Skip to main content
Download PDF

Epigenetic age and long-term cancer risk following a stroke

Genome Medicine volume 16, Article number: 135 (2024) Cite this article

Abstract

Background

The association between increased cancer risk following a cerebrovascular event (CVE) has been previously reported. We hypothesize that biological age (B-age) acceleration is involved in this association. Our study aims to examine B-age as a novel contributing factor to cancer development post-CVE.

Methods

From our prospective stroke registry (BasicMar), we selected 940 cases with epigenetic data. For this study, we specifically analyzed 648 of these patients who had available data, no prior history of cancer, and a minimum follow-up of 3 months. The primary outcome was cancer incidence. B-age was estimated using DNA methylation data derived from whole blood samples obtained within 24 h of stroke onset, employing various epigenetic clocks (including Hannum, Horvath, PhenoAge, ZhangBLUP, ZhangEN, and the mitotic epiTOC). Extrinsic epigenetic age acceleration (EEAA) was calculated as the residuals from the regression of B-age against chronological age (C-age). For epiTOC, the age-adjusted values were obtained by regressing out the effect of age from the raw epiTOC measurements. Estimated white cell counts were derived from DNA methylation data, and these cell fractions were used to compute the intrinsic epigenetic age acceleration (IEAA). Subsequently, we evaluated the independent association between EEAA, IEAA, and cancer incidence while controlling for potential confounding variables.

Results

Among 648 patients with a median follow-up of 8.15 years, 83 (12.8%) developed cancer. Cox multivariable analyses indicated significant associations between Hannum, Zhang, and epiTOC EEAA and the risk of cancer after CVE. After adjusting for multiple testing and competing risks, EEAA measured by Hannum clock maintained an independent association with cancer risk. Specifically, for each year increase in Hannum’s EEAA, we observed a 6.0% increased incidence of cancer (HR 1.06 [1.02–1.10], p value = 0.002).

Conclusions

Our findings suggest that epigenetic accelerated aging, as indicated by Hannum’s EEAA, may play a significant role in the increased cancer risk observed in CVE survivors.

Background

Cancer and stroke are leading causes of morbidity and mortality worldwide, constituting significant public health challenges [1]. These conditions are often closely related, with 6% of cancer patients developing a stroke during their lifetime, representing a risk two times higher than in the general population [2, 3]. Additionally, stroke—both ischemic and hemorrhagic—may precede the diagnosis of cancer, with a malignancy prevalence ranging between 6 and 12% in cross-sectional studies of stroke patients [4,5,6]. Moreover, patients with stroke show a 2.4-fold higher cancer incidence, as reported in previous studies that compare stroke patients with control groups [7, 8].

However, despite these entities having common risk factors, the mechanisms underlying these associations remain unknown. DNA methylation (DNAm) varies significantly throughout lifespan and represents the main source of information in the construction of epigenetic clocks, which estimate biological age (B-age) based on CpG (cytosine phosphate guanine) methylation patterns [9]. These clocks, by assessing the methylation status of specific CpG sites across the genome, can predict an organism’s B-age more accurately than chronological age (C-age), taking into consideration both genetic predispositions and environmental factors [10]. Individuals with an accelerated B-age have an increased risk of cancer and mortality, in a dose-responsive manner [11, 12], and they have a higher incidence of cerebrovascular diseases, which are associated with poorer outcomes and increased recurrence of these conditions [13,14,15,16]. Despite evidence linking accelerated B-age with both cancer and stroke, the specific association of B-age with cancer risk after stroke has not been analyzed. We hypothesize that the onset of cancer following an acute cerebrovascular event (CVE, ischemic or hemorrhagic stroke) is associated with an accelerated B-age. Our study aims to investigate the relationship between B-age at the time of CVE and the subsequent incidence of cancer over a long-term follow-up period.

Methods

Setting and participants

The BasicMar cohort is an ongoing, prospective, and observational registry of patients experiencing acute CVE who were attended at the tertiary stroke center Hospital del Mar in Barcelona, Spain. For the present study, we reanalyzed data from 940 cases recruited from the aforementioned cohort between April 2005 and April 2014. These cases were selected based on inclusion criteria detailed in previous epigenetic publications conducted by our group [13,14,15, 17]. Exclusion criteria for analysis were as follows: patients who died during hospitalization, those without a minimum 3-month follow-up, or those with a history of cancer or concurrent cancer diagnosis at the time of CVE onset (defined as any cancer diagnosis made within the first 3 months following CVE). We specifically analyzed 648 patients according to the aforementioned criteria, and a flowchart detailing the study is presented in Fig. 1. Briefly, a structured questionnaire was utilized to collect comprehensive data encompassing demographic characteristics, vascular risk factors, and clinical variables. These factors, defined in our previous studies [13,14,15, 17], include age, sex, hypertension, diabetes mellitus, dyslipidemia, ischemic heart disease, smoking habit, excessive alcohol consumption, peripheral arterial disease, atrial fibrillation, carotid atherosclerosis, previous stroke, antiplatelet treatment, body mass index, National Institute of Health Stroke Scale (NIHSS) scores at baseline and 24 h post-CVE, and poor outcome, which was defined as a Modified Rankin Scale score greater than 2 at 3 months, indicating moderate to severe disability [18]. CVE subtypes were categorized into hemorrhagic and ischemic events. The latter were divided into etiological subtypes according to the Trial of Org 10,172 in Acute Stroke Treatment (TOAST) criteria [19]. The data collection and recording process was carried out by the neurologists from our center who attended to each patient. Our study group then conducted a further review of the recorded data.

Fig. 1
figure 1

Study flowchart. A total of 648 patients were included in the study. Concomitant cancer diagnosis was considered when the diagnosis of cancer was made in the same moment of stroke diagnosis (i.e., during the hospital admission). FU, follow-up; DNAm, DNA methylation

Full size image

Study endpoint and follow-up

The primary endpoint of the study was any cancer diagnosis throughout the follow-up period, which concluded on January 1, 2023. Following-up data were extracted from patients’ hospital registries and outpatient medical records, all of which integrated within the Història Clínica Compartida de Catalunya known as HC3. This system aggregates data from primary care centers and hospital records throughout Catalonia, providing a comprehensive dataset for our analysis. Each cancer diagnosis was performed by an oncologist, and a member of our research team verified every event, utilizing specific ICD-10 codes: C00–C07 for initially identified cancers and D37–D48 for neoplasms of uncertain or unknown behavior. Cancer diagnoses were categorized based on their origin into groups: urinary tract, lung, head and neck, gastrointestinal, breast and gynecological, prostate, and other origin neoplasms. Follow-up procedures included a clinical visit with a neurologist 3 months post-CVE, followed by appointments and/or telephone contacts at the physician’s discretion, every 3 to 12 months. All patient events, death records, electronic medical records, and hospital admission records were reviewed, and the primary care physician was consulted to ascertain the status of patients before classifying them as lost to follow-up. Non-cancer-related mortality was documented as a competing event, defined by the death of patients not previously diagnosed with cancer.

Array-based DNA methylation analysis

DNA methylation quantification

DNA was extracted from whole peripheral blood collected in 10 mL EDTA tubes during the first 24 h after stroke onset. DNA methylation was obtained in two different batches. In one (n = 410), we measured DNA methylation via the Human Methylation 450 K Beadchip (Illumina, Netherlands, Eindhoven; 485,577 CpGs) in two different technical runs, while in the second batch (N = 238) we used the Infinium Methylation EPIC Beadchip (Illumina, Netherlands, Eindhoven; 865,918 CpGs). Idat files were subsequently loaded using the Minfi R library [20]. We conducted several quality controls (QCs) at the probe and sample level [21, 22]. Specifically, we removed CpGs with a detection rate lower than 99% (detection p value > 0.05) or having a beadcount lower than 1% in at least 5% of samples. Additionally, we excluded CpGs at multi-hit or single-nucleotide polymorphism locations (MethylToSNP) [23]. Finally, we excluded samples having a call-rate lower than 95% or sex mismatch. DNA methylation data was expressed as β values, which range from 0 (completely unmethylated) to 1 (completely methylated). We then applied two intra-array normalizations: preprocessnoob from Minfi to account for background signal and the BMIQ methods, as described elsewhere [24].

B-age acceleration

We calculated B-age estimation using four pre-existing epigenetic clocks: Hannum’s [25], Horvath’s [26], PhenoAge [27], Zhang’s [28] clocks—both versions based on an elastic net (EN) and Best Linear Unbiased Prediction (BLUP) models—and the epigenetic Timer Of Cancer (epiTOC) mitotic clock [29]. Additional file 1: Table S1 details the number of CpGs, tissue specificity, and correlation with C-age for each clock in both 450 K and EPIC batches. Extrinsic epigenetic age acceleration (EEAA) was determined as the residuals from the regression of B-age against C-age. Regarding epiTOC, we are presenting both the raw and C-age-adjusted values, where the age-adjusted values were also obtained by regressing out the effect of age from the raw epiTOC measurements. Additionally, we derived estimated white cell counts from DNA methylation using the EpiDISH library and the reference matrices described in Salas et al., which estimate up to 12 immune-cell (Fig. 2A) [30,31,32]. The effect of these estimated cell fractions was subsequently adjusted from EEAA calculations to compute the intrinsic epigenetic age acceleration (IEAA). This adjustment effectively eliminates the possible effects of white cell counts on our results. We examined the distribution of age acceleration and epiTOC scores, and cases falling beyond |3| standard deviations (SD) from the sample mean were trimmed to the corresponding values of 3 SD or − 3 SD, as appropriate. The distributions of EEAAs estimations can be visually explored in Fig. 2A, which shows that EEAA and IEAA estimations were approximately centered around 0. Additionally, since patients in the 450 K batch were older (75 years [65–80.8] vs. 70 years [56–78], p value < 0.05), we compared the distribution of EEAA and IEAA between the 450 K and EPIC arrays. We found only mild differences in Zhang clocks (Additional file 1: Figs. S1 and S2). For epiTOC, differences between batches were corrected when we adjusted by C-age raw estimations (Additional file 1: Fig. S3).

Fig. 2
figure 2

Age acceleration and presence of cancer. A depicts the distribution of EEAA/IEAA and epiTOC estimations. B compares these estimations between patients with and without incident cancer. Lighter colors and solid contours correspond to EEAA and raw epiTOC values, while darker colors and dashed contours correspond to IEAA and age-adjusted epiTOC. *: p value < 0.05, **: p value < 0.01. BLUP, Best Linear Unbiased Prediction; EN, elastic net; epiTOC, epigenetic Timer Of Cancer

Full size image

Statistical analysis

Data are reported as medians and interquartile ranges (IQR) and percentages according to the type of each variable. Baseline characteristics between groups were compared with t-, U Mann–Whitney, or χ2 tests, as appropriate. For the survival study, we report data at 15-year follow-up. To identify the clinical variables and demographic factors associated with cancer incidence at the univariate level, we modeled the cumulative incidence function while accounting for the competing event of non-cancer-related mortality. This approach is essential given the characteristics of our cohort of stroke survivors, who are expected to have a high mortality rate. Besides, we tertile-split EEAA and epiTOC variables and tested their effect in the cumulative incidence of cancer as well. Differences between groups were compared by doing Gray’s tests and adjusting p values by multiple testing (false discovery rate, Benjamini-Hochberg [FDRBH] procedure [33]). Multivariable Cox models were employed to assess the independent association between the risk of cancer and continuous EEAA/IEAA/epiTOC values, while accounting for potential confounders. Initially, a baseline model was built, incorporating variables known to elevate the risk of incident malignancies or those identified as significant in previous univariate analyses [34,35,36]. Therefore, we considered sex, hypertension, diabetes mellitus, dyslipidemia, atrial fibrillation, smoking, and alcoholism. Besides, considering the unique characteristics of our cohort composed of stroke patients, factors such as type of CVE (hemorrhagic or ischemic), history of previous strokes, and study batch (450 K or EPIC chips) were also incorporated. Variable selection was performed using a forward stepwise algorithm based on the Akaike Information Criterion (AIC). We then entered each continuous EEAA/IEAA and epiTOC estimation into the baseline model to derive hazard ratios (HR) and their corresponding p values. For EEAA/IEAA, HRs indicate the increased risk of cancer per year of EEAA/IEAA increase. Regarding epiTOC, we present the effects of both raw and C-age-adjusted values for a one-standard-deviation increase in these estimations. We also tested whether the effect EEAA/IEAA and epiTOC values on risk of cancer was moderated by other relevant factors such as sex at birth, type of stroke, and array type (450 K and EPIC) by additionally including the interaction term between each age acceleration variable and these moderators. These interactions examined whether our results were driven by specific strata in the cohort. To ensure the validity of the results, several statistical assumptions were thoroughly examined and satisfied, including the proportional hazard assumption, absence of concerning collinearity, and identification of influential cases. Finally, we also repeated the analyses building Fine-Gray subdistribution hazard models to account for the competing risk in the multivariable analyses as well. These models represent the instantaneous rate of occurrence of the event of interest in subjects who have not yet experienced that type of event, but accounting for all previously occurring competing events. These models were adjusted for the same set of covariables used in the Cox regression models. All multivariable results were also adjusted for multiple testing with the FDRBH method [33].

Results

Descriptive analyses all cohort

We included 648 patients in our study (see flowchart in Fig. 1) with a median age of 73 years (IQR: 62–80), of these 385 (59.4%) were male. Detailed characteristics of the cohort, including demographics, vascular risk factors, and clinical data, are presented in Table 1.

Table 1 Characteristics of the CVE cohort
Full size table

Over the follow-up period, which had a median of 8.15 years (IQR: 3.84–10.94; translating to 4,921.28 patient-years of observation), incidental cancer was identified in 83 patients, representing 12.8% of the cohort. Among these cases, gastrointestinal neoplasm was the most prevalent diagnosis, accounting for 34 (41.0%) cases. Additional file 1: Table S2 provides comprehensive data on the clinical diagnoses of all other incident events for the whole sample and split by sex. In the univariate analyses, we observed that patients with incident cancer were more often males (71.1 vs. 57.7%; p value = 0.028), had a higher alcohol consumption (41.0 vs. 28.5%; p value = 0.031), and fewer poor outcome (22.9 vs. 35.8%; p value = 0.029). No other statistically significant baseline differences were observed between groups, including C-age (Table 1). In Fig. 3A, we present the distribution of the estimated immune-cell fractions, revealing that neutrophils and natural killer cells positively correlate with C-age. Conversely, CD4 + (both memory and naïve subsets), naïve CD8 + , naïve B cells, and eosinophils negatively correlated with C-age (Fig. 3B). We analyzed whether the baseline estimated immune-cell fractions differed between individuals who later developed cancer and those who did not. Patients who later developed cancer presented higher cell counts for basophils and naïve B cell fractions (Fig. 3C, p value < 0.05). We then explored the association of incident malignancies on age acceleration and epiTOC at the univariate level (Fig. 2B). These comparisons resulted significant for both Hannum and ZhangEN EEAA/IEAA estimations (all p value < 0.05), such that patients with cancer had a higher age acceleration for these epigenetic clocks.

Fig. 3
figure 3

Estimated immune-cell counts. A shows the distribution (density function) of each cell subset. B represents the correlation between C-age and each cell type, where correlation coefficients have been obtained with Spearman rho. C compares cell proportions between patients with and without incident cancer. *: p value < 0.05. Baso, basophiles; Bmem, memory B cells; Bnv, naïve B cells; CD4Tmem, memory CD4 + T cells; CD4Tnv, naïve CD4 + T cells; CD8Tmem, memory CD8 + T cells; CD8Tnv, naïve CD8 + T cells; Eos, eosinophils; Mono, monocytes; Neu, neutrophils; NK, natural killer; Treg, regulatory T cells

Full size image

Factors associated to incidence of cancer within the follow-up

The cumulative incidence of cancer in the cohort was 17.5% (13.6 to 21.8) over the 15-year follow-up (Fig. 4A) and 223 (51.1%, 95% CI: 44.3–57.4) patients experienced mortality unrelated to cancer. We first explored which demographic and vascular risk factors were associated with incidence of cancer, finding that male sex, presence of carotid atherosclerosis, and excessive alcohol consumption were all linked to a higher risk of cancer (all p value < 0.05 detailed in Additional file 1: Table S3). However, upon adjusting for multiple testing, these associations did not remain statistically significant (Additional file 1: Table S3). We continued examining the effect of tertile-split age acceleration and epiTOC on the risk of cancer, finding that Hannum’s clock was positively associated with the incidence of cancer, with no differences between EEAA and IEAA (FDRBH < 0.05; Fig. 4B and Additional file 1: Tables S4 and S5). Results of the analysis regarding competitive event (non-cancer-related mortality) are detailed in Additional file 1: Tables S3, S4, and S5.

Fig. 4
figure 4

Cumulative incidence of cancer in the cohort. Values represent the number of patients at risk and the cumulative number of events. A shows the cumulative incidence function in the whole sample, while in B we stratified the incidence by tertile-split Hannum extrinsic epigenetic age acceleration. Time is represented in years. T1, first tertile; T2, second tertile; T3, third tertile

Full size image

Independent predictors of incident cancer

After analyzing which covariables were included in a clinical model using a forward stepwise selection algorithm based on AIC (see the “Methods” section), we found that only sex was selected. Therefore, we continued studying if continuous age acceleration and epiTOC were independently associated with cancer after adjusting for sex. After accounting for multiple comparisons, our analysis revealed that higher age acceleration was an independent predictor of cancer incidence for both the Hannum’s clock, considering EEAA and IEAA, and the ZhangEN EEAA clock (Fig. 5A). Similarly, both raw and age-adjusted epiTOC estimations were also associated with incident cancer (FDRBH < 0.05). Specifically, we found that each year increase in Hannum’s EEAA corresponded to a 6.6% heightened risk of cancer, with similar effects observed for Hannum’s IEAA and ZhangEN EEAA. Regarding epiTOC, for each 1 SD increase in raw and age-adjusted epiTOC, we observed a 27.6% and 28.8% increase in cancer incidence, respectively. Given that several cancer subtypes are sex-at-birth specific, we interrogated the interaction between sex and age acceleration and epiTOC, observing no significant moderation of sex at birth on the relationship between these estimations and incident cancer after accounting for multiple testing (Fig. 5B). However, we found a marginally significant interaction for ZhangBLUP EEAA (p value = 0.043) and ZhangBLUP IEAA (p value = 0.029). As we show in Additional file 1: Table S6, ZhangBLUP age acceleration was only significant in the subset of female individuals. On the other hand, we observed no significant interaction for stroke subtype, or batch (Fig. 5B), which indicates that the effect of age acceleration and epiTOC on cancer is not significantly different within these strata of the cohort. Finally, we conducted subdistribution hazard models in which we also considered the effect of the competing event. In these models, Hannum’s EEAA was still significantly associated with the risk of cancer after accounting for multiple testing, with similar effects as those reported above (HR = 1.06, 95% CI: 1.02–1.10, p value = 0.002, FDRBH = 0.028; Fig. 5A), while Hannum’s IEAA, Zhang’s EEAA, and raw epiTOC were only nominally significant (p value < 0.05).

Fig. 5
figure 5

Independent effect of age acceleration on incident cancer. A shows the hazard ratios and confidence interval for each epigenetic clock (adjusted for sex) and model type (Cox regression and Fine-Gray models). Lighter and darker colors correspond to EEAA and IEAA for each biological age estimations, while for epiTOC represent the raw and age-adjusted values. The red dashed line corresponds to the absence of effect. For EEAA/IEAA, hazard ratios indicate the increased risk of cancer per year of EEAA/IEAA increase. Regarding epiTOC, we present the effects of both raw and C-age-adjusted values for a one-standard-deviation increase in these estimations. B represents the interaction between these estimations and relevant moderators (sex, stroke type, and array type) which have been obtained via Cox regression models. For EEAA/IEAA, HRs indicate the increased risk of cancer per year of EEAA/IEAA increase. *: FDRBH < 0.05

Full size image

Discussion

In our comprehensive, long-term follow-up study of a cohort post-CVE, we observed a cumulative cancer incidence of 17.5%, consistent with findings from other stroke series [7, 8, 37]. Importantly, our study contributes to the understanding of this association by demonstrating for the first time that age acceleration plays a significant role in explaining the risk of subsequent cancer development following a stroke. This discovery underscores the potential of epigenetic markers in predicting health outcomes following CVE. The relationship between C-age and the incidence of cancer following a CVE remains controversial. In contrast to previous research that observed increased cancer rates in older [7, 38,39,40] and others with younger individuals [8, 41, 42], our results reveal no association between C-age and cancer incidence. On the other hand, we identified for the first time an independent association between Hannum’s age acceleration and an increased risk of cancer following a CVE. Specifically, for each year increase in EEAA, the likelihood of developing cancer increases by 6.6%. Therefore, while C-age might be confounded by potential life exposures linked to cancer, B-age is accounting for them, partially explaining the discrepancies described in previous literature along with factors such as survival bias, where individuals who live longer have more opportunities to develop cancer, and the specific type of cancer [43], which may have different associations with chronological and biological aging.

To provide a comprehensive overview of biological aging in the context of cancer risk after a CVE, our study employed multiple epigenetic clocks. This approach acknowledges the unique insights offered by each clock, enhancing the depth of our analysis. Our findings show an independent association with Hannum’s clock, indicating that older biological age correlates with increased cancer risk—a trend similarly observed with Horvath’s and PhenoAge clocks, though without reaching statistically significance. This discrepancy may stem from the specific nature of our data, which were derived from blood samples. Notably, the Hannum’s clock, designed and trained specifically on blood samples [25], stands in contrast to the multi-tissue approach of the Horvath clock in our case [26]. Similarly, the Hannum’s clock is the only one that has previously shown a relationship with outcomes after stroke and in blood samples [13,14,15]. Additionally, previous research has reported a stronger association with pan-cancer risk in the general population for Hannum’s EEAA, where a 1-year increase in EEAA was associated with a 6% increase in the risk of incident malignancies, which is a similar effect as that observed in our study [12]. In comparison, PhenoAge showed that each additional year of EEAA was associated with a 2% increase in cancer incidence, being a smaller effect [44]. This contrast may help to explain the more pronounced association observed with the Hannum’s clock in our findings. It is also possible that studies with different approaches and designs (based on different tissues, other clinical adjustments, or sample sizes) may be more suitable for detecting associations with the Horvath and PhenoAge clocks. In addition, our results were strengthened by performing an IEAA analysis, which considered the distribution of 12 immune-cell types in human blood. This comprehensive approach allowed us to account for the heterogeneity in immune cell populations and provided further evidence supporting the role of epigenetic mechanisms in the observed association between Hannum’s clock and cancer risk.

Regarding other clocks, we report for the first time a significant association between cancer and the ZhangEN clock after accounting for multiple comparisons, indicating a 7.3% increased risk of cancer per year of increase in EEAA. Additionally, we observed a 27.6% and 28.8% increase in the likelihood of incident cancer per standard deviation increase in raw and age-adjusted epiTOC values respectively, although these associations remained marginally significant after adjusting for competing risks (p value < 0.05). Notably, our study is the first to evaluate the long-term cancer risk using the epiTOC [29]. As a mitotic clock, epiTOC reflects DNAm changes associated with the cumulative number of stem cell divisions. The observed association between incident cancer after stroke and the amount of cell proliferation indicated by epiTOC suggests that epigenetic mechanisms, influenced by exposures that modulate the rate of stem-cell division, may play a significant role in the increased incidence of cancer following a CVE. Interestingly, we observed a nominally significant moderating effect of sex at birth on the relationship between incident cancer and the ZhangBLUP clock. Specifically, higher ZhangBLUP EEAA and IEAA were associated with incident cancer only in females. This finding is particularly interesting considering that some cancer subtypes are sex-specific and previous studies have shown that women with stroke are biologically younger compared to men [45]. However, this result should be interpreted with caution due to our limited sample size for studying specific cancer subtypes and the fact that this interaction was not significant after adjusting for multiple testing. Therefore, our results underscore the importance of studying sex-related differences in epigenetic aging and cancer risk in stroke survivors, but additional larger studies are needed to confirm these results. On the other hand, we did not observe significant interactions with the type of CVE (ischemic or hemorrhagic) or array type (450 K or EPIC). These findings support the idea that our results are not driven by specific strata within our cohort.

A validation of these associations is necessary to assess their reliability and generalizability and to consider the potential clinical applications of epigenetic clocks in the care of post-CVE patients. A recent study of clocks deconstruction [46] shows that although these clocks may reflect similar outcomes on B-age, they can be divided into different epigenetic modules with distinct methylation patterns. This variable module distribution among the clocks likely reflects different biological processes and may contribute to explaining the differences observed in our results. Specifically, authors reported that several modules exhibited significant acceleration in tumor tissues, with all of them being, to some extent, represented in Hannum’s clocks. Notably, Hannum’s clock showed a relatively equitable distribution between these modules and others associated with all-cause mortality. This observation further underscores Hannum’s propensity to correlate with outcomes in stroke, a complex disease influenced by numerous risk factors and life exposures. Additionally, the dynamic nature of EEAA and other epigenetic changes offers a significant opportunity for therapeutic interventions [47, 48]. Modulating B-age to modify the risk of subsequent cancer in patients who had a CVE could be a strategy to be explored in the future.

In our longitudinal analysis, factors such as male sex, atherosclerotic carotid disease, and excessive alcohol consumption were associated with an increased incidence of cancer following CVE, consistent with previous findings from the literature [4, 38, 49]. Conversely, no association was observed between cancer incidence and variables including smoking habit [7, 38, 39], diabetes mellitus [7], and ischemic heart disease [50], as previously reported. The disparity in results across studies may be largely attributed to variances in study designs, populations, and settings. Additionally, the lack of sufficient statistical power, often due to the absence of prior power calculations, can contribute to these differences. These factors underscore the importance of considering the specific characteristics and limitations of each study when interpreting findings. Notably, while most research has focused on ischemic strokes, our analysis extends to include TIAs and hemorrhagic strokes, like a small number of previous studies [41, 42, 51]. We believe that our findings, coupled with the documented older B-age in CVE cohorts compared to controls [13], help explain the previously reported increased risk of cancer following CVE. Currently, clinical practice lacks predictive models for identifying high-risk patients. In our opinion, our findings have the potential to contribute to the development of such predictive risk models by integrating epigenetic data and clinical features of patients with CVE.

One significant limitation is the sample size and the absence of a previous statistical power calculation. Moreover, due to financial constraints, we could not increase the number of cases. This limitation probably impeded finding additional potential associations between age acceleration estimations with the long-term risk of cancer. Furthermore, our sample size certainly limited our ability to conduct stratified analyses by cancer subtypes, resulting in a constrained capacity for detailed subgroup exploration. Another limitation of our study is that DNAm was exclusively measured in peripheral blood samples, which may not fully reflect tissue-specific methylation patterns relevant to certain cancers. Nonetheless, it is noteworthy that the majority of previous population-based studies have employed a similar methodology. Finally, blood samples taken during the acute phase might influence DNAm, potentially leading to reverse causality that impacts the estimations of B-age and immune cell counts. Future studies should collect data at multiple time points, not only during the acute phase but also throughout the chronic continuum.

As strengths, we conducted the unique study assessing the effect of B-age on cancer incidence in a specific CVE cohort, calculating the EEAA via different epigenetic clocks. Moreover, the study included a long follow-up of 15 years, in which data was censored to avoid higher number of drop-outs during the follow-up. Finally, at each step of the analysis, we accounted for the effect of competing events and multiple testing, bringing robustness to our conclusions.

Conclusions

Our study confirms our initial hypothesis by demonstrating that epigenetic age acceleration plays a relevant role in the increased long-term incidence of cancer among CVE patients. This highlights how epigenetic changes influence health outcomes after a cerebrovascular event. Through our analysis of multiple epigenetic clocks, we have uncovered considerable variability in the acceleration of biological aging and its correlation with cancer, underscoring the necessity for further research into specific epigenetic profiles and their potential for clinical application.

Data availability

All requests for resources and information should be directed to Joan Jiménez-Balado (jjimenez3@researchmar.net). The raw array data, including individual IDAT files and normalized beta values, are available in an open GEO repository [52, 53] (GSE280206). Phenotypic data, along with white cell counts and B-age estimations, are stored in a restricted Zenodo repository [54]. Access to these datasets will be granted for research consistent with the consent provided by participants, specifically studies related to health and disease that ensure participants remain unidentified. To request access to the phenotypic and B-age data, please use the Zenodo link [54] or directly contact the corresponding author, Joan Jiménez-Balado, at jjimenez3@researchmar.net. The code used for these analyses is available on GitHub [55].

Abbreviations

AIC:

Akaike Information Criterion

B-age:

Biological age

IQR:

Interquartile range

BLUP:

Best Linear Unbiased Prediction

BMIQ:

Beta MIxture Quantile

C-age:

Chronological age

CVE:

Cerebrovascular event

CpG:

Cytosine phosphate guanine

DNAm:

DNA methylation

EDTA:

Ethylenediaminetetraacetic acid

EEAA:

Extrinsic epigenetic age acceleration

EN:

Elastic net

EPIC:

Infinium Methylation EPIC Beadchip

epiTOC:

Epigenetic Timer Of Cancer

HC3:

Història Clínica Compartida de Catalunya

HR:

Hazard ratio

ICD-10:

International Classification of Diseases 10th Revision

IEAA:

Intrinsic epigenetic age acceleration

NIHSS:

National Institute of Health Stroke Scale

TIA:

Transient ischemic attack

TOAST:

Trial of Org 10172 in Acute Stroke Treatment

450K:

Human Methylation 450 K Beadchip

References

  1. Wang H, Naghavi M, Allen C, Barber RM, Carter A, Casey DC, et al. Global, regional, and national life expectancy, all-cause mortality, and cause-specific mortality for 249 causes of death, 1980–2015: a systematic analysis for the Global Burden of Disease Study 2015. Lancet. 2016;388:1459–544.

    Article  Google Scholar 

  2. Sun MY, Bhaskar SMM. When two maladies meet: disease burden and pathophysiology of stroke in cancer. Int J Mol Sci. 2022;23:15769.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  3. Zaorsky NG, Zhang Y, Tchelebi LT, Mackley HB, Chinchilli VM, Zacharia BE. Stroke among cancer patients Nat Commun. 2019;10:5172.

    PubMed  Google Scholar 

  4. Jiang J, Shang X, Zhao J, Cao M, Wang J, Li R, et al. Score for predicting active cancer in patients with ischemic stroke: a retrospective study. Biomed Res Int. 2021;2021:5585206.

    Article  PubMed  PubMed Central  Google Scholar 

  5. Tybjerg AJ, Skyhøj Olsen T, Andersen KK. Prevalence and risk of occult cancer in stroke. Acta Neurol Scand. 2020;141:204–11.

    Article  PubMed  Google Scholar 

  6. Wilbers J, Sondag L, Mulder S, Siegerink B, van Dijk EJ. Cancer prevalence higher in stroke patients than in the general population: the Dutch String-of-Pearls Institute (PSI) Stroke study. Eur J Neurol. 2020;27:85–91.

    Article  CAS  PubMed  Google Scholar 

  7. Rioux B, Gioia LC, Keezer MR. Risk of cancer following an ischemic stroke in the Canadian longitudinal study on aging. Can J Neurol Sci. 2022;49:225–30.

    Article  PubMed  Google Scholar 

  8. Jacob L, Kostev K. Cancer risk in stroke survivors followed for up to 10 years in general practices in Germany. J Cancer Res Clin Oncol. 2019;145:1013–20.

    Article  PubMed  Google Scholar 

  9. Jiang S, Guo Y. Epigenetic clock: DNA methylation in aging. Stem Cells Int. 2020;2020:1047896.

    Article  PubMed  PubMed Central  Google Scholar 

  10. Ryan CP. “Epigenetic clocks”: theory and applications in human biology. Am J Hum Biol. 2021;33:1–18.

    Article  Google Scholar 

  11. Fransquet PD, Wrigglesworth J, Woods RL, Ernst ME, Ryan J. The epigenetic clock as a predictor of disease and mortality risk: a systematic review and meta-analysis. Clin Epigenetics. 2019;11:62.

    Article  PubMed  PubMed Central  Google Scholar 

  12. Zheng Y, Joyce BT, Colicino E, Liu L, Zhang W, Dai Q, et al. Blood epigenetic age may predict cancer incidence and mortality. EBioMedicine. 2016;5:68–73.

    Article  PubMed  PubMed Central  Google Scholar 

  13. Soriano-Tárraga C, Giralt-Steinhauer E, Mola-Caminal M, Vivanco-Hidalgo RM, Ois A, Rodríguez-Campello A, et al. Ischemic stroke patients are biologically older than their chronological age. Aging (Albany NY). 2016;8:2655.

    Article  PubMed  Google Scholar 

  14. Soriano-Tárraga C, Mola-Caminal M, Giralt-Steinhauer E, Ois A, Rodríguez-Campello A, Cuadrado-Godia E, et al. Biological age is better than chronological as predictor of 3-month outcome in ischemic stroke. Neurology. 2017;89:830–6.

    Article  PubMed  Google Scholar 

  15. Jiménez-Balado J, Giralt-Steinhauer E, Fernández-Pérez I, Rey L, Cuadrado-Godia E, Ois Á, et al. Epigenetic clock explains white matter hyperintensity burden irrespective of chronological age. Biology (Basel). 2023;12:33.

    Google Scholar 

  16. Macias-Gómez A, Jiménez-Balado J, Fernández-Pérez I, Suárez-Pérez A, Vallverdú-Prats M, Guimaraens L, et al. The influence of epigenetic biological age on key complications and outcomes in aneurysmal subarachnoid haemorrhage. J Neurol Neurosurg Psychiatry. 2024;95:675–81.

    Article  PubMed  Google Scholar 

  17. Fernández-Pérez I, Jiménez-Balado J, Lazcano U, Giralt-Steinhauer E, Rey Álvarez L, Cuadrado-Godia E, et al. Machine learning approximations to predict epigenetic age acceleration in stroke patients. Int J Mol Sci. 2023;24:2759.

    Article  PubMed  PubMed Central  Google Scholar 

  18. Van Swieten JC, Koudstaal PJ, Visser MC, Schouten H, Van Gijn J. Interobserver agreement for the assessment of handicap in stroke patients. Stroke. 1988;19:604–7.

    Article  PubMed  Google Scholar 

  19. Adams HP, Bendixen BH, Kappelle LJ, Biller J, Love BB, David, et al. Classification of subtype of acute ischemic stroke definitions for use in a multicenter clinical trial. Stroke. 1993;24:35–41.

  20. Aryee MJ, Jaffe AE, Corrada-Bravo H, Ladd-Acosta C, Feinberg AP, Hansen KD, et al. Minfi: a flexible and comprehensive Bioconductor package for the analysis of Infinium DNA methylation microarrays. Bioinformatics. 2014;30:1363–9.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  21. Fernández-Pérez I, Jiménez-Balado J, Macias-Gómez A, Suárez Pérez A, Vallverdú-Prats M, Pérez-Giraldo A, et al. Blood DNA methylation analysis reveals a distinctive epigenetic signature of vasospasm in aneurysmal subarachnoid hemorrhage. Transl Stroke Res. 2024. https://doiorg.publicaciones.saludcastillayleon.es/10.1007/s12975-024-01252-x.

  22. Jiménez-Balado J, Fernández-Pérez I, Gallego-Fábrega C, Lazcano U, Soriano-Tárraga C, Vallverdú-Prats M, et al. DNA methylation and stroke prognosis: an epigenome-wide association study. Clin Epigenetics. 2024;16:75.

    Article  PubMed  PubMed Central  Google Scholar 

  23. Labarre BA, Goncearenco A, Petrykowska HM, Jaratlerdsiri W, Bornman MSR, Hayes VM, et al. MethylToSNP: identifying SNPs in Illumina DNA methylation array data. Epigenetics Chromatin. 2019;12:79.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  24. Teschendorff AE, Marabita F, Lechner M, Bartlett T, Tegner J, Gomez-Cabrero D, et al. A beta-mixture quantile normalization method for correcting probe design bias in Illumina Infinium 450 k DNA methylation data. Bioinformatics. 2013;29:189–96.

    Article  CAS  PubMed  Google Scholar 

  25. Hannum G, Guinney J, Zhao L, Zhang L, Hughes G, Sadda SV, et al. Genome-wide methylation profiles reveal quantitative views of human aging rates. Mol Cell. 2013;49:359–67.

    Article  CAS  PubMed  Google Scholar 

  26. Horvath S. DNA methylation age of human tissues and cell types. Genome Biol. 2013;14:R115.

    Article  PubMed  PubMed Central  Google Scholar 

  27. Levine ME, Lu AT, Quach A, Chen BH, Assimes TL, Bandinelli S, et al. An epigenetic biomarker of aging for lifespan and healthspan. Aging. 2018;10:573–91.

    Article  PubMed  PubMed Central  Google Scholar 

  28. Zhang Q, Vallerga CL, Walker RM, Lin T, Henders AK, Montgomery GW, et al. Improved precision of epigenetic clock estimates across tissues and its implication for biological ageing. Genome Med. 2019;11:54.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  29. Yang Z, Wong A, Kuh D, Paul DS, Rakyan VK, Leslie RD, et al. Correlation of an epigenetic mitotic clock with cancer risk. Genome Biol. 2016;17:1–8.

    Article  CAS  Google Scholar 

  30. Salas LA, Zhang Z, Koestler DC, Butler RA, Hansen HM, Molinaro AM, et al. Enhanced cell deconvolution of peripheral blood using DNA methylation for high-resolution immune profiling. Nat Commun. 2022;13:761.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  31. Luo Q, Dwaraka VB, Chen Q, Tong H, Zhu T, Seale K, et al. A meta-analysis of immune-cell fractions at high resolution reveals novel associations with common phenotypes and health outcomes. Genome Med. 2023;15:59.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  32. Zheng SC, Breeze CE, Beck S, Teschendorff AE. Identification of differentially methylated cell types in epigenome-wide association studies. Nat Methods. 2018;15:1059–66.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  33. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J Roy Stat Soc: Ser B (Methodol). 1995;57:289–300.

    Article  Google Scholar 

  34. Stocks T, Van Hemelrijck M, Manjer J, Bjørge T, Ulmer H, Hallmans G, et al. Blood pressure and risk of cancer incidence and mortality in the metabolic syndrome and cancer project. Hypertension. 2012;59:802–10.

    Article  CAS  PubMed  Google Scholar 

  35. Lau ES, Paniagua SM, Liu E, Jovani M, Li SX, Takvorian K, et al. Cardiovascular risk factors are associated with future cancer. JACC CardioOncol. 2021;3:48.

    Article  PubMed  PubMed Central  Google Scholar 

  36. Vinter N, Christesen AMS, Fenger-Grøn M, Tjønneland A, Frost L. Atrial fibrillation and risk of cancer: a Danish population-based cohort study. J Am Heart Assoc. 2018;7:e009543.

    Article  PubMed  PubMed Central  Google Scholar 

  37. Tanislav C, Adarkwah CC, Jakob L, Kostev K. Increased risk for cancer after stroke at a young age: etiological relevance or incidental finding? J Cancer Res Clin Oncol. 2019;145:3047–54.

    Article  PubMed  Google Scholar 

  38. Aarnio K, Joensuu H, Haapaniemi E, Melkas S, Kaste M, Tatlisumak T, et al. Cancer in young adults with ischemic stroke. Stroke. 2015;46:1601–6.

    Article  PubMed  Google Scholar 

  39. Selvik HA, Thomassen L, Bjerkreim AT, Næss H. Cancer-associated stroke: the Bergen NORSTROKE Study. Cerebrovasc Dis Extra. 2015;5:107.

    Article  PubMed  PubMed Central  Google Scholar 

  40. Quintas S, Rogado J, Gullón P, Pacheco-Barcia V, Dotor García-Soto J, Reig-Roselló G, et al. Predictors of unknown cancer in patients with ischemic stroke. J Neurooncol. 2018;137:551–7.

    Article  PubMed  Google Scholar 

  41. Vaz CG, Rodrigues J, Pereira D, Matos I, Oliveira C, Bento MJ, et al. The crosstalk between stroke and cancer: incidence of cancer after a first-ever cerebrovascular event in a population-based study. Eur Stroke J. 2023;8:792–801.

    Article  PubMed  PubMed Central  Google Scholar 

  42. Verhoeven JI, Fan B, Broeders MJM, Driessen CML, Vaartjes ICH, Klijn CJM, et al. Association of stroke at young age with new cancer in the years after stroke among patients in the Netherlands. JAMA Netw Open. 2023;6:e235002–e235002.

    Article  PubMed  PubMed Central  Google Scholar 

  43. Miranda-Filho A, Bray F, Charvat H, Rajaraman S, Soerjomataram I. The world cancer patient population (WCPP): an updated standard for international comparisons of population-based survival. Cancer Epidemiol. 2020;69: 101802.

    Article  PubMed  PubMed Central  Google Scholar 

  44. Lu AT, Quach A, Wilson JG, Reiner AP, Aviv A, Raj K, et al. DNA methylation GrimAge strongly predicts lifespan and healthspan. Aging. 2019;11:303–27.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  45. Gallego-Fabrega C, Muiño E, Cullell N, Cárcel-Márquez J, Lazcano U, Soriano-Tárraga C, et al. Biological age acceleration is lower in women with ischemic stroke compared to men. Stroke. 2022;53:2320–30.

    Article  CAS  PubMed  Google Scholar 

  46. Levine ME, Higgins-Chen A, Thrush K, Minteer C, Niimi P. Clock work: deconstructing the epigenetic clock signals in aging, disease, and reprogramming. bioRxiv. 2022;2022.02.13.480245.

  47. Noroozi R, Rudnicka J, Pisarek A, Wysocka B, Masny A, Boroń M, et al. Analysis of epigenetic clocks links yoga, sleep, education, reduced meat intake, coffee, and a SOCS2 gene variant to slower epigenetic aging. Geroscience. 2024;46:2583.

    Article  CAS  PubMed  Google Scholar 

  48. Noroozi R, Ghafouri-Fard S, Pisarek A, Rudnicka J, Spólnicka M, Branicki W, et al. DNA methylation-based age clocks: from age prediction to age reversion. Ageing Res Rev. 2021;68: 101314.

    Article  CAS  PubMed  Google Scholar 

  49. Cook MB, Dawsey SM, Freedman ND, Inskip PD, Wichner SM, Quraishi SM, et al. Sex disparities in cancer incidence by period and age. Cancer Epidemiol Biomark Prev. 2009;18:1174–82.

    Article  Google Scholar 

  50. Qureshi AI, Malik AA, Saeed O, Adil MM, Rodriguez GJ, Suri MFK. Incident cancer in a cohort of 3,247 cancer diagnosis free ischemic stroke patients. Cerebrovasc Dis. 2015;39:262–8.

    Article  PubMed  Google Scholar 

  51. Erichsen R, Sværke C, Sørensen HT, Sandler RS, Baron JA. Risk of colorectal cancer in patients with acute myocardial infarction and stroke: a nationwide cohort study. Cancer Epidemiol Biomark Prev. 2013;22:1994–9.

    Article  Google Scholar 

  52. Barrett T, Wilhite SE, Ledoux P, Evangelista C, Kim IF, Tomashevsky M, et al. NCBI GEO: archive for functional genomics data sets—update. Nucleic Acids Res. 2013;41:D991–5.

    Article  CAS  PubMed  Google Scholar 

  53. Jiménez-Balado J, Suárez-Pérez A, Jiménez-Conde J, Ois Á. Raw data for: Epigenetic Age and Long-Term Cancer Risk Following a Stroke. NCBI GEO; 2024. Available from: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE280206.

  54. Jiménez-Balado J, Suárez-Pérez A, Jiménez-Conde J, Ois Á. Data for: Epigenetic Age and Long-Term Cancer Risk Following a Stroke. Zenodo; 2024. Available from: https://zenodo.org/records/13981377.

  55. Jiménez-Balado J. BageCancerStroke. GitHub; 2024. Available from: https://github.com/JoanBalado/BageCancerStroke.

Download references

Acknowledgements

The authors would like to thank Carolina Soriano for initiating this project, Jaume Roquer for his invaluable teachings and stimulating interest in clinical research, and all the attending physicians of Hospital del Mar Neurology Department who were involved on BasicMar data recording.

Funding

This work was supported by grants from the Spanish Ministry of Science and Innovation, Instituto de Salud Carlos III with the grants (PI15/00451), (PI18/00022), (PI21/00593), (P19/00011); and Fondos FEDER/EDRF Spanish stroke research network INVICTUS+ (RD16/0019/0002) and Grant “RICORS-ICTUS” (RD21/0006/0021) funded by Instituto de Salud Carlos III (ISCIII), and by the European Union NextGenerationEU, Mecanismo para la Recuperación y la Resiliencia (MRR). Rio Hortega (CM22/00009, A.S.P) and Sara Borrell program, funded by Instituto de Salud Carlos III (CD22/00001, J.J.B).

Author information

Author notes

  1. Jordi Jiménez-Conde and Angel Ois are last authors with equal contributions.

Authors and Affiliations

  1. Neurovascular Research Group, Hospital del Mar Medical Research Institute, Barcelona, Spain

    Antoni Suárez-Pérez, Adrià Macias-Gómez, Isabel Fernández-Pérez, Marta Vallverdú-Prats, Elisa Cuadrado-Godia, Eva Giralt-Steinhauer, Maia Campanale, Ana Rodríguez-Campello, Joan Jiménez-Balado & Jordi Jiménez-Conde

  2. Department of Neurology, Hospital Del Mar and Universitat Pompeu Fabra, Barcelona, Spain

    Antoni Suárez-Pérez, Adrià Macias-Gómez, Isabel Fernández-Pérez, Elisa Cuadrado-Godia, Eva Giralt-Steinhauer, Daniel Guisado-Alonso, Ana Rodríguez-Campello, Jordi Jiménez-Conde & Angel Ois

  3. Medicine Department, Universitat Autònoma de Barcelona, Barcelona, Spain

    Antoni Suárez-Pérez, Adrià Macias-Gómez & Isabel Fernández-Pérez

  4. Universitat Pompeu Fabra, Barcelona, Spain

    Elisa Cuadrado-Godia, Eva Giralt-Steinhauer, Ana Rodríguez-Campello, Jordi Jiménez-Conde & Angel Ois

Authors
  1. Antoni Suárez-Pérez
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Adrià Macias-Gómez
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Isabel Fernández-Pérez
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Marta Vallverdú-Prats
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Elisa Cuadrado-Godia
    View author publications

    You can also search for this author inPubMed Google Scholar

  6. Eva Giralt-Steinhauer
    View author publications

    You can also search for this author inPubMed Google Scholar

  7. Maia Campanale
    View author publications

    You can also search for this author inPubMed Google Scholar

  8. Daniel Guisado-Alonso
    View author publications

    You can also search for this author inPubMed Google Scholar

  9. Ana Rodríguez-Campello
    View author publications

    You can also search for this author inPubMed Google Scholar

  10. Joan Jiménez-Balado
    View author publications

    You can also search for this author inPubMed Google Scholar

  11. Jordi Jiménez-Conde
    View author publications

    You can also search for this author inPubMed Google Scholar

  12. Angel Ois
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

ASP, JJB, and AO contributed to the conception and design of the study. ASP, JJB, and AO contributed to the acquisition and analysis of data. ASP, JJB, and AO drafted a significant portion of the manuscript and figures. ASP and JJB performed the statistical analysis. All authors interpreted the data, reviewed the manuscript, and approved the final version.

Corresponding author

Correspondence to Joan Jiménez-Balado.

Ethics declarations

Ethics approval and consent to participate

All the cohorts and samples involved in the study followed the national and international guidelines (Deontological Code, Declaration of Helsinki) and complied with the current personal data protection regulations, The Regulation (EU) 2016/679 of the European Parliament, and Ley Orgánica 3/2018 on protection of digital rights (LOPDPGDD). Local Institutional Review Boards approved all study aspects (CEIm-PSMAR, 2008/3083/l). Informed written consent was obtained from all patients or their relatives to be included in the study.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

13073_2024_1408_MOESM1_ESM.docx

Additional file 1: Table S1. B-age estimations in the sample. Fig. S1. EEAA distribution by batch. Fig. S2 IEAA distribution by batch. Fig. S3. epiTOC distribution by batch. Table S2. Distribution of cancer diagnoses. Table S3. Associations between clinical variables and incidence of cancer and non-cancer-related mortality over 15-year follow-up. Table S4. Associations between EEAA and incidence of cancer and non-cancer-related mortality over 15-year follow-up. Table S5. Associations between IEAA and incidence of cancer and non-cancer-related mortality over 15-year follow-up. Table S6. Moderation effect of sex at birth on the relationship between Zhang-BLUP and incident cancer.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Suárez-Pérez, A., Macias-Gómez, A., Fernández-Pérez, I. et al. Epigenetic age and long-term cancer risk following a stroke. Genome Med 16, 135 (2024). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13073-024-01408-2

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13073-024-01408-2

Keywords

Genome Medicine

ISSN: 1756-994X

Contact us