max of dexamphetamine after surgery in the four patients using lisdexamphetamine. In contrast, such changes were not seen after intake of dexamphetamine, although this is based on one participant only. There was no clear‐cut reduction in t_max after surgery. For methylphenidate and atomoxetine, no changes in AUC_0–24 were observed, but for atomoxetine a higher C_max and a shorter t_max were observed postoperatively.

The increased concentrations of dexamphetamine after intake of lisdexamphetamine post‐surgery cannot readily be explained. One possibility could be at the level of drug absorption. Lisdexamphetamine is well‐absorbed from the gut, ¹⁴ and changes in gastric pH do not alter its absorption. ⁸ Lisdexamphetamine is absorbed from the small intestine by active transport, likely via the peptide transporter 1 (PEPT1) ¹⁵ and thereafter completely converted to d‐amphetamine in the circulation. ⁹ , ¹⁰ In contrast, dexamphetamine is absorbed from the small intestine by passive diffusion.

PEPT‐1 may be regulated at the transcriptional and posttranscriptional levels by numerous factors. Physiological changes in obesity, such as low‐grade inflammation, may influence the activity of transporters. Multiple non‐human studies have revealed that the expression of PEPT‐1 is increased during undernutrition. ¹⁶ After bariatric surgery patients are recommended avoiding meals high in calories, fat and sugar. In theory, this could increase the absorption of lisdexamphetamine during weight loss.

Amphetamine is mainly excreted unchanged in the urine, but the proportion of an amphetamine dose excreted in unchanged form is very dependent on the urine pH, ranging from 1% in alkaline urine to 70% in acidic urine. ¹⁷ The rest of the ingested amphetamine is converted in the liver by hydroxylation and deamination, and CYP2D6 is involved in one of these minor pathways. ¹⁸

Urinary pH is inversely correlated with body weight. ¹⁹ , ²⁰ Thus, weight loss could lead to increased urinary pH and decreased excretion of dexamphetamine. Although this effect would be expected after ingestion of both lisdexamphetamine and dexamphetamine, it should be noted that the participant treated with dexamphetamine lost less weight compared to the lisdexamphetamine participants during the follow‐up year.

In the previously published experimental study on lisdexamphetamine, no significant differences were observed in any of the pharmacokinetic variables studied between subjects who had undergone RYGB and nonsurgical controls with the same BMI. ⁸ However, as this was a cross‐sectional study, possible intraindividual changes over time caused by, for example, weight loss after surgery could not be detected. Of note, both the patient group and the control group in that study had BMI values of about 30 kg/m², whereas mean BMI at inclusion in the four patients using lisdexamphetamine in our study was 38 kg/m², dropping to 28 kg/m² after 1 year.

Methylphenidate is a weak base and has a low bioavailability with a large interindividual variability. ²¹ Methylphenidate is primarily metabolized to the inactive metabolite ritalinic acid by the enzyme carboxylesterase 1 (CES1), which is found mainly in the liver. ²² About 60%–80% of the administered dose is recovered as ritalinic acid in urine. ²³ , ²⁴ , ^{25 .}

It has been shown that weight loss decreases CES1 expression in human adipose tissue ²⁶ and that reduced DNA methylation of the CES1 gene is associated with childhood obesity; ²⁷ hence, one could postulate decreased levels of ritalinic acid after weight loss, although the extent of these changes would be speculative. Both toxic symptoms and reduced effects have been reported after bariatric surgery, ⁶ , ⁷ but in these studies, serum concentrations were not measured. We did not find any obvious changes in the disposition of ritalinic acid after surgery in our patient.

Atomoxetine has a bioavailability of more than 60% and approximately 90% in CYP2D6 extensive and poor metabolizers, respectively. Its primary metabolite, 4‐hydroxyatomoxetine, which is formed by CYP2D6, is also active ²⁸ but is rapidly glucuronidated and circulating levels are low. ²⁹ In our study, no changes in AUC_0–24 were observed, but a higher C_max and a shorter t_max were seen postoperatively. Accelerated gastric emptying causing a more rapid absorption has been described after SG and RYGB for many drugs. ⁴ The observed increase in atomoxetine C_max could possibly worsen concentration‐related adverse reactions. ³⁰ The patient with decreased CYP2D6 activity had an AUC_0–24 that was three–five times higher than the patient showing normal CYP2D6 enzyme activity, with corresponding increases in C_max and t_1/2. These differences are in accordance with what has been reported in the literature previously. ²⁹ , ³⁰ , ³¹ Our results are also consistent with previous studies showing that weight loss as such does not seem to affect the metabolic activity of CYP2D6. ⁴

Our study has several strengths. We obtained a complete pharmacokinetic profile through an interval of 0–24 h, preoperatively as well as at three time points postoperatively. Measuring serum concentrations multiple times up to a year after surgery allowed us to study both short‐term and long‐term effects of bariatric surgery on pharmacokinetic variables. To our knowledge, there are no other studies comparing serum concentrations of ADHD medication longitudinally in the same subjects both before and after bariatric surgery. Finally, the changes in body composition of the participants were thoroughly described in our study, and we genotyped CYP2D6, which is the main CYP isoenzyme involved in atomoxetine metabolism.

There are also limitations of this study that should be considered. The study was conducted with a rather limited sample size. A larger sample size could have provided more reliable and conclusive results. It cannot be determined whether the observed pharmacokinetic changes are caused by the bariatric surgery per se or the resultant weight loss and altered body composition, as we did not include a control group losing weight by other means. Finally, we did not measure methylphenidate, but the inactive metabolite ritalinic acid. This was performed because methylphenidate undergoes spontaneous hydrolysis to ritalinic acid in samples while stored, ¹¹ which makes measurement of methylphenidate itself unreliable.

In general, data on pharmacokinetic changes after bariatric surgery are limited, and the lack of existing guidelines could lead patients to experience adverse drug reactions or therapeutic failure after surgery. Until more evidence emerges, particularly patients treated with lisdexamphetamine should be followed up frequently after surgery, and doses adjusted in accordance with changes in drug response. Therapeutic drug monitoring could be of help for clinicians to individualize the dose post‐surgery.

AUTHOR CONTRIBUTIONS

Drs. Spigset, Helland and Strømmen conceived and/or designed the study. Drs. Strømmen and Krabseth performed data collection and organization. Drs. Spigset, Helland and Krabseth analysed data and contributed models. The first draft of the manuscript was written by Drs. Krabseth and Spigset and all authors commented on preliminary versions of the manuscript. All authors read and approved the final version of the manuscript.

CONFLICT OF INTERESTS STATEMENT

Authors report no competing interests with this work.

Supporting information

Table S1. Clincial biochemistry test results at baseline. Results for C‐reactive protein are presented at baseline (preoperatively) as well as at 1, 6 and 12 months postoperatively.

Table S2 Patient 1. Drug: Lisdexamphetamine. Surgical procedure: Sleeve gastrectomy. Pharmacokinetic variables based upon the concentrations of dexamphetamine. All values are adjusted to a dose of 100 mg x 1.

Table S3 Body composition variables preoperatively and at 1, 6 and 12 months postoperatively. All data are presented as means ± standard deviations.

Krabseth H‐M, Strømmen M, Helland A, Spigset O. Effect of bariatric surgery on the pharmacokinetics of drugs used for attention‐deficit hyperactivity disorder—A case series. Basic Clin Pharmacol Toxicol. 2025;136(1):e14099. doi: 10.1111/bcpt.14099

DATA AVAILABILITY STATEMENT

The dataset generated and analysed during the current study is available from the corresponding author on reasonable request.