Abstract
Purpose:
The role of growth factors and inflammation in the onset of glioma is poorly understood, and conflicting reports of associations of circulating IGF-1 and inflammatory biomarkers with glioma risk exist in the literature. We examined associations between C-reactive protein (CRP), white blood cell count (WBC), neutrophil-to-lymphocyte ratio (NLR), and insulin-like growth factor-1 (IGF-1) and glioma risk in the UK Biobank cohort.
Methods:
Hazard ratios (HR) and 95% confidence intervals (CI) for glioma according to circulating biomarkers concentrations were calculated using Cox proportional hazards regression, adjusted for age, sex, race, and education. Analyses were conducted separately for glioma overall and by glioma subtype.
Results:
We identified 417 incident glioma cases among 428,537 participants with 3,255,815 person-years of follow up. Weak, non-significant associations were observed with increasing levels of these biomarkers for risk of glioma overall or by glioma subtype. Among women only, IGF-1 in the highest quartile was positively associated with glioma risk compared to the lowest quartile (HR=1.64, 95%CI: 1.03–2.60, p-trend=0.08), as was NLR (HR=1.54, 95%CI: 1.00–2.39, p-trend=0.05).
Conclusion:
In this prospective cohort, we found no significant associations between the inflammatory biomarkers CRP and WBC and the development of glioma. NLR and IGF-1 were associated with risk in women, but not men. When considered with previous studies, further investigation of NLR and IGF-1 as markers of glioma risk appears warranted, particularly in women.
Keywords: glioma, glioblastoma, cohort, epidemiology, inflammation, UK Biobank
Introduction
Gliomas are tumors arising from the glial cells of the brain and comprise 80% of all primary adult brain tumors. The etiology of gliomas is not well understood, and ionizing radiation remains the only well-validated environmental risk factor for glioma [1, 2]. Men are at substantially higher risk of developing glioma than women, with a male:female ratio of approximately 1.58 for glioblastoma, the highest grade form [3].
Chronic inflammation has been linked to a higher risk of some cancer sites; however, the causal mechanisms underlying these relationships are not fully understood [4, 5]. The role of inflammation in the incidence of glioma remains poorly studied [1]. Blood levels of C-reactive protein (CRP) are a general marker of systemic inflammation, and CRP has been positively associated with increased risk of cancers of other sites, including prostate and colorectal, in some, though not all, studies [6, 7]. Most studies of CRP and glioma were retrospective in design and subject to bias, or examined associations with prognosis after diagnosis rather than associations with risk [8–10].
Increases in white blood cell count (WBC) due to injury or stress represent an additional marker of inflammation; the neutrophil-to-lymphocyte ratio (NLR) may better reflect the balance of innate and adaptive immunity and may be more reproducible than WBC [11]. NLR has been positively associated with risk of other cancers, including lung [12], breast [13], and overall cancer risk [14]. Elevated NLR has been associated with poor overall survival in glioma patients [15, 16], though no study has evaluated its relationship with glioma risk. Insulin-like growth factor-1 (IGF-1) is part of a family of hormones that regulate normal cell growth and development. Some studies have suggested that IGF-1 also plays a role in suppressing inflammation [17, 18]. The role of IGF-1 in glioma development is unclear, with mixed results in previous case-control studies [19, 20]. In a prospective, nested case-control study of 280 patients with glioma in the European Prospective Investigation of Cancer and Nutrition (EPIC) cohort, higher levels of IGF-1 were positively associated with low-grade glioma (LGG) risk, but not with glioma overall [19].
To shed further light on the role of IGF-1 and systemic inflammation in glioma risk, we examined the association of related circulating biomarkers, specifically CRP, WBC, NLR, and IGF-1, with glioma risk in the UK Biobank.
Methods
Study Design and Participants
The UK Biobank is a prospective, population-based cohort comprised of approximately 500,000 volunteers from the United Kingdom between the ages of 40 and 69 who were recruited from 2006 to 2010. Participants were sought using National Health Service patient registries. At recruitment, patients provided written informed consent, completed questionnaires covering demographics, lifestyle factors and health/medical conditions, and provided biological specimens including a blood sample.
Outcome Ascertainment
Participant’s records were linked with the National Health Service (NHS) Central Registers for information on cancer diagnoses, which were coded based on the WHO International Classification of Diseases (ICD). Incident primary glioma cases (ICD 9 – 191 or ICD 10 - C71) were diagnosed from 2007 to 2016 and consisted of both glioblastomas (GBM) (9440–9442) and LGG (9380–9383, 9400–9401, 9410–9411, 9450–9451).
Biomarker Assessment
Blood was collected from UK Biobank participants and samples were shipped frozen to a central processing facility within 24 hours of collection [21]. High sensitivity CRP levels were measured using an assay manufactured by Beckman Coulter (UK), Ltd. and conducted on the Beckman Coulter AU5800 analytical platform [21]. Beckman Coulter LH750 instruments using the LH750 Haematology Analyser directly measured WBC while neutrophil and lymphocyte counts (used to derive NLR) were calculated using the same system [22]. IGF-1 was measured from a Diasorin S.p.A assay using the DiaSorin Liaison XL analytic platform [21]. The analytic range for CRP is 0.8–80.00 mg/L, for WBC 0.00–400.00 × 109 cells/L, and for IGF-1 1.3–195 nmol/L [21, 22]. The coefficients of variation (CV) reported by the UK Biobank were 1.70% for CRP and 5.29% for IGF-1 at intermediate concentrations of the biomarker, and were not reported for WBC or NLR [23].
Statistical Analysis
From the 502,536 total UKB participants, analyses were restricted to participants not closely genetically related to any other participants (genetically related pairs were identified by the UKB and based on identity-by-state [IBS] 0 values calculated from genotypes), with no history of cancer at baseline, and with at least one of the studied biomarkers (CRP, WBC, NLR, IGF-1) available, resulting in 428,541 participants [24]. Measurements for CRP were missing for 18,870 (4.40%) participants, 10,451 (2.44%) were missing WBC, 11,211 (2.62%) were missing NLR, and 20,195 (4.71%) were missing IGF-1. Follow-up time was defined as the time from recruitment to cancer diagnosis (other than basal and squamous cell carcinomas), last linkage, or death. Analyses were performed overall and by glioma subtype (GBM and LGG), sex, and age at enrollment (<57, ≥57 years, the median age at baseline) given reduced immunity with age [25]. Hazard ratios (HR) and 95% confidence intervals (CI) for glioma according to CRP, WBC, NLR, and IGF-1 were estimated using Cox proportional hazards regression. CRP, WBC, NLR, and IGF-1 were categorized based on quartiles in non-cases. Tests of linear trend in glioma risk for higher levels of biomarkers were assessed by assigning the median biomarker value for each category to all participants in that category, and treating those as a single continuous variable in the regression models. Continuous analyses were computed with a single linear term, and are presented per one standard deviation change. Models included terms for age (continuous), gender (male vs. female), race (white vs. non-white), and education (secondary schooling, vocational training, some college, completed college or none of the above). Models examining associations with NLR were further adjusted for WBC (continuous). We also constructed a single multivariable model adjusting for all of the above factors, and mutually adjusting for each inflammatory biomarker. Lagged analyses at three years were also performed to reduce potential effects of protopathic bias on results.
A p-value less than 0.05 was considered statistically significant. Analyses were performed using R (version 4.0.0).
Results
The 428,537 UK Biobank participants were followed for a median of 8 years post-enrollment, with 3,255,815 total person-years recorded (Table 1). A total of 417 incident glioma cases (316 GBM and 101 LGG) were diagnosed a median of 3.8 years after enrollment.
Table 1.
Baseline characteristics of UK Biobank participants included in the analysis of inflammatory biomarkers and growth factors and glioma risk.
| UK Biobank |
||
|---|---|---|
| Characteristics | Full Cohort (n=428,537) | Glioma Cases (n=417) |
|
| ||
| Age (years, median, IQR) | 57 (38–73) | 61 (40–70) |
| Gender (n, %) | ||
| Female | 229,280 (53.5) | 158 (37.9) |
| Male | 199,257 (46.5) | 259 (62.1) |
| Race (n, %) | ||
| White | 402,522 (93.9) | 405 (97.1) |
| Non-White | 24,527 (5.7) | 8 (1.9) |
| Education (n, %) | ||
| Some college (A/AS levels or equivalent) | 76,273 (17.8) | 70 (16.6) |
| Vocational training (teaching, nursing, etc.) | 21,881 (5.1) | 19 (4.5) |
| Secondary schooling | 114,035 (26.6) | 96 (22.8) |
| Completed college | 141,341 (33.0) | 139 (33.0) |
| None of the above | 70,500 (16.5) | 89 (21.1) |
| C Reactive Protein (mg/L, median, IQR) | 1.31 (0.65–2.71) | 1.46 (0.69–2.80) |
| White blood cell count (109 cells/L, median, IQR) | 6.65 (5.63–7.85) | 6.70 (5.82–7.74) |
| NLR (median, IQR) | 2.13 (1.66–2.76) | 2.25 (1.69–2.97) |
| IGF-1 (nmol/L, median, IQR) | 21.32 (17.61–24.88) | 21.31 (17.59–24.82) |
Abbreviations: IGF-1, insulin-like growth factor 1; NLR, neutrophil to lymphocyte ratio; IQR, interquartile range. Continuous variables presented as median (IQR) and categorical variables as count (%).
No statistically significant associations were observed between levels of CRP, WBC, NLR, or IGF-1 and glioma risk overall (Table 2). Multivariable HRs ranged from 1.01–1.14 comparing the highest versus lowest quartile and 0.94–1.06 for a 1 standard deviation increment across the four biomarkers. When stratified by glioma subtype, no statistically significant associations were found.
Table 2.
Associations between biomarker levels and glioma risk in the UK Biobank.
| All Glioma | GBM | LGG | |||||||
|---|---|---|---|---|---|---|---|---|---|
|
| |||||||||
| n-cases | HR | (95% CI)1 | n-cases | HR | (95% CI)1 | n-cases | HR | (95% CI)1 | |
|
| |||||||||
| CRP (mg/L) | |||||||||
| <0.64 | 92 | Ref. | 71 | Ref. | 21 | Ref. | |||
| 0.64–<1.29 | 92 | 0.89 | (0.66–1.18) | 62 | 0.77 | (0.55–1.08) | 30 | 1.30 | (0.74–2.28) |
| 1.29–<2.68 | 106 | 0.96 | (0.72–1.28) | 83 | 0.94 | (0.68–1.30) | 23 | 1.01 | (0.56–1.84) |
| ≥2.68 | 112 | 1.03 | (0.78–1.37) | 87 | 1.01 | (0.73–1.40) | 25 | 1.10 | (0.61–1.99) |
| P-trend | 0.66 | 0.60 | 0.99 | ||||||
| Continuous2 | 402 | 0.94 | (0.84–1.06) | 303 | 0.96 | (0.84–1.09) | 99 | 0.88 | (0.66–1.17) |
|
| |||||||||
| WBC (109cells/L) | |||||||||
| <5.640 | 88 | Ref. | 66 | Ref. | 22 | Ref. | |||
| 5.640–<6.660 | 109 | 1.21 | (0.91–1.61) | 88 | 1.29 | (0.93–1.78) | 21 | 0.98 | (0.54–1.80) |
| 6.660–<7.860 | 116 | 1.24 | (0.94–1.65) | 89 | 1.23 | (0.89–1.71) | 27 | 1.27 | (0.72–2.25) |
| ≥7.860 | 92 | 1.01 | (0.75–1.36) | 63 | 0.90 | (0.63–1.28) | 29 | 1.38 | (0.78–2.43) |
| P-trend | 0.92 | 0.50 | 0.18 | ||||||
| Continuous2 | 405 | 1.00 | (0.91–1.11) | 306 | 0.95 | (0.84–1.07) | 99 | 1.11 | (0.99–1.24) |
|
| |||||||||
| NLR3 | |||||||||
| <1.66 | 91 | Ref. | 72 | Ref. | 19 | Ref. | |||
| 1.66–<2.13 | 92 | 0.90 | (0.67–1.21) | 73 | 0.89 | (0.64–1.24) | 19 | 0.94 | (0.50–1.78) |
| 2.1–<2.75 | 95 | 0.93 | (0.70–1.25) | 69 | 0.86 | (0.61–1.20) | 26 | 1.22 | (0.67–2.22) |
| ≥2.75 | 127 | 1.14 | (0.86–1.51) | 92 | 1.03 | (0.74–1.42) | 35 | 1.56 | (0.88–2.77) |
| P-trend | 0.32 | 0.88 | 0.07 | ||||||
| Continuous2 | 405 | 1.03 | (0.97–1.09) | 306 | 1.02 | (0.96–1.10) | 99 | 1.04 | (0.95–1.13) |
|
| |||||||||
| IGF-1 (nmol/L) | |||||||||
| <17.66 | 102 | Ref. | 74 | Ref. | 28 | Ref. | |||
| 17.66–<21.38 | 102 | 1.03 | (0.78–1.37) | 76 | 1.07 | (0.77–1.48) | 26 | 0.95 | (0.55–1.63) |
| 21.38–<24.93 | 102 | 1.09 | (0.82–1.44) | 79 | 1.15 | (0.83–1.60) | 23 | 0.91 | (0.52–1.60) |
| ≥24.93 | 93 | 1.13 | (0.84–1.51) | 71 | 1.20 | (0.86–1.68) | 22 | 0.94 | (0.52–1.67) |
| P-trend | 0.38 | 0.25 | 0.79 | ||||||
| Continuous2 | 399 | 1.06 | (0.96–1.18) | 300 | 1.08 | (0.96–1.21) | 99 | 1.02 | (0.83–1.26) |
Abbreviations: CRP, C reactive protein; IGF-1, insulin-like growth factor 1; NLR, neutrophil to lymphocyte ratio, WBC, white blood cell count; HR, hazard ratio; CI, confidence interval, GBM, glioblastoma; LGG, lower grade glioma; SD, standard deviation
Adjusted for age (continuous), gender (male, female), race (white, non-white) and education (some college, vocational training, completed college, secondary schooling, other).
Per SD change.
Additional adjustment for white blood cell count (continuous).
When examined according to gender (Table 3), in women, the highest quartile of IGF-1 was associated with an increased risk of glioma (HR=1.64, 95%CI: 1.03–2.60) compared to the lowest quartile, with a suggestive positive trend with increasing IGF-1 (p-trend in women=0.08, p-trend in men=0.82, p-interaction=0.15). Similarly, among women, the highest quartile of NLR was positively associated with glioma when compared to the lowest quartile (HR=1.54, 95%CI: 1.00–2.39), with a significant test for trend (p-trend in women=0.05, p-trend in men=0.84, p-interaction=0.19). Among men, non-significant associations of similar magnitude as the overall associations were observed.
Table 3.
Associations between inflammatory biomarker levels and glioma risk in the UK Biobank, stratified by gender.
| Males | Females | p-interaction4 | |||||
|---|---|---|---|---|---|---|---|
|
| |||||||
| n-cases | HR | (95% CI)1 | n-cases | HR | (95% CI)1 | ||
|
| |||||||
| CRP (mg/L) | |||||||
| <0.64 | 60 | Ref. | 32 | Ref. | |||
| 0.64–<1.29 | 52 | 0.72 | (0.50–1.05) | 40 | 1.21 | (0.76–1.94) | |
| 1.29–<2.68 | 70 | 0.92 | (0.65–1.31) | 36 | 1.02 | (0.63–1.66) | |
| ≥2.68 | 66 | 1.01 | (0.71–1.45) | 46 | 1.09 | (0.68–1.74) | |
| P-trend | 0.61 | 0.94 | 0.69 | ||||
| Continuous2 | 248 | 0.95 | (0.81–1.10) | 154 | 0.93 | (0.77–1.13) | 0.84 |
|
| |||||||
| WBC (109 cells/L) | |||||||
| <5.640 | 50 | Ref. | 38 | Ref. | |||
| 5.640–<6.660 | 70 | 1.39 | (0.96–2.01) | 39 | 0.99 | (0.63–1.55) | |
| 6.660–<7.860 | 68 | 1.27 | (0.87–1.85) | 48 | 1.23 | (0.80–1.89) | |
| ≥7.860 | 63 | 1.15 | (0.78–1.68) | 29 | 0.83 | (0.51–1.35) | |
| P-trend | 0.70 | 0.75 | 0.47 | ||||
| Continuous2 | 251 | 1.04 | (0.92–1.17) | 154 | 0.94 | (0.79–1.11) | 0.23 |
|
| |||||||
| NLR3 | |||||||
| <1.66 | 50 | Ref. | 41 | Ref. | |||
| 1.66–<2.13 | 58 | 0.95 | (0.65–1.40) | 34 | 0.83 | (0.52–1.31) | |
| 2.13–<,2.75 | 62 | 0.92 | (0.63–1.34) | 33 | 0.95 | (0.60–1.50) | |
| ≥2.75 | 81 | 0.96 | (0.67–1.39) | 46 | 1.54 | (1.00–2.39) | |
| P-trend | 0.84 | 0.05 | 0.20 | ||||
| Continuous2 | 251 | 0.98 | (0.86–1.12) | 154 | 1.07 | (1.01–1.14) | 0.34 |
|
| |||||||
| IGF-1 (nmol/L) | |||||||
| <17.66 | 62 | Ref. | 40 | Ref. | |||
| 17.66–<21.38 | 58 | 0.84 | (0.59–1.21) | 44 | 1.38 | (0.89–2.12) | |
| 21.38–<24.93 | 70 | 1.02 | (0.72–1.45) | 32 | 1.14 | (0.71–1.84) | |
| ≥24.93 | 56 | 0.89 | (0.62–1.30) | 37 | 1.64 | (1.03–2.60) | |
| P-trend | 0.82 | 0.08 | 0.17 | ||||
| Continuous2 | 246 | 1.02 | (0.89–1.16) | 153 | 1.14 | (0.97–1.34) | 0.37 |
Abbreviations: CRP, C reactive protein; IGF-1, insulin-like growth factor 1; NLR, neutrophil to lymphocyte ratio; WBC, white blood cell count; HR, hazard ratio; CI, confidence interval; SD, standard deviation
Adjusted for age (continuous), gender (male, female), race (white, non-white) and education (some college, vocational training, completed college, secondary schooling, other).
Per SD change.
Additional adjustment for white blood cell count (continuous).
P-value for interaction term between gender and each marker.
When considered by age (Table 4), no statistically significant trends were identified either among participants aged 57 or older or less than 57 at enrollment for any of the four biomarkers.
Table 4.
Associations between inflammatory biomarker levels and glioma risk in the UK Biobank, stratified by age at enrollment.
| Age < 57 | Age ≥ 57 | p-interaction4 | |||||
|---|---|---|---|---|---|---|---|
|
| |||||||
| n-cases | HR | (95% CI)1 | n-cases | HR | (95% CI)1 | ||
|
| |||||||
| CRP (mg/L) | |||||||
| <0.64 | 32 | Ref. | 60 | Ref. | |||
| 0.64–<1.29 | 25 | 0.85 | (0.50–1.44) | 67 | 0.91 | (0.64–1.29) | |
| 1.29–<2.68 | 37 | 1.35 | (0.83–2.18) | 69 | 0.85 | (0.60–1.20) | |
| ≥2.68 | 26 | 1.08 | (0.64–1.82) | 86 | 1.03 | (0.74–1.45) | |
| P-trend | 0.39 | 0.88 | 0.46 | ||||
| Continuous2 | 120 | 0.93 | (0.73–1.19) | 282 | 0.95 | (0.83–1.08) | 0.91 |
|
| |||||||
| WBC (109 cells/L) | |||||||
| <5.640 | 33 | Ref. | 55 | Ref. | |||
| 5.640–<6.660 | 34 | 1.20 | (0.74–1.96) | 75 | 1.25 | (0.88–1.77) | |
| 6.660–<7.860 | 29 | 1.04 | (0.62–1.73) | 87 | 1.38 | (0.98–1.94) | |
| ≥7.860 | 28 | 1.04 | (0.62–1.74) | 64 | 1.03 | (0.71–1.49) | |
| P-trend | 0.99 | 0.76 | 0.91 | ||||
| Continuous2 | 124 | 1.02 | (0.85–1.22) | 281 | 1.00 | (0.89–1.13) | 0.85 |
|
| |||||||
| NLR3 | |||||||
| <1.66 | 33 | Ref. | 58 | Ref. | |||
| 1.66–<2.13 | 25 | 0.71 | (0.42–1.20) | 67 | 1.02 | (0.71–1.46) | |
| 2.13–<2.75 | 31 | 0.85 | (0.51–1.42) | 64 | 0.99 | (0.69–1.42) | |
| ≥2.75 | 35 | 1.04 | (0.63–1.71) | 92 | 1.23 | (0.87–1.74) | |
| P-trend | 0.74 | 0.25 | 0.81 | ||||
| Continuous2 | 124 | 1.06 | (0.97–1.14) | 281 | 1.02 | (0.95–1.11) | 0.46 |
|
| |||||||
| IGF-1 (nmol/L) | |||||||
| <17.66 | 14 | Ref. | 88 | Ref. | |||
| 17.66–<21.38 | 29 | 1.44 | (0.76–2.76) | 73 | 0.94 | (0.69–1.29) | |
| 21.38–<24.93 | 34 | 1.48 | (0.79–2.77) | 68 | 0.97 | (0.70–1.34) | |
| ≥24.93 | 42 | 1.61 | (0.88–2.96) | 51 | 0.93 | (0.66–1.33) | |
| P-trend | 0.16 | 0.74 | 0.15 | ||||
| Continuous2 | 119 | 1.13 | (0.95–1.34) | 280 | 1.00 | (0.88–1.13) | 0.16 |
Abbreviations: CRP, C reactive protein; IGF-1, insulin-like growth factor 1; NLR, neutrophil to lymphocyte ratio; WBC, white blood cell count; HR, hazard ratio; CI, confidence interval; SD, standard deviation
Adjusted for age (continuous), gender (male, female), race (white, non-white) and education (some college, vocational training, completed college, secondary schooling, other).
Per SD change.
Additional adjustment for white blood cell count (continuous).
P-value for interaction term between gender and each marker.
Results for three-year lagged analyses were also generally null, and all tests for trend with increasing biomarkers concentrations were non-significant (Supplementary Table 1, two-year lagged data not shown). Among women, the previously identified positive associations between NLR and IGF-1 and glioma risk were attenuated and non-significant. Lastly, we constructed a mutually adjusted multivariable model including all of the examined inflammatory markers, which demonstrated no significant associations between any biomarker and glioma risk.
Discussion
In this prospective study of UK residents, we found weak, non-significant associations between CRP and WBC concentrations and glioma risk. Higher levels of IGF-1 and NLR were associated with higher risk of glioma overall in women only, although the p for interaction by gender was non-significant.
Previous glioma studies examining these biomarkers have had mixed results, and have been limited by retrospective design or small sample size. Importantly, much of the published literature on these biomarkers has examined associations between post-diagnosis concentrations and survival in individuals with glioma, rather than pre-diagnostic levels and incidence. In part, this may be due to the difficulty in accruing sufficient glioma cases in individual prospective cohort studies, compared with the ease of evaluating these markers after diagnosis.
The largest study to date to examine the associations of pre-diagnostic circulating IGF-1 and IGFBP-3 with glioma incidence was based on 280 incident gliomas in the EPIC cohort [19]. The authors reported a modest non-significant positive association with IGF-I concentrations (OR=1.75, 95%CI: 0.68–4.48; highest vs. lowest quartile) and inverse association with IGFBP-3 concentrations (OR=0.57, 95%CI: 0.22–1.52; highest vs. lowest quartile) and risk of LGG (n=74 cases). No statistically significant associations were observed for high-grade gliomas (which combined GBM and grade III anaplastic tumors). Of note, results for IGF-1 and IGFBP-3 for LGG became stronger and statistically significant with mutual adjustment (OR= 3.60, 95%CI: 1.11–11.74 IGF-I and 0.28 (0.09–0.84) for IGFBP-3, highest vs. lowest quartile), although results were attenuated after excluding the first year of follow up, suggesting the possibility of reverse causation.
A separate nested case-control study of 22 glioma cases and 400 controls in the Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study (ATBC), on the other hand, examined associations between IGF-1 and IGFBP-3 serum levels and glioma risk and identified a lower risk of glioma with higher serum concentrations of IGF-1. Compared to the lowest tertile, participants in the two highest tertiles of IGF-1 concentration had lower risk of glioma (OR=0.3, 95%CI: 0.1–0.7), but the study was limited by a low number of cases.
In the present study, we found no association between IGF-1 (IGFBP-3 was not available) with glioma overall or with GBM or LGG, separately. When we examined associations using the same definition of LGG as in the EPIC report (including ICD-O codes: M9382/3, M9383/1, M9384/1, M9390/0, M9394/1, M9400/3, M9411/3, M9420/3, M9421/3, M9450/3, M9505/1, M9391/3), results remained non-significant. However, we did find a significant positive association between higher levels of IGF-1 for total glioma limited to female participants, with similar though nonsignificant results also observed in 2- and 3-year lagged analyses. In contrast, the EPIC study reported no effect modification by gender [19]. Both studies were limited by small numbers, especially at extreme levels of IGF-1. Data were too sparse in the current study to examine associations simultaneously by gender and glioma subtype, but our observation of an association of IGF-1 with glioma in women is of interest, given the differences in glioma incidence between men and women.
Some evidence suggests that higher CRP levels are associated with poorer outcomes among glioma patients [9, 10]. One case-control study demonstrated significantly increased CRP in glioma cases relative to non-glioma controls (17.0 vs. 1.8, p<0.001), but used post-diagnosis CRP levels [26]. We are unaware of any prior studies using prospectively collected samples. The null findings for CRP reported in this prospective investigation, with blood draws occurring substantially prior to diagnosis on average, suggest no association between CRP and glioma risk.
Measures of WBC are used clinically for evaluation of acute inflammation, and NLR is a reliable predictor of solid tumor survival [11]. As with CRP, several studies have demonstrated associations between higher NLR measured pre-operatively and shorter survival after diagnosis of glioma [15, 27]. A meta-analysis of 16 studies evaluating this association yielded a summary hazard ratio for death of 1.43 (95%CI: 1.27–1.62) comparing NLR above and below different clinical cutoffs in each study [15]. In this study, whereas WBC was not associated with risk of glioma, higher NLR was associated with higher risk of incident glioma, among women only. To our knowledge, no study has examined associations between pre-diagnostic NLR and WBC and glioma incidence, although a recent comprehensive study of inflammatory biomarkers and cancer risk in the UKB reported a positive association between NLR and overall brain tumor risk (including glioma and other histologies, such as meningioma) [5]. In that study, the authors did not stratify findings for brain tumors by sex. Strengths of this study include the prospective study design and reasonably large number of incident gliomas available for analysis. However, several limitations are acknowledged. First, stratified analyses by glioma subtype resulted in relatively few glioma cases, particularly with LGG, resulting in limited power to detect associations. Furthermore, the cohort was predominantly Caucasian, limiting generalizability to other racial and ethnic groups. We also conducted multiple statistical tests, and some of the results, such as those stratified by gender and glioma subtype, were not based on prior hypotheses and may be due to chance. In addition, the study was based on a single measurement of each of the biomarkers, and it was therefore not possible to differentiate between those with acute and chronic long-term inflammation in our analyses. Lastly, data were available for only a limited set of biomarkers, and therefore our ability to replicate previously reported findings (i.e., those for IGFBP-3), or to use these additional markers for adjustment, was limited.
Conclusion
In this prospective cohort, we found weak, non-significant associations between CRP and WBC and risk of glioma. However, our suggestive finding of a higher risk of glioma in women with higher levels of IGF-1 and NLR warrants further study.
Supplementary Material
Highlights.
This was a large study of more than 400,000 participants in the UK Biobank
There were no significant associations between CRP and WBC and later risk of glioma
Higher NLR and IGF-1 were associated with higher risk in women, but not in men
Acknowledgements
The authors would like to thank the participants and study investigators and staff of the UK Biobank.
Funding:
National Institutes of Health (NIH) F30 CA235791 (DJC).
Footnotes
Conflict of Interest: The authors declare that they have no conflict of interest.
Availability of Data: The datasets analyzed during the current study are available from the UK Biobank with an approved protocol. This work is based on the UK Biobank Resource under application number 16944.
Code Availability: Code is available from the investigators upon reasonable request.
Ethics Approval: N/A
Consent to Participate: Informed consent was obtained from all individual participants included in the study.
Consent for Publication: N/A
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Ostrom QT, Bauchet L, Davis FG, Deltour I, Fisher JL, Langer CE, Pekmezci M, Schwartzbaum JA, Turner MC, Walsh KM, Wrensch MR, Barnholtz-Sloan JS (2014) The epidemiology of glioma in adults: a “state of the science” review. Neuro-oncology 16: 896–913 doi: 10.1093/neuonc/nou087 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Ostrom QT, Adel Fahmideh M, Cote DJ, Muskens IS, Schraw JM, Scheurer ME, Bondy ML (2019) Risk factors for childhood and adult primary brain tumors. Neuro-oncology 21: 1357–1375 doi: 10.1093/neuonc/noz123 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Ostrom QT, Gittleman H, Liao P, Vecchione-Koval T, Wolinsky Y, Kruchko C, Barnholtz-Sloan JS (2017) CBTRUS Statistical Report: Primary brain and other central nervous system tumors diagnosed in the United States in 2010–2014. Neuro-oncology 19: v1–v88 doi: 10.1093/neuonc/nox158 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Hunter P (2012) The inflammation theory of disease. The growing realization that chronic inflammation is crucial in many diseases opens new avenues for treatment. EMBO reports 13: 968–970 doi: 10.1038/embor.2012.142 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Nøst TH, Alcala K, Urbarova I, Byrne KS, Guida F, Sandanger TM, Johansson M (2021) Systemic inflammation markers and cancer incidence in the UK Biobank. European journal of epidemiology 36: 841–848 doi: 10.1007/s10654-021-00752-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Shiels MS, Katki HA, Hildesheim A, Pfeiffer RM, Engels EA, Williams M, Kemp TJ, Caporaso NE, Pinto LA, Chaturvedi AK (2015) Circulating Inflammation Markers, Risk of Lung Cancer, and Utility for Risk Stratification. Journal of the National Cancer Institute 107 doi: 10.1093/jnci/djv199 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Wang X, Dai JY, Albanes D, Arndt V, Berndt SI, Bézieau S, Brenner H, Buchanan DD, Butterbach K, Caan B, Casey G, Campbell PT, Chan AT, Chen Z, Chang-Claude J, Cotterchio M, Easton DF, Giles GG, Giovannucci E, Grady WM, Hoffmeister M, Hopper JL, Hsu L, Jenkins MA, Joshi AD, Lampe JW, Larsson SC, Lejbkowicz F, Li L, Lindblom A, Le Marchand L, Martin V, Milne RL, Moreno V, Newcomb PA, Offitt K, Ogino S, Pharoah PDP, Pinchev M, Potter JD, Rennert HS, Rennert G, Saliba W, Schafmayer C, Schoen RE, Schrotz-King P, Slattery ML, Song M, Stegmaier C, Weinstein SJ, Wolk A, Woods MO, Wu AH, Gruber SB, Peters U, White E (2019) Mendelian randomization analysis of C-reactive protein on colorectal cancer risk. International journal of epidemiology 48: 767–780 doi: 10.1093/ije/dyy244 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Feng Y, Wang J, Tan D, Cheng P, Wu A (2019) Relationship between circulating inflammatory factors and glioma risk and prognosis: A meta-analysis. Cancer medicine 8: 7454–7468 doi: 10.1002/cam4.2585 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Bunevicius A, Radziunas A, Tamasauskas S, Tamasauskas A, Laws ER, Iervasi G, Bunevicius R, Deltuva V (2018) Prognostic role of high sensitivity C-reactive protein and interleukin-6 in glioma and meningioma patients. Journal of neuro-oncology 138: 351–358 doi: 10.1007/s11060-018-2803-y [DOI] [PubMed] [Google Scholar]
- 10.Strojnik T, Smigoc T, Lah TT (2014) Prognostic value of erythrocyte sedimentation rate and C-reactive protein in the blood of patients with glioma. Anticancer research 34: 339–347 [PubMed] [Google Scholar]
- 11.Kim S, Eliot M, Koestler DC, Wu WC, Kelsey KT (2018) Association of Neutrophil-to-Lymphocyte Ratio With Mortality and Cardiovascular Disease in the Jackson Heart Study and Modification by the Duffy Antigen Variant. JAMA cardiology 3: 455–462 doi: 10.1001/jamacardio.2018.1042 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Grieshober L, Graw S, Barnett MJ, Thornquist MD, Goodman GE, Chen C, Koestler DC, Marsit CJ, Doherty JA (2018) Methylation-derived Neutrophil-to-Lymphocyte Ratio and Lung Cancer Risk in Heavy Smokers. Cancer prevention research (Philadelphia, Pa) 11: 727–734 doi: 10.1158/1940-6207.capr-18-0111 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Fang Q, Tong YW, Wang G, Zhang N, Chen WG, Li YF, Shen KW, Wu BW, Chen XS (2018) Neutrophil-to-lymphocyte ratio, obesity, and breast cancer risk in Chinese population. Medicine 97: e11692 doi: 10.1097/md.0000000000011692 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Rimini M, Casadei-Gardini A, Ravaioli A, Rovesti G, Conti F, Borghi A, Dall’Aglio AC, Bedogni G, Domenicali M, Giacomoni P, Tiribelli C, Bucchi L, Falcini F, Foschi FG, Bagnacavallo Study G (2020) Could Inflammatory Indices and Metabolic Syndrome Predict the Risk of Cancer Development? Analysis from the Bagnacavallo Population Study. Journal of clinical medicine 9 doi: 10.3390/jcm9041177 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Lei YY, Li YT, Hu QL, Wang J, Sui AX (2019) Prognostic impact of neutrophil-to-lymphocyte ratio in gliomas: a systematic review and meta-analysis. World journal of surgical oncology 17: 152 doi: 10.1186/s12957-019-1686-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Howard R, Scheiner A, Kanetsky PA, Egan KM (2019) Sociodemographic and lifestyle factors associated with the neutrophil-to-lymphocyte ratio. Annals of epidemiology 38: 11–21.e16 doi: 10.1016/j.annepidem.2019.07.015 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Sukhanov S, Higashi Y, Shai SY, Vaughn C, Mohler J, Li Y, Song YH, Titterington J, Delafontaine P (2007) IGF-1 reduces inflammatory responses, suppresses oxidative stress, and decreases atherosclerosis progression in ApoE-deficient mice. Arteriosclerosis, thrombosis, and vascular biology 27: 2684–2690 doi: 10.1161/atvbaha.107.156257 [DOI] [PubMed] [Google Scholar]
- 18.Tien LT, Lee YJ, Pang Y, Lu S, Lee JW, Tseng CH, Bhatt AJ, Savich RD, Fan LW (2017) Neuroprotective Effects of Intranasal IGF-1 against Neonatal Lipopolysaccharide-Induced Neurobehavioral Deficits and Neuronal Inflammation in the Substantia Nigra and Locus Coeruleus of Juvenile Rats. Developmental neuroscience 39: 443–459 doi: 10.1159/000477898 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Rohrmann S, Linseisen J, Becker S, Allen N, Schlehofer B, Overvad K, Olsen A, Tjønneland A, Melin BS, Lund E, Vineis P, Grioni S, Tumino R, Palli D, Mattiello A, Bonet C, Chirlaque MD, Sánchez MJ, Rodríguez L, Dorronsoro M, Ardanaz E, Lagiou P, Trichopoulou A, Trichopoulos D, Dossus L, Grote VA, Boeing H, Aleksandrova K, Bueno-de-Mesquita HB, van Duijnhoven FJ, Peeters PH, Khaw KT, Wareham NJ, Key TJ, Rinaldi S, Romieux I, Gallo V, Michaud DS, Riboli E, Kaaks R (2011) Concentrations of IGF-I and IGFBP-3 and brain tumor risk in the European Prospective Investigation into Cancer and Nutrition. Cancer epidemiology, biomarkers & prevention : a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology 20: 2174–2182 doi: 10.1158/1055-9965.epi-11-0179 [DOI] [PubMed] [Google Scholar]
- 20.Lönn S, Inskip PD, Pollak MN, Weinstein SJ, Virtamo J, Albanes D (2007) Glioma risk in relation to serum levels of insulin-like growth factors. Cancer epidemiology, biomarkers & prevention : a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology 16: 844–846 doi: 10.1158/1055-9965.epi-06-1010 [DOI] [PubMed] [Google Scholar]
- 21.(2019) Biomarker assay quality procedures: approaches used to minimise systematic and random errors (and the wider epidemiological implications). UK Biobank. https://biobank.ctsu.ox.ac.uk/crystal/crystal/docs/biomarker_issues.pdf.
- 22.(2017) UK Biobank Haematology Data Companion Document. UK Biobank. http://biobank.ctsu.ox.ac.uk/crystal/crystal/docs/haematology.pdf.
- 23.(2019) Companion document to accompany serum biomarker data. UK Biobank. https://biobank.ctsu.ox.ac.uk/crystal/crystal/docs/serum_biochemistry.pdf.
- 24.Nait Saada J, Kalantzis G, Shyr D, Cooper F, Robinson M, Gusev A, Palamara PF (2020) Identity-by-descent detection across 487,409 British samples reveals fine scale population structure and ultra-rare variant associations. Nature communications 11: 6130 doi: 10.1038/s41467-020-19588-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Castelo-Branco C, Soveral I (2014) The immune system and aging: a review. Gynecological endocrinology : the official journal of the International Society of Gynecological Endocrinology 30: 16–22 doi: 10.3109/09513590.2013.852531 [DOI] [PubMed] [Google Scholar]
- 26.Reynés G, Vila V, Martín M, Parada A, Fleitas T, Reganon E, Martínez-Sales V (2011) Circulating markers of angiogenesis, inflammation, and coagulation in patients with glioblastoma. Journal of neuro-oncology 102: 35–41 doi: 10.1007/s11060-010-0290-x [DOI] [PubMed] [Google Scholar]
- 27.Tan Z, Shen L, Wu H, Deng L, Li Z, Huang X (2019) Preoperative Neutrophil/Lymphocyte Ratio Is an Independent Prognostic Biomarker in Patients with Low-Grade Gliomas. World neurosurgery 132: e585–e590 doi: 10.1016/j.wneu.2019.08.068 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.