Cancer is a major cause of morbidity and mortality in the United States (US), with an estimated 1,958,310 cases and 609,820 deaths in 2023 [1]. The cancer mortality rate, which reflects both cancer incidence and survival after a cancer diagnosis, is 177.5 per 100,000 persons for males and 128.7 per 100,000 persons for females [1]. This sex difference in overall cancer mortality also holds for most cancers at shared anatomic sites (excluding sex-specific cancers), where males typically have higher cancer mortality, with the highest male-to-female mortality ratios seen for melanoma of the skin, head and neck, and thyroid cancers [2, 3].

Importantly, males are more likely than females to die from their cancer after diagnosis [3]. Increased mortality among males with cancer may be due to weaker immune control of cancer after diagnosis compared to females, a process referred to as immunosurveillance or immunocontainment [4]. In support of this hypothesis, females exhibit a more robust cell-mediated and humoral immune response to a host of antigenic challenges [5]. Sex steroid hormones may mediate these immune differences [6]. Further, many genes related to immunity and tumor suppressor genes, escape X-inactivation, resulting in biallelic expression in females [7]. Females also have increased response rates and survival after chemotherapy compared to males [8], which may be because the half-life of many chemotherapeutics is longer in females [9]. Mortality from bladder cancer is an exception, as males have better 5-year survival than females after adjustment for age, stage, and grade [10].

Exploring sex differences in cancer mortality among solid organ transplant recipients can help clarify the role of immunity in driving sex differences in cancer mortality. Transplant recipients are immunosuppressed due to the requirement for lifelong use of post-transplantation immunosuppressive medications to prevent graft rejection. Transplant recipients with cancer have elevated cancer-specific mortality compared to other individuals with cancer, which could be due to the effects of this immunosuppression [11]. If sex differences in mortality among cancer patients in the general population are explained by differences in immunity, then the female advantage may be reduced in a population of solid organ transplant recipients.

We used data from the Transplant Cancer Match (TCM) Study, which links the US national solid organ transplant registry (Scientific Registry of Transplant Recipients [SRTR]) to 30 population-based cancer registries [12]. The TCM Study includes cancer registry data on cancer diagnoses for two-thirds of US solid organ transplant recipients since 1995. Along with demographic and transplant data, the TCM Study includes data on dates of cancer diagnoses, first course of cancer treatment (chemotherapy, radiation, and surgery), vital status, and cause of death. This study was not considered human subjects research by the National Cancer Institute. The TCM Study was approved by institutional review boards at participating cancer registries as required.

From TCM cancer registries, we identified cancer patients diagnosed between January 1, 1995 and December 31, 2019, and we used the linked SRTR data to select those who had had a solid organ transplant before their cancer diagnosis. We excluded individuals outside the four race/ethnic groups of non-Hispanic White, non-Hispanic Black, Hispanic, and Asian American/Pacific Islander due to small numbers. We included first malignant cancer diagnoses and evaluated non-sex-specific solid tumors with at least 200 cases among transplant recipients. The examined cancers included lip, stomach, colorectum, anus, liver (including intrahepatic bile duct), pancreas, larynx, lung, melanoma of the skin, kidney (including renal pelvis), urinary bladder, brain, and thyroid. For comparison, we also evaluated individuals in the general US population with the same types of cancer, using data from 12 Surveillance, Epidemiology, and End Results Program (SEER) cancer registries for the same calendar years [13]. Cancers diagnosed at autopsy were excluded in both datasets. Due to the rarity of organ transplantation, the general population patients can be considered to lack a prior transplant.

We identified deaths due to cancer among transplanted and general population cancer patients based on cause of death information as described by Howlader [14], and calculated corresponding cancer-specific mortality rates (we note that these mortality rates are restricted to cancer patients and thus reflect only mortality after cancer diagnosis). Separately for cancer patients with a prior transplant in the TCM study and cancer patients in the general population in SEER, we compared cancer-specific mortality in males and females using a Cox proportional hazards model. Specifically, we calculated male-to-female hazard ratios (M:F HRs) for death due to cancer, with follow-up from the time of the cancer diagnosis until death from cancer or censoring due to death from other causes or end of cancer registry follow-up (the latest follow-up date was December 31, 2019).

To detect a difference in the M:F HR between cancer patients in the transplant and general populations, we combined the TCM and SEER populations and tested for an interaction in the Cox model between sex and cohort (TCM vs. SEER) using the P-value from a Wald chi-square test. To understand whether an effect of estrogen on immunity may play a role in cancer mortality, we also performed separate analyses according to age at cancer diagnosis (<50 vs ≥50 years based on the typical age of menopause). We also conducted an analysis restricted to cancer patients diagnosed with localized stage cancer who had received treatment (one or more of chemotherapy, radiation, and surgery). To examine whether the degree of immunosuppression affects the results, we also stratified by transplanted organ. Liver transplant recipients receive the lowest dose of immunosuppressant medications, followed by kidney recipients, with recipients of lung and heart transplant receiving the largest doses. If immunity explains the sex difference in cancer-specific mortality, we would hypothesize that the M:F HR would be greatest among liver recipients and more attenuated for heart/lung recipients. Differences between the M:F HR among the three organ groups were tested using an interaction in the Cox model between sex and organ using the P-value from a Wald chi-square test. All models were adjusted for age at cancer diagnosis (<40 years, 40–79 years in 10-year increments, and 80+ years), race and ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic, and Asian-American Pacific Islander), stage of cancer (localized, regional, distant, or unknown), year of diagnosis (1995–1999, 2000–2004, 2005–2009, 2010–2014, and 2015–2019), and first course of cancer treatment based on receipt of chemotherapy (yes, no/unknown), radiation (yes, no/unknown), and surgery (yes, no/unknown) where applicable.

All analyses were conducted in SAS version 9.4 (The SAS System, Cary, NC). We present 95% confidence intervals (CIs) based on a level of statistical significance of P <0.05.

The analysis cohort comprised 17,048 cancer patients with a prior transplant (11,535 males and 5,513 females) and 1,221,914 patients from the general population (702,758 males and 569,156 females) with cancers of the lip, stomach, colorectum, anus, liver, pancreas, larynx, lung, skin (melanoma), bladder, kidney, brain, and thyroid (Table 1). The most common cancer was lung cancer among patients both with a prior transplant and in the general population. Cancer patients with a prior transplant were younger than patients in the general population (median age 60 vs. 65–69 years). The majority of both groups were non-Hispanic White (67.5% in the transplant population vs. 70.1% in the general population), and the cancer stage distribution was similar.

Cancer-specific mortality rates were highest for patients with pancreatic, lung, stomach, and liver cancers (Table 2). Among cancer patients with a prior transplant, males had significantly higher cancer-specific mortality than females for cancers of the lung (M:F HR: 1.20; 95% CI: 1.11, 1.29), kidney (M:F HR: 1.33; 95% CI:1.11, 1.60), and thyroid (M:F HR: 2.75; 95% CI: 1.10, 6.85). Among cancer patients in the general population, males had higher cancer-specific mortality than females for cancers of the lip, stomach, colorectum, anus, liver, lung, skin, and thyroid, with M:F HRs ranging from 1.06 to 1.59, while females had higher cancer-specific mortality for bladder cancer (M:F HR: 0.95; 95% CI: 0.93, 0.98).

When comparing the M:F HRs for cancer-specific mortality between cancer patients with a prior transplant and those in the general population, we found a significant interaction only for colorectal and kidney/renal pelvis cancers. Specifically, for colorectal cancer, males had increased cancer-specific mortality compared with females among patients in the general population, but no sex difference in cancer-specific mortality was observed among patients with a prior transplant (M:F HR_GenPop: 1.07; 95% CI: 1.06, 1.08; M:F HR_Transplant: 0.89; 95% CI: 0.77, 1.03; P-interaction=0.007). In contrast, for kidney cancer, there was no sex difference for patients in the general population, but cancer-specific mortality was higher for males than females among patients with a prior transplant (M:F HR_GenPop: 1.02; 95% CI: 0.99, 1.04; M:F HR_Transplant: 1.33, 95% CI: 1.11, 1.60; P-interaction=0.003).

We next looked at results stratified by age at cancer diagnosis (Table 3). Among cancer patients aged <50 years in the general population, males had worse survival than females for cancers of the lip, colorectum, anus, liver, pancreas, lung, skin, kidney, and thyroid with M:F HRs ranging from 1.11 to 4.61. The number of cancer patients with a prior transplant in this age group was somewhat limited, ranging from 25 to 1,240 patients across the cancer types. In this age group, only the M:F HRs for anal cancer differed significantly between the two populations, with a two-fold increase in cancer-specific mortality for males in the general population but no sex difference among those with a prior transplant (M:F HR_GenPop: 2.02; 95% CI: 1.71, 2.39 vs. M:F HR_Transplant: 0.67; 95% CI: 0.24, 1.89; P-interaction=0.024).

In the general population, M:F HRs were mostly weaker among patients aged ≥50 years at diagnosis compared with younger patients (Table 3). Among cancer patients in the general population aged ≥50 years, males had increased mortality compared to females for cancers of the stomach, colorectum, anus, liver, lung, skin, and thyroid with M:F HRs ranging from 1.03 to 1.50. The M:F HR estimates for colorectal and kidney cancers differed significantly between the two populations. Specifically, for colorectal cancer, there was an increased risk of mortality among male patients in the general population but no sex difference in cancer-specific mortality among patients with a prior transplant (M:F HR_GenPop: 1.07; 95% CI: 1.05, 1.08 vs. M:F HR_Transplant: 0.87; 95% CI: 0.75, 1.02; P-interaction=0.018). For kidney cancer, in contrast, there was no sex difference in cancer-specific mortality among patients the general population but a male bias for patients with a prior transplant (M:F HR_GenPop: 1.00; 0.97, 1.03 vs. HR_Transplant: 1.31; 95% CI: 1.07, 1.61; P-interaction=0.025).

When we restricted to cancer patients in the transplant and general populations who were diagnosed with localized stage cancer receiving treatment (Supplemental Table 1), we did not see substantial differences from the main results. The exception is for larynx cancer where females had worse cancer-specific mortality than males in the general population (M:F HR_GenPop: 0.89, 95% CI: 0.80, 0.99) and this association was stronger in the transplant population (M:F HR_Transplant: 0.42; 95% CI: 0.18, 0.97; P-interaction=0.035). As shown in Supplemental Table 2, we found interactions between sex and transplanted organ type (a proxy for degree of immunosuppression) for only two cancer sites: melanoma of the skin (P-interaction=0.016) and brain tumors (P-interaction=0.037). For melanoma of the skin, males had increased cancer-specific mortality compared to females among heart and/or lung recipients (M:F HR: 5.44, 95% CI: 1.22, 24.19), but no differences in cancer-specific mortality between males and females were seen among kidney and liver recipients. For brain cancer, females had increased cancer-specific mortality compared to males among liver recipients (M:F HR: 0.03, 95% CI: 0.01, 0.25), but there was no association among kidney or heart and/or lung recipients.

For most non-sex-specific cancer sites, males have increased cancer-specific mortality compared to females in the general population [3]. One possible explanation for this difference has been stronger immunosurveillance or immunocontainment of cancer among females. Under this model, we hypothesized that females would lose this immunologic advantage in an immunosuppressed cohort of solid organ transplant recipients, manifesting as an attenuation in the M:F HR_Transplant compared to the M:F HR_GenPop. Indeed, transplant recipients with cancer have higher cancer-specific mortality than patients in the general population for a range of cancers, including colorectal, breast, kidney, and bladder cancers as well as melanoma [11], implying that immunity is important for control of cancer. Notably, however, we did not find a consistent attenuation of the M:F HR_Transplant across cancer types even when analyses were restricted to patients diagnosed with localized stage cancers receiving treatment or when stratified by transplanted organ (reflecting the degree of immunosuppression). These results suggest that immunity is not the general explanation for the female advantage in survival seen among cancer patients in the general population.

In healthy populations, females exhibit a stronger innate and adaptive immunity than males due to the make-up and function of immune cells, resulting in increased incidence of autoimmune disorders among females and higher cancer incidence among males [6]. Among transplant recipients, who receive the same amount of immunosuppression regardless of sex, females are at higher risk of graft failure [15]. Previous research by our group has shown that while the male excess in cancer incidence is attenuated among transplant recipients for many cancer sites, it is not completely null [16].

The only cancer for which we saw an attenuation in the female mortality advantage among transplant recipients was colorectal cancer. In light of the overall null pattern for other cancers, the reasons for the decreased advantage for females diagnosed with colorectal cancer in the transplant population are not clear. Colorectal cancer incidence is modestly increased in transplant recipients [12, 17], and regular screening is recommended [18, 19]. A systematic review of colorectal cancer screening trials found that males experience greater mortality benefits from screening, with a 33% reduction in cancer mortality as compared to a 15% reduction for females [20], perhaps because males have a higher baseline incidence of colorectal cancer and increased mortality after cancer diagnosis [21].

Contrary to our hypothesis, the M:F HR_Transplant for cancer of the kidney (including the renal pelvis) was higher than the M:F HR_GenPop. Here, underlying end-stage kidney disease may play a role. Males have a higher incidence of end-stage kidney disease than females, and this difference increases with age [22]. However, evidence suggests that females spend more time on dialysis before being placed on the transplant waitlist [22] and have higher estimated glomerular filtration rate when listed [23, 24]. Most kidney recipients retain their damaged native kidneys after transplantation of a donor kidney, and most kidney cancers in kidney recipients arise in the native kidneys [25]. Among transplant recipients in our study, 85% of kidney cancers occurred in kidney recipients, and 70% of kidney recipients with kidney cancer were male. In an analysis restricted to kidney recipients, the M:F_Transplant HR was 1.33 (95% CI: 1.09, 1.62), similar to our main result.

To examine whether an effect of estrogen on immunity may explain the female survival advantage among cancer patients in the general population, we stratified the analyses by the typical age of menopause [26, 27]. Broadly, estrogen and progesterone are immunoprotective, and androgens are immunosuppressive [28]. Stimulation of estrogen receptor alpha has been shown to reduce both pro-inflammatory and tumor-promoting properties of macrophages [28]. In the age-stratified analysis, the M:F HRs for cancer patients in the general population were often larger at younger ages, in line with the hormone hypothesis. However, we did not see a consistent attenuation across cancer types in the M:F HR_Transplant compared to the M:F HR_GenPop among patients <50. Anal cancer was the only cancer in this age group for which the M:F HR_Transplant was attenuated compared to the M:F HR_GenPop. These results may suggest that any benefits of sex hormones on survival are not mediated by their effects on immunity. Another possibility is that estrogen does not confer additional benefit in an immunosuppressed setting.

This study used linked data from population-based registries of solid organ transplant recipients and cancer registries in the US. The large sample size allowed us to examine mortality for 13 cancer sites, although the number of cancer patients was more limited for our analysis stratified by age at diagnosis for some of the cancers. The use of population-based cancer registries allowed us to obtain reliable information on cancer diagnoses including stage at diagnosis.

We note several study limitations. We did not restrict the TCM cohort to the same registries included in SEER to maximize the number of cancer patients with a previous transplant. Consequently, the TCM cancer patients in non-SEER states may be different from the general population included in SEER. We could not adjust for details of cancer treatment as registries only collect limited information on the first course of therapy. We also lacked data on menopausal status, including use of hormone replacement therapy, so we could not directly explore the effects of estrogen on cancer mortality. We note that transplant recipients differ from people in the general population in additional ways beyond immune status, notably including the presence of significant medical co-morbidity and complications of transplantation (e.g., severe infections, kidney dysfunction due to toxicity from calcineurin inhibitors). We did not have data on these factors but do not believe that they would have systematically led to sex differences in cancer-specific mortality that could have biased our results. Finally, we did not have data on lifestyle factors for which the prevalence may differ by sex and affect mortality. However, national estimates of current smoking and physical activity are comparable between male and female cancer survivors, with females slightly more likely to smoke and report lower physical activity after a cancer diagnosis than males [29].

Overall, we did not find consistent evidence that the female advantage in cancer mortality is weakened or lost among immunosuppressed transplant recipients. These results suggest that factors other than immunosurveillance or immunocontainment may contribute to the female survival advantage seen among cancer patients in the general population.