Brain glucose metabolism and temperament in relation to severe somatization

Abstract

Abstract  Little is known about the pathophysiology of somatization. The authors’ aim was to explore associated factors with somatoform disorders. The authors studied 10 female patients with a diagnosis of somatization disorder or undifferentiated somatoform disorder with no comorbid current Diagnostic and Statistical Manual of Mental Disorders (4th edn) Axis I disorder and 12 healthy female volunteers. The predicting variables were temperament factors of the 240-item Temperament and Character Inventory instrument and regional brain glucose metabolism. Low novelty-seeking and high harm avoidance temperament traits and low caudate and low putamen glucose metabolism were statistically significantly associated with severe somatization (P < 0.05). In the present study, severe somatization associates with both altered brain glucose metabolism and temperament factors. No other studies on association of somatization with brain glucose metabolism and temperament have been published. The results are still considered exploratory due to the small number of subjects.

INTRODUCTION

Little is known about the pathophysiology of somatization. Somatization has been considered to be a ‘psychogenic’ state, due to the lack of objective findings. In the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) somatoform disorders are defined by the presence of physical symptoms that suggest a general medical condition but that are not fully explained by a general medical condition, by the direct effects of a substance, or by another mental disorder. The symptoms must cause clinically significant distress or impairment in social, occupational, or other areas of functioning. In contrast to factitious disorders and malingering, the physical symptoms are not under voluntary control.¹ Grabe et al. found somatization to be associated with difficulties in emotional processing and high harm-avoidance temperament trait among psychiatric patients.² High harm-avoidance has been reported to be positively related to severity of somatization by using Tridimensional Personality Questioinnaire (TPQ), the predecessor of the Temperament and Character Inventory (TCI).³ In cognitive performance tests somatizing patients performed at a lower level in semantic memory, verbal episodic memory and visuospatial tasks compared to healthy controls. The patients also were slower in attentional tasks.⁴ Gordon et al. reported an evoked response potential finding which was interpreted that somatization disorder is associated with a disturbed capacity to filter afferent stimuli.⁵ Rief et al. found heart rate deceleration and re-acceleration to be more pronounced in healthy controls compared to somatizing patients in attentional tasks and they suggested that the results reflected a habituation disorder of the patients.⁶ High Clara cell protein (CC16) scores, an endogenous anticytokine, were found in patients with somatization syndrome as well as lower serum interleukin (IL)-6 values. These immunological differences between groups differentiated somatizing patients from depressed patients and healthy volunteers.⁷ Reduced level of blood tryptophan was reported in patients with somatoform symptoms, even though depression was not present. This possibly reflects a monoaminergic transmitter system alteration in somatization.⁸ In a single photon emission computed tomography (SPECT) study, seven out of 11 somatization disorder patients showed hypoperfusion in different brain areas.⁹ Hakala et al. found severe somatization associating with lower cerebral glucose metabolism in both caudate nuclei, left nucleus putamen and right precentral gyrus compared to healthy volunteers.¹⁰ Hakala et al. also reported an increased volume of both caudate nuclei among somatizing women.¹¹

The present study describes the association of temperament traits and regional brain glucose metabolism among women with severe somatization and is one of a series of researches for the same subjects in the authors’ previous studies.^10,11 Based on the authors earlier observations, they chose to evaluate four brain glucose metabolism regions of interest (ROI), namely caudate and putamen nuclei, thalamus and frontal cortex.^10,11 The frontal cortex ROI consists of inferior, medial and superior frontal lobe ROI from the authors’ earlier research.¹⁰ These regions were selected for evaluation, because the authors observed a trend towards hypometabolism.¹⁰ ROI are presented in Fig. 1.

Figure 1. Regions of interests bilaterally drawn on the magnetic resonance imaging plane from top of the picture to bottom: superior, medial and inferior frontal lobe, nucleus caudate, putamen and thalamus.

METHODS

Subjects

Before entering the study, informed written consent according to the Declaration of Helsinki was obtained from each participant. The investigation had the approval of the ethics committees of Turku University, Turku University Central Hospital and Turku City Hospital, Turku, Finland. A total of 10 female patients had somatization disorder (n = 6) or undifferentiated somatoform disorder (n = 4), and no comorbid psychiatric illnesses according to the DSM-IV criteria.¹ The remainder were 12 female healthy volunteers. Mean age of the patients was 46.6 years (standard devaiton [SD], 8.91; range, 26–58) and mean age of the controls was 46.6 years (SD, 4.34; range, 40–54), respectively (two-sample t-test,P = 1.00). The characteristics of the patients are presented in Table 1.

Table 1.  Patient characteristics

Patient	Age	Diagnosis	Symptoms
1	44	Somatization disorder	Symptoms since adolescence, including, for example, lower abdominal pains, edema, low back pain, headaches, leukorrhea, menstrual problems, twitching in the temporal area. Investigated thoroughly, no formal somatic diagnosis
2	45	Undifferentiated somatoform disorder	Symptoms for 6 years, including, for example, headaches, menstrual pains, pain in the left eye, problems with urinary continence, joint pains, nausea, muscle weakness. Diagnosis: fibromyalgia
3	47	Undifferentiated somatoform disorder	Symptoms since childhood, including, for example, stomach pains, vertigo, headaches, palpitations, trembling of hands, problems with vision. Diagnosis: fibromyalgia
4	58	Undifferentiated somatoform disorder	Symptoms for over 10 years, including, for example, weakness, chest pains, pains in the eyes, insomnia, stomach pains, vertigo, back pains. No definite somatic diagnosis
5	55	Undifferentiated somatoform disorder	Symptoms for over 15 years, including, for example, headaches, fatigue, pains in the eyes, stomach aches, vertigo, cough, diffuse pains, problems with bowel evacuation. Diagnosis: fibromyalgia
6	53	Somatization disorder	Symptoms since the age of 13, including, for example, vertigo, weakness of the hands, stomach pains, menstrual problems, fatigue, tremor, problems with breathing. Diagnosis: Lactose intolerance
7	26	Somatization disorder	Symptoms since adolescence, including, for example, vomiting, headaches, problems withvision, low back pains, dyspepsia, menstrual problems. No definite somatic diagnosis
8	49	Somatization disorder	Symptoms since childhood, including, for example, headaches, menstrual problems, stomach pains, vomiting, chest and neck pain, urinary urgency. Somatic diagnosis: allergic rhinitis
9	41	Somatization disorder	Symptoms for at least 12 years, including, for example, problems with vision, headaches, weakness, stomach pains, urinary incontinence, pain during intercourse, severe menstrual pains. No somatic diagnosis
10	48	Somatization disorder	Symptoms for over 10 years, including, for example, numbness of the extremities, muscle and joint pains, stomach pains, problems with vision in the right eye, lack of libido, menstrual problems, diarrhea. Somatic diagnosis: fibromyalgia

Methods

Case identification

The authors studied 10 female patients and 12 healthy female controls with positron emission tomography (PET) using 2-[18F]-fluoro-2-deoxy-d-glucose (FDG) as a tracer. The patients were recruited through general practitioners, consultation-liason psychiatrists and other health-care specialists working in the catchment area of Turku University Central Hospital. All patients had been investigated in several somatic clinics during the years, but no formal somatic diagnoses accounting for their symptoms had been established. To minimize the confounding effect of another psychiatric disorder, a careful investigation was made to exclude comorbid Axis I psychiatric illness. Potential cases were interviewed by a psychiatrist resident, who elicited the basic information about each patient’s illness history, symptoms, symptom attribution, social background and relationships as well as treatments and diagnostic procedures. The medical records of the patients were reviewed as part of the diagnostic procedure by the resident. The Symptom Checklist (SCL-90) was obtained from each patient.¹² The patients scored near the range of a Finnish normal population in all but somatization subscale.¹³ Patients’ mean scores of the SCL are presented in Table 2. The first phase results were audited by a research psychiatrist, who performed clinical interview and confirmed or rejected the diagnosis of somatization disorder or undifferentiated somatoform disorder. A few patients with comorbid Axis I disorder were excluded at this second selection phase from the study. The third phase consisted of two psychiatrists, who independently and blind from each other, set the diagnosis for the selected patient on the basis of all the collected written information. A patient who was classified as suffering from undifferentiated somatoform disorder or somatization disorder by both evaluators was included. The controls were volunteers without any history of illness or long-lasting vague symptoms causing distress or functional impairment. Their current state of health was good.

Table 2.  SCL-90 scores

	Mean	SD
SCL, Symptom Checklist.
SCL somatization	1.50	0.81
SCL obsessive-compulsive	1.02	0.57
SCL anxiety	0.62	0.44
SCL interpersonal sensitivity	0.96	0.56
SCL depression	1.04	0.72
SCL hostility	0.58	0.38
SCL phobic anxiety	0.27	0.33
SCL paranoid ideation	0.69	0.60
SCL psychotism	0.45	0.43

The 240-item Finnish version of the TCI was obtained from the study subjects (n = 22) and scores reflecting different temperament traits were calculated and included in the statistical analysis. The raw scores of the study subjects’ TCI temperament traits as well as mean and SD are presented in Table 3. TCI is a self-report test evaluating temperament and character dimensions of the personality. The TCI has been studied in clinical and non-clinical samples and empiric data supports its psychometric characteristics.¹⁴ Four temperament traits, namely Novelty-seeking (NS), Harm avoidance (HA), Reward dependence (RD) and Persistence (P), are thought to be part of the psychobiological model of personality.¹⁵ Those traits are assumed to be genetically independent from each other, moderately heritable, and relatively stable over a lifetime as well as universal across different groups.^14,16 Novelty-seeking is defined as a tendency to react actively to novel stimuli. Harm avoidance is viewed as a bias in the inhibition of behaviors, such as pessimistic worry, passive dependent behaviors or fatigability. Reward dependence is defined as a bias in the maintenance or continuation of ongoing behaviors. Persistence is a tendency to perseverance despite frustration and fatigue.

Table 3.  Temperament raw scores, mean and standard deviation of temperament and character inventory

Patient	Temperament trait
	Novelty-seeking	Harm avoidance	Reward dependence	Persistence
Patient group
1	20	18	15	5
2	13	22	14	4
3	18	15	12	3
4	16	7	13	5
5	15	26	12	4
6	16	17	16	4
7	22	18	12	3
8	21	18	13	4
9	15	25	14	6
10	20	19	8	3
Mean ± SD	17.6 ± 3.0	18.5 ± 5.4	12.9 ± 2.2	4.1 ± 1.0
Control group
1	25	9	17	1
2	25	17	12	6
3	20	27	11	5
4	12	6	19	7
5	27	8	18	3
6	21	13	19	1
7	16	3	12	4
8	21	22	13	4
9	27	5	14	7
10	21	10	15	3
11	27	12	19	3
12	25	10	16	4
Mean ± SD	22.3 ± 4.7	11.8 ± 7.1	15.4 ± 3.0	4.0 ± 2.0

Image acquisition and processing

The subjects selected for the study stopped psychopharmacologic medication, enzyme-inducing agents, caffeine and alcohol intake as well as smoked tobacco at least 7 days prior to the PET scan. All subjects fasted 4 h or more before the PET scan.

On the day of the PET study, an intravenous line was inserted in each forearm, one for venous blood assessment (three blood glucose and 21 radioactivity samples) and the other for tracer (FDG). A warming cushion was bound around the forearm to ensure proper blood flow during the collection of blood samples. Subject positioning was handled with 3-D laser alignment with reference to the orbitomeatal line parallel to the detector rings.

All the subjects were scanned with the GE Advance tomography (General Electrics Medical Systems, Milwaukee, USA) at Turku PET Center, Turku, Finland. The scanning was performed in a dimly lit room with minimal auditory stimulation. The subjects were awake during the scan. At the time of the tracer injection the scanning started taking 50 min of 2-D dynamic imaging in 10–5 min frames. The time course of fluorine-18 radioactivity was determined from 21 venous samples (1, 1.5, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 15, 20, 25, 30, 35, 40, 45, 50 min after tracer injection), and blood glucose level from the three venous samples (0, 30 and 50 min after tracer injection).

FDG was synthesized with slight modifications according to the method described by Hamacher et al.¹⁷ The radiochemical purity exceeded 99%, and the specific radioactivity at the end of the synthesis was >75 GBq (2 Ci/µmol). The subjects were given 5 mL of the solution with a dose of about 3.7 MBq/kg (0.1 mCi/kg) of [18]-FDG intravenously in 1 min.

A total of 35 continuous transaxial slices (4.25 mm) were acquired with 18-ring bismuth germinate whole-body scanner with an axial field of view of 14.8 cm. The resolution was 4.2 mm full width at half maximum in all directions. All data were corrected for dead time, tracer decay, and measured photon attenuation and reconstructed with fitted activity curve by patpar in 128 × 128 pixel by pixel matrices.

Magnetic resonance imaging (MRI) was done on all participants as anatomic reference (1.5 Tesla instrument). All MRI were macroscopically stated as normal by a neuroradiologist – no infarctions, atrophy, dilatation of ventricles, hyperintensities or other abnormalities were seen. A reconstructed PET image and MRI were fitted with Amirfit software (Turku PET Centre, Turku, Finland) to obtain equal slice levels. This program fits the images according to the surface of the brain.

ROI were drawn individually at least on four consecutive slices in the gray matter structures (frontal cortex) and basal ganglia (putamen, caudate) as well as thalamus. The ROI were identified by visual inspection with reference to the neuroanatomical atlas of Aquilonius and Eckernäs and to the fitted MRI.¹⁸ The average glucose consumption was calculated for each ROI. The image was scaled to a global cerebral glucose metabolic rate of µmol/100 g tissue/minute by multiplying with blood glucose. The lumped constant value used was 0.81.¹⁹ This constant corrects the difference between glucose and fluorodeoxyglucose. The average regional cerebral metabolic rate of glucose (MRGlc) of left and right hemisphere ROI (e.g. left + right caudate nucleus/2) was used in the statistical analysis. Every MRI reference image was viewed by a neuroradiologist not knowing the subjects status (patient/volunteer).

Statistical analysis

Data was analyzed using logistic regression. Somatization group (patient or control) was dichotomic dependent variable and temperament traits and glucose metabolism regions were continuous independent variables in logistic models. First, univariate logistic regression analyses were done and then significant independent variables (P < 0.05) were analyzed using stepwise logistic regression (inclusion and exclusion criteria, P < 0.10). Results were quantified using odds ratios and their 95% confidence intervals. Statistical analyses were done with SAS System for Windows, release 8.02 (SAS Institute Inc., Cary, NC, USA.

RESULTS

Odds ratios, their 95% confidence intervals and P are presented in Table 4. Lower novelty-seeking scores, higher harm avoidance scores, lower caudate glucose metabolism and lower putamen glucose metabolism were statistically significantly associated with severe somatization. In stepwise logistic regression analysis, caudate glucose metabolism (P = 0.040) and novelty-seeking scores (P = 0.068) were included in the final model.

Table 4.  Univariate logistic regression analyses

	Odds ratio^† (95%CI)	P
CI, confidence interval. Somatization group (patient or control) was a binary dependent variable and temperament traits and glucose metabolism were continuous independent variables. ^†  Odds ratio for one unit increase in continuous independent variable.
Temperament trait
Novelty-seeking	0.75 (0.58–0.98)	0.031
Harm avoidance	1.18 (1.01–1.39)	0.040
Reward dependence	0.68 (0.45–1.02)	0.061
Persistence	1.04 (0.61–1.79)	0.880
Glucose metabolism
Caudate	0.67 (0.47–0.94)	0.021
Putamen	0.73 (0.54–0.97)	0.030
Thalamus	0.75 (0.57–1.00)	0.053
Frontal cortex	0.77 (0.59–1.01)	0.057

DISCUSSION

General

High harm-avoidance as temperament trait has been described across different psychiatric disorders in association with anxiety, depression, interpersonal sensitivity and phobic anxiety as measured with SCL90-R scale.² However, it is also being found among somatizing patients.³ Battaglia et al. reported high NS scores for a group suffering from both somatization and panic disorders compared to control subjects or patients with panic disorder only.²⁰ Due to the above, the authors’ hypothesis is that temperament traits, especially HA, may play a contributing or associated role in development of somatoform disorders. Futhermore, the role of NS in somatization must be considered unclear until more confirmatory data of its role is available.

In somatization, immunological abnormalities and changes both in the brain structure, metabolism and cognitive performance as well as alterations in physiological reactions and the monoaminergic system have been reported.^4–11 These results are in line with the authors’ hypothesis that central nervous system pathology is involved in somatizing conditions.

The authors findings specify the features of somatoform disorders – low novelty-seeking, high harm avoidance as temperament traits, lower glucose metabolism in both caudate and putamen nuclei. Both metabolic and temperament results converge with the clinical picture of the patients: low energy level, fatigability, tendency to avoid new stimuli, need of rest and restriction of additional activities.

Somatoform disorders do not constitute a homogenous group and considerable comorbidity with depression and anxiety disorders may confound the clinical picture. However, the results from different studies above indicate that somatoform disorders are not sharing all the findings observed in other psychiatric disorders.

Limitations and strengths

The small sample size is a possible source of bias and causes the lack of power in statistical analyses. Also, the lack of structured clinical interview, Structured Clinical Interview for DSM-IV or equivalent, may be considered a limitation in patient selection. In contrast, careful clinical evaluation of the nature of the somatic symptoms and significance of the physical findings was considered essential to assess the diagnosis correctly and rule out the possibility of any other condition which could contaminate the patient selection. To exclude Axis I disorders multiphase clinical evaluation and symptoms, a questionnaire was used in order to ensure the validity of correct psychiatric Axis I diagnosis. Four patients had a formal diagnosis of fibromyalgia. In Finland, formal diagnosis is required for social welfare compensations (e.g. sickleave economic compensation). Fibromyalgia did not exclusively explain patients’ symptoms pattern, though it is less stigmatizing and often used in clinically unclear situations. In addition, different functional somatic syndromes (e.g. fibromyalgia, irritable bowel syndrome, chronic fatigue syndrome, multiple chemical sensitivity etc.) often share more similarities than differences.²¹ One option is that these disease entities fall into the category of somatoform illness.

CONCLUSION

In the present study, severe somatization seems strongly to associate with lowered cerebral glucose metabolism rate in nucleus caudate and, according to the authors’ hypothesis, temperamental features are possibly non-specific factors predisposing subjects to somatization. All in all, there is still an inadequate amount of data to make definitive conclusions concerning the pathophysiological processes of somatization, and, therefore, further investigations are encouraged.

ACKNOWLEDGMENTS

The study was supported by a grant from the Signe and Ane Gyllenberg Foundation, the Lundbeck Foundation and the Satakunta Hospital District (EVO). Jarmo Hietala and Harry Vilkman, Turku PET Centre, Turku, Finland, are acknowledged for the Finnish version of the TCI.