Reduced olfactory performance in patients with major depression

Abstract

The aim of the present study is to investigate olfactory sensitivity and odor evaluations in a homogenous sample of unipolar depressive patients using pure olfactory odors. Twenty-four in-patients with major depressive disorder (MDD) were investigated during their acute depressive phase. Eighteen of them participated a second time after successful treatment. A group of healthy subjects, matched by age, sex, and smoking behavior, served as a control. Olfactory sensitivity, as measured by threshold tests, was strongly reduced in patients with severe depression. Additional correlative analyses revealed that the lowered sensitivity could partly be predicted by high depression scores. After successful medical treatment, these sensitivity differences were reduced and did not reach the significance level. The subjective odor evaluations (valence and intensity ratings) were not markedly changed in general. The results reveal that olfactory performance in MDD patients is reduced at an early perceptional level of stimulus processing. It is discussed whether this effect can be attributed to the close functional connection between the main olfactory bulb and the amygdala.

Introduction

It is generally assumed that the perception and processing of emotional stimuli is altered in patients with major depression (MDD). They show a greater anticipation of negative experiences and a reduced anticipation of positive experiences (MacLeod and Byrne, 1996); they show an enlarged reaction time for negative words in the Emotional Stroop paradigm (Segal and Gemar, 1997) and they have been found to have a facilitated ability to recall emotionally negative information (Mathews and MacLeod, 1994). While most laboratory studies used either language or visual stimuli as emotional cues, two studies reported a deviant emotional perception of odors in MDD patients (Pause et al., 2000a, Steiner et al., 1993). In fact, the presentation of odors as emotional stimuli is supported by the findings of a differential startle reflex modulation by pleasant and unpleasant odors (Miltner et al., 1994, Ehrlichman e al., 1997). Steiner and colleagues found that the facial reactions of depressive patients to pleasant and unpleasant odors were significantly shorter than those of controls, but no group difference was observed for valence ratings. Pause et al. compared the ratings to odors and emotional pictures in MDD patients and healthy controls. They found that patients responded with a higher arousal to negative stimuli, independent of the stimulus modality.

Considering the close neuroanatomical connection between emotional and odor processing it is proposed that, except for these modality non-specific effects, odors might have additional unique characteristics for the investigation of limbic functions: During the last 10 years, imaging techniques have allowed the brain areas belonging to the primary and secondary olfactory cortex in humans to be described: an early positron emission tomography (PET) study outlined the importance of the piriform, orbitofrontal and insular cortex in human odor perception (Zatorre et al., 1992). In particular, later studies were able to confirm the activation of the orbitofrontal cortex (Small et al., 1997, Yousem et al., 1997, Zald & Pardo, 1997, Sobel et al., 1998) and the insula during smelling (Kettenmann et al., 1997, Fulbright et al., 1998), whereas the piriform cortex seems to be activated during sniffing rather than during smelling (Sobel et al., 1998). Focusing on the emotional value of olfactory stimulation, Small et al. (1997) and Zald and Pardo (1997) found the amygdala to be activated (as shown by PET) when aversive or novel odors were presented. According to Doty et al. (1997) the piriform cortex and the amygdala are structures of the primary olfactory cortex, while the insula and the orbitofrontal cortex belong to the secondary olfactory cortices.

In studying emotional processing in humans, PET studies have been carried out with healthy volunteers in different emotional states. While the orbitofrontal cortex seems to generally participate in the processing of emotional information (Pardo et al., 1993, Lane et al., 1997b, Rolls, 1998), the insula has been found to be activated during negative emotions, like sadness or disgust (Lane et al., 1997a). The amygdala seems to play a key role in the encoding, learning and modulation of emotions (Armony and LeDoux, 2000): whereas the amygdala is activated during the encoding of pleasant and unpleasant emotions (Hamann et al., 1999), it has sometimes been found to only be involved in the processing of negative emotions (Lane et al., 1997b), especially sadness (Schneider et al., 1995). An abnormal activation of the amygdala and the orbitofrontal region could be confirmed in depressive patients (Drevets et al., 1992, Lesser et al., 1994, Brody et al., 1999).

Considering these neuroanatomical findings, it is hypothesized that a deviant limbic functionality in depressive patients could be associated with an altered chemosensory perception on a sensory level. In their methodological considerations on the measurement of olfactory dysfunction in neuropsychiatric disorders, Martzke et al. (1997) recently pointed out the important differentiation of primary (sensory level) and secondary stimulus processing (higher order level). They suggested using threshold tests to obtain information about the former processing level, while identification and discrimination tasks should be implemented to measure the later level. Moreover, they remembered to use pure olfactory odorants, which stimulate only the first cranial (olfactory) nerve without activating the fifth cranial (trigeminal) nerve.

So far, odor perception in patients with major depression (MDD) has been investigated by only a few laboratories. Measuring the ability to identify odors, the most comprehensive study (Amsterdam et al., 1987) found depressive patients (n=51) to be similar to controls. However, two smaller studies revealed conflicting results: while Warner et al. (1990) could confirm the results from Amsterdam et al. in 6 MDD patients, Serby et al. (1990) reported lower odor identification scores in depressed patients. Two recent studies (Solomon et al., 1998, McCaffrey et al., 2000) compared MDD patients with patients with Alzheimer‘s disease: they could confirm the lowered olfactory identification ability, often reported in Alzheimer’s disease, by using the depressive patients as the reference group. Olfactory sensitivity in MDD patients was measured in two studies by means of a threshold test: Serby et al., 1990, Serby et al., 1992 found nine MDD patients to have slightly (P<0.10) elevated olfactory thresholds to geraniol (a predominately olfactory odorant according to Doty et al., 1978). Testing the patients a second time, after recovery from their depressive symptoms, revealed no difference in olfactory performance due to medication or health status. In the second study (Gross-Isseroff et al., 1994), olfactory sensitivity to isoamyl acetate (a predominately trigeminal odorant) and androstenone was measured in nine MDD out-patients. In this study, the patients showed an increased sensitivity to isoamyl acetate after successful treatment. However, correlations between olfactory sensitivity and the depression score were not significant. It should be noted that the Hamilton Depression Scale (HDS) scores of these patients were less than half of the HDS scores of the patients of the aforementioned study by Serby and colleagues.

Altogether, theoretical considerations, as well as empirical findings, contribute to the hypothesis that olfactory sensitivity might be reduced in MDD patients with relatively severe symptoms (Serby et al., 1990, Serby et al., 1992). While the emotional reactivity to odors seems to be altered in MDD patients (Pause et al., 2000a, Steiner et al., 1993), it is proposed that higher order cognitive odor evaluation is not altered in depressive patients (see valence ratings in Steiner et al., 1993; identification scores in Amsterdam et al., 1987). The present study aimed to examine olfactory sensitivity in patients with MDD (according to the DSM IV criteria). The odors chosen for the present study (eugenol; phenyl-ethylalcohol, PEA) have been proven to activate only the olfactory and not the trigeminal nerve (Doty et al., 1978). Besides olfactory thresholds, intensity and valence ratings for a set of 10 different odorants were obtained. These odorants varied in pleasantness, as well as in their olfactory properties. While thresholds were used as a measure of sensory performance, the ratings were used as indicators of odor perception on a cognitive evaluative level. Subjects were tested at the beginning of a psychiatric therapy and a second time after successful treatment.