Skip main navigation

Socioeconomic Burden of Rising Methamphetamine-Associated Heart Failure Hospitalizations in California From 2008 to 2018

Originally publishedhttps://doi.org/10.1161/CIRCOUTCOMES.120.007638Circulation: Cardiovascular Quality and Outcomes. 2021;14

Abstract

Background:

Methamphetamine-associated cardiomyopathy/heart failure (MethHF) is an increasingly recognized disease entity in the context of a rising methamphetamine (meth) epidemic that most severely impacts the western United States. Using heart failure (HF) hospitalization data from the Office of Statewide Health Planning and Development, this study aimed to assess trend and disease burden of MethHF in California.

Methods:

Adult patients (≥18 years old) with HF as primary hospitalization diagnosis between 2008 and 2018 were included in this study. The association with Meth (MethHF) and those without (non-MethHF) were determined by meth-related International Classification of Diseases-based secondary diagnoses. Statistical significance of trends in age-adjusted rates of hospitalization per 100 000 adults were evaluated using nonparametric analysis.

Results:

Between 2008 and 2018, 1 033 076 HF hospitalizations were identified: 42 565 were MethHF (4.12%) and 990 511 (95.88%) were non-MethHF. Age-adjusted MethHF hospitalizations per 100 000 increased by 585% from 4.1 in 2008 to 28.1 in 2018, while non-MethHF hospitalizations decreased by 6.0% from 342.3 in 2008 to 321.6 in 2018. The rate of MethHF hospitalization increase more than doubled that of a negative control group with urinary tract infection and meth-related secondary diagnoses (7.82-fold versus 3.48-fold, P<0.001). Annual inflation–adjusted hospitalization charges because of MethHF increased by 840% from $41.5 million in 2008 to $390.2 million in 2018, as compared with an 82% increase for all HF hospitalization from $3.503 billion to $6.376 billion. Patients with MethHF were significantly younger (49.64±10.06 versus 72.20±14.97 years old, P<0.001), predominantly male (79.1% versus 52.4%, P<0.001), with lower Charlson Comorbidity Index, yet they had longer length of stay, more hospitalizations per patient, and more procedures performed during their stays.

Conclusions:

MethHF hospitalizations increased sharply during the study period and contributed significantly to the HF hospitalization burden in California. This emerging HF phenotype, which engenders considerable financial and societal costs, calls for an urgent and concerted public health response to contain its spread.

WHAT IS KNOWN

  • Methamphetamine-associated cardiomyopathy and heart failure (MethHF) is increasingly encountered clinically in the context of a rapidly spreading methamphetamine epidemic.

WHAT THE STUDY ADDS

  • MethHF hospitalization rate in California rose by 585% from 2008 to 2018 (compared with 6.0% decrease in non-MethHF and 0.92% increase for all heart failure hospitalizations) and was associated with an 840% increase in hospitalization charges (compared with 82% for all heart failure hospitalizations).

  • When compared with non-MethHF, MethHF was associated with higher hospitalization charges, longer length of stay, and more procedures performed despite a younger demographic (94.5% <65 year old) with fewer traditional cardiovascular comorbidities.

  • Our findings highlight the gravity of MethHF burden in the state of California and possibly other parts of the United States hard hit by high methamphetamine use. Urgent efforts are needed to counter its detriment on clinical outcomes as well as impact on resource utilization.

See Editorial by Reddy and Elkayam

The use of methamphetamine (meth), a potent and highly addictive stimulant, remains an extremely serious problem in the United States and worldwide.1 Amphetamine-related hospitalizations in the United States tripled between 2008 and 2015, with most of these hospitalizations occurring in the western United States.2 California, a state that has been the focus of considerable meth research, is seeing a resurgence of wide-spread meth use.3

Other than acute overdose, long-term meth use has been associated with various cardiac pathologies1,4: malignant hypertension, tachyarrhythmias, myocardial ischemia, and most notably, meth-associated cardiomyopathy/heart failure (MethHF), which has emerged as a new phenotype of heart failure (HF) with a substantial increase in HF clinical encounters.5–8

There remains a paucity of data regarding the scope of MethHF as well as its impact on cost and resource utilization in California. We hypothesized that MethHF hospitalizations increased significantly over the past decade and may now feature more prominently in the statewide HF hospitalization burden than previously anticipated. We set out to study the trend, burden, and spread pattern of MethHF through the inpatient hospitalization data made available by State of California Health and Human Services Agency’s Office of Statewide Health Planning and Development (OSHPD).

Methods

Availability of Data

In accordance with the American Heart Association’s Transparency and Openness Promotion Guidelines, methods and aggregate data are available from the corresponding author upon reasonable request. Because of the sensitive nature of the data collected for this study, request to access the dataset from qualified researchers trained in human subject confidentiality protocols may be sent to OSHPD at: [email protected].

Study Design and Data Sources

California requires all California-licensed hospitals in the state to report all hospitalizations and emergency department visits as well as ambulatory surgical encounters to OSHPD.

Patient discharge data (PDD) files are administrative files maintained by OSHPD that provide each inpatient hospitalization’s basic demographics such as age, sex, county, zip code or residence, race/ethnicity, hospital county/zip codes, adjusted length of stay, primary and other procedures performed, source of payment and total charges (US $) for the stay, as well as International Classification of Diseases (ICD) based primary diagnoses and 24 other diagnoses, patient disposition including in-hospital mortalities. These files support various research activities and public health endeavors by providing comprehensive access to temporal relationship, clinical outcomes, and resource utilization related to specific disease entities.9–11

Study Cohort

We conducted a retrospective cohort study analyzing data from OSHPD’s PDD database on all adult patients (aged ≥18 years) discharged between 2008 and 2018 with a diagnosis of HF with ICD codes for HF (Table I in the Data Supplement). OSHPD transitioned from ICD-9 to ICD-10 in the fourth quarter of 2015. Accordingly, both ICD-9 and ICD-10 codes were used. A primary HF hospitalization was defined as any HF ICD-9 or ICD-10 code used as the principal diagnosis (field name diag_p in PDD), the condition established to be the chief cause of the admission of the patient to the hospital for care.

HF ICD codes were cross-referenced with meth-related ICD codes in any diagnosis position: ICD 9 codes 304.4x (Amphetamine and other psychostimulant dependence), 305.7x (Nondependent amphetamine or related acting sympathomimetic abuse), and 969.72 (Poisoning by amphetamine) were used from 2008 to the first 3 quarters of 2015; ICD 10 code F15 were used from fourth quarter of 2015 to 2018. These codes include other psychostimulants such as amphetamines, MDMA, phenmetrazine, and other unspecified amines and related drugs, but are likely dominated by meth. According to Treatment Episode Data Set-Admissions, a national data system of annual admissions to substance abuse treatment facilities, treatment episodes listing meth as the primary drug of choice among all stimulant uses grew from 75% in 1995 to 93% in 2008 to 96% in 2017, whereas the number of treatment episodes for non-meth stimulants declined during this time.12–14 As meth dwarfs other stimulants in its prevalence, these ICD codes have been used in multiple prior studies to denote meth.8,12,15

The aggregate, nonpatient-level data of hospitalizations with meth at any diagnosis position, and a condition with no direct pathological relationship with meth use, in this case, urinary tract infection, were also separately obtained from OSHPD, to serve as negative controls for MethHF.

Outcomes

We studied the following outcomes of statewide MethHF as compared with non-MethHF and all HF hospitalizations: (1) age-adjusted HF hospitalization rates per 100 000 adults; (2) crude HF hospitalization rate based on race/sex subgroups; (3) inflation-adjusted annual HF hospitalization charges; and (4) temporal and geospatial patterns of distribution of MethHF hospitalizations.

Age-adjusted HF hospitalization rates per 100 000 adults were developed per CDC methodology16 using the US standard population available at National Institute of Health/National Cancer Institute.17 California projected and estimated statewide populations were obtained from California Department of Finance websites: population projections18 for 2010 to 2018 and population estimates for years 2008 and 2009.19

Direct age standardization for different racial groups was not performed due to the lack of detailed age/sex/racial data in the projected population from 2010 to 2018. Therefore, crude rates (defined as the number of count, divided by the mid-year total population of the selected demographic and multiplied by 10 000), rather than age-adjusted rates, were reported.

Demographic and clinical features of each HF hospitalization were analyzed as per standard. We used PDD’s normalized race group (race_grp) variable to report a patient’s race, which was based on a combination (merged) of their self-reported race and ethnicity.20 If a patient’s ethnicity was “Hispanic”, then the normalized race group was assigned “Hispanic”. For all other ethnicity, the normalized race group was the same as self-reported race categories: American Indian or Alaska Native, Asian, Black or African-American, Native Hawaiian or Other Pacific Islander, Other, Unknown/blank/invalid; due to the small numbers of MethHF in the smaller race groups, we consolidated the data into 4 major race groups: White, Black, Hispanic and others.

Three indices of comorbidities were computed: (1) a simple count of all ICD-based secondary diagnoses for each hospitalization; (2) Charlson Comorbidity Index; and (3) Medicare Severity-adjusted Diagnosis Related Group severity code: The Medicare-Severity-adjusted Diagnosis Related Group system uses “major complication/comorbidity” and “complication/comorbidity” diagnosis codes to identify the severity of inpatient cases. PDD data specify that complications/comorbidities are reserved for patients with significant acute diseases, acute exacerbation of chronic diseases, advanced or end-stage chronic diseases, or chronic diseases associated with extensive debility. Major complication/comorbidities are reserved for the more severely ill patients with life-threatening conditions.

Because OSHPD data do not provide actual reimbursed dollars for services, we used submitted charges as an estimated cost of HF hospitalizations. We computed 2 types of hospital charges: (1) annual hospitalization charges for MethHF and All HF; and (2) mean charge per patient over the 11-year time span. Annual charges were Consumer Price Index inflation adjusted to 2018 value by using the California Department of Industrial Relations’ California Consumer Price Index.21 For example, the inflation adjusted value for year 2016=actual value for year 2016 x CPI2018/CPI2016.

To determine the geospatial spread and hot spots of MethHF hospitalizations, we used the ARC GIS system (Global Information System, ArcMAP 10.8, Esri, Redlands, CA) for thermatic mapping and its Getis-Ord Gi* analysis tool to identify spatial clustering of high values, with a prespecified fixed-distance band of 10 miles and an alpha level of 0.05.22 The Getis-Ord Gi* statistic looks at statistically significant spatial clustering (either high or low value) of MethHF hospitalizations.23 To be a statistically significant hot spot, a feature (in this case, population adjusted MethHF hospitalization) will not only have a high value at a particular location but will be surrounded by neighboring high values as well. The local sum for the feature and its neighbors is compared proportionally to the sum of all features in the whole state dataset. Getis-Ord Gi* statistics generates Z scores (standard deviations), P (statistical probabilities), and confidence levels that indicate whether feature values are statistically clustered. The outcome Z-score denotes how many standards of deviation the observed sum is different from the expected sum. The higher the Z-score, the less likely the observed hotspot clustering is due to random chance. For example, Z-score of 1.645 (1 SD above the mean) indicates a 90% confidence that the hotspot cluster is not due to chance (corresponding to P of 0.10), 1.960 (2 SD above the mean) being 95% confident (corresponding to P of 0.05), etc.

Statistical Analyses

Descriptive statistics were performed at the hospitalization level unless otherwise indicated for patient characteristics, comorbidities, and outcomes data. All continuous variables, presented as mean±SD, were compared by using the Student t test. Categorical variables were expressed as percentages and were compared with the use of χ2 test. For all tests, significance was accepted as a P<0.05.

Analysis of the statistical significance for trends was performed using nonparametric analysis with the Mann-Kendall trend test, implemented in R using the trend analysis and Kendall packages. Slope estimation was calculated using the Theil-Sen estimator, implemented in R using the trend analysis package.

Statistical analyses were performed using R version 3.3.3 software (R project, Vienna, Austria) and SPSS version 22 (IBM, Armonk, New York).

This study was approved by the Committee for the Protection of Human Subjects of the State of California’s Health and Human Services Agency.

Results

Querying OSHPD’s PDD from 2008 to 2018 yielded 4 808 506 adult (≥18 years old) HF hospitalizations, out of which 1 033 076 hospitalizations had HF listed as the primary diagnosis (Figure 1, study flow chart) based on ICD coding. Of these, 42 565 HF hospitalizations were associated with meth (4.12%, MethHF), while 990 511 (95.88%, non-MethHF) were not. Detailed data are available in Table II in the Data Supplement.

Figure 1.

Figure 1. Study flow chart. HF indicates heart failure; Meth, methamphetamine; and OSHPD, Office of Statewide Health Planning and Development.

Age- and population-adjusted annual HF hospitalizations by meth status are shown in Figure 2.

Figure 2.

Figure 2. Trend of age-adjusted heart failure (HF) hospitalizations per 100 000 population in the state of California, 2008–2018, by methamphetamine (meth) status.

Age-adjusted MethHF hospitalizations per 100 000 increased from 4.1 in 2008 to 28.1 in 2018 (relative percent change of 585%); or from 1.2% of All HF hospitalizations in 2008 to 8.0% in 2018 (relative percent change of 567%), trend slope 2.5, P trend, <0.001.

For all HF hospitalizations, age-adjusted hospitalization rate declined initially from 2008 to 2014 (slope −11.6, P trend 0.007), and then rose from 2014 to 2018 (slope 17.2, P trend 0.027), with an overall nonsignificant change from 346.4 to 349.6 per 100 000 population between 2008 and 2018 (overall trend slope was −1.4, with P trend, 0.87, relative percent change 0.92%).

Non-MethHF hospitalizations followed a similar pattern as all HF: a declining trend from 2008 to 2014 (slope −12.4, P trend 0.007) followed by a rise from 2014 to 2018 (slope 13.2, P trend 0.027), while overall there was no significant change from 342.3 to 321.6 per 100 000 population between 2008 and 2018 (slope of −4.1; P trend, 0.28 and relative percent change −6.0%).

With 2014 as the inflection point, all HF hospitalization reversed its decline in the preceding years and had a net Δ of 65.5 per 100 000 (from 284.1 in 2014 to 349.6 in 2018). During the same period, MethHF hospitalization increased by 14.4 per 100 000 (from 13.7 in 2014 to 28.1 in 2018). Therefore, 22% of the rise in all HF hospitalization was attributable to MethHF (14.4/65.5×100=22%).

To compare the increase in MethHF with the background increase in meth prevalence, urinary tract infection was chosen as a control group as it bears no apparent pathophysiological relationship with meth. As shown in Figure 3, hospitalizations with urinary tract infection as primary diagnosis and concurrent meth use closely resembled the trend of meth hospitalizations (meth ICD codes present at any diagnosis position), with a rise to 3.48- and 3.82-fold those of their respective 2008 hospitalizations at year 2018. The trend of MethHF hospitalizations, however, deviated significantly from these 2 control arms with 2018 hospitalizations reaching 7.82-fold that of 2008 (P<0.001).

Figure 3.

Figure 3. MethHF (HF as primary diagnosis with meth at any diagnosis position), methUTI (urinary tract infection as primary diagnosis with meth at any diagnosis position), and meth (meth at any diagnosis position) hospitalizations, as indexed to their respective 2008 baselines, in California, from 2008 to 2018.

MethHF group, as shown in Table 1, was different from non-MethHF group in terms of demographic and clinical characteristics: patients with MethHF hospitalizations were significantly younger (mean age 49.64±10.06 versus 72.20±14.97 in non-MethHF, P<0.001; with 94.5% younger than 65 year old versus 30% in the non-MethHF group, P<0.001), predominantly male (79.1% versus 52.4%, P<0.001), and with fewer comorbidities based on all 3 indices evaluated: average number of secondary diagnoses 13.3±5.5 versus 14.4±5.8, P<0.001; 44.2% with Charlson Comorbidity Index ≥4 versus 88.9% in the non-MethHF group, P<0.001; and 20.0% without either complication/comorbidity or major complication/comorbidity based on the medicare-severity-adjusted diagnostic related groups severity coding system, versus 18.6% in the non-MethHF group, P<0.001. Consistent with this demographic profile, the MethHF group had significantly less severe traditional cardiovascular comorbidities including atrial fibrillation/atrial flutter, coronary artery disease, diabetes, hyperlipidemia, cerebrovascular disease/transient ischemic attack as well as peripheral vascular disease, with the only exception being hypertension, which was significantly higher in the MethHF group (32.7%) when compared with the non-MethHF group (30.5%, P<0.001). Consequently, unadjusted in-hospital mortality was lower for the MethHF cohort (0.66% versus 1.17%, P<0.001, Table 2). Social determinants of health, on the other hand, revealed notably higher rates of all substance use disorder categories: tobacco smoking, alcoholism, and use of marijuana as well as cocaine in the MethHF group, which was also significantly more likely to be homeless (14.5% versus 1.2%, P<0.001) and predominantly enrolled in California’s state-sponsored Medi-Cal program (63.7% of the MethHF cohort), while 69.6% of the non-MethHF group was enrolled in the federal Medicare program.

Table 1. Baseline Characteristics of HF Hospitalizations With- or Without-Associated Methamphetamine Use in the State of California, 2008–2018

MethHF
(n=42 565)
Non-MethHF (n=990 511)All HF
(n=1 033 076)
P value
Age, y49.64±10.0672.20±14.9771.27±15.46<0.001
Age group, y<0.001
 15–24262 (0.62%)2085 (0.21%)2347 (0.23%)
 25–343263 (7.67%)11 094 (1.12%)14 357 (1.39%)
 35–448911 (20.94%)31 056 (3.14%)39 967 (3.87%)
 45–5415 801 (37.12%)89 432 (9.03%)105 233 (10.19%)
 55–6412 003 (28.20%)163 433 (16.50%)175 436 (16.98%)
 65–742168 (5.09%)201 262 (20.32%)203 430 (19.69%)
 75–84141 (0.33%)255  579 (25.80%)255 720 (24.75%)
 85+16 (0.04%)236 570 (23.88%)236 586 (22.90%)
Men33 662 (79.1%)518 580 (52.4%)552 242 (53.5%)<0.001
Race<0.001
 White21 152 (49.7%)501 896 (50.7%)523 048 (50.6%)
 Hispanic10 345 (24.3%)217 719 (22.0%)228 064 (22.1%)
 Black7069 (16.6%)147  168 (14.9%)154 237 (14.9%)
 Other3999 (9.4%)123 728 (12.5%)127 727 (12.4%)
Medical history
 Cardiogenic shock1325 (3.1%)18 613 (1.9%)19 938 (1.9%)<0.001
 Atrial fibrillation/atrial flutter9772 (23.0%)455 943 (46.0%)465 715 (45.1%)<0.001
 CAD14 671 (34.5%)551 077 (55.6%)565  748 (54.8%)<0.001
 DM12 084 (28.4%)471 140 (47.6%)483 224 (46.8%)<0.001
 HTN13 921 (32.7%)301 827 (30.5%)315 748 (30.6%)<0.001
 HLD11 109 (26.1%)471 572 (47.6%)482 681 (46.7%)<0.001
 CVA/TIA4 643 (10.9%)111 871 (11.3%)116 514 (11.3%)0.014
 CKD11 193 (26.3%)490 795 (49.5%)501 988 (48.6%)<0.001
 PVD5937 (13.9%)241 429 (24.4%)247 366 (23.9%)<0.001
 COPD15 320 (36.0%)354 933 (35.8%)370 253 (35.8%)0.510
 Tobacco21 766 (51.1%)97 843 (9.9%)119 609 (11.6%)<0.001
 Alcohol8213 (19.3%)37 322 (3.8%)45 535 (4.4%)<0.001
 Marijuana6340 (14.9%)9456 (1.0%)15 796 (1.5%)<0.001
 Cocaine3840 (9.0%)15 269 (1.5%)19 109 (1.8%)<0.001
 Homelessness6193 (14.5%)11 829 (1.2%)18 022 (1.7%)<0.001
 Average number of secondary diagnoses13.3±5.514.4±5.814.3±5.8<0.001
 CCI≥418 821 (44.2%)880 443(88.9%)899 264(87.0%)<0.001
MS-DRG severity code<0.001
 Absence of CC or MCC8521 (20.0%)183 945 (18.6%)192 466 (18.6%)
 MCC15 950 (37.5%)489 800 (49.4%)505 750 (49.0%)
 CC18 094 (42.5%)316 766 (32.0%)334 860 (32.4%)
Primary payer type<0.001
 Medicare6630 (15.6%)689 833 (69.6%)696 463 (67.4%)
 Medi-Cal27 095 (63.7%)159 255 (16.1%)186 350 (18.0%)
 Private2528 (5.9%)101 903 (10.3%)104 431 (10.1%)
 Other6312 (14.8%)39 520 (4.0%)45 832 (4.4%)

Values are mean±SD or n (%). CAD indicates coronary artery disease; CC, complication/comorbidity; CCI, Charlson Comorbidity Index; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; CVA, cerebral vascular accident; DM, diabetes; HF, heart failure; HLD, hyperlipidemia; HTN, hypertension; MCC, major complication/comorbidity; MS-DRG, Medicare Severity-adjusted Diagnosis Related Group severity code; PVD, peripheral vascular disease; and TIA, transient ischemic attack.

Table 2. Clinical Outcomes and Resource Utilization of HF Hospitalizations With or Without Associated Methamphetamine Use in the State of California, 2008–2018

MethHF (n=42  565)Non-MethHF (n=990 511)All HF (n=1 033 076)P value
No. hospitalizations/patient7.77±10.543.98±6.274.13±6.54<0.001
LOS, d/patient12.47±19.239.79±32.49.86±32.11<0.001
Charge ($)/patient123 374±211 53587 896±189 03789 076±189 935<0.001
No. procedures performed/patient2.53±4.112.09±3.622.11±3.63<0.001
Disposition (per hospitalization)<0.001
 In-hospital mortality283 (0.66%)11 605 (1.17%)11 888 (1.15%)
 Routine or self-care32 551 (76.47%)540 058 (54.52%)572 609 (55.43%)
 Transferred to other hospitals430 (1.01%)10 947 (1.11%)11 377 (1.10%)
 Skilled nursing facility or custodial/supportive care1006 (2.36%)63 447 (6.41%)64 453 (6.24%)
 Home with home health care1920 (4.51%)92 428 (9.33%)94 348 (9.13%)
 Hospice192 (0.45%)11 880 (1.20%)12 072 (1.17%)

Values are mean±SD or n (%). HF indicates heart failure; and LOS, length of stay.

Case rate trends in MethHF hospitalizations differed significantly among the eight race/sex subgroups. After adjusting for demographic change in California, that is, decline in White population, growth in Hispanics with the overall Black population remaining steady, Black males demonstrated the steepest rise over the 11-year study span with a 12.48-fold increase (slope 10.12, P trend <0.001), followed by White males with a 6.35-fold increase (slope 3.7, P trend <0.001; Figure 4). Unlike White females, Hispanic females, and females of other race groups, which did not show any significant increase of MethHF hospitalization rate, Black females (slope 2.51, P trend <0.001) overtook Hispanic males (slope 2.0, P trend <0.001) as the third most rapidly growing demographic sub-group. The overall Black/White rate ratio rose from 1.38 in 2008 to 2.30 in year 2014 and has since plateaued at around this level (slope 0.11, P trend <0.001). Hispanic/White and Other race/White rate ratios did not change significantly.

Figure 4.

Figure 4. Methamphetamine-associated cardiomyopathy/heart failure hospitalizations per 100 000 population by race and sex subgroups.

In PDD files, each individual patient was assigned a unique 9-digit alphanumeric record linkage number, which was an encrypted form of the patient’s social security number. After excluding 1718 (4.0%) entries without a record linkage number in the MethHF group, 19 214 individual patients were identified, representing 7.77±10.54 hospitalizations/patient (Table 2). In the non-MethHF group, after 38 850 (3.9%) entries without unique record linkage number were excluded, a mean of 3.98±6.27 hospitalizations/patient was identified over the 11-year study period (P<0.001). The MethHF cohort had significantly longer length of stay, more procedures performed, and higher hospitalization charges than the non-MethHF cohort ($123 374±211 535 versus $87 896±189 037, P<0.001; Table 2).

Consumer Price Index-adjusted annual HF hospitalization charges rose from $3.503 billion in 2008 to $6.376 billion in 2018 for All HF hospitalizations (an 82% increase) and from $41.5 million in 2008 to $390.2 million in 2018 for MethHF hospitalizations (an 840% increase; Figure 5).

Figure 5.

Figure 5. Trends in California annual all heart failure (HF) and methamphetamine-associated cardiomyopathy/heart failure (MethHF) hospitalization charges (inflation-adjusted to 2018 $).

Due to the 9.6% of MethHF and 7.6% of non-MethHF patients with missing residential zip codes, we used county level data to construct thermatic maps of MethHF hospitalizations in all 58 California Counties in 3 representative years: 2008, 2013, and 2018 (Figure 6). Population-adjusted annual MethHF hospitalizations (Figure 6A) grew significantly statewide with the exception of a few sparsely populated counties. Higher density shifted away from urban population centers toward rural and agricultural regions of the State, with the highest density in the Butte/Yuba/Nevada tri-county area, followed by Sacramento/San Joaquin valley as well as the Fresno/Central valley areas. This observation was confirmed by Getis Ord Gi* hotspot analysis (Figure 6B), demonstrating the geographic shift of statistically significant hotspots (Orange color: 2 SDs above with 95% confidence; Red color: 3 SDs above with 99% confidence) of population-adjusted MethHF hospitalizations from the coastal areas in 2008 to more rural areas in Central California in 2013 and 2018.

Figure 6.

Figure 6. Geo-temporal spread of methamphetamine-associated cardiomyopathy/heart failure (MethHF) in California. A, Thermatic maps of population-adjusted annual MethHF hospitalizations by California counties, in 3 representative years: 2008, 2013, and 2018. B, Hot spot analysis using Getis Ord Gi* tool of population-adjusted annual MethHF hospitalizations by California counties, in 3 representative years: 2008, 2013, and 2018.

Discussion

Methamphetamine, the archetypal amphetamine-type stimulant, is one of the most commonly used illicit substances in the United States.24,25 With mounting evidence that long-term meth use leads to development of a severe form of dilated cardiomyopathy,1,26,27 MethHF is poised to become more prevalent in the midst of a growing methamphetamine epidemic worldwide. A recent analysis has shown that the US nation-wide MethHF hospitalizations increased by 204% between 2006 and 2014, in contrast to a decline in cocaine, and a marginal increase in alcohol related HF admissions.28 The highest disease burden was found to be concentrated in the Western Pacific region, which was responsible for 74.3% of national cases in 2014. Indeed, multiple case series describing the characteristics and outcomes of MethHF have come out of California including from our hospital system in the San Francisco Bay Area,5–8 yet statewide data are lacking about the disease burden, economic cost as well as temporal and spatial pattern of spread of this surging epidemic of methamphetamine-associated cardiomyopathy/heart failure. Research is urgently needed to address the scope, severity, and trajectory of this new disease entity in an effort to improve its recognition, management and prevention.

In this large, contemporary, population-based analysis of hospitalized adult HF patients in the state of California from 2008 to 2018, our findings confirm what clinicians across the state have long suspected: MethHF hospitalizations have increased by a striking 585% in the 11-year time span, in sharp contrast to the 6.0% decrease in non-MethHF, and far exceeding the insignificant 0.92% increase in all HF hospitalizations during the same period. The magnitude of the rise of statewide MethHF hospitalizations was consistent with a previous study from the San Diego Veteran’s Administration Medical Center, where primary MethHF encounters (both inpatient and outpatient) increased from 3.4% in 2006 to 6.7% in 2015.7 Similarly, another case series from Southern California has reported that the percentage of MethHF among new cardiomyopathy cases steadily increased from 1.8% in 2009 to 5.6% in 2014.29

To ascertain whether the rise in MethHF hospitalizations was largely attributable to an increase in the prevalence of meth coding in hospitalized patients, we compared MethHF to a negative control group of primary urinary tract infection hospitalizations with concurrent meth use and found that MethHF rose far in excess of what would have been expected from the increase in meth prevalence alone. By 2018, unadjusted MethHF hospitalization was 7.82-fold of its 2008 level, as opposed to 3.82-fold for MethUTI, the trajectory of which tracked nearly superimposably with that of meth hospitalizations (meth diagnosis code at any diagnosis position). The fact that the rate of increase in MethHF more than doubled that of background meth hospitalizations provides strong support of the association between meth use and heart failure.

Despite the considerable health and economic burden of HF hospitalizations and the intensive legislative, policy and public health attention in reducing them, there is no comprehensive national surveillance system to track HF-related health care and mortality burden in the United States. The CDC tracking system only reports HF hospitalization rates for Medicare beneficiaries over the age of 65,30 while 94.5% of the MethHF patients in our study were younger than age 65 and only 15.6% of them were enrolled in Medicare. MethHF hospitalizations, therefore, have long been “flying under the radar”, evading scrutiny and corrective measures. During the 2014 to 2018 period when the sharpest rise in MethHF coincided with the unexpected reversal of years of decline in all HF hospitalization rate, an alarming 22% of the increase of HF hospitalization was attributable to MethHF. Our data hopefully will bring the much-needed spotlight and sense of urgency to this under-appreciated severe form of cardiomyopathy and heart failure.

Commensurate with this surge in case numbers, inflation-adjusted MethHF hospitalization charges rose steadily from $41.5 million in 2008 (1.2% of all HF hospitalization charges) to $390.2 million in 2018 (6.1% of all HF hospitalization charges), an 840% increase over time (Figure 5), largely due to the high recidivism rate with a mean 7.77±10.54 hospitalizations per patient. In contrast, All HF hospitalization charges rose only by 82% during the same period. Additionally, despite being younger (94.5% of MethHF patients were younger than 65 year old as compared with 30% in the non-MethHF group) and with fewer traditional comorbidities, MethHF had longer hospital stay, more procedures performed and generated significantly higher charges per individual patient than non-MethHF (Table 2). Precision targeting this group of MethHF patients for HF readmission prevention by integrating substance use counseling and behavioral health services with existing infrastructure such as Hospital Readmission Reduction Program may prove to be a cost-effective and high-yield measure to lower HF event rate, reduce cost burden, and improve outcomes.

In addition to direct costs related to HF hospitalizations, our data showed that MethHF struck individuals in their most productive years. The largest age group in MethHF was 45 to 54 year olds (37.1% of the total MethHF cohort), followed by 55 to 64 year olds (28.2%), then 35 to 44 (20.9%), as opposed to the non-MethHF cohort where the largest age group was 75 to 84 year olds (25.8%), followed by 85+ age group (23.9%), followed by 65 to 74 year olds (20.3%) (Table 1). Though beyond the scope of this study, indirect costs due to loss of productivity and quality of life years incurred by MethHF were likely substantial and in urgent need of exploration and validation by future investigations.

Other than the younger age of patients with MethHF, prior case series have also reported a preponderance of cases among men, ranging from 64% to 86%.5,8,31,32 Our data are in support of this demographic pattern for MethHF, with men making up 79.1% of the MethHF cohort. Unlike many other illegal drugs, the rates of meth dependence seem to be roughly equivalent in men compared with women.1,24 It is therefore unclear whether the observed lower representation of females in MethHF was due to lower exposure from lower doses used, alternate route of administration, or the postulated protective effects of estrogen on meth metabolism and toxicity, as both animal studies and clinical observations have reported the greater male susceptibility to the effects of meth on the heart.33 Female meth users were also shown to have a more favorable response to behavioral therapy than male meth users34 and were more likely to reverse and normalize their heart function once abstinent.6 The traditional view that meth-related diseases affect predominantly low-income White men needs to be revised lest a growing trend of female MethHF patients go unnoticed and underserved, particularly in light of potentially more favorable outcomes in response to therapeutic and behavioral interventions in females.6,34

By absolute numbers, it may seem as if MethHF was a disease involving a predominantly White or Hispanic population, which made up 49.7% and 24.3% of the MethHF cohort, respectively (Table 1). After adjusting for CA state population demographics, however, the racial/ethnic disparity of MethHF became apparent: Black males led the 8 racial/sex subgroups with the steepest increase (Figure 4). In year 2018, MethHF hospitalization per 100 000 population in Black males was 2.44 times that of White males, 4.5 times that of Hispanic males, and 5.10 times that of males of other racial groups. Black female MethHF hospitalization rate far exceeded White, Hispanic, and other female groups and even surpassed that of Hispanic males to become the third rapidly rising demographic subgroup for MethHF hospitalizations. The burden of HF for Black Californians starts early, persists into older age, and has remained disproportionately higher than other racial groups since first reported in 1999.9,35 Although Non-Hispanic Black race/ethnicity was still less associated with past-year methamphetamine use (odds ratio, 0.29 [0.20–0.42]) when compared with non-Hispanic Whites race/ethnicity according to the 2015 to 2018 National Surveys on Drug Use and Health data,36 our findings highlight the rapidity of the rise and are consistent with previously reported racial/ethnic disparities in HF disease burden.37,38 The urgency for preventative programs and clinical services focusing on minority at-risk subgroups cannot be overstated. Through enhancing outreach and expanding access to behavioral therapies and other nonpharmacological approaches,25 the detriment of MethHF on minority groups can hopefully be preemptively contained. These demographic differences across age, sex, and race are striking but will require further confirmation as this study relied on administrative data. Administrative data are subject to ascertainment bias where certain groups may be tested for meth more frequently than others based on these characteristics.

Last, given the geographic, economic, and demographic diversity in the different regions of California, we also evaluated the spatial pattern of MethHF spread at the county level. Population-adjusted annual MethHF hospitalizations demonstrated alarming spatial temporal spread across the state as shown by Figure 6A. By absolute case numbers, it appeared to be a predominantly urban problem, with notable clustering around major urban centers (data not shown). With population adjustment, it became apparent that the highest density (MethHF cases per 100 000 population) was in rural parts of Central and Northern California. This pattern was confirmed statistically by the Getis Ord Gi* Hot Spot analysis (Figure 6B) and consistent with a prior study conducted by Gruenewald et al,12 which showed that between 1995 and 2008, meth spread most rapidly through low-income White and Hispanic populations living outside dense urban areas. The shifting in the mathematical hotspot of MethHF followed a similar spatial pattern of statewide meth growth, which ramped up first in urban/suburban areas with ready supply networks. This acceleration eventually declined as the market matured and the population segment most likely to adopt the drug had already been saturated, followed by the hotspot shifting from early-adopting to later-adopting communities.15 It is likely that we are seeing a 5- to 10-year delay in the spread of MethHF cases after the spread of meth, as heart failure and dilated cardiomyopathy usually develop after chronic, repeated exposures to meth.

Study Limitations

This study is a retrospective analysis of an administrative database with ICD-based case classification, which is susceptible to biases and confounding inherent to this method. While current clinical and animal studies provide compelling evidence of cardiotoxic effects of chronic meth use in the pathogenesis of a dilated cardiomyopathy phenotype with severely diminished contractile function, our study was not designed to prove a causal relationship. Rather, an association of meth use disorder and heart failure events has been corroborated by this real-world dataset in a region hard hit by the meth epidemic, supporting findings from prior studies sounding the alarm of this surging HF entity. As detailed in the Methods section, the ICD codes used in our study were not exclusive for meth. Epidemiological studies have confirmed the vast majority of stimulant use in hospitalized patients was indeed meth related, with the other stimulants making up <5% of cases. Furthermore, Thomas et al8 in their study reported that diagnosis of methamphetamine use via ICD-9 codes and/or urine toxicology corresponded with documented meth use in 84% of charts manually reviewed. This suggests that ICD-based case classification could potentially underestimate the true scope of MethHF problem. Other limitations also exist: no clinical data were available regarding temporal sequence of meth usage and heart failure diagnosis, severity, or duration of illnesses or symptoms upon which the diagnoses were based; total hospitalization charges rather than actual cost or recovered claim amount were used to approximate the financial burden of these hospitalizations; as well as potential ascertainment biases in testing for meth in subgroups with perceived high prevalence. Despite the limitations as discussed above, our findings confirm previous reports of the rising burden of heart failure hospitalizations attributable to chronic meth use in the real world and provide useful insights into the impact of age, sex, race/ethnicity on MethHF hospitalizations in California and other regions, particularly the Pacific West,2,28 which are most impacted by the rising prevalence of methamphetamine use.

Conclusions

MethHF hospitalizations in California rose by 585% from 2008 to 2018, which exceeded that of all HF hospitalizations that grew by only 0.92% and contrasted with the 6.0% decrease in non-MethHF hospitalizations in the same period. MethHF hospitalization associated charges increased by 840%, as compared with 82% for All HF hospitalizations. Patients with MethHF were significantly younger, predominantly male with fewer traditional cardiovascular comorbidities, but with overall worse profiles of social determinants of health including homelessness and other substance use disorders. The alarming rise in MethHF hospitalizations and its associated cost burden, and the complex geospatial spread patterns call for a concerted public health and clinical care response campaign to combat this multifaceted, fast-moving epidemic. A state-wide meth dashboard, like the California Opioid Overdose Dashboard, may be a prudent first step in this direction by raising awareness in both the general public and the medical communities.

Acknowledgments

We thank Aaron Maggetti of Office of Statewide Health Planning and Development for data curation; and Anandi Sujeer of Epidemiology and Data Management at Santa Clara County Public Health Department for her valuable inputs in study design and planning.

Supplemental Materials

Tables I and II

Disclosures None.

Footnotes

The Data Supplement is available at https://www.ahajournals.org/doi/suppl/10.1161/CIRCOUTCOMES.120.007638.

For Sources of Funding and Disclosures, see page 799.

Correspondence to: Susan X. Zhao, MD, Santa Clara Valley Medical Center, 751 S Bascom Ave, Suite No. 340, San Jose, CA 95128. Email

References

  • 1. Courtney KE, Ray LA. Methamphetamine: an update on epidemiology, pharmacology, clinical phenomenology, and treatment literature.Drug Alcohol Depend. 2014; 143:11–21. doi: 10.1016/j.drugalcdep.2014.08.003CrossrefMedlineGoogle Scholar
  • 2. Winkelman TNA, Admon LK, Jennings L, Shippee ND, Richardson CR, Bart G. Evaluation of amphetamine-related hospitalizations and associated clinical outcomes and costs in the United States.JAMA Netw Open. 2018; 1:e183758. doi: 10.1001/jamanetworkopen.2018.3758CrossrefMedlineGoogle Scholar
  • 3. United States Drug Enforcement Administration. 2019 National Drug Threat Assessment.2020. https://www.dea.gov/documents/2020/01/30/2019-national-drug-threat-assessment Accessed August 11, 2020.Google Scholar
  • 4. Kevil CG, Goeders NE, Woolard MD, Bhuiyan MS, Dominic P, Kolluru GK, Arnold CL, Traylor JG, Orr AW. Methamphetamine Use and Cardiovascular Disease.Arterioscler Thromb Vasc Biol. 2019; 39:1739–1746.LinkGoogle Scholar
  • 5. Zhao SX, Kwong C, Swaminathan A, Gohil A, Crawford MH. Clinical characteristics and outcome of methamphetamine-associated pulmonary arterial hypertension and dilated cardiomyopathy.JACC Heart Fail. 2018; 6:209–218. doi: 10.1016/j.jchf.2017.10.006CrossrefMedlineGoogle Scholar
  • 6. Zhao SX, Seng S, Deluna A, Yu EC, Crawford MH. Comparison of clinical characteristics and outcomes of patients with reversible versus persistent methamphetamine-associated cardiomyopathy.Am J Cardiol. 2020; 125:127–134. doi: 10.1016/j.amjcard.2019.09.030CrossrefMedlineGoogle Scholar
  • 7. Nishimura M, Ma J, Fox S, Toomu A, Mojaver S, Juang DK, Maisel AS, Thomas IC. Characteristics and outcomes of methamphetamine abuse among veterans with heart failure.Am J Cardiol. 2019; 124:907–911. doi: 10.1016/j.amjcard.2019.05.068CrossrefMedlineGoogle Scholar
  • 8. Thomas IC, Nishimura M, Ma J, Dickson SD, Alshawabkeh L, Adler E, Maisel A, Criqui MH, Greenberg B. Clinical characteristics and outcomes of patients with heart failure and methamphetamine abuse.J Card Fail. 2020; 26:202–209. doi: 10.1016/j.cardfail.2019.10.002CrossrefMedlineGoogle Scholar
  • 9. Husaini BA, Levine RS, Norris KC, Cain V, Bazargan M, Moonis M. Heart failure hospitalization by race/ethnicity, gender and age in California: implications for prevention.Ethn Dis. 2016; 26:345–354. doi: 10.18865/ed.26.3.345CrossrefMedlineGoogle Scholar
  • 10. Iannuzzi JC, Stapleton SM, Bababekov YJ, Chang D, Lancaster RT, Conrad MF, Cambria RP, Patel VI. Favorable impact of thoracic endovascular aortic repair on survival of patients with acute uncomplicated type B aortic dissection.J Vasc Surg. 2018; 68:1649–1655. doi: 10.1016/j.jvs.2018.04.034CrossrefMedlineGoogle Scholar
  • 11. Srivatsa UN, Xing G, Amsterdam E, Chiamvimonvat N, Pezeshkian N, Fan D, White RH. California Study of Ablation for Atrial Fibrillation:Re-hospitalization for Cardiac Events (CAABL-CE).J Atr Fibrillation. 2018; 11:2036. doi: 10.4022/jafib.2036CrossrefMedlineGoogle Scholar
  • 12. Gruenewald PJ, Ponicki WR, Remer LG, Waller LA, Zhu L, Gorman DM. Mapping the spread of methamphetamine abuse in California from 1995 to 2008.Am J Public Health. 2013; 103:1262–1270. doi: 10.2105/AJPH.2012.300779CrossrefMedlineGoogle Scholar
  • 13. Substance Abuse and Mental Health Services Administration, Center for Behavioral Health Statistics and Quality. Treatment Episode Data Set (TEDS): 2004-2014.National Admissions to Substance Abuse Treatment Services. BHSIS Series S-84, HHS Publication No. (SMA) 16-4986.Google Scholar
  • 14. Substance Abuse and Mental Health Services Administration, Treatment Episode Data Set(TEDS): 2017. Rockville, MD. 2019.Google Scholar
  • 15. Ponicki WR, Waller LA, Remer LG, Gruenewald PJ. Exploring the spread of methamphetamine problems within California, 1980 to 2006.GeoJournal. 2013; 78:451–462. doi: 10.1007/s10708-011-9428-4CrossrefMedlineGoogle Scholar
  • 16. Klein RJ, Schoenborn CA. Age adjustment using the 2000 projected U.S. population.Healthy People 2010 Stat Notes. 2001; 20:1–10.Google Scholar
  • 17. National Cancer Institute, Surveillance, Epidemiology and End Results Program. Single Ages - Standard Populations - SEER Datasets. SEER.2000. Accessed July 29, 2020. https://seer.cancer.gov/stdpopulations/stdpop.singleages.htmlGoogle Scholar
  • 18. California Department of Finance, Demographic Research Unit. Population Projections by Individual Year of Age, California, 2010-2060 (Baseline 2019 Population Projections; Vintage 2020 Release). Sacramento: California. 2019. Accessed July 29, 2020. http://www.dof.ca.gov/Forecasting/Demographics/Projections/Google Scholar
  • 19. California Department of Finance, Demographic Research Unit. State of California, Department of Finance, Race/Ethnic Population with Age and Sex Detail, 2000-2010, Sacramento, California.2012. Accessed July 29, 2020. http://www.dof.ca.gov/Forecasting/Demographics/Estimates/Race-Ethnic/2000-2010/Google Scholar
  • 20. Office of Statewide Health Planning and Development, Information Services Division. California Inpatient Data Reporting Manual, 8th Edition (Rev. 7/2020).Accessed February 13, 2021. https://oshpd.ca.gov/data-and-reports/submit-data/patient-data/inpatient-reporting/Google Scholar
  • 21. State of California, Department of Industrial Relations, Office of the Director-Research Unit, San Francisco, California.California Consumer Price Index (1955-2020) Accessed July 29, 2020. https://www.dir.ca.gov/OPRL/capriceindex.htmGoogle Scholar
  • 22. Anand S, Staniec A, Montez-Rath M, Vlahos P. Using GIS mapping to track hot spots of kidney disease in California.N Engl J Med. 2020; 382:2265–2267. doi: 10.1056/NEJMc2001023CrossrefMedlineGoogle Scholar
  • 23. Getis A, Ord JK. The analysis of spatial association by use of distance statistics.Geogr Anal. 1992; 24:189–206.CrossrefGoogle Scholar
  • 24. Gonzales R, Mooney L, Rawson RA. The methamphetamine problem in the United States.Annu Rev Public Health. 2010; 31:385–398. doi: 10.1146/annurev.publhealth.012809.103600CrossrefMedlineGoogle Scholar
  • 25. National Institute on Drug Abuse. Methamphetamine Research Report [Internet].National Institute on Drug Abuse. 2020 Accessed July 29, 2020. https://www.drugabuse.gov/publications/research-reports/methamphetamine/overviewGoogle Scholar
  • 26. Schürer S, Klingel K, Sandri M, Majunke N, Besler C, Kandolf R, Lurz P, Luck M, Hertel P, Schuler G, et al. Clinical characteristics, histopathological features, and clinical outcome of methamphetamine-associated cardiomyopathy.JACC Heart Fail. 2017; 5:435–445. doi: 10.1016/j.jchf.2017.02.017CrossrefMedlineGoogle Scholar
  • 27. Reddy PKV, Ng TMH, Oh EE, Moady G, Elkayam U. Clinical characteristics and management of methamphetamine-associated cardiomyopathy: state-of-the-art review.J Am Heart Assoc. 2020; 9:e016704. doi: 10.1161/JAHA.120.016704LinkGoogle Scholar
  • 28. Dickson SD, Thomas I, Nishimura M, Ma J, Arend W, Adler E, Greenberg B, Alshawabkeh L. Increasing methamphetamine-associated heart failure hospitalizations in the United States.J Am Coll Cardiol. 2019; 73:710.CrossrefGoogle Scholar
  • 29. Sliman S, Waalen J, Shaw D. Methamphetamine-associated congestive heart failure: increasing prevalence and relationship of clinical outcomes to continued use or abstinence.Cardiovasc Toxicol. 2016; 16:381–389. doi: 10.1007/s12012-015-9350-yCrossrefMedlineGoogle Scholar
  • 30. Centers for Disease Control and Prevention. Interactive Atlas of Heart Disease and Stroke Tables.Accessed August 11, 2020. https://nccd.cdc.gov/DHDSPAtlas/Reports.aspxGoogle Scholar
  • 31. Yeo KK, Wijetunga M, Ito H, Efird JT, Tay K, Seto TB, Alimineti K, Kimata C, Schatz IJ. The association of methamphetamine use and cardiomyopathy in young patients.Am J Med. 2007; 120:165–171. doi: 10.1016/j.amjmed.2006.01.024CrossrefMedlineGoogle Scholar
  • 32. Ito H, Yeo KK, Wijetunga M, Seto TB, Tay K, Schatz IJ. A comparison of echocardiographic findings in young adults with cardiomyopathy: with and without a history of methamphetamine abuse.Clin Cardiol. 2009; 32:E18–E22. doi: 10.1002/clc.20367CrossrefMedlineGoogle Scholar
  • 33. Marcinko MC, Darrow AL, Tuia AJ, Shohet RV. Sex influences susceptibility to methamphetamine cardiomyopathy in mice.Physiol Rep. 2019; 7:e14036. doi: 10.14814/phy2.14036CrossrefMedlineGoogle Scholar
  • 34. Hser YI, Evans E, Huang YC. Treatment outcomes among women and men methamphetamine abusers in California.J Subst Abuse Treat. 2005; 28:77–85. doi: 10.1016/j.jsat.2004.10.009CrossrefMedlineGoogle Scholar
  • 35. Alexander M, Grumbach K, Remy L, Rowell R, Massie BM. Congestive heart failure hospitalizations and survival in California: patterns according to race/ethnicity.Am Heart J. 1999; 137:919–927. doi: 10.1016/s0002-8703(99)70417-5CrossrefMedlineGoogle Scholar
  • 36. Jones CM, Compton WM, Mustaquim D. Patterns and characteristics of methamphetamine use among adults - United States, 2015-2018.MMWR Morb Mortal Wkly Rep. 2020; 69:317–323. doi: 10.15585/mmwr.mm6912a1CrossrefMedlineGoogle Scholar
  • 37. Gerber Y, Weston SA, Redfield MM, Chamberlain AM, Manemann SM, Jiang R, Killian JM, Roger VL. A contemporary appraisal of the heart failure epidemic in Olmsted County, Minnesota, 2000 to 2010.JAMA Intern Med. 2015; 175:996–1004. doi: 10.1001/jamainternmed.2015.0924CrossrefMedlineGoogle Scholar
  • 38. Joynt KE, Orav EJ, Jha AK. Thirty-day readmission rates for Medicare beneficiaries by race and site of care.JAMA. 2011; 305:675–681. doi: 10.1001/jama.2011.123CrossrefMedlineGoogle Scholar