Diabetes mellitus is independently associated with adverse clinical outcome in soft tissue sarcoma patients

Diabetes mellitus is independently associated with adverse clinical outcome in soft tissue sarcoma patients

In the present study, we demonstrated that DM represents an independent prognostic marker for shorter CSS and OS in a large cohort of STS patients after curative resection. However, the inclusion of mean pre-surgery serum glucose values to well-established prognostic factors in STS patients did not add prognostic information in people without established DM.The negative impact of DM on clinical outcome was reported in many tumor entities, including breast cancer, endometrial cancer and colon cancer15,16,17. Data regarding the influence of DM on clinical outcome in STS patients is sparse. Kang et al. evaluated the prognostic impact of comorbidities including DM on clinical outcome in 349 STS patients that had undergone surgery for high-grade localized disease and found that the presence of comorbidity was independently associated with poor local recurrence-free survival (LRFS) and disease-specific survival (DSS)23. However, as the effect of DM on survival was not separately analyzed in this study, the influence of DM on clinical outcome in STS patients remains unclear23. To the best of our knowledge, the impact of pre-existing DM or pre-surgery serum glucose levels on survival of patients with STS has not been studied yet. Due to the rarity and heterogeneity of STS, which account for approximate 1% of all adult malignancies, the treatment of these tumors presents a major challenge in the daily clinical routine24. Therefore, it is of utmost importance to find markers that will guide clinical decision making.The present study shows that in patients with localized STS, pre-existing DM is a prognostic marker for inferior outcome. As expected, the OS was significantly lower in STS patients with DM. However, we also found, that in patients with DM, the CSS was significantly reduced. This association could also be found in multivariable modeling, even after adjusting for various established risk factors, indicating that the presence of DM is an independent prognostic marker of worse CSS. The circumstance that the association between DM and adverse DFS was not statistically significant is likely due to the relatively small sample size of patients with DM and the low event rate. However, as the Kaplan–Meier curve (Fig. 2D) shows a trend towards decreased DFS, future studies with more participants might uncover a relation between DM and cancer recurrence. The finding that the administration of adjuvant chemotherapy was associated with an inferior CSS in non-DM patients in unadjusted analysis can be explained by the fact that all of these patients had tumors with poor prognosis as estimated by established markers like tumor size greater than 5 cm, deep tumor location and high tumor grade (G2/3). After adjusting for these approved prognostic factors, this association did not remain.Inferior OS in STS patients with DM can be explained merely by the coexistence of two diseases. However, decreased CSS suggests that the relation between DM and STS nullifies pure cumulative effects of comorbidities, thus suggesting that DM per se is a negative prognostic marker for tumor specific survival in patients with STS.It has long been recognized that tumor cells metabolize glucose in an extensive way, as postulated in 1924 by Warburg et al.25 Unlike most tissues, tumor cells seem to metabolize glucose via aerobic glycolysis instead of oxidative phosphorylation, even in an oxygen rich environment26. It seems paradoxical that tumor cells switch to a biochemically less energy efficient metabolic pathway. One possible explanation is that for rapid cell proliferation like in cancer cells, it is important to provide large amounts of lipids, nucleotides and amino acids which can be synthesized from glucose26. Aerobic glycolysis creates only 2 molecules of ATP per molecule of glucose whereas oxidative phosphorylation creates 36 molecules of ATP per molecule of glucose. The presumed lack of ATP could be compensated by the continual uptake of glucose from the bloodstream26. This massive cancer glucose consumption can be used for diagnostic purposes via FDG-PET scans. In STS patients receiving neoadjuvant chemotherapy FDG-PET allows an even more accurate prediction of the histopathologic response than change in tumor size does27.Despite this plausible biologic rationale, we could not find a significant association of elevated mean pre-surgery blood glucose levels with inferior OS, CSS, and DFS. This could likely be due to the noisiness of the glucose values. All of the patients have been in a pre-operative setting and the glucose levels were assessed at various times of the day. As we collected retrospective glucose values, there was no way to determine whether the measured glucose values were collected during fasting or non-fasting states. In order to account for these differences, we took the mean of all available pre-surgery values from a time frame of 2 weeks before surgery. Previous studies have already used this method to examine the impact of elevated glucose levels on cancer progression28,29,30,31. For example Derr et al. found a significant association between inferior outcome and elevated mean glucose values in patients with glioblastoma28. In their study, the glucose values not only included measures from fasting and non-fasting states, but also measurements from the time period between diagnosis and censor date28. This suggests major bias, as it is known that critically ill patients usually develop hyperglycemia at the end of life32.A potential explanation for the discrepancy of DM being associated with inferior outcome while we could not find this correlation with elevated mean glucose, might be due to the DM associated tissue- and organ damage in these patients. Additionally, patients usually are monitored more closely for hyperglycemia in the preoperative timeframe and hence might be under more intense glycemia control. Also, the stress induced by the cancer diagnosis as well as the hospital stay itself might have influenced glucose values in these patients33,34. Although this current study was not able to find an association between mean glucose levels and outcome, prospectively and standardized blood glucose assessments are needed in order to truly investigate the relevance and clinical impact of blood glucose levels as a prognostic marker in STS patients. Instead of a mean glucose value, future studies could investigate the impact of the hemoglobin A1C as a marker for elevated glucose values over a period of time. HbA1C has the advantage of being an indicator of elevated blood glucose values over a longer period of time35. We did not investigate the impact of HbA1C as a prognostic marker due to the lack of this value in most of our patients.Another limitation of the present study is the limited availability of information regarding other factors which might be responsible for worse prognosis and survival outcomes in the current sample. These factors include obesity, diet, physical activity, smoking status and medications with influence on serum glucose levels. Also, we were not able to support the diagnosis of preexisting DM by extracting data on the use of antidiabetic drugs in every single DM patient. While we adjusted our analysis for the significant baseline differences (age and adjuvant chemotherapy) in the DM and non-DM patients, there still might be unmeasured impact on outcome like undocumented organ damage or yet undiagnosed comorbidities. However, the strength of this study is the large sample size and the long follow-up period.In conclusion, this is the first study indicating that DM is an independent predictor of adverse clinical outcome in STS patients who underwent curative surgical resection. Large scale prospective studies are warranted to confirm our results.

Via Source link