Alcohol consumption is an important lifestyle factor that seems to be associated with the risk of physical and mental disabilities, including type 2 diabetes mellitus (T2DM) (O'Connor and Chottenfeld, 1998). Several studies have examined the relationship between alcohol consumption and incidence of T2DM in recent years. Interestingly, some studies showed that moderate alcohol consumption had a protective effect on T2DM, but others demonstrated that high alcohol intake increases the risk of diabetes (i.e., a U-shaped relationship) (Baliunas et al., 2009; Kao et al., 2001). The possible reasons for a diabetogenic effect after high alcohol intake include development of pancreatic β-cell dysfunction, obesity, and impairment of liver function in glucose metabolism. Additionally, large amounts of alcohol consumption have an influence on increased insulin resistance and reduced insulin sensitivity, which could lead to impairment of glucose homeostasis (Hanley et al., 2004; Kim et al., 2010; Poonawala et al., 2000). These pathophysiological profiles of alcohol dependence are linked to the effects of T2DM, affected by blood glucose level, which in turn are influenced by appetite-regulating peptides and neurotrophic factors.
Previous studies have found that brain-derived neurotrophic factor (BDNF), ghrelin, and leptin all mediate alcohol cravings and repetitive behaviors in alcohol dependence during alcohol abstinence. Moreover, BDNF and ghrelin secretion were decreased, and leptin release was elevated significantly by chronic alcohol intake (Badaoui et al., 2008; Joe et al., 2007; Nicolás et al., 2001).
Meanwhile, T2DM causes a reduction in BDNF levels, and diabetic neuropathies and low BDNF levels accompany impaired glucose homeostasis (Krabbe et al., 2007; Nitta et al., 2002). The prevalence of T2DM and insulin resistance was associated with decreased circulating ghrelin concentrations (Pöykkö et al., 2003). In contrast, an additional study showed that the increased leptin secretion can be a risk factor for the T2DM in men (Reilly et al., 2004).
BDNF is the most abundant neurotrophin in the brain and modulates neuronal development, synaptic plasticity, and cognition (Lebrun et al., 2006). Ghrelin and leptin are the hormones that regulate appetite and energy balance in the hypothalamus. Ghrelin is secreted primarily from the oxyntic cells of the stomach and has orexigenic effects, while leptin acts as an appetite suppressant, which is released from adipose tissue. An inverse correlation has previously been reported between ghrelin and leptin (Klok et al., 2007).
In this aspect, progression of diabetes might be impacted by these factors after abstinence, which leads to the improvement of diabetes-related parameters in alcohol dependence. Besides, the onset of diabetes might be linked to the diabetogenic effects of alcohol that causes dysregulation of various metabolic processes. Thus, alcohol abstinence could have a broad influence on the association between alcohol consumption and the incidence of diabetes.
Therefore, the present study tested the hypothesis that alcohol abstinence can alter the risk of glucose intolerance in alcohol-dependent patients. We also investigated whether the changes of BDNF, ghrelin, and leptin secretions are related to alcohol abstinence in alcohol-dependent patients who have glucose intolerance.
- Top of page
- Materials and Methods
Among the 64 subjects, 30 were diagnosed as NGT, 26 were pre-DM (they were only diagnosed with IFG), and 8 were DM patients. Clinical characteristics of subjects are summarized in Table 1. There was no difference in age among the 3 groups. Overall, average plasma AST (both NGT and pre-DM, p < 0.001), ALT (both NGT and pre-DM, p < 0.001), and γ-GTP (NGT, p < 0.01; pre-DM, p < 0.05) levels were significantly lowered after alcohol abstinence when compared with baseline levels of each value. Moreover, concentrations of fasting glucose in both the pre-DM (p < 0.05) and DM groups (p < 0.05) were significantly reduced. But, significantly, mean HOMA-IR was decreased (p < 0.05), and HOMA-β was increased (p < 0.01) only in the pre-DM group.
Table 1. Clinical Characteristics of Baseline and After 30 Days of Alcohol Abstinence
|Number of cases||30||30||26||26||8||8|
|Age (years)||47 ± 7.7||NA||47.1 ± 11.1||NA||51.4 ± 6.9||NA|
|Height (cm)||170.1 ± 4.9||NA||166.2 ± 5.9||NA||167.3 ± 5.8||NA|
|Weight (kg)||67.9 ± 9.4||68.4 ± 9.6||65.8 ± 9.7||66.1 ± 9.1||61.9 ± 10.2||62.4 ± 9.0|
|BMI (kg/m2)||23.4 ± 2.7||23.6 ± 2.7||23.6 ± 2.6||23.7 ± 2.4||22.2 ± 2.3||22.4 ± 2.0|
|AST (IU/l)||91.5 ± 90.1||19.0 ± 11.2**||69.0 ± 48.1||33.0 ± 22.6**||41.1 ± 22.8||34.3 ± 31.3|
|ALT (IU/l)||64.5 ± 68.4||12.6 ± 11.4**||46.1 ± 33.3||15.2 ± 12.1**||25.9 ± 21.6||14.9 ± 11.8|
|γ-GTP (IU/l)||159.8 ± 177.8||59.8 ± 49.1**||206.1 ± 234.0||96.3 ± 93.7*||56.1 ± 57.1||42.8 ± 30.7|
|HDL-cholesterol (mg/dl)||57.3 ± 16.3||51.1 ± 14.7*||51.6 ± 13.9||48.6 ± 16.3||50.9 ± 13.6||45.9 ± 11.3|
|LDL-cholesterol (mg/dl)||108.8 ± 25.3||114.6 ± 26.7||99.8 ± 29.4||106.2 ± 34.4||89.0 ± 32.6||90.1 ± 37.8|
|Triglyceride (mg/dl)||114.3 ± 79.5||122.5 ± 82.0||117.4 ± 85.6||126.5 ± 90.8||117.2 ± 61.7||107.7 ± 62.1|
|C-peptide (ng/ml)||1.6 ± 1.2||1.6 ± 0.6||1.8 ± 0.9||1.9 ± 0.6||1.6 ± 1.1||1.7 ± 0.9|
|Fasting insulin (μIU/ml)||5.3 ± 10.1||5.9 ± 3.5||7.0 ± 4.7||6.4 ± 4.9||7.8 ± 10.5||7.4 ± 8.5|
|Fasting glucose (mg/dl)||88.0 ± 12.6||86.1 ± 7.9||93.5 ± 10.7||89.4 ± 9.5*||125.7 ± 39.0||92.0 ± 19.4*|
|HOMA-IR||0.8011 ± 0.8||0.9375 ± 0.9||1.2197 ± 0.7||0.9387 ± 0.5*||2.2058 ± 2.6||1.7714 ± 2.3|
|HOMA-β||64.5 ± 57.5||66.8 ± 48.1||59.0 ± 42.6||88.9 ± 61.8**||77.6 ± 121.7||140.5 ± 173.3|
|QUICKI||0.4649 ± 0.1||0.4314 ± 0.1||0.3949 ± 0.07||0.4086 ± 0.09||0.3850 ± 0.07||0.4025 ± 0.05|
|BDNF (pg/ml)||1,628.8 ± 1,427.4||1,881.0 ± 1,047.7||1,316.7 ± 1,104.1||3,421.8 ± 1,118.7**||1,363.8 ± 1,164.6||2,052.1 ± 980.6|
|Ghrelin (pg/ml)||30.2 ± 24.8||154.4 ± 59.2**||23.7 ± 2.4||263.4 ± 205.5**||22.4 ± 2.0||223.0 ± 114.7**|
|Leptin (pg/ml)||3,215.6 ± 2,044.2||2,024.2 ± 2,137.8*||4,906.3 ± 5,426.0||1,927.0 ± 1,554.4**||11,247.5 ± 8,790.2||3,191.2 ± 4,482.2*|
Also BDNF, ghrelin, and leptin levels before and after alcohol abstinence in each group are shown in Table 1. All of the groups showed significantly reduced ghrelin levels (NGT and pre-DM, p < 0.001; DM, p < 0.01) and elevated leptin levels (NGT and DM, p < 0.05; pre-DM, p < 0.01). Only in the pre-DM group was there increased BDNF levels (p < 0.001).
Figure 1 shows pre–post differences in HOMA-IR, fasting glucose, leptin, BDNF, and ghrelin levels as well as a comparison of those differences among the 3 groups using repeated measures ANOVA. The results showed significant group (HOMA-IR, p < 0.05; glucose, leptin, BDNF, ghrelin, p < 0.001) and time (HOMA-IR, p < 0.05; glucose, leptin, BDNF, ghrelin, p < 0.001) effect on factor levels. Reduced levels of HOMA-IR (Fig. 1A), fasting glucose (Fig. 1B), and leptin (Fig. 1C) were significantly greater in the DM group than in NGT group, and increased BDNF (Fig. 1D) and ghrelin (Fig. 1E) levels were significantly higher in pre-DM group compared with the NGT group. Repeated measures ANOVA showed a significant interaction among the 3 groups and time after alcohol abstinence (HOMA-IR, p < 0.05, fasting glucose, p < 0.01; leptin and BDNF, p < 0.001; ghrelin, p < 0.05).
Figure 1. (A) HOMA-IR, (B) fasting glucose, (C) leptin, (D) BDNF, and (E) ghrelin level changes after alcohol abstinence in the 3 groups. Three groups are represented as white circle, solid line—NGT group (n = 30); black circle, dashed line—pre-DM group (n = 26); black square, dotted line—DM group (n = 8). The repeated measures ANOVA demonstrated significant effects of group (HOMA-IR, p < 0.05; glucose, leptin, BDNF, ghrelin, p < 0.001), and time (HOMA-IR, p < 0.05; glucose, leptin, BDNF, ghrelin, p < 0.001), as well as significant interactions (HOMA-IR, p < 0.05, fasting glucose, p < 0.01; leptin and BDNF, p < 0.001; ghrelin, p < 0.05). p By repeated-measures ANOVA, vs. NGT group. *The level which is more differed between pre- and post-abstinence than the NGT group. BDNF, brain-derived neurotrophic factor; HOMA, homeostasis model assessment; NGT, normal glucose tolerance.
Download figure to PowerPoint
Interestingly, the percent change of leptin was shown to have a positive correlation with percent change in fasting glucose (p < 0.05, r = 0.937) and HOMA-IR (p < 0.01, r = 0.728) in the DM group. Otherwise, the pre-DM group showed positive correlation between percent change in BDNF and ghrelin (p < 0.01, r = 0.550).
- Top of page
- Materials and Methods
In this study, we found that alcohol abstinence could improve diabetes-related factors (fasting glucose level, insulin resistance, and insulin secretion index) in alcohol-dependent patients with glucose intolerance. Interestingly, the pre-DM and DM groups had similarities and differences in results with respect to the significantly changed values and correlations.
The pre-DM group exhibited significantly reduced fasting glucose, HOMA-IR, and leptin levels as well as elevated HOMA-β, ghrelin, and BDNF levels. In particular, differences in ghrelin and BDNF levels were the greatest among the 3 groups, and changes in ghrelin and BDNF levels were positively correlated.
Similarly to the pre-DM group, the DM group showed the follow-up levels of HOMA-IR and HOMA-β were considerably altered compared with the baseline levels of them, although the changes were not statistically significant. However, the DM group had significantly increased ghrelin and reduced leptin and fasting glucose levels, in addition to the greatest decreases in HOMA-IR, fasting glucose, and leptin levels. Further, decreases in leptin were positively correlated with HOMA-IR and fasting glucose levels.
Several studies have suggested that obesity is strongly linked to T2DM (McKeigue et al., 1991; Mokdad et al., 2003), and heavy drinking increases body mass index (BMI) and obesity (Arkwright et al., 1982). But in this study, BMI and lipid profiles (HDL cholesterol, LDL cholesterol, and triglyceride) of the pre-DM and DM groups did not change after alcohol abstinence. Further, lipid profile levels before and after abstinence were lower than normal reference levels (HDL cholesterol, 40 to 60 mg/dl; LDL cholesterol, ≤130 mg/dl; triglyceride, ≤150 mg/dl). This indicates that there were no changes of obesity-related parameters, which were caused by chronic alcohol consumption, but alcohol dependence may be one of the risk factors for T2DM.
In addition, plasma liver enzyme (AST, ALT, and γ-GTP) levels were significantly reduced in both the NGT and pre-DM groups, while baseline and decreased concentrations of liver enzymes in DM group were the lowest but not statistically significant. This might be because both alcohol dependence and diabetes influence liver damage, so liver function in the DM group was most impaired. In fact, elevated liver enzymes are known predictors of not only alcohol dependence but also diabetes, and chronic high glucose intake might induce liver cell apoptosis (Kim et al., 2005b; Nannipieri et al., 2005).
It has been suggested that chronic alcohol consumption decreases ghrelin secretion in plasma and the fundus in stomach, its main site of secretion. This is to the fact that alcohol might affect ghrelin-producing cells in the stomach directly (Badaoui et al., 2008; Calissendorff et al., 2005). Further, effects of ghrelin in glucose and insulin homeostasis have been actively researched. Ghrelin deficiency was shown to be associated with hyposomatotropism, and this may cause raised insulin resistance and reduced insulin sensitivity levels (Buijs et al., 2003), which could play a role in developing T2DM. After alcohol abstinence, repair and proliferation of ghrelin-producing cells may increase, reduced leptin secretion induces ghrelin production inversely, and lowered plasma glucose levels are likely to stimulate ghrelin secretion as well. Furthermore, induced ghrelin secretion was shown to influence insulin resistance and sensitivity; consequently, these changes could affect glucose metabolic balance.
Several studies have suggested that BDNF secretion is affected by chronic alcohol intake and interacts with glucose metabolism. BDNF is a target of ethanol toxicity, and long-term alcohol ingestion leads to decreases in hippocampal BDNF expression, neuronal survival, and differentiation (MacLennan et al., 1995; Tateno et al., 2004). Human studies have shown similar findings where BDNF levels in alcohol-dependent patients were lower than normal controls (Joe et al., 2007). BDNF is also known to control glucose homeostasis and maintain cellular organization of β cells in pancreatic islets (Yamanaka et al., 2006). Moreover, cerebral output of BDNF was also negatively regulated by high glucose concentrations; hence, low circulating BDNF levels have been found in T2DM subjects (Krabbe et al., 2007). Consequentially, alcohol abstinence restores BDNF homeostasis, resulting in increased plasma BDNF levels. As not only alcohol addiction but also diabetes induces neurodegeneration (Nitta et al., 2002), our results showed that BDNF might be synthesized more in the pre-DM group compared with the NGT group because of neuronal compensation. Then perhaps elevated BDNF levels participate in glucose metabolism and beta cell function.
Most of the subjects in the pre-DM group had increased ghrelin levels after alcohol abstinence, which is accompanied by elevated BDNF levels. It can be explained by hyperleptinemia, which results from chronic alcohol intake. In hyperleptinemia, BDNF suppresses leptin secretion (Wang et al., 2010), and elevated BDNF secretion could accelerate this phenomenon. As leptin is inversely regulated with ghrelin (Klok et al., 2007), inhibition of leptin secretion results in increased ghrelin secretion. Even though it was not significantly correlated, changes in BDNF and leptin levels were negatively connected.
Actually, BDNF and ghrelin have similar roles in the pathophysiology of alcohol dependence and the ability to maintain abstinence. Many studies have demonstrated that BDNF seems to control the gating of alcohol intake and inhibit relapse in alcohol intake, thereby avoiding the development of alcohol addiction (Ghitza et al., 2010; Jeanblanc et al., 2009). Moreover, maybe BDNF disinhibition has a defensive role against neuronal damage and affects synaptic reorganization during abstinence from alcohol addiction (Ghitza et al., 2010). Likewise, ghrelin might also play a role in alcohol-seeking behavior during the period of abstinence. Our previous study has found that ghrelin secretion is positively correlated with the period of abstinence (Kim et al., 2005a,b), and this indicates that increased ghrelin release could help to prevent relapse in alcohol dependence.
There is another interesting issue that BDNF and ghrelin are also related in hyperglycemia in addition to alcohol relapse. Many cross-sectional studies have found that hyperglycemia appears to be positively connected to high alcohol drinking and that intake of alcohol causes dose-dependent, long-lasting hyperglycemia (Erwin and Towell, 1983; Yoon et al., 2004). In addition, hyperglycemia suppresses BDNF gene expression through inhibition of neuronal excitability (Uchino et al., 1997). From the viewpoint of deficiency of ghrelin in hyperglycemia, on the other hand, insulin resistance might induce the down-regulation of ghrelin secretion (Ukkola et al., 2006). In our study, alcohol abstinence might have improved blood glucose control, and this could also increase BDNF and ghrelin levels indirectly. Moreover, these elevated levels possibly alleviated hyperglycemia independently. Previous research has shown that administration of ghrelin prevents hyperglycemic effects and that BDNF stimulates glucose utilization and uptake (Gauna et al., 2004; Nakagawa et al., 2002). Although the direct mechanism between BDNF and ghrelin has not been fully identified, these factors seem to synergically participate in neuroadaptation and alleviating hyperglycemia.
Alcohol-dependent patients tend to release more leptin regardless of body fat content, usually because their fat tissue is more sensitized (Nicolás et al., 2001). Previously, our studies have reported that elevated leptin concentrations are associated with insulin resistance, independent of BMI or body fat mass (Ju et al., 2011). In fact, elevated insulin resistance could also lead to increased leptin concentrations, but the exact mechanism of insulin resistance–induced leptin secretion is not fully understood. Another study has demonstrated that long periods of leptin treatment resulted in reduced insulin-stimulated glucose utilization in skeletal muscle. This is because insulin-stimulated p38 mitogen-activated protein kinase activation was inhibited, and glucose transporter 4 activation was decreased by leptin (Sweeney et al., 2001). Leptin also could interfere with insulin signaling and induce gluconeogenesis in hepatocytes directly (Cohen et al., 1996).
During alcohol abstinence, decreases in fat tissue sensitivity likely leads to reduced leptin secretion (Nicolás et al., 2001). Besides, our findings showed that elevated leptin levels were not significantly correlated with changes in fat content levels. Lowered leptin levels may influence insulin resistance and glucose metabolism, and our data revealed that decreased leptin levels were accompanied by reduced HOMA-IR or fasting glucose concentrations.
Two limitations deserve consideration. First of all, our study included only men, because we did not find alcohol dependence among women during our investigation to perform analyses similar to those reported herein. In fact, gender differences in alcohol dependence and T2DM have been reported; for instance, women were shown to have a more rapid development of alcohol dependence (Mann et al., 2005), and sex effects were noted in DM because fat deposition and insulin sensitivity are different between men and women (Gale and Gillespie, 2001). Second, the medical management of symptomatic alcohol withdrawal associated with benzodiazepine could influence alcohol abstinence. Therefore, there is a need to test alcohol-dependent women with glucose intolerance to conclude whether our findings also apply to women, and further studies should be conducted systematically without using medical prescriptions.
In conclusion, to our knowledge, this is the first study to show that 1 month of alcohol abstinence can decrease fasting glucose levels in alcohol-dependent patients with glucose intolerance. Importantly, T2DM-related parameters could be improved just by alcohol abstinence without antidiabetic agents.
We demonstrated that the pre-DM and DM groups improved T2DM-related parameters, along with alteration of BDNF, ghrelin, and leptin levels. Moreover, BDNF, ghrelin, and leptin might differentially influence alterations in values depending on the stage of T2DM with alcohol-dependent patients, extrapolating from our results. Leptin may directly regulate glucose uptake and gluconeogenesis, so decreases in fasting glucose concentrations were the greatest in the DM group following alcohol abstinence. Further, the pre-DM group exhibited a positive correlation between increase in BDNF and ghrelin levels during the alcohol abstinence, which might prevent relapse of alcohol dependence and relieve hyperglycemia together. Accordingly, the assessment of metabolic parameters, BDNF, ghrelin, and leptin is a crucial consideration in overall evaluation of alcohol abstinence and therapeutic intervention for alcohol dependence with glucose intolerance.