**3. Results**

Table 3 shows the comparisons of biological variables between bipolar I patients and controls.


\* Lipid profile parameters, folatemia, vitamin B12 and uric acid were adjusted for gender, BMI, cigarette smoking, alcoholic beverages, diabetes and hypertension

\*Hcys was adjusted for gender, BMI, cigarette smoking, alcoholic beverages, diabetes, hypertension, folatemia and vitamin B12

Table 3. Comparisons of biological variables between bipolar I patients and controls.

Compared with controls, patients had significantly higher triglycerides (1.95 ± 1.55 *Vs* 1.23 ± 0.81 mmol/L; p < 0.001), Lp(a) (243 ± 223 *Vs* 87 ± 129 mg/L; p < 0.001), homocysteine levels (15.8 ± 8.9 *Vs* 11.5 ± 5.0 µmol/L; p < 0.001) and uric acid (311 ± 99 *Vs* 250 ± 107 µmol/L; p = 0.001 in men; 246 ± 97 *Vs* 197 ± 73 µmol/L; p *=* 0.012 in women), and significantly lower ApoA1 (1.20 ± 0.23 *Vs* 1.40 ± 0.67 g/L; p < 0.001) and folate (3.3 ± 0.9 *Vs* 5.1 ± 2.8 µg/L; p < 0.001) levels. After adjustment for potential confounder factors, these differences remained significant for all of these parameters except for uric acid which is remained significantly higher only for men (table 3).

Cardiovascular Risk in Tunisian Patients with Bipolar I Disorder 327

We showed significant association between bipolar I disorder and some cardiovascular risk factors: obesity (33.1% *Vs* 8%, OR = 5.68, IC 95% = 2.94-10.96; p < 0.001), hyperLp(a) (47.7% *Vs* 14.8%, OR = 5.25, IC 95% = 3.04-9.07; p < 0.001), hypertriglyceridemia (53.1% *Vs* 17.7%, OR = 4.10, IC 95% = 2.44-6.90; p < 0.001), hypofolatemia (66.2% *Vs* 36.2%, OR = 3.44, IC 95% = 2.13-5.54; p < 0.001), hyperhomocysteinemia (39.2% *Vs* 18%, OR = 2.80, IC 95% = 1.66-4.72; p < 0.001) and cigarette smoking (52.3% *Vs* 39.4%, OR = 1.68, IC 95% = 1.06-2.66; p = 0,025). After adjustment for potential confounder factors, these associations remained

Alcoholic beverage, diabetes and hyperuricemia were not significantly associated with this illness but we showed that they were more frequents in patients than controls (13.1% *Vs* 6.9%, p = 0.067; 16.1% *Vs* 9.7%, p = 0.325; 10.8% *Vs* 4.4%, p = 0.439; respectively). Additionally, the risk of diabetes and hyperuricemia were respectively multiplied by 1.5 in patients (16.1% *Vs* 9.7%, OR = 1.60, IC 95% = 0.62-4.12; p = 0.325; 10.8% *Vs* 4.4%, OR = 1.58, IC 95% = 0.49-5.08; p = 0.439) and the risk of alcoholic beverage by two (13.1% *Vs* 6.9%, OR =

On the contrary, this disease was not associated with hypertension (5.4% *Vs* 16%, OR = 0.43, IC 95% = 0.14-1.29; p = 0.136) nor with hyperLDL (13.1% *Vs* 26.9%, OR = 0.48, IC 95% = 2.53-

Fig.1. illustrates the receiver Operating Characteristic (ROC) of three index of atherogenicity

Fig. 1. Receiver Operating Characteristic (ROC) of three index of atherogenicity as predictive

significant (table 4).

7.95; p < 0.001) (table 4).

TG: triglycerides; CT: cholesterol

factors of cardiovascular risk.

2.04, IC 95% = 0.94-4.44; p = 0.067) (table 4).

as predictive factors of cardiovascular risk.


Table 4 reports the association between bipolar I disorder and cigarette smoking, alcoholic beverages, obesity, diabetes, hypertension, lipid profile parameters, hyperhomocysteinemia, hypofolatemia, hypovitamin B12 and, hyperuricemia.

\* Lipid profile parameters, folatemia, vitamin B12 and uric acid were adjusted for gender, BMI, cigarette cigarette smoking, alcoholic beverages, diabetes and hypertension; \* Hcys was adjusted for gender, BMI, cigarette smoking, alcoholic beverages, diabetes, hypertension, folatemia and vitamin B12; **\***Diabetes was adjusted for gender, BMI, cigarette smoking, alcoholic beverages, hypertension and dyslipidemia; \*Obesity was adjusted for gender, cigarette smoking, alcoholic beverages, hypertension, diabetes and dyslipidemia; **\***Hypertension was adjusted for gender, cigarette smoking, alcoholic beverages, diabetes and dyslipidemia; \ c-HDL <1.1 mmol/L (in men) and < 0 .9 (in women); £ uric acid: 210-420 µmol/L (in men) and 150-360 µmol/L ( in women)

Table 4. Association between bipolar I disorder and cigarette smoking, alcoholic beverages, obesity, diabetes, hypertension, lipid profile parameters, hyperhomocysteinemia, hypofolatemia, hypovitamin B12, and hyperuricemia.

Table 4 reports the association between bipolar I disorder and cigarette smoking, alcoholic beverages, obesity, diabetes, hypertension, lipid profile parameters, hyperhomocysteinemia,

*mmol/L* 16.1% 9.7% 1.79 0.90-3.55 0.092 1.60 0.325

*(*≥ 3.4 mmol/L) 13.1% 26.9% 0.39 0.22-0.73 0.002 0.48 < 0.001 *¥HypoHDL* 59.2% 58.3% 1.00 0.63-1.59 0.975 0.78 0.359

*(*≥ 200 mg/L) 47.7% 14.8% 5.25 3.04-9.07 **< 0.001** 4.48 **< 0.001** 

(< 3.7 µg/L) 66.2% 36.2% 3.44 2.13-5.54 **< 0.001** 3.69 **< 0.001** *£Hyperuricemia* 10.8% 4.4% 2.05 0.71-5.91 0.176 1.58 0.439 \* Lipid profile parameters, folatemia, vitamin B12 and uric acid were adjusted for gender, BMI, cigarette cigarette smoking, alcoholic beverages, diabetes and hypertension; \* Hcys was adjusted for gender, BMI, cigarette smoking, alcoholic beverages, diabetes, hypertension, folatemia and vitamin B12; **\***Diabetes was adjusted for gender, BMI, cigarette smoking, alcoholic beverages, hypertension and dyslipidemia; \*Obesity was adjusted for gender, cigarette smoking, alcoholic beverages, hypertension, diabetes and dyslipidemia; **\***Hypertension was adjusted for gender, cigarette smoking, alcoholic beverages, diabetes and dyslipidemia; \ c-HDL <1.1 mmol/L (in men) and < 0 .9 (in women); £ uric

Table 4. Association between bipolar I disorder and cigarette smoking, alcoholic beverages,

obesity, diabetes, hypertension, lipid profile parameters, hyperhomocysteinemia,

**(n = 175) OR IC 95% p OR\* p\***

10.96 **< 0.001** 8.69 **< 0.001** 

52.3% 39.4% 1.68 1.06-2.66 **0.025** - -

13.1% 6.9% 2.04 0.94-4.44 0.067 - -

5.4% 16% 0.34 0.15-0.78 0.008 0.43 0.136

26.2% 26.9% 0.96 0.57-1.61 0.891 0.99 0.987

53.1% 17.7% 4.10 2.44-6.90 **< 0.001** 3.71 **< 0.001** 

39.2% 18% 2.80 1.66-4.72 **< 0.001** 1.95 **0.038** 

21.2% 14.9% 0.69 0.34-1.38 0.296 0.62 0.215

hypofolatemia, hypovitamin B12 and, hyperuricemia.

*30 kg/m²)* 33.1% 8% 5.68 2.94-

acid: 210-420 µmol/L (in men) and 150-360 µmol/L ( in women)

hypofolatemia, hypovitamin B12, and hyperuricemia.

**Controls**

**(n = 130)**

**Parameters Patients**

*Cigarette smoking*

*Alcoholic beverages* 

*Obesity (BMI ≥*

Diabetes *≥ 6.1* 

*Hypertension*  (≥ 130/85 mm

*Hypercholesterol emia (*≥ *5.17* mmol/L*)* 

*Hypertriglycerid emia (*≥ 1.7 mmol/L)

*HyperLDL* 

*HyperLp(a)* 

*inemia*  (≥15 µmol/L)

*B12* 

*Hyperhomocyste*

*Hypovitamin* 

(< 187 ng/L)

*Hypofolatemia* 

Hg)

We showed significant association between bipolar I disorder and some cardiovascular risk factors: obesity (33.1% *Vs* 8%, OR = 5.68, IC 95% = 2.94-10.96; p < 0.001), hyperLp(a) (47.7% *Vs* 14.8%, OR = 5.25, IC 95% = 3.04-9.07; p < 0.001), hypertriglyceridemia (53.1% *Vs* 17.7%, OR = 4.10, IC 95% = 2.44-6.90; p < 0.001), hypofolatemia (66.2% *Vs* 36.2%, OR = 3.44, IC 95% = 2.13-5.54; p < 0.001), hyperhomocysteinemia (39.2% *Vs* 18%, OR = 2.80, IC 95% = 1.66-4.72; p < 0.001) and cigarette smoking (52.3% *Vs* 39.4%, OR = 1.68, IC 95% = 1.06-2.66; p = 0,025). After adjustment for potential confounder factors, these associations remained significant (table 4).

Alcoholic beverage, diabetes and hyperuricemia were not significantly associated with this illness but we showed that they were more frequents in patients than controls (13.1% *Vs* 6.9%, p = 0.067; 16.1% *Vs* 9.7%, p = 0.325; 10.8% *Vs* 4.4%, p = 0.439; respectively). Additionally, the risk of diabetes and hyperuricemia were respectively multiplied by 1.5 in patients (16.1% *Vs* 9.7%, OR = 1.60, IC 95% = 0.62-4.12; p = 0.325; 10.8% *Vs* 4.4%, OR = 1.58, IC 95% = 0.49-5.08; p = 0.439) and the risk of alcoholic beverage by two (13.1% *Vs* 6.9%, OR = 2.04, IC 95% = 0.94-4.44; p = 0.067) (table 4).

On the contrary, this disease was not associated with hypertension (5.4% *Vs* 16%, OR = 0.43, IC 95% = 0.14-1.29; p = 0.136) nor with hyperLDL (13.1% *Vs* 26.9%, OR = 0.48, IC 95% = 2.53- 7.95; p < 0.001) (table 4).

Fig.1. illustrates the receiver Operating Characteristic (ROC) of three index of atherogenicity as predictive factors of cardiovascular risk.

TG: triglycerides; CT: cholesterol

Fig. 1. Receiver Operating Characteristic (ROC) of three index of atherogenicity as predictive factors of cardiovascular risk.

Cardiovascular Risk in Tunisian Patients with Bipolar I Disorder 329

Table 6 reports the specificity and sensibility of Lp(a) and homocysteine as predictive factors

**Parameters AUC (95% CI) Cut off Specificity Sensibility p Lp(a) (mg/L)** 0.80 [0.75-0.85] 168 0.75 0.74 **< 10-3 Homocysteine (µmol/L)** 0.52 [0.44-0.57] 13.4 0.60 0.62 **< 10-3**

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, 0.75; respectively) at

The prevalence of metabolic syndrome (modified NCEP-ATP III) and its profile in bipolar I

**Total of the association of 5 criteria 2 5.9** 

**Total of the association of 4 criteria 3 8.8** 

**Total of the association of 3 criteria 29 85.3 At least three or more criteria 34 100**  Table 7. Prevalence of metabolic syndrome (modified NCEP-ATP III) and its profile in

The prevalence of metabolic syndrome in bipolar I patients was 26.1% (N = 34). The highest prevalence of this syndrome was obtained by the association between obesity, low c-HDL

Table 8 reports the Prevalence of the components of metabolic syndrome in the total sample

Table 8. Prevalence of the components of metabolic syndrome in the total sample (N=130).

**Criteria N (%) c-HDL < 1.1 mmol/L (men) and < 0 .9 (women) 77(59.2) TG ≥ 1.7 mmol/L 69(53.1) BMI ≥ 28.5 kg/m² 44 (33.8) Fasting blood glucose ≥ 6.1 mmol/L 21 (16.1) Blood pressure ≥ 130/85 mm Hg 7(5.4)** 

 Diabetes, Obesity, Hypertriglyceridemia, Low c-HDL 1 Diabetes, Obesity, Low c-HDL, High blood pressure 1 Diabetes, Hypertriglyceridemia, Low c-HDL, High blood pressure 1

 Obesity, Hypertriglyceridemia, Low c-HDL 15 Diabetes, Low c-HDL, Hypertriglyceridemia 6 Obesity, Hypertriglyceridemia, Diabetes 4 Obesity, Hypertriglyceridemia, High blood pressure 2 Diabetes , High blood pressure, Obesity 1 Hypertriglyceridemia, Low c-HDL, High blood pressure 1

N %

Table 6. Specificity and sensibility of Lp(a) and homocysteine as predictive factors of

threshold of 1.12 and 168 mg/L, respectively (tables 5, 6; fig 1, 2).

of cardiovascular risk.

AUC; Area under the curve

patients are shown in table 7.

cardiovascular risk.

bipolar I patients.

of bipolar I patients.

and hypertrilyceridemia (44.1 %) (Table 7).


The specificity and sensibility of three index of atherogenicity as predictive factors of cardiovascular risk are shown in table 5.

TG: triglycerides; CT: cholesterol; AUC; Area under the curve

Table 5. Specificity and sensibility of three index of atherogenicity as predictive factors of cardiovascular risk

Fig. 2. Illustrates the Receiver Operating Characteristic (ROC) of Lp(a) and homocysteine as predictive factors of cardiovascular risk.

Fig. 2. Receiver Operating Characteristic (ROC) of Lp(a) and homocysteine as predictive factors of cardiovascular risk.

Table 6 reports the specificity and sensibility of Lp(a) and homocysteine as predictive factors of cardiovascular risk.


AUC; Area under the curve

328 Dyslipidemia - From Prevention to Treatment

The specificity and sensibility of three index of atherogenicity as predictive factors of

Table 5. Specificity and sensibility of three index of atherogenicity as predictive factors of

Fig. 2. Receiver Operating Characteristic (ROC) of Lp(a) and homocysteine as predictive

Fig. 2. Illustrates the Receiver Operating Characteristic (ROC) of Lp(a) and homocysteine as

**(95% CI) Cut off Specificity Sensibility p** 

[0.59-0.71] 1.12 0.63 0.62 **< 10-3**

[0.44-0.57] 3.93 0.53 0.57 0.661

[0.49-0.62] 0.66 0.54 0.55 0.070

cardiovascular risk are shown in table 5.

predictive factors of cardiovascular risk.

TG: triglycerides; CT: cholesterol; AUC; Area under the curve

**Parameters AUC** 

**TG/HDL** 0.65

**CT/HDL** 0.52

**ApoB/ApoA1** 0.56

factors of cardiovascular risk.

cardiovascular risk

Table 6. Specificity and sensibility of Lp(a) and homocysteine as predictive factors of cardiovascular risk.

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, 0.75; respectively) at threshold of 1.12 and 168 mg/L, respectively (tables 5, 6; fig 1, 2).

The prevalence of metabolic syndrome (modified NCEP-ATP III) and its profile in bipolar I patients are shown in table 7.


Table 7. Prevalence of metabolic syndrome (modified NCEP-ATP III) and its profile in bipolar I patients.

The prevalence of metabolic syndrome in bipolar I patients was 26.1% (N = 34). The highest prevalence of this syndrome was obtained by the association between obesity, low c-HDL and hypertrilyceridemia (44.1 %) (Table 7).

Table 8 reports the Prevalence of the components of metabolic syndrome in the total sample of bipolar I patients.


Table 8. Prevalence of the components of metabolic syndrome in the total sample (N=130).

Cardiovascular Risk in Tunisian Patients with Bipolar I Disorder 331

The variations of lipid profile parameters according to the illness episode and therapeutic

0.37

0.28

0.41

0.32

0.40

0.36

0.33

0.36

Our study failed to show any significant association between lipid profile parameters, illness episode and treatment, while euthymic patients were found to have the highest levels of Lp(a) and depressive patients had the highest levels of ApoB/ApoA1 ratio (table 10). Additionally, we showed that women taking lithium had the lowest c-HDL values and

Table 10. Variations of lipid profile parameters according to the illness episode and

patients taking carbamazepine had the highest values of Lp(a) (table 10).

**c-HDL (mmol/L)**

**Men Women**

1.17 ± 0.36

0.99 ± 0.36

1.26 ± 0.33

1.10 ± 0.11

1.13 ± 0.24

0.86 ± 0.49

1.44 ± 0.40

0.28 0.79 2.10 ±

1.26 ± 0.40

**c-LDL (mmol/L)**

> 2.14 ± 1.10

> 2.14 ±

2.18 ± 1.04

2.08 ± 1.25

2.11 ± 1.12

2.17 ±

1.75 ± 0.74

2.30 ± 1.23

**Lp (a) (mg/L)** 

243 ±

1.09 158 ± 91 0.82 ± 0.41

268 ±

240 ±

240 ±

0.82 170 ± 93 0.78 ± 0.29

1.22 153 ± 69 0.76 ± 0.23

271 ±

293 ±

**ApoB/Apo A1** 

<sup>223</sup>0.71 ± 0.23

<sup>242</sup>0.68 ± 0.19

<sup>228</sup>0.70 ± 0.26

<sup>232</sup>0.70 ± 0.24

<sup>221</sup>0.66 ± 0.22

<sup>248</sup>0.71 ± 0.30

characteristics of bipolar I patients are shown in table 10.

(n = 130) 1.95 ± 1.55 4.42 ± 0.99 1.04 ±

(n = 21) 1.95 ± 1.32 4.49 ± 0.96 1.12 ±

(n = 73) 1.94 ± 1,80 4.53± 0.96 1.05 ±

(n = 36) 1.97 ± 1.25 4.14 ± 1.0 0.98 ±

(n = 12) 1.43 ± 0.95 4.40 ± 0.85 1.10 ±

**(**n =6) 2.33 ± 1.07 3.93 ± 0.40 1.10 ±

therapeutic characteristics of bipolar I patients.

1.78 ± 1.20 4.25 ±1.00 1.05 ±

2.24 ± 3.25 4.60 ± 1.10 0.97 ±

2.26 ± 1.64 4.74 ± 1.00 1.03 ±

**Cholesterol (mmol/L)** 

**Triglycerides (mmol/L)** 

*Patients* 

*Illness episode*

Depressive

Euthymic

Manic

*Treatment*

Valproic acid (n = 64)

Lithium

Carbamaze pine (n = 10)

AVP and Li

Antipsychotics (n = 38)

AVP: valproic acid; Li: lithium

The prevalence of individual diagnostic components, in the total sample, was as follows: 59.2% for low c-HDL, 53.1% for hypertriglyceridemia, 33.8% for obesity (BMI ≥ 28.5 kg/m²), 16.1% for high fasting glucose and 5.4% for hypertension (Table 8).


Table 9 reports the characteristics of patients with or without metabolic syndrome.

\* n = 23 (11 diabetic patients were excluded);

\*\* n = 86 (10 diabetic patients were excluded

**<sup>a</sup>**Statistical analysis was detected using Fisher's exact test

Table 9. Characteristics of patients with or without metabolic syndrome.

We found that gender was not associated with metabolic syndrome, 24.7% in men and 28.9% in women. As to age, we found that patients with metabolic syndrome were older than metabolic syndrome free patients (40.4 ± 8.8 years *V*s 37.0 ± 12.7 years), but this difference was not significant (table 9).

Our present data showed that there is no difference in metabolic syndrome prevalence between patients receiving antipsychotic and mood stabilizers treatment. However, we noted that patients treated with lithium had the highest prevalence of metabolic syndrome (table 9).

Our study failed to show any significant association between metabolic syndrome and illness episode, whereas, manic patients had the highest prevalence of this disorder (30.5%) (table 9).

Patients with metabolic syndrome had significantly higher levels of uric acid (p < 0.001) than metabolic syndrome free patients (table 9).

Concerning HOMA-IR analysis, after diabetic patients exclusion (n = 21), we noted that patients with metabolic syndrome had significantly higher levels of HOMA-IR (p < 0.001) than metabolic syndrome free patients (table 9).

The prevalence of individual diagnostic components, in the total sample, was as follows: 59.2% for low c-HDL, 53.1% for hypertriglyceridemia, 33.8% for obesity (BMI ≥ 28.5 kg/m²),

**Without MS** 

**Mean ± SD Mean ± SD** p

**[N= 96 (73.9%)] p values** 

0.9574**<sup>a</sup>**

Table 9 reports the characteristics of patients with or without metabolic syndrome.

Men 21 (24.7) 64 (75.3) 0.60 Women 13 (28.9) 32 (71.1)

Depressive 4 (19) 17 (81) 0.651**<sup>a</sup>**

*Age (years)* 40.4 ± 8.8 37.0 ± 12.7 0.11 *HOMA-IR* 6.0 ± 4.3\* 2.4 ± 1.7\*\* **< 0.001**  *Uric acid (µmol/L)* 335 ±117 272 ± 92 **< 0.001** 

We found that gender was not associated with metabolic syndrome, 24.7% in men and 28.9% in women. As to age, we found that patients with metabolic syndrome were older than metabolic syndrome free patients (40.4 ± 8.8 years *V*s 37.0 ± 12.7 years), but this

Our present data showed that there is no difference in metabolic syndrome prevalence between patients receiving antipsychotic and mood stabilizers treatment. However, we noted that patients treated with lithium had the highest prevalence of metabolic syndrome

Our study failed to show any significant association between metabolic syndrome and illness episode, whereas, manic patients had the highest prevalence of this disorder (30.5%) (table 9). Patients with metabolic syndrome had significantly higher levels of uric acid (p < 0.001)

Concerning HOMA-IR analysis, after diabetic patients exclusion (n = 21), we noted that patients with metabolic syndrome had significantly higher levels of HOMA-IR (p < 0.001)

Table 9. Characteristics of patients with or without metabolic syndrome.

16.1% for high fasting glucose and 5.4% for hypertension (Table 8).

**Variables N (% ) N (%)** 

Euthymic 19 (26) 54 (74) Manic 11 (30.5) 25 (69.5)

Antipsychotics 10 (26.3) 28 (73.7)

Mood stabilizers 24 (26.1) 68 (74.9) Lithium 4 (33.3) 8 (66.7) Valproic acid 17 (26.6) 47 (73.4) Carbamazepine 2 (20) 8 (80) Lithium and valproic acid 1 (16.7) 5 (83.3)

\* n = 23 (11 diabetic patients were excluded); \*\* n = 86 (10 diabetic patients were excluded

difference was not significant (table 9).

than metabolic syndrome free patients (table 9).

than metabolic syndrome free patients (table 9).

**<sup>a</sup>**Statistical analysis was detected using Fisher's exact test

*Gender*

*Illness episode* 

*Treatment*

(table 9).

**With MS [N= 34 (26.1%)]**  The variations of lipid profile parameters according to the illness episode and therapeutic characteristics of bipolar I patients are shown in table 10.


AVP: valproic acid; Li: lithium

Table 10. Variations of lipid profile parameters according to the illness episode and therapeutic characteristics of bipolar I patients.

Our study failed to show any significant association between lipid profile parameters, illness episode and treatment, while euthymic patients were found to have the highest levels of Lp(a) and depressive patients had the highest levels of ApoB/ApoA1 ratio (table 10). Additionally, we showed that women taking lithium had the lowest c-HDL values and patients taking carbamazepine had the highest values of Lp(a) (table 10).

Cardiovascular Risk in Tunisian Patients with Bipolar I Disorder 333

Our study failed to show any significant association between the BMI and to the illness episode and, therapeutic characteristics. However, we found that obesity was more frequent in depressive patients than in those with manic episode (38.1% *Vs* 27.8%). In addition, obesity and overweight were more frequent (72% and 52%; respectively) in patients taking

**Characteristics n = 47 (%)** 

(n = 21) 8 (38.1) 4 (19.1) 9 (42.8)

(n = 73) 25 (34.2) 23 (31.5) 25 (34.3)

(n = 36) 10 (27.8) 13 (36.1) 13 (36.1)

(n = 64) 25 (39.1) 17 (26.6) 22 (34.3)

(n = 12) 6 (50) 4 (33.3) 2 (16.7)

(n = 10) 2 (20) 5 (50) 3 (30)

(n = 6) 2 (33.3) 3 (50) 1 (16.7)

(n = 38) 8 (21.1) 11 (28.9) 19 (50)

Table 12. Distribution of BMI according to the illness episode and therapeutic characteristics.

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

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).

**25 ≤ BMI < 30 n = 40 (%)** 

**BMI < 25** 

valproic acid or lithium (table 12).

*Illness episode* Depressive

Euthymic

*Treatment* Valproic acid

Lithium

Carbamazepine

Antipsychotics

**4. Discussion** 

AVP: valproic acid; Li: lithium

95% = 2.13-6.46; p < 0.001).

AVP and Li

Manic

**BMI (kg/m²) BMI ≥ 30** 

**n = 43 (%)** 

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 patients.


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

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

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 (table 11).

The distribution of BMI according to the illness episode and therapeutic characteristics is shown in table 12.

Our study failed to show any significant association between the BMI and to the illness episode and, therapeutic characteristics. However, we found that obesity was more frequent in depressive patients than in those with manic episode (38.1% *Vs* 27.8%). In addition, obesity and overweight were more frequent (72% and 52%; respectively) in patients taking valproic acid or lithium (table 12).


AVP: valproic acid; Li: lithium

Table 12. Distribution of BMI according to the illness episode and therapeutic characteristics.
