**4. Discussion**

332 Dyslipidemia - From Prevention to Treatment

Table 11 reports the variations of uric acid, homocysteine, folate and vitamin B12 concentrations according to the illness episode and therapeutic characteristics of bipolar I

(n = 130) 311 ± 99 246 ± 97 15.8 ± 8.9 3.3 ± 0.9 356 ± 198

(n = 21) 228 ± 79 205 ± 128 15.2 ± 6.7 3.4 ± 1.3 481 ± 299a

(n = 73) 309 ± 110 271 ± 95 16.1 ± 10.1 3.4 ± 0.7 322 ± 165

(n = 36) 327 ± 87 217 ± 57 15.5 ± 7.7 3.3 ± 1.2 352 ± 158

(n = 64) 328 ± 91 279 ± 109 16.3 ± 10.0 3.5 ± 0.9 361 ± 91

(n = 12) 331 ± 79 219 ± 88 17.1 ± 12.7 2.8 ± 0.6 399 ± 79

(n = 10) 278 ± 160 207± 61 16.9 ± 9.4 3.2 ± 0.8 240 ± 160**\***

(n = 6) 343 ± 76 288 15.5 ± 2.9 3.0 ± 0.8 518 ± 236**\***

(n = 38) 269 ± 96 218 ± 83 14.2 ± 5.9 3.4 ± 1.5 338 ± 233

We found a significant association between vitamin B12 values and illness episode (F 2-130 = 5.688, p = 0.004). Manic patients had lower values of this parameter than depressive patients. Moreover, we showed that vitamin B12 was significantly associated with the therapeutic characteristics. Indeed, patients taking carbamazepine had significantly lower values of this parameter than those taking valproic acid and lithium (p = 0.04) (table 11). In patients, there was no significant change in homocysteine, folate and uric acid values in relation to illness episodes and the treatment, whereas the lowest values of uric acid were seen in depressive patients (both in men and women) compared to manic patients and in men taking antipsychotics and women taking carbamazepine compared to the other groups

The distribution of BMI according to the illness episode and therapeutic characteristics is

Table 11. Variations of uric acid, homocysteine, folate and vitamin B12 concentrations according to the illness episode and therapeutic characteristics of bipolar I patients.

AVP: valproic acid; Li: lithium; \*Carb Vs AVP/Li, p = 0.04; aF 2-130 = 5.688, p = 0.004

**Homocysteine (µmol/L)** 

**Folate (µg/L)** 

**Vitamin B12 (ng/L)** 

**Uric acid (µmol/L)** 

**Women (n = 45)** 

**Men (n = 85)** 

patients.

*Patients* 

Euthymic

*Treatment* V*alproic acid* 

Lithium

Carbamazepine

AVP and Li

A*ntipsychotics*

(table 11).

shown in table 12.

Manic

*Illness episode* Depressive

> Our study showed that patients had significantly higher levels of triglycerides and Lp(a), and significantly lower levels of ApoA1 than control subjects. Furthermore, bipolar I disorder was showed to have significant association with hyperLp(a) (47.7% *Vs* 14.8%, OR = 4.48, IC 95% = 2.53-7.95; p < 0.001) and hypertriglyceridemia (53.1% Vs 17.7%, OR = 3.71, IC 95% = 2.13-6.46; p < 0.001).

> In patients, the TG/HDL ratio and Lp(a) were found as the best predictive factors of cardiovascular risk in terms of sensibility (0.62, 0.74; respectively) and specificity (0.63, 0.74; respectively) at threshold of 1.12 and 168 mg/L, respectively. These results reflect a high risk of cardiovascular disease and may explain the high rates of morbidity and mortality in this population. Several studies have found mortality rates between 1.5 and 2.5 times higher in bipolar patients than the general population. After suicide and accidents, cardiovascular and all vascular diseases are the leading causes of death in these patients, with standardized mortality ratios ranging from 1.47 to 2.6. (Garcia-Portilla et al., 2009; Sicras et al., 2008).

Cardiovascular Risk in Tunisian Patients with Bipolar I Disorder 335

manic, mixed, or depressive episode. These differences could be due to ethnicity and

About therapeutic characteristics, any significant association was shown between lipid profile parameters and treatment, while, women taking lithium had the lowest c-HDL values and patients taking carbamazepine had the highest values of Lp(a). The mechanism(s) by which these drugs exert weight gain are not well known, but are presumed to involve increased energy intake (e.g., overeating), decreased energy expenditure (e.g., reduced resting metabolic rate, reduced physical activity, or reduced diet-

Additionally, we found that the prevalence of metabolic syndrome was 26.1% among patients, 24.7% in men and 28.8% in women. These prevalences were definitely higher than those reported in the Tunisian general population (13% in men and 18% in women) using a

Compared with other studies, the prevalence of metabolic syndrome in our patients is included between those in Spanish patients (22.4%), Italian patients (25.3%) and US patients (30%) (Garcia-Portilla et al., 2008; Salvi et al., 2008; Fagiolini et al., 2005). The increasing prevalence of metabolic syndrome is important because it confers greater cardiovascular morbidity and mortality. Prospective observational studies have demonstrated an association between metabolic syndrome and development of type II diabetes (Hanson et al., 2002; Resnick et al., 2003; Klein et al., 2002; Sattar et al., 2003), cardiovascular disease

Our study showed that the highest prevalence of metabolic syndrome was obtained by the association between obesity, low c-HDL and hypertriglyceridemia. Moreover, the most individual components of this syndrome, in the total sample of patients, was low c-HDL (59.2%), hypertriglyceridemia (53.1%) and obesity (BMI ≥ 28.5 kg/m²) (33.8%), confirming in part the higher risk of dyslipidemia and obesity in bipolar I patients and in other hand the

We found no significant difference in the prevalence of metabolic syndrome among gender

We noted that there was no significant change in the prevalence of metabolic syndrome in relation to illness episode; however, manic patients had the highest prevalence. This may explain the high risk of cardiovascular disease in manic patients compared with depressive one (Murray et al., 2009). Additionally, Angst et al. (2002) showed that individuals with bipolar I disorder are at greater risk for cardiovascular mortality than individuals with bipolar II disorder. However, the difference in cardiovascular mortality between the two bipolar subtypes reflects the manic symptom burden, which predicts cardiovascular mortality independently of diagnosis and cardiovascular risk factors at intake. The results suggest that mania, either directly (through factors intrinsic to illness) or indirectly (through

other mediators or associated variables), may itself influence cardiovascular disease.

Our study failed to show any significant association between metabolic syndrome and treatment. However, we noted that patients treated with lithium had the highest prevalence of metabolic syndrome. The increased risk to develop metabolic syndrome during treatment with lithium is in part related to its propensity to induce weight gain. According to Casey, lithium has been shown to stimulate appetite and increase calorie intake through different

induced thermogenesis), or a combination of the two (Malhotra & McElroy, 2002).

eating habits.

mechanisms.

previous criteria (Bouguerra et al., 2006).

(Lakka et al., 2002; Kip et al., 2004), and stroke (Kurl et al.,2006).

and age. This is in line with results reported by Yumru et al., (2007).

higher risk of cardiovascular disease in this population.

The exact mechanisms increasing the incidence of cardiovascular risk in bipolar patients remain to be clarified, but they possibly include industrialisation, stress, lack of exercise, dietary lipids (that is, omega-3 fatty acid deficiency), increasing incidence of smoking and alcohol consumption and other factors (Ezzaher et al., 2010). These hypotheses will be, in part, justified later in this study.

Several investigators have been hypothesized that abnormalities in fatty acid composition may play a role in psychiatric disorders (Horrobin & Bennett, 1999). Maes et al. (1996, 1999) reported that patients with major depression had a significantly elevated ratio of ecosapentaenoic acid (EPA; 20: 5n-3)/docosahexaenoic acid (DHA; 22: 6n-3), lower level of EPA and total n-3 Omega-3 polyunsaturated fatty acids, in both serum cholesteryl esters and phospholipids when compared to patients with minor depression and normal controls. Similar findings were revealed in terms of fatty acid compositions of the erythrocyte membrane (Adams et al., 1996; Edwards et al., 1998; Peet et al., 1998; Chiu et al., 2003).

Moreover, many prospective and case-control studies have shown a positive association between serum triglycerides and coronary artery disease risk and demonstrated the importance of fasting triglycerides level as an independent risk factor. A number of clinical trials including the Framingham Heart Study have concluded that a low HDL cholesterol level predicts the risk for coronary artery disease independently of other risk factors. Each 1 mg/dL decrease in HDL cholesterol has been shown to increase risk for coronary artery disease by 2% in men and 3% in women. The Veterans Affairs High-Density Lipoprotein Cholesterol Interventional Trial, investigating the impact of fibrate therapy on cardiovascular risk, demonstrated that 6% increase in HDL cholesterol was associated with a 22% decrease in coronary events (Kabakci et al., 2008). In addition, Lp(a) has been shown to be an independent risk factor for atherosclerosis (Hakim et al., 2008) and has been found to exert a broad variety of pro-atherogenic and pro-thrombotic properties (von Eckardstein et al., 2001). Elevated plasma Lp(a) has been shown also to be associated with premature cardiovascular disease, premature cerebrovascular disease and premature peripheral vascular disease (Valentine et al., 1996).

The underlying mechanism for the altered lipid status in bipolar patients is unclear. A possible explanation might be found in the patient's nutritional status, the decrease in physical activity and the medications used (Ezzaher et al., 2010). Additionally, Chung et al. (2007) reported that bipolar disorder is associated with perturbations in lipid profile which play an important role in the pathophysiology of mood disorders, particularly in bipolar disorders. Indeed, cholesterol is one component of circulating lipoprotein particles that, besides handling cholesterol, carries micronutrients such as vitamins A and E as well as triglycerides and phospholipids. The latter compounds give rise to substrates such as fatty acids and choline, which are used in both the structural lipids of neuronal membranes and intercellular communication. Therefore, higher levels of one or more compounds of lipoprotein particles circulating in the bloodstream may produce subtle but measurable enhancements of mental processes by influencing the supply of fat-soluble micronutrients, specific fatty acids, or structural lipids (Ezzaher et al., 2010).

Our study failed to found any significant association between lipid profile parameters and illness episode, while euthymic patients were found to have the highest levels of Lp(a). Additionally, depressive patients had the highest levels of ApoB/Apo A1 ratio. However, some authors (Sagud et al., 2009) showed that serum cholesterol and LDL values were significantly lower in manic patients and others (Chung et al., 2007) showed that there was no difference in mean serum level of cholesterol or triglycerides among patients with

The exact mechanisms increasing the incidence of cardiovascular risk in bipolar patients remain to be clarified, but they possibly include industrialisation, stress, lack of exercise, dietary lipids (that is, omega-3 fatty acid deficiency), increasing incidence of smoking and alcohol consumption and other factors (Ezzaher et al., 2010). These hypotheses will be, in

Several investigators have been hypothesized that abnormalities in fatty acid composition may play a role in psychiatric disorders (Horrobin & Bennett, 1999). Maes et al. (1996, 1999) reported that patients with major depression had a significantly elevated ratio of ecosapentaenoic acid (EPA; 20: 5n-3)/docosahexaenoic acid (DHA; 22: 6n-3), lower level of EPA and total n-3 Omega-3 polyunsaturated fatty acids, in both serum cholesteryl esters and phospholipids when compared to patients with minor depression and normal controls. Similar findings were revealed in terms of fatty acid compositions of the erythrocyte membrane (Adams et al., 1996; Edwards et al., 1998; Peet et al., 1998; Chiu et al., 2003). Moreover, many prospective and case-control studies have shown a positive association between serum triglycerides and coronary artery disease risk and demonstrated the importance of fasting triglycerides level as an independent risk factor. A number of clinical trials including the Framingham Heart Study have concluded that a low HDL cholesterol level predicts the risk for coronary artery disease independently of other risk factors. Each 1 mg/dL decrease in HDL cholesterol has been shown to increase risk for coronary artery disease by 2% in men and 3% in women. The Veterans Affairs High-Density Lipoprotein Cholesterol Interventional Trial, investigating the impact of fibrate therapy on cardiovascular risk, demonstrated that 6% increase in HDL cholesterol was associated with a 22% decrease in coronary events (Kabakci et al., 2008). In addition, Lp(a) has been shown to be an independent risk factor for atherosclerosis (Hakim et al., 2008) and has been found to exert a broad variety of pro-atherogenic and pro-thrombotic properties (von Eckardstein et al., 2001). Elevated plasma Lp(a) has been shown also to be associated with premature cardiovascular disease, premature cerebrovascular disease and premature peripheral

The underlying mechanism for the altered lipid status in bipolar patients is unclear. A possible explanation might be found in the patient's nutritional status, the decrease in physical activity and the medications used (Ezzaher et al., 2010). Additionally, Chung et al. (2007) reported that bipolar disorder is associated with perturbations in lipid profile which play an important role in the pathophysiology of mood disorders, particularly in bipolar disorders. Indeed, cholesterol is one component of circulating lipoprotein particles that, besides handling cholesterol, carries micronutrients such as vitamins A and E as well as triglycerides and phospholipids. The latter compounds give rise to substrates such as fatty acids and choline, which are used in both the structural lipids of neuronal membranes and intercellular communication. Therefore, higher levels of one or more compounds of lipoprotein particles circulating in the bloodstream may produce subtle but measurable enhancements of mental processes by influencing the supply of fat-soluble micronutrients,

Our study failed to found any significant association between lipid profile parameters and illness episode, while euthymic patients were found to have the highest levels of Lp(a). Additionally, depressive patients had the highest levels of ApoB/Apo A1 ratio. However, some authors (Sagud et al., 2009) showed that serum cholesterol and LDL values were significantly lower in manic patients and others (Chung et al., 2007) showed that there was no difference in mean serum level of cholesterol or triglycerides among patients with

part, justified later in this study.

vascular disease (Valentine et al., 1996).

specific fatty acids, or structural lipids (Ezzaher et al., 2010).

manic, mixed, or depressive episode. These differences could be due to ethnicity and eating habits.

About therapeutic characteristics, any significant association was shown between lipid profile parameters and treatment, while, women taking lithium had the lowest c-HDL values and patients taking carbamazepine had the highest values of Lp(a). The mechanism(s) by which these drugs exert weight gain are not well known, but are presumed to involve increased energy intake (e.g., overeating), decreased energy expenditure (e.g., reduced resting metabolic rate, reduced physical activity, or reduced dietinduced thermogenesis), or a combination of the two (Malhotra & McElroy, 2002).

Additionally, we found that the prevalence of metabolic syndrome was 26.1% among patients, 24.7% in men and 28.8% in women. These prevalences were definitely higher than those reported in the Tunisian general population (13% in men and 18% in women) using a previous criteria (Bouguerra et al., 2006).

Compared with other studies, the prevalence of metabolic syndrome in our patients is included between those in Spanish patients (22.4%), Italian patients (25.3%) and US patients (30%) (Garcia-Portilla et al., 2008; Salvi et al., 2008; Fagiolini et al., 2005). The increasing prevalence of metabolic syndrome is important because it confers greater cardiovascular morbidity and mortality. Prospective observational studies have demonstrated an association between metabolic syndrome and development of type II diabetes (Hanson et al., 2002; Resnick et al., 2003; Klein et al., 2002; Sattar et al., 2003), cardiovascular disease (Lakka et al., 2002; Kip et al., 2004), and stroke (Kurl et al.,2006).

Our study showed that the highest prevalence of metabolic syndrome was obtained by the association between obesity, low c-HDL and hypertriglyceridemia. Moreover, the most individual components of this syndrome, in the total sample of patients, was low c-HDL (59.2%), hypertriglyceridemia (53.1%) and obesity (BMI ≥ 28.5 kg/m²) (33.8%), confirming in part the higher risk of dyslipidemia and obesity in bipolar I patients and in other hand the higher risk of cardiovascular disease in this population.

We found no significant difference in the prevalence of metabolic syndrome among gender and age. This is in line with results reported by Yumru et al., (2007).

We noted that there was no significant change in the prevalence of metabolic syndrome in relation to illness episode; however, manic patients had the highest prevalence. This may explain the high risk of cardiovascular disease in manic patients compared with depressive one (Murray et al., 2009). Additionally, Angst et al. (2002) showed that individuals with bipolar I disorder are at greater risk for cardiovascular mortality than individuals with bipolar II disorder. However, the difference in cardiovascular mortality between the two bipolar subtypes reflects the manic symptom burden, which predicts cardiovascular mortality independently of diagnosis and cardiovascular risk factors at intake. The results suggest that mania, either directly (through factors intrinsic to illness) or indirectly (through other mediators or associated variables), may itself influence cardiovascular disease.

Our study failed to show any significant association between metabolic syndrome and treatment. However, we noted that patients treated with lithium had the highest prevalence of metabolic syndrome. The increased risk to develop metabolic syndrome during treatment with lithium is in part related to its propensity to induce weight gain. According to Casey, lithium has been shown to stimulate appetite and increase calorie intake through different mechanisms.

Cardiovascular Risk in Tunisian Patients with Bipolar I Disorder 337

higher only during the manic phase of bipolar disorder but not during the depressive or euthymic phases (De Berardis et al., 2008). Additionally, lithium was found to low uric acid plasma levels and to have uricosuric effects in mania. Carbamazepine and phenytoin similarly decreased uric acid levels; in contrast, valproate appeared to have the opposite effect. However, it is important to note that the effect of these drugs on uric acid levels in relationship to clinical improvement in patients with bipolar disorder has not been

Compared with controls, patients had significantly higher levels of homocysteine and significantly lower levels of folatemia. Additionally, significant associations were showed between bipolar I disorder and hyperhomocysteinemia (39.2% *Vs* 18%, OR = 1.95, IC 95% = 1.04-3.69; p = 0.038) and hypofolatemia (66.2% *Vs* 36.2%, OR = 3.69, IC 95% = 2.20-6.19; p < 0.001). Homocysteine is an intermediary metabolite of the essential amino acid methionine. Folate and vitamin B12 are required for remethylation of homocysteine to methionine

According to Reynolds (2006), hyperhomocysteinaemia has long been identified as a risk factor for vascular disease and the lowering of homocysteine concentrations by the treatment with folic acid, or possibly vitamin B12 and vitamin B6 which might reduce the risk of both cardiovascular and cerebrovascular diseases. Moreover, the association between increased circulating homocysteine concentrations and premature vascular thrombotic events in individuals with hereditary homocystinuria is well established. This process may include platelet activation, smooth muscle cell proliferation, and enhanced leukocyte binding to the endothelium. In recent years, a relationship between milder degrees of hyperhomocysteinaemia and vascular disease has emerged, and this has been the subject of intense research. Hyperhomocysteinemia can be caused by a wide range of disorders, the most important of which are genetic defects of the enzymes involved in homocysteine metabolism and/or deficiencies of their co-factors: folate (former vitamin B9), vitamin B12

Our study showed a significant association between bipolar I disorder and hyperhomocysteinemia. The exact mechanisms underlying the hyperhomocysteinemia in this disease are not completely understood and controversed among studies. Several hypotheses have been postulated including nutritional folate and vitamin B deficiency, and/or reduced glomerular filtration rate in bipolar patients (Vuksan-Ćusa et al., 2011). En effect, we found a significant association between this disease and hypofolatemia. Furthermore, some authors (Atmaca et al., 2005) showed that at a high concentration, homocysteine is considered to be a neurotoxic substance, causing activation of NMDA (Nmethyl D-aspartate) receptors and leading to excitotoxicity. By impairing the neural plasticity and promoting neuronal degeneration, homocysteine could contribute to the pathogenesis of neurodegenerative and psychiatric disorders (Ipcioglu et al., 2008). Additionally, homocysteine is a methyl donor when activated to S-adenosylmethionine. So aberrant DNA methylation due to hyperhomocysteinemia also may be involved in the

In the other hand, folate appears to influence the synthesis rate of tetrahydrobiopterin, a cofactor in the hydroxylation of phenylalanine and tryptophan, rate-limiting steps in the biosynthesis of dopamine, norepinephrine, and serotonin, neurotransmitters postulated to play a role in the monoamine hypothesis of affective disorders. In addition, methyl tetrahydrofolate has been shown to bind to presynaptic glutamate receptors, where it may

pathogenesis of bipolar disorder as well as schizophrenia (Mill et al., 2008).

systematically evaluated (Salvadore et al., 2010).

and vitamin B6 (Haj mouhamed et al., 2011).

(Hankey & Eikelboom, 1999).

HOMA-IR is significantly higher in patients with metabolic syndrome than others. This increase in HOMA-IR values reflects an insulin resistance and is associated with two to three fold increases in cardiovascular disease independent of classical risk factors (Toalson et al., 2004). In addition, uric acid levels were significantly higher in patients with metabolic syndrome. According to Vuorinen-Markkola et al. (1994), hyperuricemia forms another consistent feature of the metabolic syndrome what led to the suggestion of uric acid being a new component of the syndrome.

In addition, Chien et al., (2008) reported that metabolic syndrome induces high oxidative stress and the accompanying hyperuricemia worsens this stress. Furthermore, uric acid stimulates vascular smooth muscle proliferation, induces endothelial dysfunction, decreases endothelial nitric oxid production, and consequently, makes peripheral tissue resistant to insulin effects and results in endothelial dysfunction (Chien et al., 2008). High levels of uric acid are associated with increased renal glomerular pressure and sodium reabsorption, enhanced by high insulin concentrations (Alkerwi et al., 2009). In addition, hyperuricemia was associated with insulin resistance markers, including triglycerides, microalbuminuria and impaired glucose tolerance. These disturbances contribute to increase cardiovascular risk (Chien et al., 2008). This insulin- resistance causes steatosis, which is associated with hyper secretion of hepatic enzymes (Fromenty et al., 2004).

In men, uric acid was significantly higher in patients than controls. Additionally, the risk of hyperuricemia in bipolar I patients was approximately multiplied by 1.5 (10.8% *Vs* 4.4%, OR = 1.58, IC 95% = 0.49-5.08; p = 0.439). Many, but not all, epidemiological studies have suggested that high plasma uric acid is a risk factor for cardiovascular diseases. This raised level of plasma uric acid, parallel to an increased risk of cardiovascular diseases, could be either primary or secondary to the underlying causes of the cardiovascular diseases. However, the specific role of plasma uric acid in this constellation remains uncertain, although it may be involved in the platelet adhesiveness, aggregation, or inflammation and it may be implicated in the genesis of hypertension. In contrast, there is some evidence that the increase of plasma uric acid is protective against the cardiovascular diseases, since uric acid acts as an endogenous antioxidant, and the higher plasma uric acid levels found in cardiovascular diseases patients suggest that any protective antioxidant effect of uric acid is hidden by other negative effects in these pathogeneses (Haj mouhamed et al., 2010).

Additionally, Torres et al. (2007), reported that hyperuricemia which implicated in the oxidative stress plays an important role in the pathophysiology of bipolar disorders. The idea that the purinergic system might be involved in bipolar disorder dates back to Kraepelin, who was the first to describe an association between manic symptoms, uric acid excretion, hyperuricemia, and gout. In fact, the purinergic system modulates sleep, motor activity, cognition, attention, behavior, and mood. Even in the absence of a psychiatric diagnosis, individuals with higher uric acid levels are more likely to show higher drive, disinhibition, hyperthymia, or irritable temperament (Lorenzi et al., 2010). Similarly, diseases characterized by purinergic turnover dysfunction and uric acid overproduction (e.g., Lesch–Nyhan syndrome) are associated with impulsive/aggressive behavior, disinhibition, and increased sexual drive (Salvadore et al., 2010).

Among clinical and therapeutic characteristics, we found that there was no significant change in uric acid values in relation to illness episodes and the treatment. This finding is not in agreement with the previous studies that reported that plasma uric acid levels were

HOMA-IR is significantly higher in patients with metabolic syndrome than others. This increase in HOMA-IR values reflects an insulin resistance and is associated with two to three fold increases in cardiovascular disease independent of classical risk factors (Toalson et al., 2004). In addition, uric acid levels were significantly higher in patients with metabolic syndrome. According to Vuorinen-Markkola et al. (1994), hyperuricemia forms another consistent feature of the metabolic syndrome what led to the suggestion of uric acid being a

In addition, Chien et al., (2008) reported that metabolic syndrome induces high oxidative stress and the accompanying hyperuricemia worsens this stress. Furthermore, uric acid stimulates vascular smooth muscle proliferation, induces endothelial dysfunction, decreases endothelial nitric oxid production, and consequently, makes peripheral tissue resistant to insulin effects and results in endothelial dysfunction (Chien et al., 2008). High levels of uric acid are associated with increased renal glomerular pressure and sodium reabsorption, enhanced by high insulin concentrations (Alkerwi et al., 2009). In addition, hyperuricemia was associated with insulin resistance markers, including triglycerides, microalbuminuria and impaired glucose tolerance. These disturbances contribute to increase cardiovascular risk (Chien et al., 2008). This insulin- resistance causes steatosis, which is associated with

In men, uric acid was significantly higher in patients than controls. Additionally, the risk of hyperuricemia in bipolar I patients was approximately multiplied by 1.5 (10.8% *Vs* 4.4%, OR = 1.58, IC 95% = 0.49-5.08; p = 0.439). Many, but not all, epidemiological studies have suggested that high plasma uric acid is a risk factor for cardiovascular diseases. This raised level of plasma uric acid, parallel to an increased risk of cardiovascular diseases, could be either primary or secondary to the underlying causes of the cardiovascular diseases. However, the specific role of plasma uric acid in this constellation remains uncertain, although it may be involved in the platelet adhesiveness, aggregation, or inflammation and it may be implicated in the genesis of hypertension. In contrast, there is some evidence that the increase of plasma uric acid is protective against the cardiovascular diseases, since uric acid acts as an endogenous antioxidant, and the higher plasma uric acid levels found in cardiovascular diseases patients suggest that any protective antioxidant effect of uric acid is hidden by other negative effects in these

Additionally, Torres et al. (2007), reported that hyperuricemia which implicated in the oxidative stress plays an important role in the pathophysiology of bipolar disorders. The idea that the purinergic system might be involved in bipolar disorder dates back to Kraepelin, who was the first to describe an association between manic symptoms, uric acid excretion, hyperuricemia, and gout. In fact, the purinergic system modulates sleep, motor activity, cognition, attention, behavior, and mood. Even in the absence of a psychiatric diagnosis, individuals with higher uric acid levels are more likely to show higher drive, disinhibition, hyperthymia, or irritable temperament (Lorenzi et al., 2010). Similarly, diseases characterized by purinergic turnover dysfunction and uric acid overproduction (e.g., Lesch–Nyhan syndrome) are associated with impulsive/aggressive behavior,

Among clinical and therapeutic characteristics, we found that there was no significant change in uric acid values in relation to illness episodes and the treatment. This finding is not in agreement with the previous studies that reported that plasma uric acid levels were

new component of the syndrome.

hyper secretion of hepatic enzymes (Fromenty et al., 2004).

pathogeneses (Haj mouhamed et al., 2010).

disinhibition, and increased sexual drive (Salvadore et al., 2010).

higher only during the manic phase of bipolar disorder but not during the depressive or euthymic phases (De Berardis et al., 2008). Additionally, lithium was found to low uric acid plasma levels and to have uricosuric effects in mania. Carbamazepine and phenytoin similarly decreased uric acid levels; in contrast, valproate appeared to have the opposite effect. However, it is important to note that the effect of these drugs on uric acid levels in relationship to clinical improvement in patients with bipolar disorder has not been systematically evaluated (Salvadore et al., 2010).

Compared with controls, patients had significantly higher levels of homocysteine and significantly lower levels of folatemia. Additionally, significant associations were showed between bipolar I disorder and hyperhomocysteinemia (39.2% *Vs* 18%, OR = 1.95, IC 95% = 1.04-3.69; p = 0.038) and hypofolatemia (66.2% *Vs* 36.2%, OR = 3.69, IC 95% = 2.20-6.19; p < 0.001). Homocysteine is an intermediary metabolite of the essential amino acid methionine. Folate and vitamin B12 are required for remethylation of homocysteine to methionine (Hankey & Eikelboom, 1999).

According to Reynolds (2006), hyperhomocysteinaemia has long been identified as a risk factor for vascular disease and the lowering of homocysteine concentrations by the treatment with folic acid, or possibly vitamin B12 and vitamin B6 which might reduce the risk of both cardiovascular and cerebrovascular diseases. Moreover, the association between increased circulating homocysteine concentrations and premature vascular thrombotic events in individuals with hereditary homocystinuria is well established. This process may include platelet activation, smooth muscle cell proliferation, and enhanced leukocyte binding to the endothelium. In recent years, a relationship between milder degrees of hyperhomocysteinaemia and vascular disease has emerged, and this has been the subject of intense research. Hyperhomocysteinemia can be caused by a wide range of disorders, the most important of which are genetic defects of the enzymes involved in homocysteine metabolism and/or deficiencies of their co-factors: folate (former vitamin B9), vitamin B12 and vitamin B6 (Haj mouhamed et al., 2011).

Our study showed a significant association between bipolar I disorder and hyperhomocysteinemia. The exact mechanisms underlying the hyperhomocysteinemia in this disease are not completely understood and controversed among studies. Several hypotheses have been postulated including nutritional folate and vitamin B deficiency, and/or reduced glomerular filtration rate in bipolar patients (Vuksan-Ćusa et al., 2011). En effect, we found a significant association between this disease and hypofolatemia. Furthermore, some authors (Atmaca et al., 2005) showed that at a high concentration, homocysteine is considered to be a neurotoxic substance, causing activation of NMDA (Nmethyl D-aspartate) receptors and leading to excitotoxicity. By impairing the neural plasticity and promoting neuronal degeneration, homocysteine could contribute to the pathogenesis of neurodegenerative and psychiatric disorders (Ipcioglu et al., 2008). Additionally, homocysteine is a methyl donor when activated to S-adenosylmethionine. So aberrant DNA methylation due to hyperhomocysteinemia also may be involved in the pathogenesis of bipolar disorder as well as schizophrenia (Mill et al., 2008).

In the other hand, folate appears to influence the synthesis rate of tetrahydrobiopterin, a cofactor in the hydroxylation of phenylalanine and tryptophan, rate-limiting steps in the biosynthesis of dopamine, norepinephrine, and serotonin, neurotransmitters postulated to play a role in the monoamine hypothesis of affective disorders. In addition, methyl tetrahydrofolate has been shown to bind to presynaptic glutamate receptors, where it may

Cardiovascular Risk in Tunisian Patients with Bipolar I Disorder 339

In bipolar I patients, the prevalences of obesity and overweight were respectively 33.1 % and 30.7 %. These findings were similar to those reported by Elmslie et al (2000) and Fagiolini et al (2002) (36 % and 32 %). However, higher values were reported by McElroy et al (2004) (44

For this population, we found that the prevalence of obesity greatly exceeded that found in controls (12.3%) and in the general population (20%) (Haddad et al., 2006). Obesity in patients with bipolar I disorder thus constitutes a major public health problem and suggests that the development and testing of specific interventions that target the obesity epidemic in this particular population are urgently needed. Bipolar disorder and obesity both have tremendous impact on the physical and mental well-being of affected individuals. Therefore, both illnesses should be treated with a coordinated intensive and multifaceted

Moreover, the risk of obesity in these patients is approximately multiplied by nine (33.1% *Vs* 8%, OR = 8.69, IC 95% = 3.61-20.87; p < 0.001). This could be one of the missing factors in understanding the relationship between psychiatric disorders and increased cardiovascular risk. In fact, some studies have reported that psychiatric disorders, particularly bipolar disorder, are significantly associated with adverse cardiovascular events and coronary heart disease (Garcia-Portilla et al., 2009). The mechanisms through which obesity leads to coronary heart disease remain hotly debated, but the accumulation, particularly, of visceral fat is widely favoured as the primary mechanism, leading, through the release of fatty acids and other mediators, to insulin resistance, dyslipidaemia, and a pro-inflammatory state. However, obesity in general, and central obesity in particular (ie. excessive visceral intraabdominal fat) have been under-recognised as risk factors for coronary heart disease in the population, where most attention has been placed on smoking and cholesterol (Pinkney,

According to Raji et al (2009), the cardiovascular afflictions including obesity, diabetes, hypertension and stroke increase the risk for cognitive decline and dementia, but it is unknown whether these factors, specifically obesity and type 2 diabetes mellitus, are associated with specific patterns of brain atrophy. Obesity and type 2 diabetes mellitus may amplify the risk for dementia by worsening cerebral atrophy even in cognitively intact individuals, raising their vulnerability to future Alzheimer's disease neuropathology. The same authors, mostly in subjects younger than 65 years, suggest also that increased body tissue fat content (adiposity) is correlated with atrophy in the temporal cortex, frontal lobes, putamen, caudate, precuneus, thalamus, and white matter. It is unknown, but of great interest, whether high tissue fat content, as measured by BMI, is associated with differences

Additionally, some studies showed that obesity has psychosocial consequences, including discrimination and stigmatization, which may contribute to the severity of bipolar disorder by negatively impacting patients' general physical health and functioning, quality of life, self-esteem, and psychological well-being. Obese patients have an increased risk of sleep apnea, which causes sleep disruptions and may lead to mood destabilization in patients with bipolar disorder. Obesity may also impact effectiveness of pharmacotherapies by altering the distribution and elimination of medications. Truncal obesity, which is most common, increases the risk of type 2 diabetes mellitus, dyslipidemia, hypertension, stroke, ischemic heart disease, and early death (Cheymol, 2000; Fagiolini et al., 2003; Plante &

in brain structure in cognitively normal elderly (Raji et al., 2009).

% and 20 %).

2001).

Winkelman, 2008).

treatment (Fagiolini et al., 2003).

potentially modulate the release of other neurotransmitters, including the monoamines (Atmaca et al., 2005).

Moreover, some studies showed that lower folatemia in patients with psychiatric disorders can be due to their nutritional status (Reif et al., 2005). Indeed, poor appetite as a symptom of bipolar disorder could lead to decreased intake of B vitamins which could then lead to elevated homocysteine concentrations (Tolmunen et al., 2004).

We found a significant association between vitamin B12 values and illness episode. Manic patients had lower values of this parameter than depressive patients. This can be explained by the eating habits of bipolar patients. Indeed, Parikh et al. (2000) found that manic episode is often associated with weight loss.

About therapeutic characteristics, we showed that only vitamin B12 was significantly associated with the medication use. Indeed, patients taking carbamazepine had significantly lower values of this parameter than those taking valproic acid and lithium. These findings are not in agreement with others studies. In fact, Derkes and Westphal (2005) showed that carbamazepine can cause elevated homocysteine concentrations. Although, according to Ozbek et al (2008), homocysteine, folate and vitamin B12 were not related to drug usage. Additionally, Osher et al (2008) reported that there were no significant differences in homocysteine levels between patients receiving versus not receiving lithium, neuroleptic or valproate. However, Sener et al. (2006) suggested that carbamazepine, as enzyme inducer, can directly modulate the activity of different liver enzymes. Liver enzyme induction may cause depletion of the cofactor involved, folic acid, pyridoxal 5'-phosphate and vitamin B12, leading to the alterations in homocysteine status.

Our study showed that bipolar I patients are so much more likely to be smokers than controls (52.3% *Vs* 39.4%, OR = 1.68, IC 95% = 1.06-2.66; p = 0.025). An association between smoking and bipolar I disorder has been established and prevalence rates for lifetime and current smoking have been shown to be as high as 82.5% and 68.8% respectively (Lasser et al., 2000). The possible explanations for the high rates of smoking include an increased genetic vulnerability, a greater susceptibility to addiction because of a greater subjective experience of reward or pleasure, or that tobacco helps relieve some of the symptoms related to a behavioural disorder. For example, cigarette smoking may be an attempt to selfmedicate symptoms of depression, anxiety, boredom or loneliness. Other possible explanations for continuing to smoke include increased withdrawal symptoms and reduced side effects from psychiatric medication (Williams & Ziedonis, 2004). Additionally, it has been reported that nicotine stimulates the brain to release dopamine, which is associated with pleasurable feelings, and smokers quickly develop regular smoking patterns. Eventually, smokers need increasing levels of nicotine to feel 'normal'. In the other hand, cigarette smoking is known to contribute to many diseases, including cancer, chronic obstructive pulmonary disease, stroke, cardiovascular diseases, and peptic ulcers. Investigators have attempted to elucidate the mechanisms of the pathogenesis associated with cigarette smoking, but the conclusions were not consistent. A basic hypothesis is that free radicals cause oxidative damage to macromolecules such as lipids, proteins, and DNA. Therefore, these radicals play an important role in the pathogenesis of these diseases (Haj mouhamed et al., 2010).

In this study, the prevalence of obesity is higher in patients with bipolar I disorder than in controls. Moreover, the risk of obesity in these patients is approximately multiplied by nine (33.1% *Vs* 8%, OR = 8.69, IC 95% = 3.61-20.87; p < 0.001).

potentially modulate the release of other neurotransmitters, including the monoamines

Moreover, some studies showed that lower folatemia in patients with psychiatric disorders can be due to their nutritional status (Reif et al., 2005). Indeed, poor appetite as a symptom of bipolar disorder could lead to decreased intake of B vitamins which could then lead to

We found a significant association between vitamin B12 values and illness episode. Manic patients had lower values of this parameter than depressive patients. This can be explained by the eating habits of bipolar patients. Indeed, Parikh et al. (2000) found that manic episode

About therapeutic characteristics, we showed that only vitamin B12 was significantly associated with the medication use. Indeed, patients taking carbamazepine had significantly lower values of this parameter than those taking valproic acid and lithium. These findings are not in agreement with others studies. In fact, Derkes and Westphal (2005) showed that carbamazepine can cause elevated homocysteine concentrations. Although, according to Ozbek et al (2008), homocysteine, folate and vitamin B12 were not related to drug usage. Additionally, Osher et al (2008) reported that there were no significant differences in homocysteine levels between patients receiving versus not receiving lithium, neuroleptic or valproate. However, Sener et al. (2006) suggested that carbamazepine, as enzyme inducer, can directly modulate the activity of different liver enzymes. Liver enzyme induction may cause depletion of the cofactor involved, folic acid, pyridoxal 5'-phosphate and vitamin B12,

Our study showed that bipolar I patients are so much more likely to be smokers than controls (52.3% *Vs* 39.4%, OR = 1.68, IC 95% = 1.06-2.66; p = 0.025). An association between smoking and bipolar I disorder has been established and prevalence rates for lifetime and current smoking have been shown to be as high as 82.5% and 68.8% respectively (Lasser et al., 2000). The possible explanations for the high rates of smoking include an increased genetic vulnerability, a greater susceptibility to addiction because of a greater subjective experience of reward or pleasure, or that tobacco helps relieve some of the symptoms related to a behavioural disorder. For example, cigarette smoking may be an attempt to selfmedicate symptoms of depression, anxiety, boredom or loneliness. Other possible explanations for continuing to smoke include increased withdrawal symptoms and reduced side effects from psychiatric medication (Williams & Ziedonis, 2004). Additionally, it has been reported that nicotine stimulates the brain to release dopamine, which is associated with pleasurable feelings, and smokers quickly develop regular smoking patterns. Eventually, smokers need increasing levels of nicotine to feel 'normal'. In the other hand, cigarette smoking is known to contribute to many diseases, including cancer, chronic obstructive pulmonary disease, stroke, cardiovascular diseases, and peptic ulcers. Investigators have attempted to elucidate the mechanisms of the pathogenesis associated with cigarette smoking, but the conclusions were not consistent. A basic hypothesis is that free radicals cause oxidative damage to macromolecules such as lipids, proteins, and DNA. Therefore, these radicals play an important role in the pathogenesis of these diseases (Haj

In this study, the prevalence of obesity is higher in patients with bipolar I disorder than in controls. Moreover, the risk of obesity in these patients is approximately multiplied by nine

elevated homocysteine concentrations (Tolmunen et al., 2004).

(Atmaca et al., 2005).

is often associated with weight loss.

leading to the alterations in homocysteine status.

(33.1% *Vs* 8%, OR = 8.69, IC 95% = 3.61-20.87; p < 0.001).

mouhamed et al., 2010).

In bipolar I patients, the prevalences of obesity and overweight were respectively 33.1 % and 30.7 %. These findings were similar to those reported by Elmslie et al (2000) and Fagiolini et al (2002) (36 % and 32 %). However, higher values were reported by McElroy et al (2004) (44 % and 20 %).

For this population, we found that the prevalence of obesity greatly exceeded that found in controls (12.3%) and in the general population (20%) (Haddad et al., 2006). Obesity in patients with bipolar I disorder thus constitutes a major public health problem and suggests that the development and testing of specific interventions that target the obesity epidemic in this particular population are urgently needed. Bipolar disorder and obesity both have tremendous impact on the physical and mental well-being of affected individuals. Therefore, both illnesses should be treated with a coordinated intensive and multifaceted treatment (Fagiolini et al., 2003).

Moreover, the risk of obesity in these patients is approximately multiplied by nine (33.1% *Vs* 8%, OR = 8.69, IC 95% = 3.61-20.87; p < 0.001). This could be one of the missing factors in understanding the relationship between psychiatric disorders and increased cardiovascular risk. In fact, some studies have reported that psychiatric disorders, particularly bipolar disorder, are significantly associated with adverse cardiovascular events and coronary heart disease (Garcia-Portilla et al., 2009). The mechanisms through which obesity leads to coronary heart disease remain hotly debated, but the accumulation, particularly, of visceral fat is widely favoured as the primary mechanism, leading, through the release of fatty acids and other mediators, to insulin resistance, dyslipidaemia, and a pro-inflammatory state. However, obesity in general, and central obesity in particular (ie. excessive visceral intraabdominal fat) have been under-recognised as risk factors for coronary heart disease in the population, where most attention has been placed on smoking and cholesterol (Pinkney, 2001).

According to Raji et al (2009), the cardiovascular afflictions including obesity, diabetes, hypertension and stroke increase the risk for cognitive decline and dementia, but it is unknown whether these factors, specifically obesity and type 2 diabetes mellitus, are associated with specific patterns of brain atrophy. Obesity and type 2 diabetes mellitus may amplify the risk for dementia by worsening cerebral atrophy even in cognitively intact individuals, raising their vulnerability to future Alzheimer's disease neuropathology.

The same authors, mostly in subjects younger than 65 years, suggest also that increased body tissue fat content (adiposity) is correlated with atrophy in the temporal cortex, frontal lobes, putamen, caudate, precuneus, thalamus, and white matter. It is unknown, but of great interest, whether high tissue fat content, as measured by BMI, is associated with differences in brain structure in cognitively normal elderly (Raji et al., 2009).

Additionally, some studies showed that obesity has psychosocial consequences, including discrimination and stigmatization, which may contribute to the severity of bipolar disorder by negatively impacting patients' general physical health and functioning, quality of life, self-esteem, and psychological well-being. Obese patients have an increased risk of sleep apnea, which causes sleep disruptions and may lead to mood destabilization in patients with bipolar disorder. Obesity may also impact effectiveness of pharmacotherapies by altering the distribution and elimination of medications. Truncal obesity, which is most common, increases the risk of type 2 diabetes mellitus, dyslipidemia, hypertension, stroke, ischemic heart disease, and early death (Cheymol, 2000; Fagiolini et al., 2003; Plante & Winkelman, 2008).

Cardiovascular Risk in Tunisian Patients with Bipolar I Disorder 341

Some studies have shown that alcohol directly contributes to heart disease and stroke. Heavy drinking raises levels of triglycerides circulating in the bloodstream leading to diabetes and blocked or narrowed arteries that carry blood to the heart. If coronary arteries are clogged with fats, blood cannot flow freely, resulting in heart disease or stroke. Additionally, alcohol directly contributes to heart failure by damaging the heart muscle and arteries. Cardiomyopathy, or an enlargement of the heart muscle, results from long-term alcohol use. An enlarged heart no longer works efficiently and fails to provide enough oxygenated blood to other organs of the body. Furthermore, alcohol is associated with cardiac arrhythmia (irregular heartbeat), sudden cardiac death, stroke and atrial fibrillation

In our patients, hypertension was not associated with bipolar disorder (5.4% *Vs* 16%, OR = 0.43, IC 95% = 0.14-1.29; p = 0.136). De Heart et al. (2010) explained the decrease of hypertension frequency in individuals with a mental illness by changes in lifestyle of

Several methodological limitations should be considered when interpreting these findings. First, larger sample sizes of groups would be beneficial. Second, our work is a crosssectional study that does not permit to follow up biological parameters. Third the sample of bipolar patients may not be representative of more heterogeneous populations. Finally, the diagnosis of controls was made by psychiatrists but without formal use of structured

Our results demonstrate that Tunisian bipolar I patients are exposed to higher cardiovascular risk. In fact, they had perturbations in lipid profile: significantly higher values of triglycerides and Lp(a), and significantly lower values of ApoA1, significantly hyperhomocysteinemia and hyperuricemia (in men), significantly hypofolatemia and high prevalence of metabolic syndrome. Obesity, hyperLp(a), hypertriglyceridemia, hypofolatemia, hyperhomocysteinemia and cigarette smoking were the main cardiovascular risk factors associated with bipolar I disorder. Indeed, the risk of obesity was increased approximately for nine once, hyperLp(a), hypertriglyceridemia and hypofolatemia approximately for four once and the other factors approximately for tow once. The TG/HDL ratio and Lp(a) were found as the best predictive factors of cardiovascular risk in

terms of sensibility and specificity at threshold of 1.12 and 168 mg/L, respectively.

levels of HOMA-IR and uric acid than metabolic syndrome free.

and should facilitate access to appropriate medical care.

Our findings noted a significant association between vitamin B12 values and illness episode. Manic patients had lower values of this parameter than depressive patients. Moreover, we showed that vitamin B12 was significantly associated with the therapeutic characteristics. Indeed, patients taking carbamazepine had significantly lower values of this parameter than those taking valproic acid and lithium. Additionally, there was no significant change in homocysteine, folate, uric acid values and metabolic syndrome in relation to illness episode and the treatment, whereas the patients with metabolic syndrome had significant higher

Therefore, bipolar I patients require specific care, particularly for lipid profile, vitamin status and weight; the effectiveness of this care will be evaluated during follow-up period Clinicians should track the effects of treatment on physical and the biological parameters,

(Pearson, 1996).

**5. Conclusion** 

patients such as reducing salt intake.

instruments to exclude psychiatric disorders in controls.

In addition, obesity was more frequent in depressive patients than in those with manic episode (38.1% *Vs* 27.8%). Previous studies reported that patients who had depressive symptomatology were more likely to have excessive caloric and cholesterol intake, to smoke and to be inactive than non-depressed subjects. Another explanation might involve biological mechanisms: it is ascertained that hypothalamic–pituitary- adrenal (HPA) axis dysregulation and high cortisol blood levels lead to increased visceral fat. HPA axis dysregulation has been a common finding in both unipolar and bipolar disorders; recently, some studies reported that increased cortisol blood levels correlated to the amount of intraabdominal fat in major depression (Maina et al., 2008).

About therapeutic characteristics, we found that obesity and overweight were more frequent (72% and 52%; respectively) in patients taking valproic acid or lithium. These findings are in line with those reported by De Hert et al. (2011). Moreover, Casey et al. (2005) reported that lithium have been shown to stimulate appetite through different mechanisms. The "carbohydrate craving" that is thought to be one of the mechanisms of increased calorie intake in people taking lithium is well known. In addition, it is believed that valproate also stimulates weight gain through a variety of mechanisms, especially the development of insulin resistance and diabetes mellitus type 2. In this line, our study found that this type of diabetes is frequent in patients (16.2%). Additionally, the risk of diabetes is multiplied by 1.5 in patients (16.2% *Vs* 9.7%, OR = 1.60, IC 95% = 0.62-4.12; p = 0.325).

Previous studies suggested that patients with both bipolar disorder and comorbid diabetes have more lifetime psychiatric hospitalizations than patients with bipolar patients without diabetes. The association between these two disorders underscores the importance of screening for diabetes in patients with bipolar illness, particularly because early detection and initiation of treatment to control glycemia may prevent diabetes-related complications. Moreover, other studies have demonstrated cerebrovascular lesions involving small intraparenchymal cerebral vessels and focal infarctions in patients with diabetes. These lesions predominantly occur in areas providing blood supply to the base of the pons, thalamus, and basal ganglia. Diabetes has been implicated as a risk factor for subcortical white-matter lesions observed on magnetic resonance imaging (MRI) scans; similar MRI findings have been noted in patients with bipolar disorder. Cerebral microvascular disease may lead to greater frequency of manic episodes, another reason to minimize diabetesrelated complications in patients with comorbid bipolar disorder (Cassidy et al., 1999; Holman et al., 2008).

Alcoholic beverage was not significantly associated with this illness but we showed that it was more frequent in patients than controls (13.1% *Vs* 6.9%, OR = 2.04, IC 95% = 0.94-4.44; p = 0.067). It has been well documented that bipolar disorder and alcoholism commonly cooccur. In fact, the lifetime prevalence of alcohol abuse and drug abuse in people with bipolar disorder are known to be three to nine times more frequent that of the general population (Merikangas et al., 2007; Regier et al., 1990; ten Have et al., 2002).

Additionally, some studies showed that the feelings of depression and anxiety associated with bipolar can be a factor that leads to alcoholism. People with bipolar disorder may use alcohol or other drugs to self medicate these feelings, especially in instances where the person has not been diagnosed. However, alcohol makes the symptoms of bipolar disorder worse. Anyone who shows symptoms of bipolar disorder should seek the advice of medical professionals (Le Strat, 2010).

Some studies have shown that alcohol directly contributes to heart disease and stroke. Heavy drinking raises levels of triglycerides circulating in the bloodstream leading to diabetes and blocked or narrowed arteries that carry blood to the heart. If coronary arteries are clogged with fats, blood cannot flow freely, resulting in heart disease or stroke. Additionally, alcohol directly contributes to heart failure by damaging the heart muscle and arteries. Cardiomyopathy, or an enlargement of the heart muscle, results from long-term alcohol use. An enlarged heart no longer works efficiently and fails to provide enough oxygenated blood to other organs of the body. Furthermore, alcohol is associated with cardiac arrhythmia (irregular heartbeat), sudden cardiac death, stroke and atrial fibrillation (Pearson, 1996).

In our patients, hypertension was not associated with bipolar disorder (5.4% *Vs* 16%, OR = 0.43, IC 95% = 0.14-1.29; p = 0.136). De Heart et al. (2010) explained the decrease of hypertension frequency in individuals with a mental illness by changes in lifestyle of patients such as reducing salt intake.

Several methodological limitations should be considered when interpreting these findings. First, larger sample sizes of groups would be beneficial. Second, our work is a crosssectional study that does not permit to follow up biological parameters. Third the sample of bipolar patients may not be representative of more heterogeneous populations. Finally, the diagnosis of controls was made by psychiatrists but without formal use of structured instruments to exclude psychiatric disorders in controls.
