Liver Transplant Recipients Incur More Costly Care When Undergoing Treatment for Breast Cancer

Purpose The incidence of breast cancer (BC) in solid organ transplant recipients is comparable to the age-matched general population. It is among the three top-most commonly occurring cancers in women after liver transplantation (LT). There is limited information on the management and outcomes of breast cancer in liver transplant recipients (LTR). We aim to evaluate the impact of LT on breast cancer surgery outcomes. Further we compare the outcomes after breast cancer surgery in LTR in transplant versus non-transplant centers. Methods National Inpatient Sample (NIS) database was accessed to identify LTR with BC. Mortality, complications, hospital charges and total length of stay (LOS) were evaluated with multivariate logistic regression testing. Weighted multivariate regression models were employed to compare outcomes at transplant and non-transplant centers.


Introduction
Liver transplantation (LT) is the standard treatment for those with acute and chronic liver disease, as well as those with various types of liver neoplasms. Livers are the second most commonly transplanted organ in the United States 1 . The number of liver transplants performed in the United States has steadily increased over the past twenty years. Almost 9,000 liver transplants were performed in 2020 2 . The number of women surviving liver transplant over the age of 40 in the United States is 88.2%. The one, three, and ve-year survival rate for women over the age of 40 is 88.2%, 81.4%, 76.7% 3 .
Chronic immunosuppressive therapy, essential for allograft survival, remains the most important longterm risk factor contributing to morbidity following LT. Both infectious and neoplastic complications are much more common in the immunosuppressed host. Malignancies in transplant recipients often have a more rapid progression, an unfavorable prognosis, and a poor response to standard treatment [4][5][6] .
Therefore, as both the number of liver transplants performed and survival increases, identi cation and management of complications in these patients is paramount.
Breast carcinoma (BC) is the leading cause of new cancer diagnosis in women. Approximately 13% of women in the United States will develop breast cancer during their lifetime 7 . Treatment for breast cancer is individualized. However, the current mainstay of BC treatment is breast surgery. While the outcomes of those undergoing breast surgery are well documented in the general population, little is known regarding how patients fare if they have previously undergone a liver transplant.
The rate of reported de novo breast cancer following liver transplantation varies. There is a general consensus that the risk of BC does not appear to be increased in those having undergone solid organ transplant [8][9][10][11][12][13][14][15] . Nonetheless, once cancer develops in transplant recipients, the post-treatment outcomes may be worse than expected in the general population 16 . Despite this, little is known regarding the outcomes of breast surgery in LTR. Koonce et al reported no signi cant complications following breast reconstruction in those who previously underwent a solid organ transplant. However, this cohort consisted of only 17 women, two of whom underwent a LT 17 . Similarly, in a case report by Nakakimura et al., no severe adverse events were observed in one woman who underwent breast surgery and chemotherapy following a LT 18 . Others observed higher mortality when diagnosed with late stage breast cancer after liver transplantation 19 .
Breast cancer surgery outcomes data following solid organ transplantation has largely focused on those with kidney transplantation 20 . Consequently, little attention has been afforded to LTR subsequently treated for BC. Since liver transplantation has become a common procedure and recipients live with allografts, it is imperative to develop a greater understanding of the outcomes of breast cancer surgery in this cohort. Our purpose is to evaluate the in uence of LT on the short-term outcomes of breast cancer surgery in women at transplant and non-transplant centers.

Methods
Data from the National Inpatient Sample (NIS), between 2005 to 2014 on Breast Lumpectomy and Mastectomy were isolated (ICD 9 code: 85.20-85.23 and 85.33-85.36 and 85.41 -85.48). The NIS is the largest publicly available all-payer inpatient healthcare database designed to produce U.S. regional and national estimates of inpatient utilization, access, charges, quality, and outcomes 21 . A history of liver transplant (ICD9 code: V42.7) was determined within this subset. Hospital and patient-level characteristics between breast cancer with and without liver transplant were compared with t-test, Mann-Whitney test and chi-squared test.
The Elixhauser Comorbidity Index (1988) categorized and scored comorbidities. The Elixhauser Comorbidity Index is a method of categorizing comorbidities of patients based on the International Classi cation of Diseases (ICD) diagnosis codes found in administrative data 22 . A greater score is associated with worse prognosis 23 . The in uence of LT on mortality and morbidity was evaluated with logistic regression testing. Total hospital charge and length of stay were converted to a binary variable based on their median. The role of LT on the total hospital charge and length of stay was evaluated with logistic regression testing, where the dependent variable was length of stay or total hospital charge below or above median. Similarly, the effect of LT on total hospital charge and length of stay (LOS) was measured with linear regression. Total charges were adjusted based on consumer price index (CPI) 2020. Since there was no mortality in the LT cohort, only three of the outcomes were assessed in a multivariate fashion.
Multivariate logistic regressions were performed to compare outcomes sorted by transplant center (TC), teaching centers, and patients who underwent reconstruction following breast cancer surgery. The selected co-variates were standard patient and hospital characteristics in NIS which were statistically signi cant between LT and no LT. These include race, co-morbidity, primary expected payer, zip code income quartile, hospital ownership, location/teaching status, and region. Missing values are reported in Table I and Table II and were coded for the covariates. There was no exclusion in the result of multivariate logistic regression. We identi ed TC as hospitals with at least one liver transplant performed during the timeframe. All results were calculated after applying the sampling weight built in NIS.

Results
A total of 736,626 women underwent surgery for breast cancer. Of these, 99 received liver transplantation. There was no statistical difference in terms of age at the time of diagnosis of breast cancer. The majority of women in each cohort were white, with a signi cantly higher percentage of women in the LT group being white (65.3% vs. 62.1%, p <0.001). Of the 99 LTR, 69.7% had an Elixhauser comorbidity score of 3 or greater (median score of 4), while only 21.5% of non-LT patients had a score of 3 or greater (median score of 0) (p<0.001, Table I). Socioeconomic status for the LTR cohort was higher than the non-LTR cohort, as these women belonged mostly to the higher income quartile (third quartile 35.8% vs. 24.3%, p 0.017). There was no statistical difference in the frequency or type of surgical procedure, lumpectomy or mastectomy (Table I).
The dominant payment method was private insurance (50.8%). However, Medicare was a more common method of payment for the LTR group compared to the non-LTR group (65.0 vs. 36.6%, p <0.001). Most centers were public hospitals, with large bed size, and urban teaching a liates. Although there were some statistical differences in the components of these variables, overall, these hospitals were comparable (Table II).
Liver transplant recipients underwent breast cancer surgery predominantly in transplant centers when compared to non-LTR (35.0% vs. 23.2%, p=0.004). There were no deaths in the 99 liver transplant recipients. There were signi cantly more complications in the LT cohort when compared to the non-LTR (15.0% vs. 8.2%, p=0.012). However, on multivariate analysis, undergoing LT was not an independent risk factor for post-operative complications followed breast cancer surgery (OR 1.223 p 0.480) (Table IV). Total hospital charges for breast cancer surgery were higher in the liver transplant group ($63,724 vs. $43,003, p<0.001). (Table III) LOS for breast cancer surgery in the reconstructed group was signi cantly shorter in the liver transplant group (LOS> 2 days OR 0.170, p 0.002) (Table IV).

Discussion
Organ transplantation has signi cant survival and quality of life bene ts compared to best medical (nontransplant) management. One of the most important factors that has allowed for prolonged allograft survival has been the advances in immunosuppressive regimens. Although de novo malignancies are known long-term complications of organ transplantation, breast cancer is not increased in the transplant population when compared against age-matched SEER general population data. Incident rates in published literature show age-speci c breast cancer incidence after 50 years old in those with liver transplantation similar to that of the general population. Our sample size is small when compared to the overall incidence in the literature of de novo breast cancers in those with liver transplants. Nonetheless, after weighting, our results re ect a realistic appraisal of patients with breast cancer and liver transplantation.
The care of the liver transplant recipient requires a lifelong multidisciplinary effort by a wide range of specialists. Clinicians must not only consider all of the transplant-related complications, but also typical age-related comorbidities. Moreover, chronic immunosuppressive therapy can induce or accelerate some conditions that the non-transplant patient may not be routinely monitored for, speci cally malignancy.
Centralized and specialized management of breast cancer in the liver transplant recipient is paramount. On multivariate analysis, undergoing a LT was not an independent risk factor for developing a post-operative complication (OR 1.223 p 0.480) (Table IV). This suggests that factors other than liver transplantation are associated with development of post-operative complications.
A signi cantly higher proportion of LTR had an Elixhauser comorbidity score of ≥ 3 (69.7% vs. 21.5%, p<0.001, Table I), indicating that LT patients suffered from a higher degree of co-morbidity. However, our data shows that despite the LTR having signi cantly more comorbidities, there were no differences in mortality, complication rate, total charge, or length of stay when these patients were managed at a transplant center (Table IV).
Breast cancer management in LTR at non-transplant centers incurred higher charges but no difference in complication rate nor LOS when compared to breast cancer management in LTR at transplant centers.
Of women who received a liver transplant, length of stay following breast cancer surgery was signi cantly shorter in the group which underwent breast reconstruction. (OR <1, p 0.002) (Table IV). This may be due to the fact that, in general, immediate breast reconstruction is performed by careful selection of those patients who are possibly overall healthier. We do not have knowledge of pre-and posttransplant performance status, immunosuppressive regimens, or pretransplant health that may overall lend to healthier LTR and thus ability to withstand an immediate breast reconstruction with acceptable outcomes and LOS.
Our analyses revealed no statistical difference when comparing the overall survival of the two cohorts. This mirrors previous reports. Jeong et al. compared the prognosis of post-transplant breast cancer patients receiving immunosuppressants to general breast cancer survivors. All individuals had previously undergone either a liver or kidney transplant. They discovered that after matching by tumor size, lymph node metastasis, and age, disease-free survival, breast-cancer speci c survival, and overall-survival were not signi cantly different between the two cohorts 24 .
A nal, notable point is the fact that total hospital charges for breast cancer surgery were higher in the liver transplant group, even after controlling for other variables (OR 2.621, p <0.001) (Table IV). This may be explained by LTR suffer from a higher degree of co-morbidity. An analysis of 126,664 individuals with breast cancer, revealed the average medical cost per patient with comorbidity was higher compared to the average medical cost per person without comorbidity (p<0.05) 25 . We hypothesize that increased comorbidities in the LT cohort may have played a role in these women incurring higher costs for breast cancer surgery. This analysis is not without limitations, as there is inherent weakness of large database analysis. Time between liver transplantation and breast cancer surgery was not known. Prolonged periods of immunosuppressive treatment may induce DNA damage and inhibit immune surveillance mechanisms, thus increasing risk of lymph node metastases which would require more extensive BC surgery, possibly axillary node dissection, with locally advanced disease at presentation 24 . Additionally, immunosuppressive medications are unknown. This prevented us from stratifying outcomes based on type of immunosuppressive agent. Further, information on the breast cancer stage and neoadjuvant chemotherapy treatment prior to surgery is unknown. Thus, we were unable to assess outcomes on early versus advanced disease. Similarly, long term patient outcomes are not available due to database limitations, and could not be assessed, and may differ from the reported short term outcomes in our analyses.
Additional research is needed to more comprehensively understand the di culties that post-liver transplant breast cancer survivors face following breast cancer surgery compared to the general breast cancer population. Future analyses should consider factors such as breast cancer stage, type of immunosuppressive therapy, and time since LT.
The strengths of this manuscript are re ected in this being the largest and rst reported analyses that determines that prior liver transplantation does not increase morbidity nor mortality in women undergoing surgery for breast cancer. However, we found total hospital charges for breast surgery were signi cantly higher in LTR. These results may be used to guide clinical practice when treating women for breast cancer who have undergone a liver transplant. Code availability: Data and coding available upon request.
Authors' contributions All authors contributed to the study conception and design. All authors read and approved the nal manuscript.