The long-term effects of illicit drug use (IDU) on the clinical outcome of patients with heart failure (HF) are not well described. The objective of the present study was to describe the characteristics of patients with HF who used illicit drugs and to determine the effects of IDU on the clinical outcomes such as in-hospital mortality and hospital readmission for HF. A retrospective cohort study was conducted that included all patients admitted with HF from June 2003 to September 2004 and followed up until 2008 at a university hospital serving an at-risk population. The patients were divided into 2 groups: IDU and non-IDU according to self-reported use or positive laboratory results. The outcome measures were in-hospital mortality, HF readmission rate, interval to readmission for HF, and average brain natriuretic peptide and troponin levels throughout the follow-up period. Of 646 reviewed records, 542, representing 357 patients, were included in the present analysis. Of the 357 patients, 53 patients were in the IDU group and 304 were in the non-IDU group. Kaplan-Meier log-rank analysis and Cox proportional hazard analysis showed that IDU was associated with a shorter interval to readmission for HF (hazard ratio 3.8, 95% confidence interval 2.3 to 10.7, p <0.0001) but not with in-hospital mortality (hazard ratio 0.7, 95% confidence interval 0.3 to 1.7, p = 0.4). Multiple linear regression analysis identified IDU as an independent variable for the HF readmission rate (p = 0.0001) but not for average brain natriuretic peptide or average troponin levels. In conclusion, the results of the present study have demonstrated that IDU was associated with a decreased interval to readmission for HF and greater HF readmission rates.
The effect of illicit drug use (IDU) on clinical outcomes in HF is not well studied. In 2010, 8.9% of Americans aged ≥12 years reported IDU in the previous month. The objective of the present study was to describe the characteristics of patients with HF who used illicit drugs and to determine the effects of IDU on the clinical outcomes of patients admitted with HF at a university hospital serving an at-risk population. We reviewed the data collected for the Joint Commission (TJC) from patients admitted with HF at a university hospital serving an at-risk population in Louisiana to determine the effects of IDU on the clinical outcomes of HF. The population studied was characterized by a high proportion of uninsured or underinsured patients (66%), a high proportion of blacks (59%), a low literacy rate, a high disease burden, and greater inpatient mortality for HF than reported by the national registries.
Methods
The study protocol has been previously published in detail. The institutional review board at Louisiana State University Health Science Center–Shreveport approved the study. A retrospective cohort study of all patients admitted to a single, tertiary care center in Louisiana with a diagnosis of HF from June 1, 2003 through September 30, 2004, was performed. Data were collected according to the University Health System Consortium protocol using the prespecified International Classification of Diseases, 9th revision, diagnosis codes for the JCT.
Patients were excluded if they were not expected to continue care in the hospital system, such as out-of-state residents or patients admitted on an emergency basis only. Because this center is the only tertiary care center caring for the uninsured or underinsured population in a circumference of about 100 miles, most of the population included in the present study obtained inpatient care at the Louisiana State University Health Science Center–Shreveport. Therefore, evaluating the clinical outcomes was possible through the institution’s medical record system. Patients who died during the index admission were also excluded.
The medical records of all patients were reviewed. Data collection was guided by the American College of Cardiology/American Heart Association Task Force on Clinical Data Standards. Key aspects of all admission and discharge summaries to the date of medical record review (May to December 2008) were recorded. Admission for HF was defined as an admission in which the main reason for admission, or a main problem during the admission, was the treatment of HF as defined by the American College of Cardiology/American Heart Association Task Force. The index admission was defined as the first identified admission during the study period. The use of illicit drugs was identified either by patient self-report as documented in any admission history and physical examination or discharge summary (a history of IDU is a routine standard question for all medical admissions in the institute) or positive findings on a drug urine screening test for any of the illicit substances identifiable by the test that was routinely used at the Louisiana State University Health Science Center–Shreveport during the study period. The drug urine screening was ordered according to the discretion of the treating physicians.
The patients were divided into 2 groups according to IDU: those with IDU (those who had used an illicit drug at any admission or follow-up visit during the study period) and those without IDU (those for whom no evidence was found in their medical records of IDU).
The following outcome measures were recorded: in-hospital mortality, interval from the index admission to readmission for HF, HF readmission rate (calculated as the number of HF admissions divided by the interval of available follow-up), average serum B-type natriuretic peptide (BNP) throughout the follow-up period, and average troponin I levels throughout the follow-up period. The interval of available follow-up was defined as the period from the index admission until death or the last record of any medical activity within the system, such as collected laboratory findings or a scheduled clinic appointment.
Continuous variables are presented as the mean ± SD and categorical variables as percentages. The mean values of the continuous numeric variables were compared using the 2-sample t test for independent means. For categorical variables, the chi-square or Fisher exact test was used. If the Shapiro-Wilk test indicated non-normality, the variables are presented as the median and interquartile ranges. Continuous variables not meeting the Shapiro-Wilk test for normal distribution were studied using parametric tests ( t test) if a histogram of the log transform showed a strong central tendency. Each parameter was evaluated using the subjects without missing data for that specific parameter. The survival analysis and interval from the index admission to readmission for HF were compared between the IDU group and non-IDU group using Kaplan-Meier log-rank analysis and Cox proportional hazard analysis. Also, the hazard ratios and 95% confidence intervals were calculated. Multiple linear regression analysis was performed to identify variables associated with the HF readmission rate, average BNP level, and average troponin level. Natural logarithm transformation was used for the HF readmission rate because the Shapiro-Wilk test indicated that it was not normally distributed and without a central tendency. On multiple regression analysis, a stepwise method was used, adding and removing factors in the order that maximized improvement in the value of R 2 . Initial multivariate stepwise regression analyses of the readmission rate included the following variables: age, atrial fibrillation, drug use, alcohol use, coronary artery disease, hypertension, diabetes, ejection fraction, hyperlipidemia, noncompliance, race, and gender. An optimal HF medication management score (0 to 3 indicating whether 1, 2, or 3 of β blockers, angiotensin-converting enzyme inhibitors, and diuretics were used at the index admission) was also included in the initial tests. Overall p values <0.05 were considered significant for all analyses. All procedures were performed using the SAS System version 8.2 (SAS Institute, Cary, North Carolina).
Results
Of the 646 discharge records reviewed, 104 were excluded (78 discharges for patients from different healthcare system who were not expected to have follow-up within our system, 14 deaths during the index admission, 9 with an incorrect International Classification of Disease, 9th revision, diagnosis, and 3 with incorrect medical record numbers). The final analysis included 542 hospital records, representing 357 patients: 53 in the IDU group and 304 in the non-IDU group. Of those in the IDU group, the urine drug screening test was positive for cocaine in 60%, cannabis in 37%, opioids in 17%, benzodiazepines in 8%, barbiturates in 6%, and methamphetamine in 0%. The baseline characteristics and outcome variables between the 2 groups are listed in Table 1 . The mean ± SD of follow-up for all patients was 2.39 ± 1.6 years. No difference was found in the risk of in-hospital mortality (hazard ratio 0.7, 95% confidence interval 0.3 to 1.7, p >0.05; Figure 1 ). The IDU group had significantly earlier readmission (hazard ratio 3.8, 95% confidence interval 2.3 to 10.7, p <0.0001; Figure 2 ).
| Variable | IDU | p Value ⁎ | |
|---|---|---|---|
| Yes (n = 53) | No (n = 304) | ||
| Age (years) | 47 ± 8 | 59 ± 13 | <0.0001 |
| Men | 42 (79%) | 163 (54%) | <0.001 |
| Race | |||
| White | 5 (9%) | 88 (29%) | <0.005 |
| Black | 48 (91%) | 216 (71%) | |
| Insurance | |||
| None | 8 (15%) | 62 (21%) | 0.02 |
| Governmental insurance for medications and services | 35 (66%) | 132 (44%) | |
| Governmental insurance for services only | 8 (15%) | 87 (29%) | |
| Commercial insurance | 2 (4%) | 22 (7%) | |
| Index admission brain natriuretic peptide (pg/ml) † | 0.03 | ||
| Mean ± SD | 2,342 ± 2,387 | 1,527 ± 1,753 | |
| Median | 861 | 905 | |
| Interquartile range | 1,410–3,425 | 391–2,088 | |
| Index admission troponin I (ng/ml) † | <0.6 | ||
| Mean ± SD | 0.17 ± 0.23 | 0.23 ± 1.6 | |
| Median | 0.08 | 0.04 | |
| Interquartile range | 0.05–0.17 | 0.02–0.09 | |
| Serum creatinine (mg/dl) ‡ | <0.14 | ||
| Mean ± SD | 2.0 ± 2.1 | 1.6 ± 1.7 | |
| Median | 1.3 | 1.2 | |
| Interquartile range | 1.05–1.85 | 0.9–1.6 | |
| Serum sodium (mEq/L) ‡ | 139.4 ± 3.4 | 139.4 ± 3.5 | 0.95 |
| Hemoglobin (g/dl) ‡ | 13.5 ± 2.6 | 12.5 ± 2.1 | 0.002 |
| Left ventricular ejection fraction ‡ | 0.30 ± 0.15 | 0.34 ± 0.17 | 0.2 |
| Follow-up duration (days) | 988 ± 690 | 993 ± 597 | 0.98 |
| Co-morbidities § | |||
| Coronary artery disease | 11 (21%) | 126 (42%) | 0.005 |
| Systolic heart failure | 42 (79%) | 234 (77%) | 0.9 |
| Cerebrovascular accident | 1 (2%) | 27 (9%) | 0.04 † |
| Hyperlipidemia | 21 (40%) | 177 (58%) | 0.02 |
| Diabetes mellitus | 8 (15%) | 146 (48%) | <0.0001 |
| Alcohol abuse | 18 (34%) | 38 (13%) | 0.0002 |
| Hypertension | 48 (91%) | 263 (87%) | 0.5 |
| Obesity | 4 (8%) | 60 (20%) | 0.03 |
| Obstructive sleep apnea | 8 (15%) | 45 (15%) | 1.0 |
| Atrial fibrillation | 4 (8%) | 62 (20%) | 0.3 † |
| Noncompliance ∥ | 28 (53%) | 51 (17%) | <0.0001 |
| Medications | |||
| Angiotensin-converting enzyme inhibitor | 41 (77%) | 268 (88%) | 0.05 |
| Aldosterone antagonist | 13 (25%) | 59 (19%) | 0.5 |
| β Blocker | 30 (57%) | 212 (70%) | 0.08 |
| Digoxin | 14 (26%) | 100 (33%) | 0.4 |
| Diuretic | 277 (94%) | 50 (91%) | 0.6 |
| Outcome measures | |||
| In-hospital mortality (n) | 4 (8%) | 32 (11%) | 0.6 |
| Average brain natriuretic peptide (pg/ml) † | 0.04 | ||
| Median | 1,161 | 1,228 | |
| Interquartile range | 2,190–3,237 | 464–2,265 | |
| Average troponin I (ng/ml) † | 0.04 | ||
| Median | 0.08 | 0.04 | |
| Interquartile range | 0.06–0.17 | 0.02–0.12 | |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
