Increases in triglyceride levels are associated with clinical response to clozapine treatment

Increases in triglyceride levels are associated with clinical response to clozapine treatment

Clozapine, the most efficacious antipsychotic medication for treatment-resistant schizophrenia currently available (Boston et al., 1996a; Dursun et al., 1999; Pande et al., 2002), is associated with substantial weight gain (Rummel-Kluge et al., 2010), dyslipidaemia (Meyer and Koro, 2004) and metabolic syndrome (De Hert et al., 2008). An association between treatment resistance to clozapine in schizophrenia and lower cholesterol levels was suggested in a cross-sectional study from the 1990s (Boston et al., 1996b), the results of which were supported by a subsequent prospective study that demonstrated that the lower levels of cholesterol in patients with treatment-resistant schizophrenia couldn’t be attributed to clozapine use. The authors postulated that lower cholesterol levels was an intrinsic metabolic dysfunction in such patients (Dursun et al., 1999). Interestingly, more recent studies have reported a possible association between weight gain (Law et al., 1994; Meltzer et al., 2003) and an association between increased triglyceride (TG) levels (Pande et al., 2002; Procyshyn et al., 2009) and symptom improvement with clozapine treatment in schizophrenia. A randomised, double-blind trial of clozapine augmentation (Procyshyn et al., 2007) found that an increase in serum TG levels was associated with a reduction in the total Positive and Negative Syndrome Scale (PANSS) score, independent of changes in weight, in patients on long-term clozapine treatment. In the present study we sought to determine if alteration in lipid levels is predictive of clinical response in a six-month observational study of patients switched to clozapine.

Methods

This study included 49 patients meeting the Diagnostic and Statistical Manual of Mental Disorders (fourth edition, text revision) (DSM-IV R) criteria for schizophrenia who were recruited over a two-year period for a prospective study of structural magnetic resonance imaging (MRI) changes with clozapine treatment (Ahmed et al., 2012). All patients had failed at least two antipsychotic trials of adequate duration and dosage. The clinical assessment included severe combined immunodeficiency (SCID), PANSS and Global Assessment of Functioning (GAF), which were conducted at baseline before commencing clozapine treatment and at six-month follow-up. Anthropomorphic measurements including weight and waist circumference, and metabolic parameters including fasting total serum cholesterol, serum TG, low-density lipoprotein (LDL) cholesterol and high-density lipoprotein (HDL) cholesterol were collected at baseline and follow-up. Serum clozapine levels were measured at follow-up for each individual patient. This was an observational open-label study, with participants as well as investigators aware of the treatment intervention. However, the PANSS rater at follow-up/endpoint was blind as to the serum lipids and weight change of the individual patients.

Informed consent was obtained from each of the patients and ethical approval was granted by the clinical research ethics committee at University Hospital Galway, Ireland.

Results

Patients had a mean age of 37.4 (SD=9.3, range 22–57) and 33 (67%) were male. The mean age of illness onset was 23.1 years (SD=5.8), the mean illness duration was 14.4 years (SD=7.7, range 3–31 years) and the mean number of psychotic episodes was 5.6 (SD=3.89). The mean TG level at baseline was 1.75 mmol/L (SD=0.91), the mean fasting serum cholesterol was 4.89 mmol/L (SD=0.98) and the mean weight was 85.55 kg (SD=17.97). Prior to switching to clozapine, all patients were taking at least one antipsychotic medication, with 95.9% (n=47) taking second-generation antipsychotic medications. Forty-nine per cent (n=24) of patients at baseline were treated with olanzapine alone and 49% (n=24) were treated with two antipsychotic medications. There was no significant difference in baseline serum TG levels between those treated with olanzapine (mean=1.66 mmol/L, SD=1.0) and those treated with other antipsychotic medications (mean=1.84, SD=0.82) (t=−0.701, df=47, p=0.49). Further, there was no significant difference in baseline serum TG levels between those patients treated with two antipsychotic medications at baseline (mean=1.76mml/L, SD=0.98) and those who were treated with antipsychotic monotherapy (mean=1.75 mmol/L, SD=0.84) (t=0.029, df=47, p=0.979). The mean follow-up period was 32.6 weeks (SD=6.6, range 23–62 weeks). At follow-up the mean fasting serum TG level was 2.35 mmol/L (SD=1.20), the mean fasting cholesterol was 5.48 mmol/L (SD=0.83), the mean weight was 90.06 kg (SD=17.26) and the mean serum clozapine level was 500 µg /L (SD=0.28, range 70–1360 µg/L). The mean PANSS improvement was 24.3 (SD=16.3) and the GAF improvement was 20.8 (SD=15.5).

We conducted bivariate analysis to assess for correlations between change in PANSS score and metabolic indices. The reduction in total PANSS score was strongly negatively correlated with increase in TG levels (r=−0.57, p<0.001), weight gain (r=−0.32, p=0.03) and increase in waist circumference (r=−0.35, p=0.01). No correlations were found between change in serum TG levels and serum clozapine levels (r=−0.04, p=0.77). We then performed multiple linear regression analysis and examined the association between change in serum TG and cholesterol levels with clinical improvement as measured by total PANSS scores, controlling for potential confounding factors of weight gain, change in waist circumference and serum clozapine levels.

This model accounts for 40% of the variance in the PANSS score (adjusted R²=0.40, F = 7.34, df=4 p<0.001). An increase in TG levels was a significant predictor of clinical improvement (B=9.33, t=3.56, df=4, p<0.001) and also of improvement in positive PANSS scores (B=2.85, t=3.61, df=4, p=0.001) as well as negative PANSS scores (B=1.93, t=2.36, df=4, p=0.02). None of the other variables in the regression equation reached statistical significance in their prediction of PANSS score change. The relationship between change in TG levels and change in PANSS scores is displayed in Figure 1.

OPEN IN VIEWER

Figure 1.

Correlation between PANSS change and serum triglyceride change.

Discussion

In this study we demonstrate a strong correlation between clinical improvement with clozapine treatment and an increase in serum TG levels, which is not accounted for by potential confounds. Treatment-emergent weight gain was not found to be an explanatory variable for clinical improvement, in contrast to other studies which have identified such an association (Dursun and Reveley, 1993; Meltzer et al., 2003).

There is no clear basis for a unitary mechanism of action for the differential metabolic and therapeutic effects of clozapine. However, it is known that clozapine is the most efficacious antipsychotic and is associated with the most substantial metabolic disturbance (De Hert et al., 2011) and hypertriglyceridaemia (Meyer and Koro, 2004) of the currently licensed antipsychotic agents. The resultant dyslipidaemia may affect the pharmacokinetics, distribution and pharmacological activity of lipophilic drugs (Procyshyn et al., 2007). One potential mechanism for clozapine-associated clinical improvement with hypertriglyceridaemia is that clozapine is redistributed from the lipoprotein deficit fraction of serum to the very low-density lipoprotein (VLDL) fraction as TG levels increase, thus altering its pharmacokinetics and allowing for a more sustained clozapine release (Procyshyn et al., 2001). It has been further hypothesised that with this partitioning of clozapine from the lipoprotein component of serum into VLDL, there is an increase in the active serum concentration of clozapine (Procyshyn et al., 2007) and that this mechanism of redistribution may facilitate the penetration of clozapine across the blood-brain barrier (Procyshyn et al., 2009). Another possible explanation for the association is that elevated TG levels are a marker of clozapine’s central effects on lipid biosynthesis with associated amelioration of disturbed myelination in schizophrenia (Le Hellard et al., 2010). This mechanistic action of clozapine is further supported by a recent animal study which demonstrated that mouse brain networks regulated by clozapine seem to be more involved in neuronal survival/repair, synaptic transmission and control of carbohydrate and fat metabolism than those regulated by haloperidol (Rizig et al., 2012). The impact of clozapine on lipid metabolism has also been postulated to lead to an increase in serotonin neural activity (Dursun et al., 1999). Reduced plasma and platelet serotonin levels have been shown to be associated with a poor clinical response to second-generation antipsychotics (SGAs) with potent antagonism at 5-HT receptors (van der Heijden et al., 2004), though this hasn’t been investigated in clozapine-treated patients (Stone et al., 2010). The increase in TG, seen with clozapine treatment, may enhance central serotonin neural activity and contribute to the serotonergic activity of clozapine in improving the clinical response of prior non-responders to other SGAs.

Limitations of this study include the open-label study design and the small sample size; further replication in a larger sample would help to confirm the study findings. This study adds to a small pool of literature which has suggested a link between increased TG levels and changes in clozapine’s pharmacological action and heightened therapeutic efficacy (Procyshyn et al., 2007). Further large-scale studies are required to fully elucidate the association and potential mechanistic relationship between clinical response to clozapine and elevated lipid levels, and to generate novel therapeutic approaches which may be fostered by a more thorough understanding of such an association.