Baseline single institutional retrospective review of body mass index (BMI) as a prognostic indicator in patients with newly diagnosed glioblastoma (GBM)

Abstract

Purpose:

Glioblastoma (GBM) is the most common primary brain cancer in adults with a very poor prognosis. Metabolic drivers of tumorigenesis are highly relevant within the central nervous system, where glucose is the critical source of energy. The impact of obesity on survival outcomes in patients with GBM is not well established. This study investigates the prognostic value of body mass index (BMI) in patients diagnosed with GBM.

Methods:

Adult patients with newly diagnosed GBM treated at Thomas Jefferson University Hospital between January 1, 2008, and December 31, 2022, were included in the study. BMI was calculated using the formula BMI = kg/m². Patients BMI groups were underweight (BMI <19.00), normal weight (BMI 19.00–24.99), overweight (BMI 25–29.99), and obese (BMI >30.00). All patients received 60Gy of radiation therapy with concurrent and adjuvant temozolomide following maximal safe resection. A difference in clinical outcomes of overall survival (OS) and progression-free survival (PFS) were evaluated between the groups using Kaplan-Meier and log-rank tests.

Results:

A total of 392 patients met inclusion criteria. The median age was 60.3 (range 18.9–86.7), with 144 females and 248 males. Median BMI was 27.0 (Range; 17.7–52.9). Non-overweight GBM patients (BMI <25.00, OS 2.1 years, CI 1.7–2.4 years) had increased overall survival compared to overweight patients (BMI ≥ 25.00, OS 1.5 years, CI 1.4–1.6 years) (p<0.001). Patients with MGMT-methylated GBM also had significantly greater OS and PFS compared to MGMT-unmethylated patients (p<0.001). Non-overweight GBM patients (BMI <25.00, median PFS 1.5 years, CI 1.3–2.0 years) also had increased progression-free survival compared to overweight patients (BMI ≥ 25.00, median PFS 1.1 years, CI 0.9–1.2 years) (p<0.001).

Conclusions:

Our study indicates normal BMI (19.00–24.99) at the time of GBM diagnosis is a favorable prognostic indicator for overall and progression-free survival. Additional studies are warranted for further analysis of BMI and survival outcomes in GBM patients.

Introduction:

Glioblastoma (GBM) is the most common and aggressive primary brain cancer in adults, characterized by its rapid progression and poor prognosis. Despite advances in treatment, survival outcomes remain dismal, underscoring the urgent need for a deeper understanding of factors influencing patient prognosis. Among these factors, obesity has emerged as a significant area of interest due to its well-established role as a risk factor for developing various cancers, including GBM (1–2).

Obesity, characterized by excess adipose tissue, influences cancer development through mechanisms such as altered insulin resistance, estrogen production, inflammatory pathways, and metabolic dysregulation. In the context of GBM, these metabolic alterations are particularly relevant given the central nervous system's reliance on glucose as its primary energy source. However, the specific impact of obesity on survival outcomes in GBM patients remains a subject of considerable debate (1–7).

The literature presents a paradoxical picture: some studies suggest that a higher BMI may be associated with longer overall survival (OS) and progression-free survival (PFS) in GBM patients. For instance, papers by Potharaju and Cha et al studied OS in newly diagnosed GBM patients and found patients with overweight and obese BMIs survived significantly longer. This trend is echoed in studies by Schneider and Valente et al assessing PFS, where higher BMI appears to correlate with a survival benefit (4,5). This prolonged survival in obese GBM patients has been supported in other studies with broader scopes as well. Guven et al. pooled analysis of five studies including 2131 patients, revealed patients with obesity survived longer than those with lower BMIs. Huang et al. also found that with a BMI above 25, patients were more likely to develop various cancers but had increased overall survival following diagnosis (6,7). Collectively, this data suggests that a higher patient BMI may paradoxically prolong GBM survival after diagnosis.

Conversely, other studies report that a high BMI at diagnosis correlates with shorter survival times and poorer outcomes. Particularly, a study by Siegel et al. found that the OS of normal weight patients was significantly greater than that of both obese patients and underweight patients (8). Similarly, in another study by Chambless et al the PFS was negatively associated with a higher BMI (9). Poorer survival outcomes with a high BMI have been established in other GBM patient groups as well. In brain stem glioma patients, both PFS and OS were significantly reduced in high vs low BMI groups (10). High BMI patients with MGMT methylated GBM also had reduced OS compared to their low BMI counterparts, which was not seen in patients with MGMT unmethylated GBM (11).

Despite this, others have found that patient BMI does not positively or negatively associate with GBM or glioma patient survival (12). Therefore, the impact of obesity on GBM patient survival outcomes has not been well determined.

Given these conflicting findings, our single institutional retrospective review aims to provide clarity by conducting a comprehensive retrospective review of the impact of BMI on survival outcomes in patients with newly diagnosed GBM. This investigation seeks to bridge the gap in current literature, offering a more definitive understanding of how BMI influences the prognosis of GBM patients.

Methods:

IRB Approval Statement:

This study was approved by the Institutional Review Board (IRB) of Thomas Jefferson University.

Patient Population:

The study cohort comprised adult patients with newly diagnosed glioblastoma (GBM) and received treatment at Thomas Jefferson University Hospital over a 14-year period, from January 1, 2008, to December 31, 2022. Inclusion criteria were specifically designed to capture a comprehensive and representative sample of GBM patients undergoing standard treatment protocols.

Exclusion criteria were applied to ensure the integrity and comparability of the data. Patients were excluded if they received short course radiation therapy of less than 60 Gy, if radiation therapy was the sole definitive treatment modality, or if there was a lack of available data regarding height and weight at the time of diagnosis. These criteria were established to focus on patients who received more standardized treatment regimens and for whom complete anthropometric data were available, thereby enhancing the reliability of our findings.

Body Mass Index (BMI):

Body Mass Index (BMI) was calculated for each patient at the time of GBM diagnosis using the standard formula: BMI = weight in kilograms divided by height in meters squared (kg/m²). To categorize patients based on their BMI, we adhered to the guidelines established by the Centers for Disease Control and Prevention (CDC). Accordingly, patients were classified into four distinct groups: Underweight (BMI < 19.00), Normal Weight (BMI 19.00–24.99), Overweight (BMI 25.00–29.99), and Obese (BMI ≥ 30.00).

For the purposes of more focused analysis, we further consolidated these categories into two broader groups: Underweight/Normal Weight (BMI ≤ 24.99) and Overweight/Obese (BMI ≥ 25.00). This bifurcation allowed for a streamlined analysis of the impact of BMI on GBM outcomes, particularly in distinguishing between patients with a BMI indicative of underweight or normal weight and those with a BMI in the overweight or obese range.

Clinical Variables:

In this study, we meticulously analyzed a range of clinical variables to comprehensively assess their impact on patient outcomes. These variables included sex, age at diagnosis, molecular characteristics (including MGMT methylation status), radiological findings, dates of disease progression, and dates of death.

As part of the standard treatment protocol, all patients included underwent maximal safe resection followed by a uniform course of radiation therapy, receiving a total dose of 60 Gy. This was complemented with concurrent and adjuvant temozolomide therapy, aligning with current best practices for GBM treatment.

Statistical Analysis:

To evaluate the clinical outcomes, specifically overall survival (OS) and progression-free survival (PFS), we employed Kaplan-Meier survival analysis coupled with log-rank tests. These statistical methods were chosen for their efficacy in comparing survival distributions across the different BMI categories. Additionally, we conducted a focused analysis on the impact of patient age at diagnosis and the molecular feature of MGMT methylation status, recognizing their potential influence on the prognosis of GBM patients.

A multivariable model was used to further investigate the relationship between PFS and OS with age, sex, MGMT methylation status, BMI, KPS, and the extent of surgery. This relationship was analyzed on patient data that had each of the variables present. To conduct our analysis, we employed a multivariable Cox proportional hazards model. The Cox model was implemented using the ‘survival’ package in R, a software environment for statistical computing.

Given the multivariable nature of our model, it was necessary to select the most relevant variables from our initial set of predictors. For this, we used a backward stepwise variable selection process guided by the Akaike Information Criterion (AIC). The variable selection process was performed using the ‘MASS’ package in R.

Results:

Patient Characteristics

Our study included a total of 392 patients who met the inclusion criteria. Thirty-two patients in our database were excluded from this analysis for either not being diagnosed with a glioblastoma, having received short-course radiation therapy, a lack of available height and weight data, or a lack of follow-up to evaluate survival. The median age of the cohort was 60.3 years, ranging from 18.9 to 86.7 years, with a distribution of 144 female and 248 male patients. The median Karnofsky Performance Status (KPS) at diagnosis was 90. Among the 221 patients who underwent documented surgical resection, 47.5% (N=105) achieved gross-total resection (GTR), 14.0% (N=31) had near-total resection (NTR), 24.4% (N=54) underwent sub-total resection (STR), and 14.0% (N=31) received a biopsy only. MGMT methylation status was known for 149 patients, with 45.0% (N=67) being MGMT methylated. The study included only two underweight patients (BMI < 19.00), leading to their exclusion from further analysis. The remaining cohort comprised 122 normal weight (BMI 19.00–24.99), 150 overweight (BMI 25.00–29.99), and 120 obese patients (BMI ≥ 30.00). The mean BMI was 28.2, with a median of 27.0 (Range: 17.7–52.9) (Table 1).

Table 1:

Patient Demographics

	N=392	BMI <25 (n=122)	BMI ≥25-<30 (n=150)	BMI ≥30 (n=120)
Median Age (years)	60.3	59.0	61.6	59.5
Sex
Female	N=144	56 (45.9%)	43 (28.7%)	45 (37.5%)
Male	N=248	66 (54.1%)	107 (71.3%)	75 (62.5%)
Median KPS	N=221	90	90	90
MGMT methylation status known	N=149	N=38	N=68	N=43
Methylated	N=67 (45.0%)	N=20 (52.6%)	N=32 (47.1%)	N=15 (34.9%)
Unmethylated	N=81	N=18	N=36	N=28
Unknown	N=243	N=84	N=82	N=77
Multi-focal disease	N=226 63 (27.9%)	N=59 18 (30.5%)	N=97 24 (24.7%)	N=70 21 (30.0%)
Extent of resection	N=221	N=57	N=96	N=68
Biopsy only	N=31 (14.0%)	10 (17.5%)	11 (11.5%)	10 (14.7%)
GTR	N=105 (47.5%)	28 (49.1%)	50 (52.1%)	27 (39.7%)
NTR	N=31 (14.0%)	7 (12.3%)	13 (13.5%)	11 (16.2%)
STR	N=54 (24.4%)	12 (21.1%)	22 (22.9%)	20 (29.4%)

BMI and Overall Survival

A significant difference in overall survival (OS) was observed between non-overweight (BMI < 25.00, OS: 2.1 years, Confidence Interval [CI]: 1.7–2.4 years) and overweight patients (BMI ≥ 25.00, OS: 1.5 years, CI: 1.4–1.6 years), with non-overweight patients demonstrating a statistically significant survival benefit (p < 0.001) (See Figure 1). This trend persisted when comparing non-overweight patients to those in the overweight (BMI 25.00–29.99, OS: 1.6 years, CI: 1.4–1.8 years) and obese categories (BMI ≥ 30.00, OS: 1.5 years, CI: 1.3–1.6 years) separately (p = 0.001) (Figure 2). No significant difference in survival was noted between the overweight and obese groups.

Figure 1. Median Kaplan-Meier OS for non-overweight (BMI ≤ 24.99) vs overweight (BMI ≥ 25.00) patients.

Median OS for non-overweight patients (BMI ≤ 24.99) was 2.1 years (CI 1.7–2.4 years) and for overweight/obese patients (BMI ≥ 25.00) was 1.5 years (CI 1.4–1.6 years). Median OS was significantly greater in non-overweight GBM patients when compared to their overweight counterparts (p<0.001).

Figure 2. Median Kaplan-Meier OS for non-overweight (BMI ≤ 24.99), overweight (BMI 25–29.99), and obese (BMI >30.00) patients.

Median OS for non-overweight patients (BMI ≤ 24.99) was 2.1 years (CI 1.7–2.4 years), overweight patients (BMI 25–29.99) was 1.6 years (CI 1.4–1.8 years), and obese patients (BMI >30.00) was 1.5 years (CI 1.3–1.6 years). Median OS was significantly greater in the non-overweight patient group (p=0.001).

Patients with MGMT methylated GBM showed a significantly longer OS (2.1 years, CI: 1.5 – 2.8 years) compared to those with MGMT unmethylated GBM (OS: 1.3 years, CI: 1.1 – 1.5 years) (p < 0.001) (See Figure 5). When stratified by both MGMT methylation status and BMI classification, significant differences in OS were observed among MGMT unmethylated non-overweight (BMI < 25.00, OS: 1.4 years, CI: 1.0 – 2.1 years), MGMT unmethylated overweight (BMI ≥ 25.00, OS: 1.2 years, CI: 1.1 – 1.4 years), MGMT methylated non-overweight (BMI < 25.00, OS: 1.8 years, CI: 1.1 - NA years), and MGMT methylated overweight patients (BMI ≥ 25.00, OS: 2.1 years, CI: 1.5 – 2.8 years) (p < 0.001) (Figure 7).

Figure 5. Median Kaplan-Meier OS in MGMT methylated vs unmethylated patients.

MGMT methylated patients (median OS 2.1 years, CI 1.5 – 2.8 years) had significantly greater OS than their MGMT unmethylated counterparts (median OS 1.3 years, CI 1.1 – 1.5 years) (p<0.001).

Figure 7. Median Kaplan-Meier OS stratified by MGMT methylation status and BMI classification.

MGMT unmethylated non-overweight patients (BMI < 25.00, median OS 1.4 years, CI 1.0 – 2.1 years), MGMT unmethylated overweight patients (BMI ≥ 25.00, median OS 1.2 years, CI 1.1 – 1.4 years), MGMT methylated non-overweight patients (BMI < 25.00, median OS 1.8 years, CI 1.1 - NA years), and MGMT methylated overweight patients (BMI ≥ 25.00, median OS 2.1 years, CI 1.5 – 2.8 years) had significantly different OS (p<0.001).

BMI and Progression Free Survival

The progression-free survival (PFS) analysis mirrored the trends observed in OS. Non-overweight GBM patients (BMI < 25.00, median PFS: 1.5 years, CI: 1.3–2.0 years) had significantly better outcomes compared to overweight patients (BMI ≥ 25.00, median PFS: 1.1 years, CI: 0.9–1.2 years) (p < 0.001) (Figure 3). This pattern was consistent when comparing non-overweight patients (BMI <25.00, median PFS: 1.5 years, CI: 1.3–2.0 years) to both overweight (BMI 25.00–29.99, median PFS: 1.0 years, CI: 0.9–1.2 years) and obese groups (BMI ≥ 30.00, median PFS: 1.1 years, CI: 0.9–1.3 years) separately (p < 0.001) (Refer to Figure 4). Similar to OS, no significant difference in PFS was found between the overweight and obese groups. MGMT methylated patients exhibited a significantly longer median PFS (1.0 years, CI: 0.9 – 1.2 years) compared to MGMT unmethylated patients (median PFS: 0.6 years, CI: 0.6 – 0.8 years) (p < 0.001) (Figure 6).

Figure 3. Median Kaplan-Meier PFS for non-overweight (BMI ≤24.99) and overweight (BMI ≥ 25.00) patients.

Median PFS for non-overweight patients (BMI ≤ 24.99) was 1.5 years (CI 1.3–2.0 years) and for overweight patients (BMI ≥ 25.00) was 1.1 years (CI 0.9–1.2 years). Median PFS was significantly greater in non-overweight patients (p < 0.001).

Figure 4. Median Kaplan-Meier PFS for non-overweight (BMI 19.00–24.99), overweight (BMI 25–29.99), and obese (BMI >30.00) patients.

Median PFS for non-overweight patients (BMI 19.00–24.99) was 1.5 years (CI 1.3–2.0 years), overweight patients (BMI 25–29.99) were 1.0 years (CI 0.9 – 1.2 years), and obese patients (BMI >30.00) was 1.1 years (CI 0.9–1.2 years). Median PFS significantly differed between groups (p<0.001).

Figure 6. Median Kaplan-Meier PFS in MGMT methylated vs unmethylated patients.

MGMT methylated patients (median PFS 1.0 years, CI 0.9 – 1.2 years) had significantly greater PFS than MGMT unmethylated patients (median PFS 0.6 years, CI 0.6 – 0.8 years) (p<0.001).

Multivariable Analysis:

Results from a Cox proportional hazards model revealed statistically significant results for Age at diagnosis, MGMT methylation status, and Extent of surgery. For age at diagnosis, the hazard ratio is 1.017 for PFS and 1.016 for OS, indicating that for each additional year of age at diagnosis, the hazard (or risk) of progression or death increases by 1.7% and 1.6%, respectively. Both these results are statistically significant (p<0.05). Regarding MGMT methylation status, the hazard ratio is 0.460 for PFS and 0.407 for OS, suggesting that the risk of progression or death is about 54% and 59% lower, respectively, for patients with methylated MGMT compared to those without. These results are highly statistically significant (p<0.001). The hazard ratios for the different surgery types (Gross total resection [GTR] vs Biopsy only, Subtotal resection [STR] vs. biopsy only, STR vs. GTR) vary. Notably, patients who underwent GTR have a significantly lower risk of death (HR=0.377, p=0.001) compared to those who only had a biopsy. Similarly, patients who underwent STR have a significantly lower risk of death (HR=0.568, p=0.049) compared to those who only had a biopsy. However, patients who underwent STR have a significantly higher risk of both progression (HR=1.548, p=0.019) and death (HR=1.505, p=0.040) compared to those who underwent GTR. (Table 2)

Table 2:

Multivariable Analysis

	Progression Free Survival Hazards Ratio (95% CI) 148 subjects (148 events)	Overall Survival Hazards Ratio (95% CI) 148 subjects (127 events)
Age at Diagnosis	1.017 (1.004–1.030) p=0.008*	1.016 (1.001–1.031) p=0.031*
MGMT Methylation Status	0.460 (0.325–0.651) p<0.001*	0.407 (0.279–0.593) P<0.001*
Extent of Surgery: GTR vs BX Only	0.873 (0.512–1.491) p=0.620	0.377 (0.214–0.665) p=0.001*
Extent of Surgery: STR vs BX Only	1.352 (0.789–2.316) p=0.272	0.568 (0.323–0.997) p=0.049*
Extent of Surgery: STR vs GTR	1.548 (1.075–2.229) p=0.019*	1.505 (1.018–2.226) p=0.040*

Discussion:

Obesity describes excessive accumulation of fat or adipose tissue in the body, and BMI is the most used clinical measure of body fatness. A high BMI (≥25) is correlated with a myriad of health consequences, including cancer risk. BMI is calculated by taking a person’s weight in kilograms and dividing it by their height in meters squared (BMI = kg/m²). According to the CDC, an individual’s BMI calculation can fall into one of four categories: Underweight (BMI <18.5), healthy/normal weight (BMI 18.5 to <25), overweight (BMI 25.0 to <30), and obese (BMI >30.0) (13). BMI tends to correlate with body fatness, and a low or high BMI has been associated with numerous negative health consequences. However, as we study specific diseases in greater depth, there may be some exceptions to this rule.

It has been suggested that newly diagnosed GBM patients with a higher BMI may have better clinical outcomes than those with a lower BMI (2–7). This phenomenon has been previously described as the obesity paradox. This counterintuitive phenomenon has been observed in various cancers, but its underlying mechanisms remain elusive (7,14,15). While some have proposed mechanisms that include changes in systematic hormone levels affecting cancer metabolism, there have not yet been preliminary studies on animal models to support this. Others propose that the obesity paradox is likely explained by methodological limitations in studies, such as BMI being a poor measure of body fatness in cancer survivors (15).

Other studies, like those of Siegal and Chambless et all mentioned previously, have demonstrated findings opposite to the obesity paradox, indicating that high BMI leads to poor outcomes in newly diagnosed GBM patients (8–11,16). These findings have been less consistent in GBM patients than those supporting the obesity paradox but cannot be overlooked given that previous research suggesting that adipocytes and lipid metabolism may play a role in the proliferation of cancers (17). In contrast to the obesity paradox, the notion that high BMI leads to poorer outcomes in cancer patients is supported by multiple plausible mechanisms, including systemic and tumor microenvironmental inflammation and immune mediated responses, insulin resistance, insulin-like growth factors (IGFs) and sex hormones pathways (15).

In this study, we examined BMI as a prognostic indicator in patients with newly diagnosed glioblastoma. It was hypothesized that there would be a difference in survival outcomes based on BMI in patients newly diagnosed with GBM. When comparing non-overweight (BMI < 25.00) to overweight (BMI 25.00–29.99) and obese (BMI ≤ 30.00) patients, we found that patients in our non-overweight group had significantly prolonged OS (2.1 vs 1.5 years, p<0.001) and PFS (1.5 vs. 1.1 years, p<0.001) when compared to their overweight counterparts. Therefore, our data supports the previous research highlighting poorer outcomes for overweight and obese patients newly diagnosed with GBM and does not support the obesity paradox.

Moreover, the presented data also illustrated that MGMT methylation status significantly affected overall and progression-free survival outcomes, with MGMT methylation demonstrating a clear survival benefit. When assessing OS in patients stratified by both MGMT methylation status and BMI classification, it is apparent that MGMT methylation is a more profound prognostic indicator than BMI (Figure 7).

While our study provides valuable insights, it is important to acknowledge its limitations. The retrospective design and single-institutional setting may limit the generalizability of our findings. Furthermore, the use of BMI as the sole measure of obesity does not account for factors like muscle mass, bone density, and overall body composition, which can be crucial in understanding the complex relationship between body fat and cancer progression.

Conclusion:

In conclusion, our study suggests that a normal BMI at the time of GBM diagnosis is a favorable prognostic indicator for survival. These findings add to the growing body of evidence that challenges the obesity paradox in GBM, highlighting the need for a more nuanced understanding of how obesity impacts cancer outcomes. Future research should focus on more comprehensive measures of body composition and explore the biological mechanisms underlying these associations. Such studies could pave the way for more personalized treatment strategies and improve the clinical management of GBM patients.

Data Availability:

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.