Original Contribution

Impact of Sex on Outcomes Among Patients With Cardiac Allograft Vasculopathy Who Undergo Percutaneous Coronary Intervention

Michael S. Lee, MD1; Subeer K. Wadia, MD1; Richard Shlofmitz, MD²

Michael S. Lee, MD1; Subeer K. Wadia, MD1; Richard Shlofmitz, MD²

Abstract: Background. Orthotopic heart transplantation (OHT) is a reasonable treatment strategy for patients with end-stage heart failure. Percutaneous coronary intervention (PCI) can be performed to treat cardiac allograft vasculopathy (CAV). We sought to examine sex-based differences and outcomes in patients undergoing PCI for CAV. Methods. This was a single-center, retrospective study of consecutive patients undergoing PCI for CAV from July 1993 to July 2017. The primary outcome was a composite of death, myocardial infarction (MI), target-vessel revascularization (TVR), or repeat OHT. Results. A total of 140 patients (39% women) who underwent PCI for CAV were studied over a median follow-up of 3.9 years. Women were more likely to be younger and had less hypertension, with a trend toward less diabetes and higher ejection fractions. No significant differences in the primary composite outcome were noted between women and men (hazard ratio, 1.16; 95% confidence interval, 0.76-1.76; P=.49), as well as the individual outcomes of death, MI, or TVR (P>.05 for all). Women were more likely to undergo repeat OHT than men (hazard ratio, 2.93; 95% confidence interval, 1.48-5.81; P<.01). After covariate adjustment, female sex and younger age were independently associated with repeat OHT for CAV (P<.05). No differences in the primary outcome as well as the individual outcomes of death, MI, or TVR were noted between sex groups in patients who underwent PCI for CAV. Conclusion. Female patients were more likely to undergo redo OHT for CAV. Younger age at time of initial OHT was associated with a need for repeat OHT.

J INVASIVE CARDIOL 2020(32):12:453-458. Epub 2020 November 22.

Key words: cardiac allograft vasculopathy, percutaneous coronary intervention, orthotopic heart transplant

Cardiac allograft vasculopathy (CAV) affects up to 47% of patients 10 years after orthotopic heart transplantation (OHT) and remains an important cause for graft failure, repeat OHT, and death.1-4 Predictors of CAV in the International Society of Heart and Lung Transplantation (ISHLT) 2019 registry1 include older donor age, donor history of hypertension, younger recipient age, greater numbers of HLA-DR mismatches, and recipient pretransplant diagnosis of ischemic heart disease.

Females who undergo OHT have longer waitlist times and receive higher-risk hearts than men.5-8 Despite the differences in OHT allocation, studies have demonstrated mixed results with respect to sex-based differences in survival following OHT. In this study, we sought to examine outcomes and sex-based differences in patients undergoing percutaneous coronary intervention (PCI) for CAV.


This was a retrospective study of consecutive patients undergoing PCI for a diagnosis of CAV at the University of California, Los Angeles (UCLA) Medical Center between July 1993 and July 2017 using a previously collected database of patients.9 Patients were included if they had ISHLT CAV grade 2 (moderate, left main 50%-69% or a single primary vessel >70%, or isolated branch stenosis >70% in branches of 2 systems) or CAV grade 3 (severe, left main >70%, or ≥2 primary vessels >70%, or isolated branch stenosis >70% in all 3 systems).10 All patients were on standard immunosuppressive therapy for the respective time. Prior to 2000, the standard immunosuppressive regimen included cyclosporine, prednisone, and mycophenolate mofetil. In 2000, tacrolimus supplanted cyclosporine. Sirolimus was added if CAV was diagnosed on coronary angiography or intravascular ultrasound. 

The initial revascularization technique was plain old balloon angioplasty. Bare-metal stents were used from 1996 to 2003. Thereafter, drug-eluting stents became the preferred stent type. Cypher sirolimus-eluting stents (Cordis Corporation) were first used in 2003 and Taxus paclitaxel-eluting stents (Boston Scientific Corporation) were first used in 2004. Everolimus-eluting stents (Abbott Vascular) were the drug-eluting stent of choice starting from 2009 until the end of the study period. All patients were treated with statin and aspirin unless contraindicated. Patients were treated with a P2Y12 inhibitor for a minimum duration of 1 month after PCI with bare-metal stent and a minimum of 6 months with drug-eluting stent. Standard PCI techniques for CAV patients have been previously described.11 

Electronic medical records were searched for baseline characteristics. Patients were followed in the heart transplant clinic after PCI. Data were collected from medical records, telephone encounters, and Social Security Death Index. Surveillance angiography with intravascular ultrasound was routinely performed at 6-12 months or sooner if clinically indicated. The UCLA Medical Center institutional review board approved the use of the database review for this study.  

Definitions. The primary outcome was a composite of death, myocardial infarction (MI), target-vessel revascularization (TVR), or repeat OHT. Secondary analyses of the individual outcomes were performed. MI was defined based on the American College of Cardiology definition.12 TVR was defined as any repeat revascularization within the target vessel. Stent thrombosis was defined as definite or probable according to the Academic Research Consortium definition.13

Statistical analysis. Descriptive baseline data are presented as median with interquartile range (IQR) or count with percentages, unless otherwise specified. Differences in the distributions of baseline continuous variables were tested for significance using the Mann-Whitney U-test with a threshold two-sided P<.05. Differences in categorical variables were analyzed with Pearson’s Chi-square or Fisher’s exact tests, as appropriate based on sample sizes. Kaplan-Meier curves using log-rank tests were constructed to graphically display time to outcome events. Predictor variables were assessed for time-dependent association with the primary outcome in a Cox proportional hazards regression model. The proportional hazards assumption was assessed graphically with log-log plots. Hazard ratios (HRs) are presented with 95% confidence intervals (CIs). 

Following the initial analysis using Cox regression, variables with significant association to outcomes were reanalyzed using the classification and regression tree (CART) methodology. Briefly, CART recursively splits a data set by independent variables (predictors) that are maximally different with respect to a dependent variable (outcome) of interest, potentially discovering complex interactions.14-16 CART analysis ultimately yields a decision tree based on the Gini index, a measure of dissimilarity. Once independent variables (predictors) are entered into the model, the dependent variable (outcome of interest) is divided by the independent variables to maximize the Gini index of dissimilarity. The order of splitting is thus determined by the Gini index to yield the most homogenously similar subgroups at each level in the decision tree. To avoid indefinite splitting, we chose a priori stopping criteria of 6 for parent nodes and 3 for terminal nodes. Statistical analyses were implemented using IBM SPSS Statistics 26 (IBM Corporation).


A total of 140 patients (39 women [28%], 101 men [72%]) who underwent PCI for CAV at our institution were studied over a combined median follow up of 3.9 years (IQR, 1.4-8.4 years). Female patients were more likely to be younger and had less hypertension, with a trend toward less diabetes and higher ejection fraction (Table 1). Procedural characteristics, including stent sizes and types, were similar (Table 2).

At 5 years post PCI for CAV, the primary composite outcome event occurred in 101 patients (72% of the study population). No significant differences in the primary composite outcome were noted between female and male patients (5-year freedom from event: female 29% [95% CI, 14-43] vs male 28% [95% CI, 18-38], log-rank P=.49; HR, 1.16 [95% CI, 0.76-1.76], P=.49) (Figure 1). No differences were noted in the individual outcomes of death, MI, or TVR (Table 3 and Figures 2-4). Female patients were more likely to undergo repeat OHT than male patients (HR, 2.93; 95% CI, 1.48-5.81; P<.01) (Figure 5). This finding remained significant after covariate adjustment for diabetes, hypertension, years after initial OHT, and age (female vs male adjusted HR, 2.46; 95% CI, 1.16-5.21; P=.02). Younger age was also independently associated with repeat OHT (younger per year adjusted HR, 1.04; 95% CI, 1.02-1.06; P<.001).

A CART analysis was conducted to further assess the relationship and interactions between patient sex, age at initial OHT, and years since initial OHT with the need for repeat OHT for CAV. Figure 6 shows the results of the CART with the “best fit” decision tree. The best first determinant in separating the group based on the risk of repeat OHT is the patient age at initial OHT, where undergoing initial OHT before the age of 60 years confers a higher risk of undergoing repeat OHT (36% of patients <60 years old vs 10% of patients ≥60 years old). Additional factors affecting splitting of the group included patient sex and years since initial OHT. As demonstrated by the decision tree, the cohort with the highest risk or most likely to undergo repeat OHT was female patients <60 years old at the time of their initial OHT and >8 years out from their initial OHT (75% of patients, annualized risk of 1.7% per patient year). The group least likely to undergo repeat OHT for CAV was male patients who were >60 years old at the time of their initial OHT (independent of years since initial OHT), with 8% of these patients undergoing repeat OHT for CAV (annualized to 0.1% per patient year). The CART decision tree correctly classified patient risk of undergoing repeat OHT in our dataset by 81%.


In patients undergoing PCI for CAV, women were more likely to be younger with fewer comorbidities than men. Following PCI, there was no significant difference in our primary composite outcome of death, MI, TVR, or repeat OHT. However, in secondary analyses, female sex was independently associated with an increase in repeat OHT over the follow-up period when compared with male sex. Female patients in our cohort were younger, with fewer comorbidities compared to male patients. This is consistent with a registry analysis,8 which reported that women who receive OHT are younger and have lower rates of hypertension, diabetes, peripheral vascular disease, tobacco use, and prior cardiac surgery when compared with men. 

Although rates of CAV after OHT have decreased slightly over time with improvement in medical therapy as well as improved awareness and surveillance,1 CAV remains the most prevalent indication for repeat OHT (79% of repeat OHT patients).17 Limited data exist on the prognosis of patients who undergo repeat OHT for CAV. Small patient series suggest a recurrence rate of CAV of 19% among patients who undergo repeat OHT.18 With limited donor availability and the high risk of redo sternotomy, ethical concerns about repeat OHT also exist, underscoring the importance of identifying patients at highest risk for repeat OHT. 

The results of the CART decision tree further refine the risk of repeat OHT among patients with CAV who undergo PCI. Our findings demonstrate an elevated risk of needing repeat OHT for female patients compared to male patients. Younger patients — especially those >8 years out from their initial OHT — were more likely to need repeat OHT. However, even within this subset, the risk of female status dramatically increased the risk of needing repeat OHT. The risks increase substantially from 39% (0.5% per patient year) to 75% (1.7% per patient year) for younger female patients who are >8 years from their initial OHT. In contrast, younger male patients had a more modest increase in risk from 22% (0.2% per patient year) to 29% (0.4% per patient year) if they were >8 years out from their initial OHT.

The reasons for the difference in risk among female and male patients remain unclear. Given that prior studies have reported women receive <25% of available heart transplants in the United States and are more often recipients of higher-risk donor hearts,5-8 the 39% female representation in our study is higher than expected. In multivariate analysis, female sex and younger age were independently associated with repeat OHT for CAV. This higher-than-expected female representation in our study, along with the independent association of female sex on repeat OHT, suggests that women have worse outcomes after initial OHT than men, potentially due to women receiving higher-risk donor hearts than men.

Alternative possibilities exist to explain the difference in repeat OHT rates between women and men, and can only be hypothesized. Multiparous women are known to have increased sensitization due to prior antibody exposures.19 Whether this plays a role in the subsequent development of CAV remains uncertain. Women are known to receive less aggressive therapy for heart failure compared with men.20 

Study limitations. This was a retrospective evaluation at a single institution. Patient selection, institutional variation, data completion percentages, and lack of standardized protocols are inherent biases. As the database was a cohort of patients with CAV undergoing PCI, no comparisons with OHT patients without CAV were performed. The data presented cannot be extrapolated to other important causes of graft failure following OHT, such as acute or chronic rejection. CART is a non-parametric technique, and caution must be exercised when generalizing the results to a broader population. Although prespecified stopping criteria help to prevent overfitting, some degree of terminal node instability is inherent to the decision tree design. While the CART analysis was useful in identifying relationships in this single-center cohort to identify risk of repeat OHT, the decision tree needs further validation and confirmation in a larger population — ideally, in a multicenter cohort. Finally, due to the application of stopping criteria to prevent overfitting, there may be additional variables/risk factors that were not provided in the decision tree, but may be important predictors for repeat OHT.


There was no significant difference in our primary composite outcome of death, MI, TVR, or repeat OHT at long-term follow-up among female and male patients with CAV who underwent PCI. However, female patients were more likely to undergo redo OHT for CAV. Younger age at time of initial OHT was associated with a need for repeat OHT.

From the ¹Division of Cardiology, UCLA Medical Center, Los Angeles, California; and the ²Division of Cardiology, St. Francis Heart Center, Roslyn, New York.

Disclosure: The authors have completed and returned the ICMJE Form for Disclosure of Potential Conflicts of Interest. The authors report no conflicts of interest regarding the content herein.

Manuscript accepted April 6, 2020.

Address for correspondence: Michael S. Lee, MD, UCLA Medical Center, 100 Medical Plaza, Suite 630, Los Angeles, CA 90095. Email: mslee@mednet.ucla.edu

  1. Khush KK, Cherikh WS, Chambers DC, et al. The International Thoracic Organ Transplant Registry of the International Society for Heart and Lung Transplantation Thirty-Sixth Adult Heart Transplantation Report – 2019; focus theme: donor and recipient size match. J Heart Lung Transplant. 2019;38:1056-1066.
  2. Costanzo MR, Naftel DC, Pritzker MR, et al. Heart transplant coronary artery disease detected by coronary angiography: a multi-institutional study of preoperative donor and recipient risk factors. Cardiac Transplant Research Database. J Heart Lung Transplant. 1998;17:744.
  3. Keogh AM, Valentine HA, Hunt SA, et al. Impact of proximal or midvessel discrete coronary artery stenoses on survival after heart transplantation. J Heart Lung Transplant. 1992;11:892.
  4. Agarwal S, Parashar A, Kapadia S, et al. Long-term mortality after cardiac allograft vasculopathy: implications of percutaneous intervention. JACC Heart Fail. 2014;2:281-288.
  5. Hsich EM, Starling RC, Blackstone EH, et al. Does the UNOS heart transplant allocation system favor men over women? JACC Heart Fail. 2014;2:347-355.
  6. Kobashigawa JA. U.S. donor heart allocation bias for men over women? A closer look. JACC Heart Fail. 2014;2:356-357.
  7. Schulze PC, Kitada S, Clerkin K, Jin Z, Mancini D. Regional differences in recipient waitlist time and pre- and post-transplant mortality following the 2006 UNOS policy changes in the donor heart allocation algorithm. JACC Heart Fail. 2014;2:166-177.
  8. Moayedi Y, Fan CP, Cherikh WS, et al. Survival outcomes after heart transplantation: does recipient sex matter? Circ Heart Fail. 2019;12:e006218. Epub 2019 Oct 10.
  9. Lee MS, Lluri G, Finch W, Park KW. Role of percutaneous coronary intervention in the treatment of cardiac allograft vasculopathy. Am J Cardiol. 2018;121:1051-1055.
  10. Mehra MR, Crespo-Leiro MG, Dipchand A, et al. International Society for Heart and Lung Transplantation working formulation of a standardized nomenclature for cardiac allograft vasculopathy – 2010. J Heart Lung Transplant. 2010;29:717-727.
  11. Lee MS, Kobashigawa J, Tobis J. Comparison of percutaneous coronary intervention with bare-metal and drug-eluting stents for cardiac allograft vasculopathy. JACC Cardiovasc Interv. 2008;1:710-715.
  12. Thygesen K, Alpert JS, Simoons ML, et al. Third universal definition of myocardial infarction. J Am Coll Cardiol. 2012;60:1581-1598.
  13. Cutlip DE, Windecker S, Mehran R, et al. Academic Research Consortium. Clinical end points in coronary stent trials: a case for standardized defintiions. Circulation. 2007;115:2344-2351.
  14. Zimmerman RK, Balasubramani GK, Nowalk MP, et al. Classification and regression (CART) analysis to predict influenza in primary care patients. BMC Infect Dis. 2016;16:503. 
  15. Strobl C, Malley J, Tutz G. An introduction to recursive partitioning: rationale, application, and characteristics of classification and regression trees, bagging, and random forests. Psychol Methods. 2009;14:323-348.
  16. Bertolet M, Brooks MM, Bittner V. Tree-based identification of subgroups for time-varying covariate survival data. Stat Methods Med Res. 2016;25:488-501.
  17. DePasquale EC, Cheng R, Nsair A, et al. Cardiac retransplantation: how far have we come? J Am Coll Cardiol. 2014;63(12 Suppl):A807.
  18. McCreath L, McCubrey R, Folsom J, et al. Cardiac allograft vasculopathy in redo-transplants: is it more or less (or) the same the second time around? J Heart Lung Transplant. 2015;34(Suppl):S62.
  19. Velez M, Johnson MR. Management of allosensitized cardiac transplant candidates. Transplant Rev (Orlando). 2009;23:235-247.
  20. Frankenstein L, Clark AL, Ribeiro JP. Influence of sex on treatment and outcome in chronic heart failure. Cardiovasc Ther. 2012;30:182-192.