Author + information
- Received February 7, 2014
- Revision received February 24, 2014
- Accepted March 7, 2014
- Published online August 1, 2014.
- Eileen M. Hsich, MD∗,†∗ (, )
- Randall C. Starling, MD, MPH∗,†,
- Eugene H. Blackstone, MD∗,†,‡,
- Tajinder P. Singh, MD, MSc§,‖,
- James B. Young, MD∗,
- Eiran Z. Gorodeski, MD, MPH∗,
- David O. Taylor, MD∗ and
- Jesse D. Schold, PhD‡
- ∗Heart and Vascular Institute, Cleveland Clinic, Cleveland, Ohio
- †Cleveland Clinic, Lerner College of Medicine of Case Western Reserve University School of Medicine, Cleveland, Ohio
- ‡Department of Quantitative Health Sciences, Cleveland Clinic, Cleveland, Ohio
- §Department of Cardiology, Boston Children’s Hospital, Boston, Massachusetts
- ‖Harvard Medical School, Boston, Massachusetts
- ↵∗Reprint requests and correspondence:
Dr. Eileen M. Hsich, Kaufman Center for Heart Failure, Heart and Vascular Institute, Cleveland Clinic, J3-4, 9500 Euclid Avenue, Cleveland, Ohio 44195.
Objectives The aim of this paper was to identify sex differences in survival of patients awaiting orthotopic heart transplantation (OHT).
Background Women have a higher mortality rate while awaiting OHT than men, and the reason has not been fully determined.
Methods We included all adult patients in the Scientific Registry of Transplant Recipients (SRTR) placed on the OHT waiting list from 2000 to 2010. The primary endpoint was all-cause mortality before receiving OHT, analyzed using time-to-event analysis. Multivariate Cox proportional hazards models were used to evaluate sex differences in survival, with data stratified by United Network for Organ Sharing (UNOS) status at time of listing.
Results There were 28,852 patients (24% women) awaiting OHT. This cohort included 6,163 UNOS status 1A (25% women), 9,168 UNOS status 1B (25% women), and 13,521 UNOS status 2 (24% women) patients. During a median follow-up of 3.7 years, 1,290 women and 4,286 men died. Female sex was associated with a significant risk of death among UNOS status 1A (adjusted hazard ratio [HR]: 1.20; 95% confidence interval [CI]: 1.05 to 1.37, p = 0.01) after adjusting for more than 30 baseline variables. In contrast, female sex was significantly protective for time to death among UNOS status 2 patients (adjusted HR: 0.75; 95% CI: 0.67 to 0.84, p < 0.001). No sex differences were noted among UNOS status 1B patients.
Conclusions There are sex differences in survival between women and men awaiting heart transplantation, and the current UNOS transplant criteria do not account for this disparity.
Women in the United States have a higher mortality rate than men while awaiting orthotopic heart transplantation (OHT) (1), which has not been fully evaluated. Based on publicly available Scientific Registry of Transplant Recipients (SRTR) data, the median OHT wait time for women during this same time period was shorter than for men (1), suggesting it was not due to availability of donors. In 1 small European study (58 women, 260 men), more women (17%) than men (12%) died awaiting OHT during a 12-month follow-up. After adjusting for age, heart failure survival score, serum creatinine, inpatient status, cardiac index, low vocational level, smoking, and low emotional support at time of transplant listing, female sex was still associated with a higher risk of death/deterioration (hazard ratio [HR]: 2.3; 95% confidence interval [CI]: 1.04 to 5.12; p = 0.04) (2). What remains unknown is whether sex differences in waitlist mortality also exist in the United States after adjusting for baseline risk factors.
The current OHT allocation system in the United States is based primarily on severity of illness (3). However criteria for OHT listing and heart failure (HF) survival models (4–6) do not distinguish women from men despite known sex differences in cause (7–9), cardiac remodeling (10–12), response to therapy (13–16), and prognosis (17–19). Therefore, advanced HF therapies such as OHT or mechanical circulatory support may be recommended with no evidence-based expectations if sex differences in prognostic risk factors are not recognized and utilized. The goal of this study was to further evaluate sex differences in mortality for HF patients awaiting OHT, using our current allocation system that stratifies patients into categories based on severity of illness: United Network for Organ Sharing (UNOS) status 1A for high-risk patients; UNOS status 1B for intermediate-risk patients; and UNOS status 2 for lower risk, ambulatory patients. To account for the limited mechanical circulatory support available to rescue women prior to April 2008, when the U.S. Food and Drug Administration approved a smaller device called HeartMate II (Thoratec Corp., Pleasanton, California) that could be implanted in petite patients (body surface area: <1.5 m2), we also assessed the importance of the era before and after that date to look for any sex interaction.
Scientific Registry of Transplant Recipients
This study used data from SRTRs. The SRTR database includes data for all donors, waitlisted candidates, and transplantation recipients in the United States submitted by members of the Organ Procurement and Transplantation Network (OPTN) and has been described elsewhere (20). The U.S. Department of Health and Human Services Health Resources and Services Administration provides oversight of the activities of OPTN and SRTR contractors. Human error in collecting data is minimized by edit checks, validation of data at time of entry, and internal verification, when there are outliers.
Patient population and UNOS status
We included all adult patients in the SRTR database who were placed on the waiting list for OHT from January 1, 2000, to December 31, 2010. Follow-up data were available until November 30, 2011. Patients were excluded if they were <18 years of age because the UNOS criteria for listing pediatric patients differs from that for patients who are adults, and the donor pools are distinguished by age (3).
Data were stratified according to UNOS status at time of waitlisting. UNOS status 1A includes patients requiring ventricular assist device (VAD), total artificial heart (TAH), extracorporeal membrane oxygenation (ECMO), intra-aortic balloon pump (IABP), mechanical ventilation, continuous intravenous high-dose inotropes, or an exemption for critical illness such as ventricular tachycardia or complications with mechanical circulatory support. UNOS status 1B is the next highest status for OHT and includes patients receiving continuous intravenous doses of inotrope support and stable VAD patients. UNOS status 2 is the least urgent status for patients actively waiting for OHT and is reserved for patients receiving standard medical therapy.
The primary endpoint was all-cause mortality, assessed as a right-censored time to death, with follow-up censored at the time of transplantation. SRTR mortality data are maintained by the transplantation centers and verified with the U.S. Social Security Administration Death Master File which was available until November 30, 2011.
Sex-specific baseline characteristics were reported according to UNOS status at the time of listing for OHT. Continuous variables were expressed as means, and categorical variables were expressed as frequencies. Chi-square and Wilcoxon rank-sum tests were used for group comparisons. Sex-specific survival analysis was performed for UNOS status 1A, 1B, and 2 patients, using the Kaplan-Meier method with censoring for OHT. The primary analysis was based on intent to treat such that deaths following removal from the waiting list were included in the primary analysis. The cumulative incidence of transplantation and death was estimated as competing risks, using the Fine and Gray method (21). Cox proportional hazard models were created to assess for the association between female sex and death according to initial UNOS status at time of listing. Two models were created. Model 1 was adjusted for the following characteristics at time of listing: age, diabetes mellitus status, dialysis, body mass index, previous OHT, race (white, black, Hispanic, Asian, other), history of cerebral vascular accident and tobacco use, inotrope use, glomerular filtration rate (GFR), ventilator status, insurance (private, Medicare/Medicaid, other), type of ventricular assist device (left ventricular assist device [LVAD] or right ventricular assist device with or without LVAD or TAH/unspecified mechanical circulatory device), antiarrhythmia, previous cardiac surgery, hypertension, malignancy, peripheral vascular disease, ECMO, IABP, era, cardiac diagnosis (dilated cardiomyopathy, ischemic cardiomyopathy, congenital heart disease, hypertrophic cardiomyopathy, restrictive cardiomyopathy, valvular cardiomyopathy, and other), ABO blood type, defibrillator, pulmonary artery mean, mean pulmonary capillary wedge pressure (PCWP), total albumin, and cardiac output with dummy variables for missing GFR, pulmonary artery mean pressure, PCWP, cardiac output, and albumin. Model 2 was performed as a sensitivity analysis that excluded variables with ahigh proportion of “missingness” (>10% that included cardiac output [12% “missingness”]), albumin (15% “missingness”), pulmonary pressure (11% “missingness”), and PCWP pressure (13% “missingness”) and including dummy variables when needed for missing variables among characteristics that had <10% of missing data. In both models, we imputed mean values for missing values and included an interaction term between coronary artery disease and sex.
In order to understand the association between baseline parameters (at the time of placement on the waiting list) and the likelihood of being placed on the waiting list as a status 1A patient, we generated a propensity score. The propensity score was derived from a multivariate logistic model that included the full study population and all parameters previously described, with the exception of initial status and candidate sex. The outcome variable of this model was whether patients were listed as status 1A. Based on the output of this model, we evaluated the probability that a patient would be placed as status 1A as predicted by the set of covariates in the model. We then compared the average probabilities between men and women (using a 2-sample t-test) to understand whether sex was associated with greater predicted likelihood of status 1A placement.
All analyses were performed using SAS version 9.2 software (SAS Institute, Cary, North Carolina). A p value of <0.05 was considered statistically significant.
Baseline characteristics of 28,852 adult HF patients (24% women) awaiting OHT are shown in Table 1. This cohort included 6,163 UNOS status 1A (25% women), 9,168 UNOS status 1B (25% women), and 13,521 UNOS status 2 (24% women) patients. Most patients were white, with a higher prevalence of blacks among women than among men in all subgroups. Women were younger and had a lower body mass index than men at time of listing for OHT in each subgroup. Most patients had an idiopathic dilated cardiomyopathy with slightly more congenital heart disease, hypertrophic cardiomyopathy, restrictive cardiomyopathy, and valvular disease among women than among men in all subgroups. Previous cardiac surgery and tobacco abuse were more likely in men than in women, whereas history of malignancy was more common in women than in men in all subgroups. Defibrillators were more likely to be present in men than in women at time of listing, with UNOS status 1A patients having a lower percentage of patients with a defibrillator at time of listing than UNOS status 1B or 2 patients. Among UNOS status 1A patients, women were more likely than men to be on a ventilator and require inotrope or ECMO support and less likely to have a TAH, LVAD support, or an IABP. Most UNOS status 1B patients were receiving inotropes at the time of listing, with a similar percentage of women and men. In all UNOS status subgroups, women had slightly lower peak VO2 values than men. Right-heart catheterization showed slightly better hemodynamics for UNOS status 2 than 1A patients, with no sex differences except for lower cardiac output among women than among men in all subgroups. ABO blood types, previous cerebral vascular accident, and history of peripheral vascular disease were similar among all subgroups. Few patients were undergoing dialysis.
There were 1,290 women and 4,286 men who died during a median follow-up of 3.7 years. Women had a statistically significant worse survival than men when initially listed for OHT as UNOS status 1A (Fig. 1) but a better survival than men when listed as UNOS status 2 (Fig. 2). There were no significant sex differences in survival for patients initially listed as UNOS status 1B.
Higher mortality in women than in men initially listed as UNOS status 1A was associated with lower likelihood for undergoing OHT (Fig 3A). Lower mortality in women than in men initially listed as UNOS status 2 was associated with higher likelihood for OHT (Fig. 3C). There were no significant sex differences in competing outcomes between transplantation and death among patients awaiting OHT as UNOS status 1B patients (Fig. 3B). Both women and men in all UNOS subgroups had the highest mortality and transplantation rate within the first year after listing. There was a plateau in the mortality and transplantation curves around the second and third years after listing, except for the mortality curve among UNOS status 2 patients, which continued to rise with time.
Figure 4 demonstrates that female sex was still associated with a significant risk of death among UNOS status 1A patients (adjusted HR: 1.20; 95% CI: 1.05 to 1.37, p = 0.01 after adjusting for >30 confounding factors including age, ABO blood type, body mass index, GFR, IABP, ECMO, defibrillator, type of VAD, and era. In contrast, female sex was significantly protective for time to death among UNOS status 2 patients (adjusted HR: 0.75, 95% CI: 0.67 to 0.84, p < 0.001). No sex differences were noted among UNOS status 1B patients. Similar results were obtained when data were reanalyzed for women and men without dummy variables for missing data. The only significant interaction we found was between sex and presence of coronary artery disease (p value for interaction = 0.03).
To evaluate whether women were sicker than men at time of listing as UNOS status 1A, we created a model to determine the most likely characteristics of patients listed as UNOS status 1A with all variables excluding sex and UNOS status, using the entire OHT cohort in SRTR from 2000 to 2010. We then assessed the likelihood that a patient listed as status 1A would be a woman versus a man. We found women slightly more likely than men to have the characteristics of a UNOS status 1A patient (22.2% females vs. 20.9% males, p < 0.001).
To further evaluate sex differences in outcome among patients initially listed as UNOS status 1A, 1B, or 2, we analyzed changes in status at end of study (censored at time of transplantation, death, or last day of study while on waiting list). Among patients initially listed as UNOS status 1A, there were slightly fewer women than men remaining at urgent status and more women temporarily inactive on UNOS waiting list (63% females vs. 67% males remained status 1A at end of study; 16% females vs. 17% males were status 1B; 2% females vs. 1% males were status 2; and 19% females vs. 15% males were inactive status 7; p < 0.0001). When the cohort was limited to only those who underwent transplantation, there were no significant sex differences among the percentages of patients who were UNOS status 1A (82% females vs. 81% males were status 1A at end of study, p = 0.87), but fewer women than men who initially did not have mechanical circulatory support at the time of listing received a VAD or TAH at the time of transplantation (31% females vs. 42% males; p < 0.0001).
Among patients initially listed as UNOS status 1B, there were slightly fewer women than men who were at a higher status at the end of the study (26% females vs. 33% males were status 1A at end of study; 55% females vs. 51% males were status 1B; 2% females vs. 2% males were status 2; and 16% females vs. 15% males were inactive status 7; p < 0.0001). When the cohort was limited to only those who underwent transplantation, women were less likely to be UNOS status 1A and more likely to be UNOS status 1B at the time of transplantation (33% females vs. 40% males were status 1A at end of study; 65% females vs. 59% males were status 1B; 2% females vs. 1% males were status 2; p < 0.0001). Again more men than women who did not initially have mechanical circulatory support at time of listing had either a VAD or TAH at the time of transplantation (21% females vs. 30% males; p < 0.0001).
Among patients initially listed as UNOS status 2, women were less likely than men to have a higher status at the end of the study (12% females vs. 17% males were status 1A at end of study; 22% females vs. 24% males were status 1B; 44% females vs. 36% males were status 2; and 22% females vs. 23% males were inactive status 7; p < 0.0001). When the cohort was limited to only those who underwent transplantation, women were less likely than men to have a higher UNOS status at the end of the study (17% females vs. 26% males were status 1A at end of study; 30% females vs. 34% males were status 1B; 53% females vs. 40% males were status 2; p < 0.0001) and less likely to have VAD or TAH at the time of transplantation (7% females vs. 15% males; p <0.0001).
In a large, national transplantation registry, we found sex differences in mortality while patients awaited OHT. After we adjusted for possible confounding variables, female sex was associated with a higher risk of death among patients listed initially as UNOS status 1A, and male sex was associated with a higher risk of death among patients listed initially as UNOS status 2. No sex differences were noted among UNOS status 1B patients.
Our study adds to the growing concern that the current OHT allocation system needs to be refined(22–24). Over the last decade, there has been no significant change in the number of OHTs in the United States annually despite a high waitlist mortality (25). To minimize death on the waiting list, the current transplantation allocation system was based primarily on severity of illness. However, recent studies raise concern regarding racial disparity (26), appropriateness of elective 30-day UNOS status 1A time for patients with an LVAD (24,27,28), and transplantation of stable UNOS status 2 HF patients (29). Our study adds to this literature suggesting that a sex-specific disparity exists in waitlist survival.
Few studies have evaluated sex differences in mortality while patients await OHT (2,30). One small European study noted women had a higher mortality rate than men awaiting OHT, which remained after adjusting for age, HF survival score, serum creatinine level, inpatient status, cardiac index, low vocational level, smoking, and low emotional support at the time of transplantation listing (HR: 2.3; 95% CI: 1.04 to 5.12; p = 0.04) (2). A larger study in the United States using UNOS data analyzed 14,153 OHT candidates listed from 2003 to 2008 to determine effectiveness of transplanting UNOS status 2 patients. Patients were stratified by UNOS status, and based on univariate logistic regression models, women initially listed as UNOS status 1A or 1B had a higher risk than men for death/delisting due to severity of illness and a lower chance than men for transplantation, whereas the opposite was true for UNOS status 2 patients. The authors concluded that women benefited from being listed as UNOS status 2 and that removing this status would result in a larger sex disparity (30). Our study found similar sex differences in mortality while patients awaited OHT as UNOS status 1A and 2 patients, even after multivariate analysis accounting for >30 possible confounders.
The disparity in survival rates between women and men is of concern given the limited number of donor hearts available every year and the limited research in this field to further evaluate cause. For patients initially listed as UNOS status 1A, more women than men died on the waiting list. Based on the characteristics available that identified UNOS status 1A patients, we found women and men to have a similar profile at time of listing (22.2% predicted vs. 21.8% actual females and 20.9% predicted vs. 21.2% actual males were UNOS status 1A at time of listing). Therefore, the higher mortality rate in women than in men was less likely due to sex differences in severity of illness at time of listing. UNOS status 1A women were less likely than men to be bridged with VAD or TAH support at time of transplantation and more likely than men to be temporarily inactivated. The data remain limited, but these findings raise concern that women were not successfully bridged to transplantation while they remained at high status and were inactivated due to worsening condition. Objective evidence to support this is limited in the SRTR database as there was a high rate of missing important variables like hemodynamics, and data were only available at discrete time points (time of listing, time of transplantation, and change in status), preventing capture of any change in variable that affects prognosis but not UNOS status. We also did not have information as to why fewer women received mechanical circulatory support, which might have been due to fewer women than men being eligible for devices or fewer women than men who consented to devices. However, it is important to mention that the risk of survival for women did not change even after adjusting for >30 variables including mechanical circulatory support and era pre- and post-FDA approval of HeartMate II. As for UNOS status 2 patients, the differences in survival between women and men likely has to do with a lack of sex-specific OHT guidelines for peak oxygen consumption (31). We and others have shown that women tend to live longer than men with the same peak oxygen consumption value (32,33). Therefore, perceived differences in survival of ambulatory patients may be due to premature listing of women as UNOS status 2 when peak oxygen consumption values are similar to those in men. Unfortunately, this hypothesis cannot be explored because data for peak oxygen consumption were missing from approximately 50% of patients in the SRTR database, even in the cohort listed as UNOS status 2. However, it is supported by the fact that women had better survival than men on the waiting list despite fewer women receiving VAD or TAH support at time of transplantation and fewer women than men at higher status at time of OHT.
Sex differences in survival while awaiting OHT is a concern, and despite not identifying the cause, a solution to reduce mortality must be found. We propose changing a “rule-based” heart transplant allocation system to a “survival model-based” allocation system to account for sex differences in survival. We also propose making the process dynamic so that as variables for an individual change, the level of risk is adjusted. The lung allocation system is the best example, whereby regression models are used to balance differences and assign a weight score that is used to rank patients. The lung transplant waitlist survival model includes variables like oxygen dependency at rest, mechanical ventilation, pulmonary artery pressures, 6-min walk distance, and underlying lung disease. Prognostic risk factors can vary based on the category that defines a patient’s underlying lung disease. To create a similar strategy for OHT, more research will be needed to further define the variables associated with mortality and to determine whether there are possible interactions with other variables. We would also propose changing the system to a dynamic process where variables can be updated daily to change an individual’s risk score. This not only will provide a better allocation system for the patients, but it will also improve our current registry. A system that depends on entering essential variables will be properly updated with information and likely have a lower rate of missing data.
Our study has several important limitations. The validity of the data is dependent on accuracy upon data entry. To minimize human error, SRTR data are assessed by edit checks, validation of data at time of entry, and internal verification when there are outliers. Despite these attempts, there are still likely database errors. For instance, among ambulatory stable UNOS status 2 candidates are patients on mechanical ventilation (women = 0.6%, men = 0.7%), inotropes therapy (women = 5.5%, men = 6.1%), use of LVADs (women = 1.2%, men = 1.5%), and use of IABPs (women = 0.6%, men = 0.9%). These appear to be errors as the level of medical support does not match the severity of illness defined by UNOS status. However, individual centers may list at a lower UNOS status than clinically indicated, so it remains unknown whether these are actual database errors or patients intentionally labeled at lower status to prevent OHT while ill. Nonetheless, the low percent of “possible errors” would not be expected to alter the data significantly. It is also important to mention that the database is not inclusive of all objective data. In our multivariate analysis some important variables were not utilized such as natriuretic peptides, panel of reactive antibody and cardiac index. The lack of these variables, which were not available, limits the analysis not only for our study but also for future studies that may help in changing the allocation system. However, for the purpose of this study, it is unlikely that the proportion of missing data elements was systematically and markedly different by candidate sex. Thus, the potential for this incomplete data to alter the qualitative findings of the study should be minimal but does prevent further understanding of sex differences in waitlist survival. Another important limitation is the fact that baseline data entered may not be standardized. For instance, peak VO2 (ml/kg/min) is calculated using an individual’s weight, but there is no requirement to use lean body mass despite sex differences in body composition. Furthermore, if lean body mass is used to calculate peak VO2, there is no requirement as to how to calculate it (i.e., estimated vs. measured). Hemodynamics and serum laboratory values should be provided at the time of listing, but which values are entered if more than 1 is obtained is, again, determined by the individual center. Finally, despite the fact that the SRTR database is the best database available with which to study patients awaiting transplantation, it captures only information at given time points such as time of listing, time of transplantation, and time of status change. Therefore, variables that affect prognosis but do not change the status of the patient are not routinely updated.
In a large, national registry, we found sex differences in survival among patients awaiting OHT even after rigorous multivariable risk adjustment. The cause remains unknown but should raise concern as the current UNOS transplant criteria does not account for this disparity.
This research was funded by Heart and Vascular Institute, Cleveland Clinic, Cleveland, Ohio. Dr. Schold is a member of the Scientific Registry of Transplant Recipients Technical Advisory Committee. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and acronyms
- extracorporeal membrane oxygenation
- glomerular filtration rate
- heart failure
- intra-aortic balloon pump
- left ventricular assist device
- orthotopic heart transplantation
- pulmonary capillary wedge pressure
- Scientific Registry of Transplant Recipients
- total artificial heart
- United Network for Organ Sharing
- ventricular assist device
- Received February 7, 2014.
- Revision received February 24, 2014.
- Accepted March 7, 2014.
- American College of Cardiology Foundation
- ↵Scientific Registry of Transplant Recipients: Heart: Candidates: Gender 2000–2009 Heart Waiting List. Available at: http://srtr.transplant.hrsa.gov/annual_reports/2010/iHR_Candidates_gender.htm?o=5&g=1&c=1. Accessed on June 19, 2014.
- ↵OPTN Policy 3.7. Available at: http://optn.transplant.hrsa.gov/PoliciesandBylaws2/policies/pdfs/policy_9.pdf. Accessed June 19, 2014.
- Aaronson K.D.,
- Schwartz J.S.,
- Chen T.M.,
- Wong K.L.,
- Goin J.E.,
- Mancini D.M.
- Levy W.C.,
- Mozaffarian D.,
- Linker D.T.,
- et al.
- Lee D.S.,
- Gona P.,
- Vasan R.S.,
- et al.
- Douglas P.S.,
- Katz S.E.,
- Weinberg E.O.,
- Chen M.H.,
- Bishop S.P.,
- Lorell B.H.
- Piro M.,
- Della Bona R.,
- Abbate A.,
- Biasucci L.M.,
- Crea F.
- Hsich E.M.,
- Pina I.L.
- Shekelle P.G.,
- Rich M.W.,
- Morton S.C.,
- et al.
- Leppke S.,
- Leighton T.,
- Zaun D.,
- et al.
- Rogers J.G.
- ↵Scientific Registry of Transplant Recipients. Heart waiting list by gender 2000–2009. Available at: http://www.srtr.org/annual_reports/2010/survival_rates.htm. Accessed June 19, 2014.
- Singh T.P.,
- Almond C.S.,
- Taylor D.O.,
- Milliren C.E.,
- Graham D.A.
- Dardas T.,
- Mokadam N.A.,
- Pagani F.,
- Aaronson K.,
- Levy W.C.
- Wever-Pinzon O.,
- Drakos S.G.,
- Kfoury A.G.,
- et al.
- Lietz K.L.,
- Miller L.W.
- Elmariah S.,
- Goldberg L.R.,
- Allen M.T.,
- Kao A.