Evaluation of cardiovascular risk factors in long-term survivors of adult- and childhood-onset brain tumours: a pilot study

Introduction

Malignant brain tumours affect both children and adults, although have markedly different long-term outcomes. Five-year survival rates of 73% have been reported following multi-modality treatment in children diagnosed with malignant brain tumours before the age of 15 years (1). Not dissimilar outcomes are observed in teenage and young adults (15–40 years), in whom 5-year survival approaches 68%. In contrast, considerably lower 5-year survival rates of up to 20% are observed in adults diagnosed with malignant brain tumours after the age of 40 years (1). With advances in cancer therapies, however, survival rates in adult brain tumour patients have continuously improved over the last four decades (2).

Increasing survival from childhood cancer has led to the recognition that exposure to multimodal cancer therapy is associated with an increased risk of long-term morbidity and late mortality. Late-onset adverse sequelae in childhood brain tumour survivors (SCBT) included increased mortality, subsequent cancers, endocrinopathies, chronic medical conditions, neurocognitive impairment, and worse socioeconomic outcomes (3, 4). The excess mortality of SCBT predominantly relates to the recurrence or progression of the primary disease; however, an increase in cardiovascular mortality (3) and risk of early and late-occurring stroke (4, 5) are also observed. There are limited cross-sectional studies evaluating conventional vascular risk factors in SCBT (6, 7); however, the data reported in these studies support a mechanism for the increased late-occurring vascular mortality. Far fewer data exist for adult brain tumour survivors, likely reflecting the guarded prognosis of these individuals. Data in teenage and young adult (15–39 years) brain tumour survivors show increased vascular morbidity (8, 9), however, there are no data in adult survivors diagnosed with malignant brain tumours at an older age. Furthermore, there are no data in either teenage and young adults, or older adults, regarding vascular risk factors.

In the current pilot study, we characterized cardiovascular risk factors in long-term brain tumour survivors of both adult-onset (AO) and childhood-onset (CO), with a secondary aim of determining if a significant dichotomy exists in the prevalence of these risk factors between brain tumour survivors of AO and CO.

Methods

Results

A total of 36 patients (15 female, mean age 30.9 ± 13.9 years) and 36 controls (15 female, mean age 31.5 ± 13.4 years) were enrolled to the study. Of the patients recruited; 16 (44.4%) had gliomas (astrocytoma, oligodendroglioma, and ependymoma), 12 (33.3%) had primitive neuroendrocrine tumours/medulloblastomas, 5 (13.9%) had pineal tumours (pineal germinomas), 2 (5.6%) had chondrosarcomas, and 1 (2.8%) had a brain stem tumour with no formal histological diagnosis. Diagnosis and treatment of the tumour was during childhood in 16 and AO in 20. The mean radiotherapy dose to the tumour was 51.9 ± 10.1 Gy. Endocrine deficits were highly prevalent within the cohort (severe growth hormone deficiency (GHD) n = 22; partial GHD (n = 10); adrenocorticotrophic hormone deficiency (n = 1); hypogonadism (n = 7); and hypothyroidism (n = 7)). All patients with deficiencies were on appropriate replacement therapy, with the exception of GHD, where 10/22 were receiving GH replacement therapy.

Discussion

In this pilot study, we show survivors of primary malignant brain tumours appear to have an increased prevalence of adverse vascular risk factors. Specifically, these patients exhibited changes in body composition defined by a greater proportion of body fat which was both central and subcutaneous; increased total and LDL-cholesterol; and insulin resistance. No concomitant increase in BP was observed. The excess of adverse vascular risk factors was observed in the overall cohort, as well as in the sub-cohorts of survivors of malignant brain tumours of both AO and CO. The degree of abnormality in vascular risk factors was similar after the division of the cohort into those of CO and AO. To our knowledge, vascular risk factors have not previously been studied in survivors of adult malignant brain tumours.

Previous studies have been derived primarily from childhood cancer survivors (CCS) of mixed aetiology or those with acute lymphoblastic leukaemia (ALL) and have consistently shown increased prevalence of vascular risk factors. Abnormalities of body composition in CCS are well documented and characterized by an increased risk of obesity (11, 12, 13, 14). Talvensaari and colleagues compared 50 CCS with age- and sex-matched controls and showed an increased prevalence of obesity (16 vs 5%) and body FM (25% vs. 20%) (15). Similarly, assessment of FM in survivors of childhood ALL showed increased total body FM, with more than one in four patients having an FM greater than the 90th centile of the reference values (16). The prevalence of truncal adiposity, which has the strongest association with components of the metabolic syndrome, is estimated to be around two-fold greater in CCS than that of healthy controls (17). There are few data in CCS regarding ambient lipid profiles. Talvesaari and colleagues described reduced HDL-C and total cholesterol levels, with a resultant increase in the TC:HDL ratio in CCS. LDL-C and triglyceride levels of the CCS were however normal (15). In 23 recipients of childhood bone marrow transplant (BMT) for haematological malignancies (n = 17) or aplastic anaemia/myelodysplasia, 9 had hypertriglyceridaemia and 5 had low HDL values. Only mean triglyceride values were, however, significantly different from the control group (18).

In a mixed cohort of CCS where BPs were compared with standards for age, gender, and height, 28% of CCS were observed to have a systolic or diastolic BP above the 90th centile (14). The highest prevalence of hypertension, unsurprisingly, occurred in patients with a primary diagnosis of Wilm’s tumour (14). A further study reported a four-fold increase in the frequency of hypertension in a mixed cohort of CCS (17), however, a number of studies have failed to detect any abnormality of BP (15). Insulin resistance is fundamental to metabolic syndrome and conveys a significantly elevated risk of vascular disease. Fasting glucose and insulin are reported to be elevated in both long-term CCS (15) and recipients of BMT during childhood (18). In a prospective study of glucose tolerance in 248 CCS, the recorded frequencies of hyperinsulinaemia, impaired glucose tolerance, and diabetes were significantly elevated at 18% (17). Where insulin resistance has been described in CCS, an association with obesity, subnormal HDL levels, or hypertriglyceridaemia is frequent (17, 18), raising the question as to whether these individuals have metabolic syndrome.

These data may, however, not necessarily be transferrable to patients treated for childhood malignant brain tumours (SCBT) given the differences in treatment regimens. In comparison with other common childhood malignancies (leukaemia, lymphoma, Wilm’s tumour, neuroblastoma, soft tissue sarcoma), children with malignant brain tumours more frequently receive surgery and cranial radiation, and traditionally chemotherapy has been used sparingly. It is important, therefore, to establish the impact of the different disease processes and treatment regimens used for malignant brain tumours on the prevalence of vascular risk factors.

Data specific to long-term SCBT has variably shown increases in the prevalence of obesity (7) or no difference to reference populations (19). When direct measures of fat mass are used, CBTS have increased total body FM%, truncal FM, and appendicular FM (7), and reduced LBM (7). A recent systematic review and meta-analysis of 17 studies including a total of 2032 SCBT failed to show an increase in the prevalence of overweight or obesity compared with non-cancer controls (20). The authors, however, went on to analyse six studies with data on adiposity and showed SCBT to have increased WHR and per cent FM. Our data are in keeping with the data on SCBT showing increases in adiposity but extend this further to show increases in both truncal and subcutaneous FM. Truncal FM was increased by just over 80% in the childhood survivors and 40% in AO survivors, the difference likely relating to higher truncal fat mass in the adult controls.

In the few studies of SCBT that have examined serum lipids, an increase in LDL-C, apolipoprotein-B, total cholesterol to HDL ratio, triglycerides, and reduction in HDL-C has been reported (6, 7). No differences in lipid levels are observed between those with exposure of the hypothalamus to external beam radiation (XRT), or not (7). Insulin resistance estimated by the HOMA-IR has been found to be higher in SCBT who received hypothalamic XRT, compared with those who were not exposed (7). Our cohort showed increases in total and LDL-C but without the accompanying changes in HDL-C or triglycerides previously reported. The latter is difficult to explain given the significant increase in truncal fat mass and insulin resistance in our cohort. Notably, similar adverse changes in lipid profile and insulin sensitivity were observed in both our CO and AO survivors. Survivors of childhood brain cancer are reported to have isolated systolic hypertension in one study (6). No difference in BP is present in those exposed and not exposed to hypothalamic XRT (7). Using 24-h ambulatory BP monitoring, we could not demonstrate a difference between patients and control subjects for our overall cohort, or after stratification by the timing of onset of the brain tumour.

The observation of increased vascular risk factors in CBTS is given importance by data showing CBTS to have a standardized mortality rate (SMR) of 4.2 for cardiac deaths (3), and increased risk of late-occurring stroke (RR 29.0) (5). Similar findings have been reported in unselected cohorts of long-term CCS for both cardiac (SMR 5.8–8.2) (21, 22, 23) and cerebrovascular disease (SMR 4.6) (22). Data in teenage and young adult cancer survivors (age 15–39 years) have shown a 30 and 40% increased rate of admission to hospital with cardiovascular disease (9) and cerebrovascular events (8) respectively, confirming excess vascular disease continues to be a significant late sequelae in cancer survivors with increasing age at diagnosis. We are unaware of similar data in purely adult brain tumour survivors and suspect this relates to the relatively poor prognosis of these individuals with age at diagnosis. With further improvements in the duration of survival of adult-onset brain tumours, however, cardiovascular and cerebrovascular diseases are likely to take on increasing relevance.

Hypothalamic damage resulting from the tumour, surgery, or radiotherapy has been implicated as the primary risk factor in the development of obesity in CBTS (24). In keeping with this finding the authors showed hypothalamic tumour location, hypothalamic irradiation doses > 51 Gy and endocrinopathies (particularly GHD) to be associated with the development of obesity in CBTS (24). Younger age at diagnosis also increased the risk of obesity, suggesting the younger brain to be more susceptible to hypothalamic damage (19, 24). Comparison of CBTS exposed to hypothalamic XRT (>1 Gy) to those not exposed to XRT show similar BMI, total body FM, and %FM, however those with exposure of the hypothalamus to XRT have greater visceral, but not subcutaneous, fat, and metabolic sequelae (7). It has been postulated that the greater visceral fat and GHD may at least in part explain the increased metabolic sequelae (7).

Examining previous data concerning vascular risk factors in brain tumour survivors (6, 7, 19, 20) and accepting that all previous data have been in SCBT, our cohort appears to show a greater degree of difference compared with the reference population. Possible explanations for this observation are that our entire cohort received high dose cranial XRT and almost all patients showed hormone deficits, in particular GHD (25). GHD individuals show increases in truncal fat mass, total and LDL-C, and insulin resistance (26, 27). Whether GH replacement can reverse the observed metabolic abnormalities in survivors of brain tumours is unclear. Our previous work showed GH replacement in adult survivors of childhood cancer of mixed aetiology led to minor improvements in body composition and serum lipids (28), suggesting GHD likely contributes to the overall adverse cardiovascular risk profile but is unlikely to be the only determinant. The primary limitation of our study relates to the small numbers within our cohort, however, provides a platform to support further examination of cardiovascular risk factors in a larger cohort of brain tumour survivors. Additionally, our cohort is heterogeneous in terms of subtypes of brain tumours and treatment regimens, though given the low incidence of many childhood brain tumours and continued evolution of multi-modality treatment regimens for all brain tumours, it is unlikely a sufficiently large study of long-term survivors with a single tumour pathology and single treatment regimen will be possible. A further confounder is that our control group may not be fully representative of the background population. The mean BMI of our control group was 24.1 kg/m² which is lower than that reported in larger population studies in the UK (29) and therefore may have amplified differences between our patients and controls. Our data are cross-sectional and would be strengthened by longitudinal data to determine timelines for change in vascular risk factors in our patients. Such a study would, however, require recruitment of patients at the time of definitive treatment of their tumour and prolonged follow-up.

In summary, in this pilot study, we have provided data that suggest survivors of malignant brain tumours have an adverse cardiovascular risk profile characterized by a combination of excess total body and truncal fat mass, elevated total and LDL-cholesterol, and insulin resistance. We found these anomalies to be present in brain tumour survivors of both CO and AO. These observations are in keeping with studies showing increases in cardiovascular and cerebrovascular disease in cancer survivors diagnosed in childhood, adolescent, and young adult life. To our knowledge, these data are the first to examine the cardiovascular risk profile of AO brain tumour survivors. As survival rates improve in AO brain tumours our findings are likely to derive increasing importance and therefore warrant further study.

Declaration of interest

The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.

Funding

This work did not receive any specific grant from any funding agency in the public, commercial, or not-for-profit sector.

Author contribution statement

SSK, RA, JL, & RDM were involved with study design and conceptualization. NK, MG, KS, JL, and RDM were involved with data curation and formal analysis. NK, SSK, KS, GKD, HR, AG, MK-W, GG, CL, AA, JL, and RDM undertook investigation and methodology (including patient recruitment, screening, consent, laboratory analyses, patient investigations). SSK, NK, KS and JL undertook day to day project administration. RA and RDM undertook overall supervision of the laboratory analysis and research team. NK produced the - original draft, which was edited and reviewed by RA and RDM before being sent to the other investigators for review. All investigators were involved with review and editing of the draft to achieve the final manuscript.