INTRODUCTION

Organic solvent abuse is a worldwide problem, particularly among adolescents (Flanagan and Ives, 1994; Neumark et al, 1998). The easy accessibility to organic solvents makes them among the first drugs tried by teenagers. Consequently, teenagers are liable to develop an addiction to organic solvents, and such use can serve as a ‘gateway’ to the ingestion of other illicit drugs (Ives, 2000). Toluene is one of the major components of many organic solvents and is highly lipid-soluble (Morton, 1987). Toluene rapidly permeates lipid-rich regions of the body such as the brain, and its recreational intake can cause chronic and persistent psychiatric symptoms, including hallucinations and alterations in personality, even after cessation of toluene use (Hormes et al, 1986; Morton, 1987). However, the mechanisms responsible for these residual psychiatric symptoms following chronic toluene use remain unknown.

In experimental animal studies, exposure to toluene has been shown to produce changes in lipid membrane fluidity and membrane function (Kyrklund et al, 1987; LeBel and Schatz, 1990; von Euler et al, 1990). Furthermore, long-term exposure to toluene induces a persistent dopaminergic dysfunction in the rat basal ganglia, and this toluene-induced dopaminergic dysfunction has been related to the behavioral changes and the persistent cognitive deficits in rats (Cintra et al, 1999; Hillefors-Berglund et al, 1995; von Euler et al, 1993,2000).

In human MR studies of toluene users, T2-weighted magnetic resonance images revealed global atrophy in the cerebrum, cerebellum, and brainstem (Rosenberg et al, 1988). Loss of differentiation between gray and white matter, increased periventricular white matter signal intensity (Aydin et al, 2002; Filley et al, 1990; Rosenberg et al, 1988,2002), and hypointensity in the thalamus and basal ganglia (Rosenberg et al, 2002; Unger et al, 1994) have also been found in toluene users. Furthermore, white matter changes have been reported to be associated with neurocognitive deficits in long-term toluene users (Filley et al, 1990). However, little is known about the effects of long-term toluene use on the human brain at the neurochemical level.

Proton magnetic resonance spectroscopy (1H MRS) offers an opportunity to evaluate in vivo neurochemical changes that might occur in toluene users. In this study, we investigated the potential of 1H MRS scanning for revealing neurochemical changes in the chemistry of the basal ganglia of abstinent toluene users; we also considered whether or not such changes, if found, could possibly be related to the clinical characteristics observed in abstinent toluene users.

METHODS

Subjects

A total of 12 toluene users and 13 healthy comparison subjects who had never used toluene were recruited for our comprehensive research project on substance-related psychiatric disorders. Candidates were excluded if they were seropositive for human immunodeficiency virus type-1 (HIV-1), if they had a history of head trauma with loss of consciousness, or if they had brain morphological abnormalities (eg atrophy or tumor, etc) on MR images. All participants showed no neurological abnormalities and had no history of stupor or coma, which may be observed during intoxication after the use of high doses of toluene. All subjects were screened using the Structured Clinical Interview for DSM-IV (SCID) (First et al, 1996). More specifically, detailed information on toluene use and the history of psychiatric symptoms was retrospectively obtained from the toluene users; SCID-based interviews with the users and their family members were conducted, and patient medical records were considered as well. None of the toluene users had used any other drugs (ie they were mono-drug users) and the subjects had had no history of DSM-IV mental disorders before the use of toluene. The comparison subjects had no experience of the use of any illicit drugs, and had no history of mental disorders, including substance-related disorders, according to the DSM-IV. Both groups reported similar occasional smoking and drinking habits, and none of the subjects matched the DSM-IV criteria for either nicotine- or the alcohol-related disorders. The toluene users were recruited from the Hamamatsu University Hospital, and from three associated hospitals (Hattori Mental Hospital, Shimada Municipal Hospital, and Hamamatsu Neurological Hospital).

The duration of toluene use was defined as the time that had intervened between the initial and the last use of toluene. In cases involving intervals of abstinence exceeding 1 month during the duration of toluene use as defined, these intervals were subtracted from the total length of toluene use. The duration of abstinence from toluene use was defined as the time between the day of the last use of toluene and the day of the 1H MRS examination. Cumulative lifetime exposure to toluene was estimated in milliliters. Among the 12 toluene users, five had been treated with neuroleptics, while the remaining seven toluene users had not previously taken neuroleptics prior to the 1H MRS examination. Current and total lifetime neuroleptic doses in the five toluene users were quantified in chlorpromazine-equivalent dosages (American Psychiatric Association, 1997). In order to establish recent abstinence from toluene use, we regularly tested the subjects for urinary hippuric acid, a biomarker of toluene use (Raikhlin-Eisenkraft et al, 2001), using high-performance liquid chromatography, which was carried out according to the standard diagnostic methods. In addition, to verify abstinence from other illicit drugs, we used a rapid immunoassay for the qualitative detection of the metabolites of the following eight classes of drugs: amphetamines (including (±)3,4-methylenedioxymethamphetamine and methamphetamine), barbiturates, benzodiazepines, cocaine, methadone, opiates, tetrahydrocannabinol, and tricyclic antidepressants (Triage 8, Biosite Diagnostics, San Diego, USA). These assessments were also performed on the same day as the 1H MRS examination. Psychiatric symptoms in the toluene users were evaluated using the Brief Psychiatric Rating Scale (BPRS) (Overall and Gorham, 1962) on the day of the 1H MRS study. The assessment was conducted by a trained research psychiatrist blind to the 1H MRS data. After providing the subjects with a complete description of the study, written informed consent was obtained from all participants before entry into the study, in compliance with the procedures of the Ethics Committee of the Hamamatsu University School of Medicine.

Image Acquisition

MRS was performed at the Hamamatsu University Hospital, using a 1.5-T whole-body MR scanner (Signa Horizon; General Electric Medical Systems, Milwaukee, WI, USA) equipped with self-shielded gradient coils. A quadrature, MR bird-cage coil (trans/receive) was used throughout the study. The localized MR spectra were recorded using the point-resolved spatially localized spectroscopy (PRESS) sequence from the bilateral basal ganglia. The RF transmitter frequency was centered at the water resonance frequency (63.88 MHz). A 90° flip angle was determined by varying the RF amplifier gain until a maximal signal consistent with a good slice profile was obtained. The B0 shim on the 27 ml cubic excited volume was optimized using linear shims to a full-width at half-maximum (FWHM) of <6.4 Hz. Voxel size was 8 cm3 (2 × 2 × 2 cm3). To establish the voxel of interest (VOI) on an MR image, scout views in the proton density images were obtained vertically in the Z-axis (Figure 1). Proton density images were acquired because gray matter contrasts well with white matter in this type of imaging study (Sekine et al, 2002).

Figure 1
figure 1

Axial slice proton density MRI showing the typical location of the MRS voxel. The right side of the figure represents the left side of the brain.

Full size image

After the location of the VOI had been verified by localized MR imaging, the magnetic field in the VOI was optimized by means of a water signal to establish a linewidth of <3 Hz. The acquisition parameters were as follows: repetition time (TR), 1500 ms; echo time (TE), 40 ms; and acquisition time, 3 min 42 s. Peaks of N-acetylaspartate (NAA) (2.0 ppm), creatine plus phosphocreatine (Cr+PCr) (3.0 ppm), choline-containing compounds (Cho) (3.2 ppm), and myo-inositol (MI) (3.6 ppm) were obtained. Data processing were performed using a PROBE/SV-Quant tool (General Electric Medical Systems, Milwaukee, WI, USA), which employs a standard nonlinear least-squares algorithm for fitting the observed data. The metabolite ratios were calculated from the fitted peak integrals in the resulting spectra using a method of automatic metabolite peak measurement (Kreis et al, 1991; Webb et al, 1994). Representative spectra for the right basal ganglia region of a comparison subject and a toluene user are shown in Figure 2. The ratios of NAA/Cr+PCr, MI/Cr+PCr, Cho/Cr+PCr, and Cho/NAA in the toluene users were compared with those of the comparison subjects.

Figure 2
figure 2

Proton MRS spectra for the right basal ganglia region of (a) a comparison subject and (b) a toluene user. MI, myo-inositol; Cho, choline-containing compounds; Cr+PCr, creatine plus phosphocreatine; NAA, N-acetylaspartate.

Full size image

It should be noted that we were unable to completely exclude potential cerebrospinal fluid (CSF) contamination. All measurements were determined by reference to a standard neuroanatomic atlas (Daniels et al, 1987), and great care was taken to minimize CSF contamination. However, measurement of tissue water signal from the voxel was too variable owing to a tissue/ventricle artifact to be used to calculate absolute concentrations of metabolites. Accordingly, only the metabolite ratios are reported here.

Statistical Analysis

Student's t-test was used for the comparison of the means of the metabolite ratios between toluene users and comparison subjects. The relationships between the metabolite ratios and clinical parameters (ie the duration of toluene use, the duration of abstinence, the lifetime amount of toluene use, the current neuroleptic dose, the lifetime neuroleptic dose, and the BPRS score) were evaluated using Pearson's correlation coefficient. We measured the metabolite ratios in the basal ganglia on both the left and right sides of each subject. The values measured in each of the hemispheres were not naturally independent within the same individual, and separate analyses of the ratios, as measured on each side, would have increased the risk of Type I error due to multiple testing. To avoid such errors, we used repeated measures analysis of variance (ANOVA), in which the repeated measures were the left and right metabolite ratios. In all of the statistical analyses, an alpha level of 0.05 was considered to be significant.

RESULTS

The demographic and clinical characteristics of the toluene users and the comparison subjects are shown in Tables 1 and 2. No significant difference was observed as regards the distribution of age or sex between the toluene users and the comparison subjects. The mean age of the first inhalation in toluene users was 15.8 years (range: 13–23). While one toluene user had a first-degree relative with alcoholism, none of the other users had any first- or second-degree relatives with any psychiatric disorders. The 12 toluene users, with or without neuroleptic treatment, showed similar clinical symptoms, such as auditory hallucinations, persecutory delusions, and emotional apathy. However, they exhibited no noticeable diminished volition, incongruity of affect, stereotypy, mannerism, or loosening of association, and their interest in social activities and interpersonal relationship were fairly maintained. Among the 12 toluene users, five had been treated with neuroleptics, including chlorpromazine, levomepromazine, haloperidol, risperidone, and quetiapine. The neuroleptics successfully improved psychotic symptoms characterized by delusions or hallucinations, but were less effective for emotional apathy.

Table 1 Demographic Characteristics of Toluene Users and Comparison Subjects
Full size table
Table 2 Clinical Characteristics of Toluene Users
Full size table

The metabolite ratios for the left and right basal ganglia are shown in Table 3. Repeated measures ANOVA revealed no significant interaction of group and laterality (left vs right) in the metabolite ratios measured, which suggested that any differences observed between the ratios of the two groups were similar in the left and right basal ganglia. There was no significant difference in the NAA/Cr+PCr ratio between the two groups (F=0.001, df=1,23, p=0.98). When the toluene users were divided into two groups, that is, medicated (n=5) and nonmedicated (n=7), no significant difference between either of these groups and the comparison subjects was found (F=0.06, df=1,16, p=0.81 for the medicated group; F=0.06, df=1,18, p=0.81 for the nonmedicated group). There was also no significant difference in the MI/Cr+PCr ratio between the toluene users and the comparison subjects (F=0.09, df=1,23, p=0.77). The difference between the medicated patients and comparison subjects was not significant, nor was that between the nonmedicated patients and comparison subjects (F=0.001, df=1,16, p=0.98 for the medicated group; F=0.57, df=1,18, p=0.11 for the nonmedicated group). However, the toluene users had significantly higher ratios of Cho/Cr+PCr than did the comparison subjects (F=9.10, df=1,23, p=0.006). A similar significance was also noted in the five medicated (F=10.19, df=1,16, p=0.006) and seven nonmedicated (F=4.50, df=1,18, p=0.05) toluene users as compared with the comparison subjects. The ratio of Cho/NAA was significantly higher in the toluene users compared with that of the comparison subjects (F=6.15, df=1,23, p=0.02). The difference in the ratio between nonmedicated toluene users and comparison subjects was of marginal significance (F=4.01, df=1,18, p=0.06), but the difference between medicated toluene users and the comparison subjects was found to be significant (F=6.24, df=1,16, p=0.02).

Table 3 Metabolite Ratios in the Left and Right Basal Ganglia of Toluene Users (n=12) and Comparison Subjects (n=13)
Full size table

There was a significant positive correlation between the Cho/Cr+PCr ratio and the BPRS score on the left side of the basal ganglia (r=0.57, df=11, p=0.05) (Figure 3), whereas such a correlation was not found on the right side (r=0.18, df=11, p=0.58). None of the other clinical variables, that is, the duration of toluene use (left: r=0.09, df=11, p=0.79; right: r=0.05, df=11, p=0.87), the duration of abstinence (left: r=0.25, df=11, p=0.44; right: r=0.18, df=11, p=0.57), the total lifetime exposure to toluene (left: r=0.42, df=11, p=0.18; right: r=0.24, df=11, p=0.46), current neuroleptic dose (left: r=0.05, df=11, p=0.87; right: r=0.09, df=11, p=0.78), or lifetime neuroleptic dose (left: r=0.006, df=11, p=0.99; right: r=0.26, df=11, p=0.42), were correlated with the Cho/Cr+PCr ratio on either side of the basal ganglia. Due to the possibility that the psychiatric symptoms were ameliorated by medication at the time of assessment, we computed the partial correlation coefficient with adjustment for the current neuroleptic dose. As anticipated, a slightly higher correlation emerged between the Cho/Cr+PCr ratio in the left basal ganglia and the BPRS score (r=0.65, df=9, p=0.03), whereas the correlation between Cho/Cr+PCr ratio in the right basal ganglia and the total BPRS score remained nonsignificant (r=0.27, df=9, p=0.42).

Figure 3
figure 3

Correlations between clinical variables and values of the Cho/Cr+PCr ratio in the (a) left and (b) right basal ganglia of toluene users (n=12). A significant correlation was found only in the left basal ganglia using Pearson's correlation coefficient.

Full size image

DISCUSSION

The present study revealed that the ratio of Cho/Cr+PCr in the bilateral basal ganglia of toluene users was significantly higher than the corresponding value in the comparison subjects. Although the possible change in the Cr+PCr level, as reference, might have modified or obscured actual alterations in the primary metabolites, the Cho/NAA ratio, as well as the Cho/Cr+PCr ratio, was significantly higher in the basal ganglia of the toluene users compared with that of the comparison subjects. Therefore, it follows that changes in the Cho/Cr+PCr ratio may arise from an elevation of Cho in the basal ganglia of toluene users. To our knowledge, this is the first examination of the effects of toluene on the MRS measures of brain chemicals. The Cho/Cr+PCr ratio may be a nonspecific marker of toluene-induced biochemical changes in the basal ganglia in long-term toluene users. However, since five of the 12 toluene users examined received neuroleptic medication, the effect of medication on the metabolite ratio should be considered. Indeed, several 1H MRS studies of medicated patients with schizophrenia have shown an increased Cho/Cr+PCr ratio in the caudate nucleus (Ando et al, 2002; Bustillo et al, 2001), although other studies have not (Heimberg et al, 1998; Ohara et al, 2000). Consequently, at this point, the effect of neuroleptic medication on the metabolite values of brain Cho/Cr+PCr remains controversial. In this study, a significant increase in the Cho/Cr+PCr ratio was observed, in the seven toluene users who had not been medicated. However, the level of significance demonstrated in the seven nonmedicated users was of borderline significance (p=0.05), whereas the result was highly significant in the five medicated toluene users (p=0.006). Therefore, the use of neuroleptic medication may have exerted some effect on the elevation of the Cho/Cr+PCr ratio in the participants of this study.

The peak of choline-containing compounds on 1H MRS reflects the concentrations of free choline, phosphocholine, and glycerophosphorylcholine (Miller et al, 1996; Stanley et al, 2000). The compounds responsible for the choline peak are associated with membrane phospholipids, and have been reported to be increased in a variety of human pathologies that are characterized by increased membrane phospholipid turnover such as demyelinating processes and gliosis; examples include multiple sclerosis (Tartaglia et al, 2002), HIV-related brain disease (Chang et al, 1999), dementia (Pfefferbaum et al, 1999), and epilepsy (Lundbom et al, 2001). In animal studies, Gerasimov et al (2002) have recently demonstrated using PET that rapid uptake and clearance of radioactivity of [11C]toluene into specific regions of the brain take place, especially in the striatum, and this effect is thought to be the high lipid solubility of this agent. In addition, both acute and long-term exposure of rats to toluene have been shown to alter membrane composition and membrane fluidity (Edelfors and Ravn-Jonsen, 1989; Kyrklund et al, 1987; Lebel and Schatz, 1990; von Euler et al, 1990). Two post-mortem studies of the brains of toluene users have reported diffuse demyelination in the subcortical white matter, in addition to the gliosis of white and/or gray matter (Escobar and Aruffo, 1980; Rosenberg et al, 1988). In human MRI studies, T2-weighted images of toluene abusers showed hypointensity in the basal ganglia and thalamus (Aydin et al, 2002; Filley et al, 1990; Rosenberg et al, 1988; Unger et al, 1994), and this type of hypointensity has been suggested to be associated with changes in myelin metabolism (Aydin et al, 2002; Rosenberg et al, 1988). In addition, demyelinating lesions are thought to contribute to the increase in the Cho/Cr ratio in a patient with multiple sclerosis (Arnold et al, 1992). Taken together, the elevated Cho/Cr+PCr ratio in the basal ganglia observed in abstinent toluene users in this study may have been reflective of a long-lasting abnormality in membrane phospholipid metabolism, in particular, demyelination.

The elevation of Cho/Cr+PCr ratios in the left, but not in the right, basal ganglia was significantly correlated with the severity of psychiatric symptoms as measured by the BPRS. The reason for this lateralized dysfunction of the brain in toluene users is not known. However, it is generally accepted that left lateralized dysfunction or structural changes can be observed in neuropsychiatric disorders such as schizophrenia (Hietala et al, 1995; Petty, 1999; Reynolds, 1983). In this study, an elevation in the Cho/Cr+PCr ratio was noted among those who had stopped using toluene for 4.9 years on average. Therefore, the abnormalities in membrane phospholipid metabolism observed in these toluene users may be persistent and may participate in the pathogenesis of protracted psychiatric symptoms. Although no prior studies have examined the relationship between psychiatric symptoms and neurochemical alterations in toluene users, there is some evidence linking membrane abnormalities with psychiatric symptoms in patients with schizophrenia (Fenton et al, 2000; Fukuzako et al, 1996; Ross et al, 1997).

In this study, no significant difference was found in NAA/Cr+PCr ratios between the toluene users and the comparison subjects. This finding was not altered when the administration of neuroleptic medication was taken into account. In addition, since NAA has been widely used as a marker of neuronal density (Miller, 1991; Simmons et al, 1991; Tsai and Coyle, 1995), our results suggest that toluene abusers may be spared from neuronal loss in the basal ganglia. Indeed, this view is in line with the finding that there was no difference in the MI/Cr+PCr ratio between toluene users and comparison subjects in this study. Since MI is considered to be a marker of glial response (Brand et al, 1993), a finding of no change in the MI/Cr+PCr ratio may represent the absence of neuronal damage in toluene users. Histological studies of rats chronically exposed to toluene have shown no significant neuronal loss in the neocortex or cerebellum (Korbo et al, 1993; Ladefoged et al, 1991), although one study has shown a reduction in the density of neurons in the hippocampus of toluene-exposed rats (Korbo et al, 1996). Furthermore, an enzyme immunoassay study using γ-enolase and calbindin-D28k as neural cell markers showed no quantitative changes in the cerebrum of rats after exposure to toluene (Huang et al, 1992). In post-mortem case reports about two toluene users who abused toluene for 12 and 20 years, respectively, one patient was reported to have neuronal loss in the cerebrum, basal ganglia, and cerebellum (Escobar and Aruffo, 1980), while in the other case, no neuronal loss was detected in the cerebrum or cerebellum (Rosenberg et al, 1988). Generally, neural cells appear less likely to be compromised by exposure to toluene, although very high doses of toluene can damage neural cells (Hansson et al, 1988). However, vulnerability to toluene exposure may vary in different regions of the brain. Therefore, the result showing no change in the NAA/Cr+PCr ratio in our study does not exclude the possibility of neuronal damage occurring in brain regions other than the basal ganglia.

A worldwide comprehensive survey has shown that the majority of toluene users start inhaling toluene when they are teenagers (Kozel et al, 1995; Neumark et al, 1998). In accordance with that survey, all but one of our toluene users (with the exception starting toluene inhalation at the age of 23 years) began to take toluene while still a teenager, that is, under 18 years of age. Furthermore, the mean age of initiation of toluene use in this study (about 16 years) was far earlier than the mean age of onset of schizophrenia, which has been shown to be in the early 20s (Bromet et al, 1995). Although we cannot rule out the possibility that the use of toluene precipitates schizophrenia in vulnerable individuals, as pointed earlier, none of the toluene users had a family history of major psychosis. However, the toluene users in the present study had psychotic symptoms akin to those related to the paranoid type of schizophrenia, namely, auditory hallucinations and persecutory delusions. Nonetheless, it may be of note that none of our toluene users showed avolition, alogia, inappropriate affect, stereotypy, mannerism, or social withdrawal, which are frequently seen in patients with schizophrenia. In general, these findings suggest that the toluene-related psychiatric condition may differ from that of schizophrenia.

In summary, we demonstrated an elevation of the Cho/Cr+PCr ratio in the basal ganglia of recreational toluene users, and this finding provides evidence of an abnormality in membrane metabolism. In addition, this membrane disturbance was associated with the emergence and persistence of psychiatric symptoms in toluene users. However, our results should be considered as preliminary due to the small sample size of the present study.