The purpose of the present study was to assess whether a low socioeconomic (SE) position is associated with outcomes in heart transplant recipients. We used the US Census 2000 database to derive a summary SE score for 520 patients who had undergone underwent a first heart transplant at 1 of 4 Boston hospitals during 1996 to 2005 and compared the outcomes in the lowest quartile SE group (n = 129) to those for the remaining patients (n = 391). The low SE group and controls were similar with respect to cardiac diagnosis, hemodynamic support, listing status, year of transplant, and initial immune suppression. Low SE patients were more likely to be nonwhite. Graft loss occurred in 142 patients (135 deaths and 7 repeat transplants). Hospital mortality after transplantation was not associated with race/ethnicity or low SE position. In patients who survived the transplant hospitalization, nonwhite ethnicity (hazard ratio 1.8, 95% confidence interval 1.1 to 2.9) and low SE group (hazard ratio 1.7, 95% confidence interval 1.1 to 2.5) were associated with a greater risk of subsequent graft loss. In the adjusted analysis, the risk of graft loss remained greater for both nonwhite race/ethnicity (hazard ratio 1.7, 95% confidence interval 1.0 to 2.9) and low SE position (hazard ratio 1.5, 95% confidence interval 1.0 to 2.4). Rejection episodes were more frequent in nonwhite transplant recipients and in those in the low SE group. In conclusion, among heart transplant recipients who survive the transplant hospitalization, nonwhite recipients and those in a low SE position are at greater risk of rejection and graft loss.
Because socioeconomic (SE) data are not routinely collected as a part of the medical records, its association with patient outcomes after transplantation has been difficult to study. Analyses using proxy SE variables such as median income in the zip code of patient residence have not demonstrated an association of SE position with the outcomes of transplant recipients. Zip codes have an average population of 30,000, are administrative units established by the US Postal Service for the most efficient delivery of mail, and can have a large internal heterogeneity in the SE position of their residents. A much smaller unit of population, a “block group,” with an average population of 1,000 residents, is the smallest geographic unit for which census SE data are tabulated. Block groups are designed to be relatively homogenous with respect to the economic status and living conditions of their residents. A block group has been described as the neighborhood of a person’s residence. Previous studies using block group SE data in population studies have demonstrated an association of low SE position with incident coronary heart disease, cancer, and all-cause mortality. The purpose of the present study was to evaluate whether low patient SE position, determined by the SE characteristics of the block group of patient residence, is associated with graft loss and risk of graft rejection in heart transplant recipients.
Methods
The present study was a multicenter, retrospective cohort study. All patients who underwent a first heart transplant at 1 of 4 Boston Transplant Centers (Children’s Hospital Boston, Massachusetts General Hospital, Brigham and Women’s Hospital, Tufts Medical Center, all Boston, Massachusetts) from January 1, 1996 to December 31, 2005 were eligible. The patients who underwent repeat transplant or were non-United States residents (international patients) who came to the United States for heart transplantation were excluded. The institutional review boards of all 4 hospitals approved the study, with a waiver of informed consent.
Each patient’s home address at transplantation was used to extract the block group of residence from the US Census Web site. Using a previously described measure of SE position from the SE characteristics of block group of residence, a summary SE score was derived for each transplant recipient. This score was used as the main indicator of the SE position of the patient. The 6 SE variables selected for the summary score were originally described by Diez Roux et al using factor analysis, a statistical technique to determine which variables of a large set (eg, a large set of Census SE variables) can be meaningfully combined into a composite score.
To determine this score, the data for the 6 SE variables for each subject’s block group were collected from the US Census 2000 Report. These variables represented 3 dimensions of wealth and income (log of the median household income, log of the median value of housing units, and the percentage of households receiving interest, dividends, or net rental income), 2 dimensions of education (the percentage of adults ≥25 years old who had completed high school and the percentage ≥25 years old who had completed college), and 1 dimension of occupation (the percentage of employed persons ≥16 years old in executive, managerial, or professional specialty occupations) for the residents of the block group. For each variable, a z-score for each block group was calculated by subtracting the overall mean of that variable (across all block groups in the sample) from the value of the variable for that block group and dividing by the standard deviation. The summary SE score for each subject was obtained by summing the 6 z-scores (1 for each of the 6 variables) for that subject. The data on a simple measure of block group SE position, the proportion of persons living in the block group who were below the federally defined poverty level, were also collected.
Race/ethnicity was defined as designated by the transplant center for the state and national databases. The patients were divided into white and nonwhite (black, Hispanic, and other) categories for data analysis.
Because the determinants of post-transplant hospital mortality and postdischarge mortality can be different, 2 primary outcome variables were analyzed: (1) the post-transplant hospital mortality (as a binary outcome), and (2) the interval to graft loss (death or retransplantation) in patients who survived the transplant hospitalization. The patients were followed up until graft loss (event) or were censored on June 1, 2008. We also evaluated the association of low SE position with 2 secondary outcomes: (1) the interval to the first rejection episode, and (2) the incidence rate of rejection (number of rejection episodes/patient-years of follow-up). We defined a rejection episode as meeting one of the following criteria: (1) endomyocardial biopsy showing International Society of Heart and Lung Transplantation grade ≥2R (old grade ≥3A), (2) antibody-mediated rejection, or (3) rejection as determined from clinical and/or echocardiographic findings of graft dysfunction resulting in acute augmentation of immune suppression.
The patients were divided into 2 groups according to their summary SE scores: the low SE group (quartile with the lowest summary SE scores, n = 129) and the control group (the remaining patients, n = 391). The decision to compare the low SE group with the remaining patients rather than to compare 4 equal-size groups (quartiles) was made both because of the study hypothesis (ie, that the low SE group would be at risk) and because a preliminary analysis suggested similar outcomes for patients in quartiles 2 to 4. The 2 groups were compared for the distribution of demographic variables using Fisher’s exact test or the Wilcoxon rank-sum test, and the SE variables (of the block groups of residence) were compared using t tests. The association of early hospital mortality (a binary outcome) with the potential risk factors was evaluated using a logistic regression model and that of subsequent graft loss in those surviving the transplant hospitalization using Kaplan-Meier survival curves (with the log-rank test) and a Cox proportional hazard model. A Cox model was also used to evaluate the interval to the first rejection episode. A Poisson regression model was used to evaluate the incidence rate of rejection. Multivariate models were constructed by evaluating all variables that had a univariate p value of <0.2; however, the SE group and race/ethnicity were included regardless of statistical significance. The models for rejection were adjusted for individual hospitals. Potential interactions of SE position with age, ethnicity, or medical insurance were explored. All tests were 2-sided. The data were analyzed using Statistical Analysis Systems, version 9.1 (SAS Institute, Cary, North Carolina) and Stata, version 10.0 (StataCorp, College Station, Texas). We had full access to the data and take responsibility for its integrity; all of us have read and agreed to the report as written.
Results
A total of 560 United States residents underwent their first heart transplant at the participating institutions from 1996 to 2005 and were eligible for the present study. Of these, 40 (7%) did not have a home address in the medical records (only a post office box number was available) to determine their block group and SE data. The remaining 520 patients formed the analytic cohort for the present study. The median age of these patients was 51 years (range 7 days to 71 years); 105 (20%) were ≤19 years old and 111 (21%) were ≥60 years old. Of the 520 patients, 397 (76%) were male, 443 (85%) were white, and 77 were nonwhite (31 black, 37 Hispanic, and 9 other). The underlying diagnosis was dilated cardiomyopathy in 273 (52%), ischemic cardiomyopathy in 148 (29%), congenital heart disease in 57 (11%), and other in 42 (8%). The medical insurance was private for 339 patients (65%) and public for 166 (32%); 15 patients had missing insurance data.
A comparison of the clinical and demographic variables between the low SE group and the control group is listed in Table 1 . The groups were similar with respect to the distribution of gender, cardiac diagnoses, diabetes, hypertension, history of smoking (ever smoked), year (era) of transplant, and management with extracorporeal membrane oxygenation or ventricular assist device before transplantation. The low SE group was slightly younger, had a greater proportion of black and Hispanic subjects, and had a lower proportion of white subjects. The low SE group patients were more likely to have public medical insurance than were the controls. The 2 groups did not differ with respect to the proportion with a positive cross-match or a primary cytomegalovirus mismatch with their donor, the proportion who received induction therapy at transplantation, or their maintenance immune suppression regimen 2 weeks after transplantation.
| Variable | Low SE Group (n = 129) | Control Group (n = 391) | p Value |
|---|---|---|---|
| Age at transplant (years) | 0.02 | ||
| Median | 47.8 | 52.0 | |
| Range | 0.1–71.2 | 0–70.4 | |
| Age at transplant (years) | 0.11 | ||
| <1 | 3 (2.3%) | 8 (2.1%) | |
| 1–9 | 17 (13.2%) | 34 (8.7%) | |
| 10–19 | 10 (7.8%) | 33 (8.4%) | |
| 20–39 | 15 (11.6%) | 37 (9.5%) | |
| 40–59 | 67 (51.9%) | 185 (47.3%) | |
| ≥60 | 17 (13.2%) | 94 (24%) | |
| Race/ethnicity | <0.001 | ||
| White | 88 (68.2%) | 355 (90.8%) | |
| Black | 19 (14.7%) | 12 (3.1%) | |
| Hispanic | 20 (15.5%) | 17 (4.3%) | |
| Other | 2 (1.6%) | 7 (1.8%) | |
| Female | 31 (24.0%) | 92 (23.5%) | 0.91 |
| Diagnosis | 0.20 | ||
| Idiopathic dilated cardiomyopathy | 76 (58.9%) | 197 (50.4%) | |
| Ischemic cardiomyopathy | 32 (24.8%) | 116 (29.7%) | |
| Congenital heart disease | 15 (11.6%) | 42 (10.7%) | |
| Other | 6 (4.7%) | 36 (9.2%) | |
| United Network of Organ Sharing status at transplantation | 0.53 | ||
| 1 | 23 (17.8%) | 63 (16.1%) | |
| 1A | 46 (35.7%) | 125 (32.0%) | |
| 1B | 26 (20.2%) | 73 (18.7%) | |
| 2 | 34 (26.4%) | 130 (33.2%) | |
| Transplant era | 0.06 | ||
| 1996–1998 | 37 (28.7%) | 142 (36.3%) | |
| 1999–2001 | 50 (38.8%) | 109 (27.9%) | |
| 2002–2005 | 42 (32.5%) | 140 (35.8%) | |
| Insurance | <0.001 | ||
| Public | 60 (46.5%) | 106 (27.1%) | |
| Private | 63 (48.8%) | 276 (70.6%) | |
| Missing | 6 (4.7%) | 9 (2.3%) | |
| Diabetes | 19 (14.8%) | 60 (15.4%) | 1.00 |
| Hypertension | 29 (22.7%) | 89 (23.5%) | 0.90 |
| History of smoking | 70 (54.3%) | 185 (47.3%) | 0.31 |
| Extracorporeal membrane oxygenation | 10 (7.8%) | 15 (3.8%) | 0.09 |
| Ventricular assist device | 12 (9.3%) | 33 (8.4%) | 0.72 |
| Panel reactive antibody >10% | 7 (8.3%) | 22 (9.4%) | 1.0 |
| Positive cross-match | 2 (1.6%) | 5 (1.3%) | 0.69 |
| Induction therapy | 13 (10.1%) | 49 (12.5%) | 0.53 |
| Early hospital death | 11 (8.5%) | 31 (7.9%) | 0.85 |
| Graft loss/death | 45 (34.9%) | 97 (24.8%) | 0.03 |
The SE characteristics of the block groups of patient residence of the low SE and control groups are compared in Table 2 . The differences between the 2 groups were statistically significant for all SE variables. Thus, patients in the low SE group lived in neighborhoods with lower median household incomes and housing values; fewer adults who had completed high school or college education; fewer workers in managerial, professional, or executive professions; and fewer households with rental, interest, or dividends as income sources (p <0.001 for all comparisons). The low SE group lived in neighborhoods with a significantly greater percentage of persons living below the poverty level.
| Variable | Low SE Group (n = 129) | Control Group (n = 391) |
|---|---|---|
| Composite socioeconomic score | −6.38 ± 2.61 | −2.10 ± 3.93 |
| Median household income ($ thousands) | 31.3 ± 8.5 | 59.9 ± 18.3 |
| Median value of houses ($ thousands) | 105 ± 39 | 180 ± 87 |
| Households with rental, dividend, or interest income | 26.0 ± 11.0% | 47.8 ± 11.5% |
| Residents >25 years old with high school graduation | 69.6 ± 12.5% | 88.5 ± 7.0% |
| Residents >25 years old with college degree | 12.8 ± 6.1% | 33.8 ± 15.5% |
| Employed as manager, professional, or executive | 22.9 ± 7.4% | 40.9 ± 11.7% |
| Those living below poverty level | 19.3 ± 12.2% | 5.1 ± 4.4% |
Cardiac allograft loss occurred in 142 patients (135 deaths and 7 repeat transplantations). The graft survival rate was 87.7% (95% confidence interval 84.6% to 90.2%), 84.6% (95% confidence interval 81.2% to 87.4%), and 79.0% (95% confidence interval 75.1% to 82.4%) 1, 3, and 5 years, respectively, after transplantation for the study cohort. Overall, graft loss was observed in 45 low SE patients (35%) and 97 control patients (25%).
A total of 42 early deaths occurred during the transplant hospitalization (8.1%). On both univariate and adjusted analyses, early hospital mortality was associated with pretransplant mechanical support (extracorporeal membrane oxygenation or ventricular assist device) and with heart transplantation during the earlier years (1996 to 2001) but not with ethnicity or low SE group ( Table 3 ). In patients who survived to discharge after the transplant hospitalization, the era of transplantation and use of pretransplantation mechanical support (extracorporeal membrane oxygenation/ventricular assist device) were not associated with subsequent graft loss. In these patients, nonwhite patients ( Figure 1 ), those in the low SE group ( Figure 2 ), and those with a history of smoking were at a greater risk of subsequent graft loss on univariate analysis ( Table 3 ). The multivariate model that best predicted the risk of graft loss in hospital survivors included nonwhite race/ethnicity, low SE position, and pretransplant history of smoking ( Table 3 ). The addition of SE group to a model that included race/ethnicity and smoking (but not SE group) lowered the hazard ratio for nonwhite ethnicity from 2.0 to 1.7. Of note, the type of medical insurance (public vs private) was not associated with either hospital deaths or subsequent graft loss in hospital survivors. Overall, a greater proportion of patients in the low SE group died from acute rejection (3.9% vs 0.8% controls, p = 0.01) or from acute rejection, coronary artery disease, or chronic graft dysfunction (10% vs 3% controls, p = 0.002).
| Predictor | Univariate Analysis | Multivariate Analysis | ||
|---|---|---|---|---|
| HR/OR (95% CI) | p Value | HR/OR (95% CI) | p Value | |
| Early hospital deaths | ||||
| Transplantation 1996–2001 ⁎ | 2.1 (1.0–4.4) | 0.06 | 4.4 (1.8–10.6) | 0.001 |
| Extracorporeal membrane oxygenation/ventricular assist device | 6.9 (3.5–13.6) | <0.001 | 11.8 (5.4–25.7) | <0.001 |
| Ethnicity (nonwhite) | 1.4 (0.6–3.1) | 0.42 | 1.4 (0.6–3.6) | 0.46 |
| Low socioeconomic group | 1.1 (0.5–2.2) | 0.83 | 0.8 (0.3–1.7) | 0.52 |
| Graft loss in hospital survivors | ||||
| Transplant 1996–2001 ⁎ | 1.2 (0.7–2.1) | 0.51 | — | — |
| Extracorporeal membrane oxygenation/ventricular assist device | 1.6 (0.8–3.1) | 0.17 | — | — |
| Low socioeconomic group | 1.7 (1.1–2.5) | 0.02 | 1.5 (1.0–2.4) | 0.06 |
| Ethnicity (nonwhite) | 1.8 (1.1–2.9) | 0.02 | 1.7 (1.0–2.9) | 0.05 |
| History of smoking | 1.5 (1.0–2.3) | 0.05 | 1.6 (1.0–2.4) | 0.03 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
