Impact of Recent Methamphetamine Use on Vancomycin Clearance

Abstract

Purpose

To determine whether recent methamphetamine use increases vancomycin clearance.

Methods

This was a multi-center, retrospective, IRB-approved study at two tertiary care medical centers. Adult patients with a urine drug screen, ≥3 consecutive vancomycin doses, and an appropriately drawn vancomycin trough were assessed and classified as amphetamine positive or amphetamine negative. The primary outcome was vancomycin clearance.

Results

88 patients were included in the analysis, with 44 patients in each group. Vancomycin clearance was greater in the amphetamine positive group (94.54 vs. 86.84 mL/min, p = 0.042, 95% CI 0.29–15.09). There was no significant difference in goal vancomycin trough achievement between groups (34.1% amphetamine positive vs. 43.2% amphetamine negative; p = 0.512). Per multifactorial logistic regression analysis, older age and male gender were associated with decreased vancomycin clearance, while higher BMI and cocaine positive urine drug screen were associated with increased vancomycin clearance.

Conclusion

Recent methamphetamine use may increase vancomycin clearance. Larger prospective trials with protocolized vancomycin dosing strategies are needed to further elucidate the impact of methamphetamine use on attainment of goal vancomycin troughs in addition to the potential impact on vancomycin clearance.

Introduction

People who inject drugs (PWID) are 16 times more likely to develop methicillin-resistant Staphylococcus aureus (MRSA) infections. The proportion of invasive MRSA cases that occurred in PWID increased from 4.1% in 2011 to 9.2% in 2016. The most common invasive MRSA infections in PWID from 2005 to 2016 were bacteremias (73.6%) and endocarditides (20.4%).¹ Vancomycin is recommended as a drug of choice for empiric and targeted coverage in both severe and non-severe MRSA infections.²

Previous studies have found that injection drug use (IDU) is associated with expedited vancomycin clearance.^3–5 A published abstract by Ferreira et al. with 47 subjects analyzed steady-state vancomycin trough concentrations from consistent dosing regimens and described vancomycin trough attainment rates in non-critically ill PWID.³ 43% of patients had used methamphetamine and heroin, and 6% had used methamphetamine only. The authors concluded that vancomycin trough concentrations were frequently below guideline-based effectiveness targets in PWID. However, they did not comment on statistical significance for any of their outcomes. Likewise, a published abstract by Farré et al. with 65 subjects sought to characterize vancomycin pharmacokinetic parameters in non-cirrhotic people with alcohol use disorder, patients with alcohol-induced cirrhosis, and PWID.⁴ Vancomycin clearance was found to be higher in patients with alcohol use disorder and PWID, but this difference was not statistically significant. A prospective study by Rybak et al. evaluated the pharmacokinetics of vancomycin in 14 PWID, 10 burn patients, and 10 controls.⁵ Vancomycin clearances averaged 142.8, 98.0, and 67.7 mL/min in burn patients, PWID, and control patients, respectively. The PWID were found to have a higher vancomycin clearance (77.4 ± 24.1 vs. 58.4 ± 15.9 mL/min) than control patients, though this difference was not statistically significant. To our knowledge, this is the only peer-reviewed prospective study evaluating the impact of IDU on vancomycin clearance. No study however has focused on the impact of specific illicit drug types on vancomycin clearance.

Since methamphetamine is injected intravenously about 13% of the time, it is one of the most common reasons for illicit-substance-related emergency department visits in Arizona, and since Arizona has the second highest rate of methamphetamine use for people aged 12 years or older in the country, we sought to investigate the impact of recent methamphetamine use on vancomycin clearance specifically.^6–8 Although our study was conducted in Arizona, this is a nationwide issue with 1.1 million people over the age of 12 years being diagnosed with a methamphetamine use disorder in 2018.⁹ We hypothesized that recent methamphetamine use would lead to increased vancomycin clearance and higher rates of subtherapeutic levels.

Methods

An IRB-approved retrospective chart review from October 1st, 2017 to July 30th, 2018 was conducted at two academic medical centers. Inclusion criteria were age greater than 18 years, confirmed amphetamine use via urine drug screen (UDS), receival of three or more vancomycin doses, and a minimum of one appropriately-drawn vancomycin trough level (defined as a level drawn prior to the third to fifth dose, 0 to 60 minutes prior to the next dose, and at least one dosing interval apart from the previous dose).^10,11 Amphetamine and para-hydroxymethamphetamine are methamphetamine’s main metabolites. Therefore, a positive UDS for amphetamine could suggest recent methamphetamine or amphetamine use. Exclusion criteria were a creatinine clearance (CrCl) less than 60 mL/min at admission per the Cockcroft-Gault calculation, and no UDS on file for the same admission. Study data were collected and managed using REDCap electronic data capture tools.^12,13 REDCap (Research Electronic Data Capture) is a secure, web-based software platform designed to support data capture for research studies.

The amphetamine-positive group consisted of patients who met inclusion criteria, and the amphetamine-negative group consisted of patients meeting inclusion criteria with a negative UDS for all drugs. Amphetamine-negative and amphetamine-positive subjects were matched 1:1 based on CrCl within ± 40 mL/min and gender.¹⁴

The primary outcome was to determine whether recent methamphetamine use increases vancomycin clearance. Secondary outcomes were achievement of goal vancomycin trough, differences in vancomycin clearance between amphetamine-positive critically ill and non-critically ill patients, and differences in vancomycin clearance between methamphetamine and amphetamine salt (Adderall^®) users. Target vancomycin trough concentrations were defined as 10–15 mcg/mL for urinary tract infections and cellulitis, and 15–20 mcg/mL for all other infection types in patients who were treated with vancomycin therapy (defined as at least three administered doses).¹⁰ Critically ill status was defined as admission to an intensive care unit (ICU), and amphetamine salt use was assessed through review of the Arizona Controlled Substances Prescription Monitoring Program (AZ PMP).¹⁵ This PMP captures all controlled substances dispensed over a specific timeframe including those for which patients paid in cash.

Recent methamphetamine use was defined as a positive amphetamine result per UDS and documentation of illicit drug use in the patient chart for the same admission. The UDS used at the academic medical centers in this project consists of individual immunoassay reagents by Abbott^® for amphetamines, barbiturates, benzodiazepines, cocaine, tetrahydrocannabinol, opiates, and phencyclidine. All patients with a positive amphetamine UDS were included in the amphetamine-positive group, and subgroup analyses were carried out for those who also had a positive cocaine UDS and for those who were found to have prescribed amphetamines through the AZ PMP. Cocaine was included to assess its potential effect on vancomycin clearance since it is also a stimulant drug.

Vancomycin was dosed based on weight, with loading doses of 20–30 mg/kg and maintenance doses of 15–20 mg/kg every 8–12 hours. Vancomycin clearance in milliliters per minute was estimated using pharmacokinetic equations and patient specific parameters, assuming a one-compartment model. Patient-specific vancomycin clearance was estimated using CrCl and the patient’s rate of elimination (K_e) of vancomycin. Serum creatinine values used were from the same day the vancomycin trough was collected. The Cockcroft-Gault equation was used to estimate CrCl.¹⁶ The weight used for the Cockcroft-Gault equations was: actual body weight if the actual weight was less than 20% of the ideal body weight (IBW), IBW if actual weight was within ±20% of IBW, and adjusted body weight if actual weight was greater than 20% of IBW.^16,17 The Matzke equation

$${\text{K}}_{\text{e}}=(0.00083*\text{\hspace{0.17em}CrCl})+0.0044$$

was used to estimate K_e for patients weighing ≤90.1 kg. K_e was then used to estimate vancomycin clearance (Cl_Vanco) in

$${\text{Cl}}_{\text{Vanco}}=\text{Vd}*{\text{K}}_{\text{e}}$$

where volume of distribution (Vd) is 0.7 L/kg for all patients.¹⁷ Units for this equation were converted from liters per hours to milliliters per minute. The Crass equation

$${\text{Cl}}_{\text{Vanco}}=9.656-(0.078*\text{age})-(2.009*\text{SCr})+(1.09*\text{gender})+(0.04*\text{total\hspace{0.17em}body\hspace{0.17em}}{\text{weight}}^{0.75})$$

was used for patients weighing >90.1 kg, where female gender equates to 0 and male gender equates to 1.¹⁸

The primary outcome of vancomycin clearance was analyzed via a paired t-test. Fisher’s exact test was used for the target trough analysis, and a student’s t-test for the comparison in vancomycin clearance among amphetamine positive patients in or outside the ICU. A multifactorial logistic regression analysis was performed to evaluate the impact of age, gender, infection types, appropriate selection and dosing of vancomycin, as well as other potential confounders (race, cocaine in UDS, BMI, and methamphetamine administration route if mentioned in any notes for that admission in the amphetamine-positive group) on increased vancomycin clearance. Since the dependent variable was continuous, results of this regression were reported as beta coefficients. These can be interpreted as degree of change in the outcome variable for every 1-unit of change in the predictor variable. If the beta coefficient is negative, the interpretation is that for every 1-unit increase in the predictor variable, the outcome variable will decrease by the beta coefficient value.¹⁹ In order to achieve 80% power for detecting a 21 mL/min difference in the primary outcome, a total of 30 subjects total were needed with an alpha of 0.05.⁴

Results

A total of 88 patients were assessed after screening for inclusion and exclusion criteria (Figure 1). The majority of patients were Caucasian males with an average weight between 73 to 77 kilograms, and an average age between 38 and 46 years old. The amphetamine positive group had statistically significant younger patients, higher baseline CrCl estimates, and higher number of skin and soft tissue infections (SSTIs) and joint infections. Other baseline demographics and characteristics of the patients included in this study are shown in Table 1. A breakdown of vancomycin trough levels achieved in each group is available in Figure 2. Differences in vancomycin troughs less than or above 9.5 are depicted in Figure 3.

Figure 1.

Patient Enrollment

Table 1. Baseline Demographics and Clinical Characteristics.

VARIABLE		AMPHETAMINE- NEGATIVE, N (%) N = 44		AMPHETAMINE- POSITIVE, N (%) N = 44		P-VALUE
Gender
Male		27 (61.4%)		28 (63.6%)		1.000
Race
Caucasian		38 (86.4%)		41 (93.2%)		0.334
American Indian or Alaskan Native		5 (11.4%)		2 (4.5%)
Black or African American		1 (2.3%)		0 (0.0%)
Other		0 (0.0%)		1 (2.3%)
Type of bacterial infectionb		n = 50		n = 69
SSTI		6 (12%)		17 (24.6%)		0.014a
UTI		1 (2%)		1 (1.5%)		1.000
Bacteremia		7 (14%)		15 (21.7%)		0.084
Endocarditis		3 (6%)		3 (4.4%)		1.000
Osteomyelitis		7 (14%)		2 (2.9%)		0.157
Pneumonia		16 (32%)		13 (18.8%)		0.651
CNS infection		3 (6%)		1 (1.5%)		0.616
Pyelonephritis		1 (2%)		0 (0%)		1.000
Joint		6 (12%)		17 (24.6%)		0.014a
Age (years), mean ± SD		46.16 ± 17.63		38.80 ± 12.25		0.025a
Weight (kg), mean ± SD		77.08 ± 20.62		73.71 ± 18.04		0.417
Height (cms), mean ± SD		168.34 ± 10.25		172.73 ± 10.23		0.048a
BMI (kg/m2), mean ± SD		27.17 ± 6.87		24.61 ± 5.53		0.058
CrCl (mL/min), mean ± SD		116.6 ± 45.13		136.32 ± 39.53		0.032a
ICU admission, no. (%)		16 (36.3%)		13 (29.6%)		0.651

^ap ≤ 0.05. SSTI: skin and soft tissue infection, UTI: urinary tract infection, CNS: central nervous system, BMI: body mass index, CrCl: creatinine clearance, ICU: intensive care unit. ^bsome subjects had more than 1 infectious diagnosis.

Figure 2.

Vancomycin Trough Ranges. p = 0.086

Figure 3.

Vancomycin Troughs (VTs) Less Than or Greater Than 9.5 mcg/mL. p = 0.009

The amphetamine-positive group on average had a higher estimated vancomycin clearance compared to the amphetamine-negative group (94 mL/min vs. 87 mL/min; p = 0.042, 95% CI 0.29–15.09). Fewer patients in the amphetamine-positive group achieved goal vancomycin trough, however this was not found to be statistically significant (43.2% vs. 34.1%; p = 0.512). However, a vancomycin trough of at least 9.5 was achieved significantly less in the amphetamine-positive group compared to the amphetamine-negative group (59.1% vs. 84.1% respectively, p = 0.009). There were no significant differences in vancomycin clearance between groups when accounting for severity of illness (critically ill vs. non-critically ill), and no difference in the number of patients requiring ICU admission. A summary of primary and secondary outcomes can be found in Table 2.

Table 2. Clinical Outcomes.

		AMPHETAMINE- NEGATIVE, N = 44		AMPHETAMINE- POSITIVE, N = 44		P-VALUE (95% CI)
ClVanco (mL/min), mean ± SD		86.84 ± 38.66		94.54 ± 31.81		0.042 a (0.29–15.09)
Achievement of goal vancomycin trough, no. (%)		19 (43.2)		15 (34.1)		0.512
		CRITICALLY ILL, N = 13		NON-CRITICALLY ILL, N = 31		P-VALUE (95% CI)
ClVanco in amphetamine-positive patients (mL/min), mean ± SD		91.96 ± 30.29		95.61 ± 32.85		0.733 (−17.78–25.07)

^ap ≤ 0.05.

Variables that were found to influence vancomycin clearance in the multifactorial logistic regression analysis were age, gender, BMI, and cocaine detection in the UDS. Older age and male gender were associated with decreased vancomycin clearance, while higher BMI and presence of cocaine in the UDS were associated with increased vancomycin clearance. These variables are listed in Table 3 along with their respective beta coefficients. There were not enough data points to include race or route of amphetamine administration in the analysis.

Table 3. Multifactorial Logistic Regression.

VARIABLE		UNSTANDARDIZED BETA COEFFICIENT		STANDARDIZED BETA COEFFICIENT		P-VALUE
Age		−1.05		−0.462		<0.001
Male gender		−23.89		−0.329		<0.001
Higher BMI		2.51		0.449		<0.001
Cocaine in UDS		58.25		0.247		<0.001

Discussion

Our findings suggest that recent methamphetamine use may augment vancomycin clearance. However, this augmentation may not be clinically significant, given the amphetamine-positive group had an average estimated vancomycin clearance of 94 mL/min vs. 87 mL/min in the amphetamine-negative group. The primary outcome may have also been affected by the difference in CrCl in both groups at baseline. Subjects in the amphetamine positive group were significantly younger than those in the amphetamine negative group, which may have contributed to a higher CrCl at baseline. However, subjects were matched 1:1 to decrease the influence of this on our primary outcome, and the difference between 116.6 vs. 132.3 mL/min is not clinically significant since both clearance rates would warrant similar dosing intervals.

The mechanism through which methamphetamine affects vancomycin clearance remains unclear. It is estimated that 70% of vancomycin is cleared renally, mainly by glomerular filtration, and 90% of this is excreted unchanged.¹⁸ Therefore, it is likely that methamphetamine affects vancomycin elimination, not its metabolism. Methamphetamine, on the other hand, is cleared via glomerular filtration as well as tubular secretion, and 37–54% is excreted unchanged.²⁰ Amphetamine and para-hydroxymethamphetamine are also eliminated via renal filtration and secretion.²⁰ More research is needed to elucidate the mechanism of how methamphetamine impacts vancomycin elimination.

Although there was no statistically significant difference in achievement of goal vancomycin troughs between groups, 40.9% of patients in the amphetamine-positive group had a trough <9.5, which is subtherapeutic for all indications of vancomycin, compared to 15.9% of the patients in the amphetamine-negative group (p = 0.009) (Figure 3). This finding is concerning from an antimicrobial stewardship standpoint. Subtherapeutic doses of vancomycin are associated with drug resistance including higher minimum inhibitory concentrations, and the development of vancomycin-intermediate S. aureus and vancomycin-resistant S. aureus strains.²¹ SSTIs were more prevalent in the amphetamine positive group. While the majority of patients had more than one infectious diagnosis at once, 16 out of 17 amphetamine positive patients had an SSTI as their sole infectious diagnosis, which would warrant less aggressive dosing to target a lower goal trough (10–15 mcg/dL). A Pearson Chi-square analysis of vancomycin troughs <9.5 between SSTI patients in both the amphetamine-negative and the amphetamine-positive group was carried out (p = 0.901). This suggests that the higher number of amphetamine-positive patients with vancomycin troughs <9.5 was not driven by the higher number of SSTIs in the amphetamine-positive group. Only 38.6% of patients in the study achieved target vancomycin troughs regardless of methamphetamine use, likely due to the medication not having reached steady state at that point. This finding could also suggest vancomycin may not have been dosed appropriately at the two medical centers where the study took place. However, this percentage of goal vancomycin trough attainment using the first trough drawn is similar to that of other academic centers in the United States.^22,23 A prospective observational study by Meng et al. at Stanford Health Care Stanford Hospital with 296 patients found a goal vancomycin trough attainment rate of 55%. A retrospective study by O’Brien et al. at Emory University Hospital and Emory University Hospital Midtown with 200 patients found 29% attainment rates for troughs between 15–20 mg/dL. However, vancomycin monitoring guidelines were updated since this study took place. It is now recommended to monitor vancomycin dosing for serious MRSA infections based on AUC/MIC and not based on troughs as a way to optimize efficacy while minimizing renal injury.²

There was no significant difference in vancomycin clearance between amphetamine-positive critically ill and non-critically ill patients. Critically ill patients have been shown to be at risk for having subtherapeutic vancomycin troughs, as well as increased renal clearance.^24–26 Therefore, patients who were both critically ill with recent methamphetamine use were expected to have a higher vancomycin clearance compared to non-critically ill patients with recent methamphetamine use. The sample size may not have been large enough to detect a difference between these two groups.

Based on the multifactorial logistic regression analysis, older age and male gender were associated with reduced vancomycin clearance, while cocaine use and higher BMI were associated with increased vancomycin clearance. Reduced vancomycin clearance is expected with older age, as renal function is reduced in this population, and vancomycin is primarily cleared renally.²⁷ Males have faster renal clearance than females, and have been shown to have increased vancomycin clearance compared to females.²⁸ On the other hand, chronic cocaine use has been linked to renal damage.^29,30 Like methamphetamine, the mechanism for possible augmentation of vancomycin clearance due to cocaine use is unknown. BMI has also been associated with increased vancomycin clearance due to high vancomycin distribution into adipose tissue.²⁸

This study has several limitations that should be taken into consideration. Its retrospective design may have allowed for the introduction of selection bias. In addition, its small sample size may not have provided enough power to detect a difference in the secondary outcome of vancomycin trough achievement between groups. Patients were also predominantly Caucasian, therefore results from this study are not generalizable to other races. The inability to differentiate between methamphetamine route of administration for all patients was also a limitation since this may be a confounder. However, route of administration is not always assessed or documented in the electronic medical record. Not all patients who presented to the emergency department received a UDS, therefore amphetamine-positive patients for whom a UDS was not done could have been missed. Timing of vancomycin initiation in relation to when methamphetamine was last used could also affect vancomycin clearance. Methamphetamine is detectable in the urine for two to three days after last use via its main metabolite, amphetamine.³¹ Therefore, even if it is detected on UDS, its effects on vancomycin clearance may vary depending on how recently it was used. Timing of vancomycin initiation relative to a positive amphetamine finding in the UDS was not collected, but this would have also been logistically difficult through a retrospective chart review. Another limitation of this study is the inclusion of troughs drawn before the third dose per clinical practice at one of the study sites, and thus no steady state levels by the time the trough was drawn. IDSA guidelines recommend that steady state troughs be drawn before the fourth to fifth dose of vancomycin.¹⁰ This study also did not assess for the appropriateness of dosing regimens, which would dictate expected attainment of goal vancomycin troughs. Lastly, patients with prescribed medications that could have caused a false positive for amphetamine on the UDS were not excluded.^32,33 However, this would have been difficult to carry out since access to external prescription history is not warranted by all insurance programs, and medication reconciliation evaluations are not always accurately documented in the electronic medical record.³⁴

Conclusion

Recent methamphetamine use may increase vancomycin clearance. Larger prospective trials with protocolized vancomycin dosing strategies are needed to further elucidate the impact of methamphetamine use on attainment of goal vancomycin troughs in addition to the potential impact on vancomycin clearance.