Effect of Type 2 Diabetes Mellitus on Prognosis of Nonmetastatic Renal Cell Cancer

Renal cell carcinoma (RCC) forms 2% to 3% of all cancers, with the highest incidence in Western Europe and the United States [1, 2]. The incidence of RCC has steadily and gradually increased throughout the world, with some exceptions such as Scandinavian countries [3, 4]. The established risk factors for RCC are smoking, obesity, and hypertension [5]. Hypertension and obesity in particular are increasing in the United States, and this increase may help to explain the rise in incidence of RCC [6, 7]. Hypertension is an independent risk factor for the development of RCC [8]. The association of RCC with the duration of hypertension and the blood pressure level was also shown in previous studies [9, 10]. Obesity is another important risk factor that may increase the relative risk for development of RCC to 2.76 in patients with a body mass index >40 kg/m² [11]. The main etiologic mechanism in these patients is hyperinsulinemia and increased insulin growth factor (IGF) secretion. The high prevalence of diabetes mellitus (DM) in patients with obesity and epidemiologic evidence for an association of DM with multiple cancers suggests a possible role of DM in the pathogenesis of RCC [12, 13]. Additionally, hypertension and obesity are highly linked with type 2 DM and metabolic syndrome. Although DM is not accepted as a risk factor for RCC development, studies have shown a slightly or significantly increased risk of RCC in subjects with DM [14-16]. However, data concerning the prognostic effect of DM on RCC are absent. In the present study, therefore, we aimed to investigate the effect of DM on the prognosis of localized RCC that was treated surgically.

We retrospectively analyzed 586 consecutive patients who had been treated surgically (by either partial or radical nephrectomy) at our department for renal tumors between 1995 and 2011. Patients with nonmetastatic RCC were eligible for inclusion in the study. After excluding patients with histopathologically diagnosed benign tumors, those who had inadequate follow-up, and those who had lymph node involvement or distant-site metastasis, 492 of the 586 patients were enrolled in the study. All patients had preoperative laboratory tests. Preoperative characteristics including age, gender, history of type 2 DM, and medication for DM were recorded. The 2009 TNM classification was used for pathologic tumor staging. The Fuhrman grading system and Heidelberg histologic classification were used to define the tumor grade and histologic subtype, respectively.

Patients were divided according to their diabetic status. Group I consisted of patients with type 2 DM and group II consisted of nondiabetic patients. Diabetic patients were defined as having a diagnosis of type 2 DM at the time of surgery. The DM diagnosis was based on preoperative fasting glucose levels >126 mg/dL [17] and receiving current medical therapy for DM, such as oral antidiabetics or insulin. In group I, 41 and 21 patients were using oral antidiabetics and insulin, respectively. Owing to inadequate data for glycated hemoglobin (HbA_1c) levels, we did not include this parameter in the statistical analysis. The clinicopathologic parameters evaluated and compared were as follows: age, gender, tumor size, pathologic stage, tumor grade, histological subtype, and multifocality. Disease-specific survival (DSS) and overall survival (OS) were assessed to determine the status of the disease in these patients.

Patient follow-up was individualized for every patient. In general, the patients were assessed four times in the first year by physical examination and abdominal ultrasonography; abdominal computed tomography (CT) and chest X-ray or thoracic CT was performed 6 months postoperatively. Depending on the patient's risk status, follow-up in the second and third years was either twice or once per year. Patients were followed up yearly thereafter if there was no sign of recurrence.

The chi-square test was applied to evaluate the association between categorical variables. The Mann-Whitney test was used to compare all mean values of continuous variables. The Kaplan-Meier estimates for DSS and OS were obtained according to DM status, and the differences were examined by the log-rank test. To determine independent prognostic factors, multivariate survival analysis was performed by a Cox regression model with respect to potential influencing factors including DM, age, grade, stage, and histological subtype. Statistical significance in this study was defined as p<0.05. All statistical analyses were performed by using the SPSS ver. 15.0 (SPSS Inc., Chicago, IL, USA).

The mean age of the patients was 56.7±12 years (range, 15 to 84 years; median, 58 years) and the mean length of follow-up was 35.9±28 months (range, 1 to 145 months; median, 34.3 months). One hundred thirty-six patients had undergone partial nephrectomy and 356 had undergone radical nephrectomy. All patients who underwent nephron-sparing surgery had negative margins. Of the 492 patients, 62 (12.6%) had a diagnosis of DM at the time of surgery (group I) and 430 did not have DM (group II). The baseline characteristics of the patients are compared in Table 1. The mean age and the incidence of clear cell RCC histological subtype were significantly higher in group I than in group II (p<0.001 and p=0.036, respectively). Other clinicopathological parameters including gender, tumor size, tumor grade, pathologic stage, and multifocality were not significantly different between the two groups.

TABLE 1
Clinicopathological parameters according to diabetic status of the patients

• Click for larger image

Kaplan-Meier analysis using a univariate log-rank test showed estimated 5-year OS as 62.9% and 77.7% for group I and group II, respectively (p=0.1). The 5-year DSS rate was 70% for group I and 82.2% for group II (p=0.3). The comparison of DSS and OS between these groups did not reveal any significant difference (Figs. 1, 2).

FIG. 1
Kaplan-Meier curve of the two groups for disease-specific survival. DM, diabetes mellitus.

• Click for larger image

FIG. 2
Kaplan-Meier curve of the two groups for overall survival. DM, diabetes mellitus.

• Click for larger image

Cox proportional models were formed according to DSS and OS. Age, tumor size, pathologic stage, tumor grade, histological subtype, and multifocality were included in the multivariate analysis. Pathologic stage, tumor grade, tumor size, and multifocality were independent prognostic parameters for DSS (Table 2). The independent prognostic factors for OS were tumor grade, pathologic stage, and multifocality.

TABLE 2
Multivariate analysis of predictors for disease-specific survival in patients with localized renal cell carcinoma

• Click for larger image

Similar to RCC, DM is also increasing worldwide. In the United States, the incidence of DM tripled from 1980 through 2007 [18]. The Western diet type and sedentary lifestyle are accepted as the major causes for the development of metabolic syndrome and type 2 DM. Meta-analyses performed for endometrial cancer, pancreatic cancer, and colorectal cancer have shown an increased incidence for these cancers in type 2 DM patients [19-21]. As mentioned before, established risk factors for RCC such as hypertension and obesity are highly linked with type 2 DM and metabolic syndrome. However, data concerning an association of RCC with DM are inadequate and conflicting [14, 15]. In a large retrospective, population-based study, patients with DM were found to have an increased risk of RCC compared with the general population [22]. Distinctly, Joh et al. [14] demonstrated an independent association between type 2 DM and RCC in women. On the other hand, Zucchetto et al. [15] reported a nonsignificantly increased risk of RCC in subjects with DM. Although making a conclusive comment on the relationship between DM and RCC is not possible at the moment, this subject deserves further study.

Several mechanisms are proposed for the association between cancers and type 2 DM. One of these mechanisms is the stimulation of cell proliferation by hyperinsulinemia and the secretion of IGF-1. Insulin and IGF-1 generate their effects through insulin receptors and IGF-1 receptors, respectively, to promote cellular proliferation and inhibit apoptosis in many tissue types. These effects might contribute to the formation of cancer. Additionally, the cancer cells may demonstrate an overexpression of insulin receptors, and activation of these receptors may favor cancer progression and facilitate the growth of tumors [23]. The other possible mechanisms are hyperglycemia, inflammatory cytokines, and reactive oxygen products [24]. These causes may also affect the clinical course of RCC and enhance the aggressiveness of the cancer. Thus, it is reasonable to question whether the prognosis of RCC is different in patients with type 2 DM than in nondiabetic patients. Regarding this issue, Washio et al. [25] using the database of the Japan Collaborative Cohort Study for Evaluation of Cancer Risk showed that DM increased the risk of kidney cancer death in a Japanese population. However, DM failed to remain as a significant risk factor for death in RCC patients after other factors were controlled for.

The present study focused on the prognostic effects of type 2 DM and its impact on OS and DSS in patients with clinically localized, surgically treated RCC. The comparison of DSS and OS between patients with and without type 2 DM did not reveal any significant difference, and type 2 DM was not detected as an independent prognostic factor in the multivariate analysis. Antonelli et al. [26] evaluated the role of preexisting type 2 DM in the prognosis of RCC. Similar to our study, they did not find DM as an adverse factor for RCC and non-RCC-related mortality. Habib et al. [27] analyzed a total of 473 cases of RCC and reported that 25.4% of RCC cases were associated with DM. The prevalence of DM in our study group was 12.6%, which agrees with the reported prevalence of type 2 DM in Turkey [28]. Antonelli et al. [26] reported a prevalence of type 2 DM of 8.9% in their cohort. In our study, the mean age of the patients with DM was significantly higher than the mean age of the nondiabetics. This finding can be explained by the increasing incidence of type 2 DM in elderly patients. In this study, the only pathologic prognostic factor that exhibited a significant difference was histological subtype and clear cell subtype, which were significantly higher in type 2 DM patients than in nondiabetic patients. In their recent study, Habib et al. [27] also reported that clear cell histology was the most common histology (92%) in diabetic RCC cases. Additionally, they found a predominance of small, localized RCC in diabetic patients. The cumulative findings of our study and Habib et al's study may suggest an increased risk of development of the clear cell histological cell type in diabetic RCC patients.

This study had some limitations. The retrospective nature of this study was a limiting factor. Although inclusion of HbA_1c levels may have allowed us to predict previous blood glucose status, this test only covers the previous 3-month period, which seems to be inadequate to demonstrate the effects of blood glucose levels on the development and character of RCC. Only serial pre-RCC era HbA_1c measurements may meet the expectations of clinicians. The duration of DM, medication type, and medication time are also important aspects for evaluating the effects of type 2 DM on RCC development and characteristics. One other limitation is the evaluation of the prognostic role of type 2 DM without the inclusion of metabolic syndrome components in the evaluation. Obesity and hypertension, which are established risk factors for RCC, may also affect the prognosis in these patients. Currently, however, there are no studies investigating the prognostic effects of these diseases on RCC.

This study investigated the role and effect of DM on the prognosis of localized RCC that was treated surgically. The present study did not detect DM as an independent prognostic factor for RCC.