Author + information
- Susan X. Zhao, MDa,∗ (, )
- Calvin Kwong, MDb,
- Aravind Swaminathan, MDa,
- Amit Gohil, MDc and
- Michael H. Crawford, MDd
- aDivision of Cardiology, Santa Clara Valley Medical Center, San Jose, California
- bDepartment of Medicine, Santa Clara Valley Medical Center, San Jose, California
- cDivision of Pulmonary and Critical Care Medicine, Santa Clara Valley Medical Center, San Jose, California
- dDivision of Cardiology, University of California-San Francisco, San Francisco, California
- ↵∗Address for correspondence:
Dr. Susan X. Zhao, Santa Clara Valley Medical Center, 751 S. Bascom Avenue, Suite 340, San Jose, California 95128.
Objectives This study sought to characterize patients with methamphetamine-associated pulmonary arterial hypertension (MA-PAH) and cardiomyopathy (MA-CMP), to compare with MA controls (MA-CTL), users with structurally normal hearts, with the aim of identifying risk factors for these conditions.
Background MA-PAH and MA-CMP are 2 poorly understood cardiac complications in MA users.
Methods We retrospectively studied the clinical characteristics and outcomes of 50 MA-PAH, 296 MA-CMP, and 356 MA-CTL patients, whom we evaluated between 2010 and 2017.
Results After a median follow-up of 20.0 months (interquartile range [IQR]: 7.6 to 42.6 months), all-cause mortality was 18.0% for MA-PAH, 15.2% for MA-CMP, and 4.5% for MA-CTL group (p < 0.001). More women (58%) were in the MA-PAH group than in the MA-CMP (14%; p < 0.001) and MA-CTL (42%; p = 0.028) groups, whereas the MA-CMP group was predominantly male (86% vs. 58% in the MA-CTL group; p < 0.001). More MA-CMP patients had hypertension (p < 0.001) or alcoholism (p < 0.001) than MA-CTL patients. Logistic regression analyses identified male sex, alcoholism, and hypertension as independent factors associated with MA-CMP with the following respective adjusted odds ratios (OR) of 3.791 (95% confidence interval [CI]: 2.508 to 5.730), OR of 2.959 (95% CI: 2.084 to 4.203), and OR of 2.111 (95% CI: 1.486 to 2.999), whereas female sex was the only factor associated with MA-PAH.
Conclusions Both MA-PAH and MA-CMP patients carried significant disease burden and mortality risk. Male sex, hypertension, and alcoholism were strongly associated with MA-CMP, whereas female sex and other unknown factors may influence development of MA-PAH. This study adds to the understanding of MA-associated cardiac complications and highlights directions for future investigation.
The use of methamphetamine (MA) is widespread, and in many countries, MA is the drug of major abuse. MA exerts its cardiotoxic effects by stimulating the release of catecholaminergic neurotransmitters in both the central and peripheral nervous systems, which modulates heart rate and blood pressure (1–4).
MA has been implicated in various cardiac pathologies, 2 of which are pulmonary arterial hypertension (PAH) and dilated cardiomyopathy (CMP). MA and other stimulant use was found in up to 30% of cases of PAH (5). MA use is also a known cause of dilated CMP (6–14). To the best of our knowledge, no studies have compared the clinical characteristics and overall impact on survival of these 2 distinct pathologies (CMP vs. PAH) among MA users.
We hypothesized that characteristics of MA users with PAH (MA-PAH) were different from those with MA-associated CMP (MA-CMP) and a control group of MA users with neither condition (MA-CTL). Information gained from this study may provide insight into the underlying mechanisms of these 2 distinct forms of MA-related cardiac complications and aid the public health efforts in battling this surging epidemic.
A retrospective chart review of the echocardiogram database and electronic medical records was performed to identify all patients with documented MA abuse, using the International Classification of Diseases-10th Revision-Clinical Modification (ICD-10-CM) code F15 (any diagnosis related to methamphetamines) and/or positive urine toxicology at Santa Clara Valley Medical Center, San Jose, California, between January 2010 and June 2017. MA-PAH was defined as right ventricular systolic pressure (RVSP) >45 mm Hg on index transthoracic echocardiogram (TTE) without other causes of pulmonary hypertension. Specifically, other causes of pulmonary hypertension (World Health Organization groups 1 to 5) including chronic thromboembolic pulmonary hypertension, connective tissue disease, congenital heart disease, human immunodeficiency virus infection, and liver cirrhosis due to hepatitis C were sought to the extent possible in this patient population, and those with established other causes were excluded. MA-CMP was defined as left ventricular ejection fraction (LVEF) <40% on index TTE. Patients with known causes of CMP, such as coronary artery disease (with history of percutaneous coronary intervention or coronary artery bypass graft surgery or significant regional perfusion defect on nuclear perfusion imaging), primary valvular heart disease (either native or prosthetic valves), active hyperthyroidism, and metastatic malignancy were excluded. Of note, patients were also excluded if CMP was present on initial TTE, yet LVEF improved to ≥50% upon follow-up. Documented MA users with normal heart structure, function, and estimated RVSP on index TTE were included in the MA-CTL group.
Clinical and demographic data collected from the 3 groups of patients (MA-PAH, MA-CMP, and MA-CTL) included age at time of index TTE, sex, ethnicity, body mass index, body surface area, urine toxicology, known route and duration of MA use, homelessness, use of tobacco products, alcohol abuse, and other drug abuse as well as other medical comorbidities (hypertension, diabetes mellitus, stroke/transient ischemic attack, atrial fibrillation/flutter, or chronic kidney disease). Therapies patients were receiving for MA-CMP (beta-blockers, loop diuretics, angiotensin-converting enzyme inhibitor/angiotensin receptor antagonists, mineralocorticoid receptor antagonists, implantable cardioverter-defibrillator placement) and for MA-PAH (phosphodiesterase-5 inhibitors, endothelin receptor antagonists, prostacyclin pathway agonists) were reported as well. Echocardiographic parameters assessed included: LVEF (%), left ventricular end-diastolic volume index (ml/m2), left ventricular end-systolic volume index (ml/m2), left atrial volume index (ml/m2), right atrial volume index (ml/m2), right ventricular end-diastolic area (cm2), right ventricular end-systolic area (cm2), right ventricular fractional area change (%), and RVSP (mm Hg) according to published standards (15,16). Death was ascertained by chart documentation as well as cross-referencing Social Security Death Index.
All continuous variables, presented as mean ± SD or medians with interquartile range (IQR) as appropriate, were compared by using analysis of variance (ANOVA) or nonparametric analysis of variance (Kruskal-Wallis test) with Bonferroni post hoc testing. Categorical variables were expressed as percentages and were compared with the use of chi-square test or Fisher exact test, as appropriate.
Multivariate logistical regression analysis with backward elimination was performed to determine covariates of MA-PAH or MA-CMP by starting with baseline univariate predictors with a p value of <0.10. Time-to-event curves were analyzed according to the Kaplan-Meier method, and group comparisons were made by applying the log-rank test. For all tests, significance was accepted as a p value <0.05. All statistical analyses were performed using SPSS version 22 software (IBM, Armonk, New York). The Institutional Review Board of the Santa Clara Valley Medical Center approved the study protocol.
A total of 4,622 consecutive patients with a history of MA use were identified during the study period. After exclusion criteria were applied, 296 patients were included in the MA-CMP group; 50 patients were included in the MA-PAH group; and 356 MA users with structurally normal heart and normal estimated RVSP on index TTE constituted the MA-CTL group. Cohort selection flowchart is shown in Figure 1.
Table 1 outlines the baseline information for these 3 groups of patients. Most MA users were mid to late 40 years of age. There were no statistically significant differences in age or body sizes as represented by body mass index. The racial makeup in all 3 MA groups shared a similar distribution within the main racial/ethnic groups in Santa Clara County, except for a paucity of Asian Americans.
There was a strong female predominance in the MA-PAH group (58%) compared to 14% in the MA-CMP group (p < 0.001) and 42% in the MA-CTL group (p = 0.028), whereas the MA-CMP group was made up of predominantly males (p < 0.001 compared to either MA-PAH or MA-CTL groups).
More MA-CMP patients had concurrent alcohol abuse (59%) than the other 2 groups (33% in the MA-CTL group; p < 0.001; and 40% in the MA-PAH group; p = 0.007). Concurrent cocaine abuse was found in 17% of MA-CMP patients, 16% of MA-PAH patients, and 10% of MA-CTL patients (MA-PAH vs. MA-CMP: p = 0.907; MA-PAH vs. MA-CTL: p = 0.184; MA-CMP vs. MA-CTL; p = 0.010). There were fewer smokers in the MA-PAH group (74%) compared to the MA-CTL group (83%; p = 0.011) or the MA-CMP group (84%; p = 0.034).
Comorbidities were prevalent within the entire cohort, with overall 32% homelessness, 56% hypertension, 21% diabetes, 12% atrial fibrillation/atrial flutter, 11% stroke/transient ischemic attack, and 20% patients with chronic kidney disease stage ≥3. More patients in the MA-CMP group carried the diagnoses of hypertension, stage 3 and greater chronic kidney disease, and atrial fibrillation/flutter than the other 2 MA groups, whereas the rates for diabetes and prior stroke/transient ischemic attack were not statistically different. Of the 296 patients with MA-CMP, 228 (77%) were hospitalized for heart failure at the time of index TTE, and 168 (57%) were found to have persistent New York Heart Association (NYHA) functional class III or IV heart failure during follow-up. Disease-specific therapies for MA-PAH and MA-CMP are listed in Table 1.
Echocardiographic and Doppler indices for the MA users are presented in Table 2. Mean LVEF for the MA-CMP group was 25.2 ± 6.5%, whereas the other 2 groups had preserved LV systolic function (p < 0.001). The left ventricles and left atria in the MA-CMP group were significantly more dilated than those in the other 2 groups, as manifested by increased LV end-diastolic volume index, LV end-systolic volume index, and LA volume index (p < 0.001). In contrast, these left-heart indices in the MA-PAH group were significantly smaller than those in the MA-CTL group. The right-heart sizes, as represented by RV end-diastolic area, RV end-systolic area, and right atrial volume index, were highest in the MA-PAH group, followed by those in the MA-CMP group, and followed by those in the MA-CTL group (Table 2). Median RVSP was 75 mm Hg (IQR: 58 to 91 mm Hg) for MA-PAH, 45 mm Hg (IQR: 37 to 53 mm Hg) for MA-CMP, and 26 mm Hg (IQR: 22 to 33 mm Hg) in the MA-CTL group (p < 0.001).
Twenty patients (15 women and 5 men) from the MA-PAH group underwent right-heart catheterization prior to being considered for PAH-specific therapy. Their ages, sex, pre- and post-inhaled nitric oxide challenge right-heart catheterization data are listed in Table 3. Median mean pulmonary arterial pressure was 46 mm Hg (IQR: 41 to 51 mm Hg), median pulmonary vascular resistance was 11.8 Wood units (IQR: 9.0 to 15.3 Wood units). No patients manifested positive vasoreactivity as previously defined (17,18).
During median follow-up of 20.0 months (IQR: 7.6 to 42.6 months), there were 16 deaths (4.5%) from any cause in the MA-CTL group, 9 (18.0%) in the MA-PAH group, and 45 (15.2%) in the MA-CMP group. Kaplan-Meier curves for cumulative survival are presented in Figure 2A (all 3 groups: p < 0.001) and Figure 2B (MA-PAH and MA-CMP groups only: p = 0.697).
Table 4 outlines univariate and multivariate logistical regression modeling of the clinical features associated with MA-CMP compared with those of the MA-CTL group. Atrial fibrillation/flutter, chronic kidney disease, and history of stroke/transient ischemic attack were not included, as these disease entities are known to occur downstream of CMP and clinical heart failure. Male sex (odds ratio [OR]: 3.791; 95% confidence interval [CI]: 2.508 to 5.730; p < 0.001), alcohol abuse (OR: 2.959; 95% CI: 2.084 to 4.203; p < 0.001) and hypertension (OR: 2.111; 95% CI: 1.486 to 2.999; p < 0.001) were strongly associated with the MA-CMP phenotype, whereas homelessness (OR: 0.433; 95% CI: 0.296 to 0.633; p < 0.001) reduced its odds. No significant interactions were found between these variables. Good model fit was supported by case classification rate of 69.8% (threshold: >65%), as well as a Hosmer-Lemeshow significance value of 0.655 (threshold: >0.05).
Table 5, similarly, compares univariate and multivariate analyses of predictors for MA-PAH with those of the MA-CTL group. In contrast to the MA-CMP model, only 2 variables (sex and diabetes) were retained by the multivariate model, of which only 1 variable (sex) achieved statistical significance. Male sex significantly reduced the odds (OR: 0.493; 95% CI: 0.268 to 0.907; p = 0.023), whereas female sex emerged as a powerful and sole predictor of developing MA-PAH. Good model fit was shown by a case classification rate of 87.9% and a Hosmer-Lemeshow significance value of 0.658.
Methamphetamine abuse is a serious problem worldwide. In the United States, methamphetamine has evolved from the favorite drug of the working poor to gradually permeating the socioeconomic spectrum, from being restricted to a few circumscribed areas in western United States to middle America and beyond. The demographic profile of MA users has also expanded to include women and minorities (19). In our cohort, 53% of all MA users were white, comparable to the overall population of 55% white in Santa Clara County; and 38% were Hispanic/Latino, compared to the general population of 26% (20). Although Asians make up 37% of Santa Clara County, only 3% of MA users were Asian. African American MA users represented 4% of the study population compared to 3% in the general population. Women accounted for 31% of all MA users, compared to 50% in the general population. Thus, in our study, white and Hispanic/Latino men still dominated among all MA users, but the gender gap is closing.
Although it is widely recognized that MA exerts its cardiotoxicity through its sympathomimetic properties, causing coronary vasoconstriction, tachycardia, hypertension, and/or direct myocardial toxicity, little is known about the pathogenesis and risk factors for dilated CMP associated with chronic MA abuse. Case series (6–10) are small, with the largest cohort including only 107 cases (6), and are descriptive. Even less in known about the PAH phenotype associated with MA abuse (5,21–23). The study by Chin et al. (5) evaluated 28 PAH patients with a history of stimulant (including MA, amphetamine, and/or cocaine) abuse. Our study is the largest cohort to date for both conditions, including 296 nonischemic, nonvalvular dilated CMP patients, and 50 PAH patients after excluding other known pulmonary hypertension causes, matched with a contemporary control group of 356 patients with documented history of MA abuse and structurally normal hearts without elevated pulmonary pressures.
Both the MA-CMP and the MA-PAH groups had advanced disease with significant morbidity and mortality. The mean LVEF in the MA-CMP group was 25.2 ± 6.5% with dilated left and right heart chambers, whereas the MA-PAH group had median RVSP of 75 mm Hg (IQR: 58 to 91 mm Hg) with dilated right heart and significantly reduced RV systolic function. At a median follow-up of 20 months, all-cause mortality was 15.2% in the MA-CMP group and 18.0% in the MA-PAH group, both of which were significantly higher than that of the control group (4.5%). While 57% of patients with MA-CMP had NYHA functional class III/IV heart failure during follow-up visit, only 14% of this group of patients received implantable cardioverter-defibrillator placement for prevention of sudden arrhythmic death, likely due to a combination of reasons including ongoing substance abuse, poor functional status/prognosis, and/or issues with compliance.
Clinical features strongly associated with MA-CMP included male sex, concurrent alcohol abuse, and presence of systemic hypertension as demonstrated by logistical regression analysis. Male predominance in MA-CMP has been consistently shown in previous series, ranging from 64% to 93% (6–10). As with all illicit drug abusers, MA users consume a wide variety of other substances (24). The importance of concomitant use of other illicit agents with MA is that, when combined with alcohol or cocaine, methamphetamine toxicity is increased (25,26). Although cocaine co-abuse was not found to be a major determinant in our cohorts of patients, alcohol abuse emerged as a strong indicator of MA-CMP phenotype with an OR of 2.959 (95% CI: 2.084 to 4.203; p < 0.001). Alcohol enhances the discriminative stimulus effects of MA primarily by the N-methyl-d-aspartate receptors, thus providing a rationale for why people may choose to co-abuse alcohol and MA (27). Indeed, 59% of MA-CMP patients also abused alcohol, significantly higher than those in the MA-PAH group (40%) and the MA-CTL group (33%). The combination of MA and alcohol does not produce a new psychoactive substance but does increase heart rate and blood pressure beyond those seen in users of methamphetamine alone (25,28). This pharmacological interaction may explain the significantly higher prevalence of systemic hypertension in the MA-CMP group.
Homeless patients had a reduced odds of developing MA-CMP, with an OR of 0.433 (95% CI: 0.296 to 0.633; p < 0.001). This finding is unexpected and at first glance appears counterintuitive as one might expect a more serious substance abuse problem in the homeless population. One plausible explanation for this observation is not biological but socioeconomic, as limited financial means probably limit access to MA, cocaine, as well as alcohol in this group of patients.
Tobacco use was higher in the MA-CMP and MA-CTL patients than in the MA-PAH subjects, but overall usage was very high (84%, 83%, and 74%, respectively) compared to that reported (10% to 21%) in the general adult population of Santa Clara County (29). However, tobacco use did not predict MA complications. The high rate of tobacco use in our cohort substantiates the observation that MA use itself was a marker for other high-risk behaviors (such as smoking, alcoholism, use of other illicit drugs, homelessness, and so forth) that led to an elevated all-cause mortality rate of 4.5% over a median of only 20 months of follow-up in the MA-CTL group.
In contrast to the MA-CMP group, the only clinical factor shown to be significantly associated with MA-PAH was female sex. Women made up 58% of the MA-PAH group, 14% of the MA-CMP group, and 42% of the MA-CTL group. This distribution is consistent with the overall two thirds predominance of females in World Health Organization group 1 PAH subjects. Unlike men, female MA users tend to report that their initial MA use stemmed from the desire to lose weight or to cope with depression (19). Due to the retrospective nature of our study, exposure to anorexic or other psychotropic agents cannot be accurately assessed for the patients in our cohort. It has been previously reported that medications known to increase the risk of PAH (e.g., fenfluramine, d-fenfluramine, and aminorex) share the common ability to serve as potent substrates for serotonin transporter protein (30). Illicit MA, when self-administered by intravenous or inhalation route, would be expected to achieve serum levels high enough to stimulate serotonin transporter protein and lead to increase in 5-hydroxytryptamine levels in vivo and eventually lead to development of PAH (31). Another mechanistic link between MA and PAH was explored by Orcholski el al. (23), who performed whole-genome sequencing in 18 MA-PAH patients (11 females, 7 males). A total of 94.4% of these patients were found to be heterozygous for a single-nucleotide variant predicted to reduce the activity of carboxylesterase 1, an enzyme involved in metabolism of MA and most amphetamine derivative drugs in pulmonary endothelial cells. Reduced expression or activity of this key enzyme may lead to altered drug metabolism and resultant chronic organ injury. Thus, there is a possible common molecular underpinning of PAH-associated with either MA or anorexigens such as fenfluramine (a chemical analogue of amphetamine) in genetically susceptible individuals.
As illustrated in Figure 3, the scale is tipped toward the MA-CMP phenotype when the MA user is male, with concomitant alcohol abuse and presence of systemic hypertension, whereas female sex and some other unknown factors, which may be clinical and/or genetic, predispose to the development of MA-PAH.
Limitations of our study include the obligatory weaknesses associated with retrospective design. Potential confounding was minimized by using a logistic regression model but not fully eliminated. Additionally, information about the duration and route of methamphetamine use, which may play a key role in determining which cardiovascular complication subtype a given patient is susceptible to, was incomplete in our cohort. Although there is no evidence to support the idea that any one route of MA administration is more strongly associated with cardiotoxicity than another, the rates of cardiovascular pathologies are likely to be higher with routes that deliver a higher dose or with frequent use (e.g., injecting and smoking) (32). To date, however, there have been no studies investigating the relative risk of cardiovascular complications associated with different forms of MA administration. Finally, data were gathered from a single center and may not be applicable to other populations. However, the advantages of a single-center study include uniformity in clinical practice, data acquisition, and interpretation.
In this retrospective analysis of MA users from a large public hospital in northern California, 296 patients with MA-CMP and 50 patients with MA-PAH were identified. Compared with 356 MA users with structurally normal hearts and normal pulmonary pressures, MA-CMP and MA-PAH patients had significantly higher morbidity and mortality. Predisposing factors for MA-CMP included male sex, history of alcohol abuse, and presence of systemic hypertension. The only risk factor associated with MA-PAH was female sex. Future research is needed to establish a mechanistic relationship between MA use and these distinct forms of cardiac pathology, as well as the interaction of MA with other concurrent risk factors (such as alcohol abuse) or pre-existing cardiac conditions (such as hypertension).
COMPETENCY IN MEDICAL KNOWLEDGE: Pulmonary arterial hypertension and dilated CMP are two cardiovascular sequelae-associated with methamphetamine abuse. Clinical features of these two groups of patients are distinct from each other, while both carry significant disease burden and mortality risk.
TRANSLATIONAL OUTLOOK: From a clinical and public health perspective, this study identifies risk factors-associated with the development of MA-PAH and MA-CMP and will help guide efforts in prevention, counseling and caring for the at-risk individuals. Future research in genetic, biochemical and molecular mechanisms may shed light on the differential pathogenesis of these two disease entities.
The authors gratefully acknowledge contributions from Jennifer Tong, MD, Lawrence Dang, PharmD, and John Guillory for querying the databases at Santa Clara Valley Medical Center.
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- left ventricular ejection fraction
- methamphetamine-associated pulmonary arterial hypertension
- methamphetamine-associated cardiomyopathy
- methamphetamine-associated control
- right ventricular systolic pressure
- transthoracic echocardiogram
- Received August 14, 2017.
- Revision received October 5, 2017.
- Accepted October 7, 2017.
- Schürer S.,
- Klingel K.,
- Sandri M.,
- et al.
- Voskoboinik A.,
- Ihle J.F.,
- Bloom J.E.,
- Kaye D.M.
- Jafari Giv M.
- Lang R.M.,
- Badano L.P.,
- Mor-Avi V.,
- et al.
- Rudski L.G.,
- Lai W.W.,
- Afilalo J.,
- et al.
- Opitz C.,
- Rosenkranz S.,
- Ghofrani H.A.,
- et al.
- McLaughlin V.V.,
- Archer S.L.,
- Badesch D.B.,
- et al.
- ↵U.S. Census Bureau. QuickFacts selected: Santa Clara County, California. Available at: https://www.census.gov/quickfacts/fact/table/santaclaracountycalifornia/PST045216. Accessed September 28, 2017.
- Toolaney G.H.
- Darke S.,
- Degenhardt L.,
- Mattick R.
- ↵Tobacco Use in Santa Clara County—tobacco-use. Available at: https://www.sccgov.org/sites/phd/hi/hd/Documents/tobacco-reports/tobacco-use.pdf. Accessed September 28, 2017.
- Rothman R.B.,
- Ayestas M.A.,
- Dersch C.M.,
- Baumann M.H.
- Kaye S.,
- McKetin R.