Treatment with paroxetine,but not amitriptyline,lowers levels of lipoprotein(a) in patients with major depression

Introduction

Lipoprotein(a) (Lp(a)) is a complex low-density lipoprotein (LDL)-like particle consisting of apolipoprotein B-100 and apolipoprotein(a). Several studies have shown that high serum Lp(a) levels constitute a major risk factor for atherosclerosis and the progression of heart disease (Bennet et al., 2008; Danesh et al., 2000) and ischaemic stroke (Erqou et al., 2009; Smolders et al., 2007), as it enters the arterial intima of humans and promotes thrombosis, inflammation and foam cell formation (Boffa et al., 2004; Nielsen et al., 1998). Major depression has been linked to heart disease (Kent and Shapiro, 2009).

Few studies have been carried out to evaluate Lp(a) concentrations in patients with major depressive disorder. Emanuele et al. (2006) compared the Lp(a) levels in four groups of psychiatric patients (schizophrenia, major depression, bipolar disorder, personality disorder) and healthy controls and found a significant difference in Lp(a) levels in schizophrenia, major depression and bipolar disorder compared with healthy controls (p = 0.021, p = 0.039 and p = 0.00027, respectively). These results were confirmed in another recent study with depressed patients (Hamidifard et al., 2009). In another study, however, no significant difference between Lp(a) levels in patients with major depression and healthy controls was observed (Sarandol et al., 2006).

In this paper, Lp(a) concentrations of 35 in-patients with major depressive disorder were assessed and compared with those of 33 healthy controls. The patients were treated for 4 weeks with either paroxetine or amitriptyline in a randomized, double-blind manner and their Lp(a) levels were re-assessed. This is the first study to investigate antidepressant treatment effects upon Lp(a).

Materials and methods

The effects of antidepressant treatment with paroxetine or amitriptyline on several lipid parameters have been examined in a previous study by our research group. In that study, all patients received randomized, double-blind antidepressant treatment with one of the aforementioned antidepressants after a 6-day wash-out period (days -6 to -1) (Kopf et al., 2004). All participants provided written informed consent and the study was approved by the local ethics committee. With the exception of lorazepam and zolpidem, no additional psychotropic medication was allowed throughout the study. Other relevant axis I psychiatric disorders led to exclusion from the study.

Remaining blood samples that had not been used in this earlier study were now used in order to investigate the concentrations of Lp(a); this explains the differing group sizes in the present study (amitriptyline: n = 14, final dosage = 150 mg, or paroxetine: n = 21, final dosage = 40 mg). Thus, we first assessed the baseline (day -1) Lp(a) concentrations in surplus samples of 35 in-patients with DSM-IV major depression and a Hamilton Depression Rating Scale (HAMD)-21 score ≥ 18 and compared them with those of 33 healthy controls. Then, we re-assessed Lp(a) levels in blood samples that had been collected after 4 weeks of treatment (day 28). Blood collection was always performed under fasting conditions. Remission to treatment was defined as a final HAMD score of ≤ 7. Lp(a) was measured using the Lp(a) 21 FS reagens (Diagnostic Systems International) in the Roche/Hitachi 912 photometric system. The reagens allows the detection of a minimum of 3 mg/dL Lp(a) and has an intra- and inter-assay variability coefficient of approximately 3%.

Student’s t-tests were used for between- and within-group comparisons. The treatment effects of medication and the change in HAMD scores (ΔHAMD) in the course of treatment were studied by means of analysis of co-variance with repeated measures (rm-ANCOVA). We used only two-sided tests and a p < 0.05 was accepted as level of significance. All findings are expressed as means ± SD.

Results

Discussion

We assessed the baseline Lp(a) concentrations of 35 in-patients with DSM-IV major depression who then received randomized, double-blind treatment with amitriptyline or paroxetine. A 4-week treatment with paroxetine, but not amitriptyline, was found to be associated with significantly lower levels of Lp(a). This is a finding that has not been described before.

Antidepressant treatment, along with remission of depression, has been repeatedly shown to improve the metabolic situation of patients. It has a beneficial effect on lipid regulation (Kopf et al., 2004) and improves glycaemic control in depressed diabetic patients (Lustman et al., 2000). An increase in insulin sensitivity has been described in non-diabetic depressive patients who remitted and responded to treatment (Weber-Hamann et al., 2008), with distinct effects of selective serotonin reuptake inhibitors (SSRIs) and tricyclic antidepressants (Weber-Hamann et al., 2006).

Current available data are consistent with the existence of a causal relationship between Lp(a) concentrations and vascular outcomes. There is an increased need for investigation of Lp(a) as a therapeutic target (Erqou et al., 2009), but a substantial modification of Lp(a) levels through lifestyle changes and without the involvement of pharmacological agents has been so far difficult to achieve (Kostner and Kostner, 2004). At the same time, no widely useable method for the pharmacologic lowering of Lp(a) is currently available, although niacin, certain inhibitors of the cholesteryl ester transfer protein (Insull et al., 2004) or even statins (Marcovina et al., 2003) have proven a certain efficacy, and large randomized trials are in progress (clinicaltrials.gov).

In the present paper, we have shown that Lp(a) changes due to antidepressant therapy with the SSRI paroxetine. Some of the findings have suggested that only very high levels of Lp(a) are associated with risk of cardiovascular ailment (Danesh et al., 2000, Kamstrup et al., 2008). This was not the case in our group of patients. There is also evidence of an interactional effect of Lp(a) and concomitantly increased LDL cholesterol levels on the risk of cardiovascular sequelae (Kamstrup, 2010). Thus, our results could support the assumption that paroxetine – in contrast to amitriptyline – might contribute to a decrease in the vascular mortality risk by reducing the Lp(a) levels of patients. Our results cannot be generalized to individuals without depression, and warrant replication in studies with larger patient numbers. Our aim was not to assess the cardiovascular risk of the participants; the effects of treatment with paroxetine or amitriptyline on lipid regulation has already been described in the study by Kopf et al. (2004).

This is the first study showing an effect of antidepressant treatment on Lp(a). The role of Lp(a) in the pathophysiology of depression, and the pathophysiological mechanisms explaining the decrease in Lp(a) levels under paroxetine treatment are unknown. Lp(a) has been described as a factor with high heritability. Up to 90% of the variation in plasma levels of Lp(a) is accounted by apo(a) polymorphisms, and 70% by the size of apo(a) isoforms (Keller, 2007). The genetic background of high Lp(a) levels in psychiatric patients, its influence upon metabolic regulatory mechanisms and interactions with different classes of antidepressants are not well understood, and warrant future research.