RESEARCH ARTICLE

Intellectual Function in Patients with Anorexia Nervosa and Bulimia Nervosa Siri Weider1,2*, Marit Sæbø Indredavik3,4, Stian Lydersen3 & Knut Hestad1,5 1

Department of Psychology, Norwegian University of Science and Technology (NTNU), Trondheim, Norway Specialised Unit for Eating Disorder Patients, Department of Psychiatry, Levanger Hospital, Health Trust Nord-Trøndelag, Levanger, Norway 3 Regional Centre for Child and Youth Mental Health and Child Welfare, NTNU, Trondheim, Norway 4 Department of Child and Adolescent Psychiatry, St. Olav’s University Hospital, Trondheim, Norway 5 Division of Mental Health, Innlandet Hospital Trust, Hamar, Norway 2

Abstract Objective: This study aimed to examine cognitive function in individuals with anorexia nervosa (AN) and bulimia nervosa (BN) on the basis of IQ measures, indexes and subtests of the Wechsler Adult Intelligence Scale – Third Edition (WAIS-III). Methods: A total of 41 patients with AN, 40 patients with BN and 40 healthy controls (HC), matched for sex, age and education, were recruited consecutively to complete the WAIS-III. Results: The AN group showed a significantly lower performance than the HC group on most global measures and on eight of the 13 administered subtests. Minor differences in verbal function were detected between the BN group and the HC group. Conclusion: The patients with eating disorders showed normal intellectual functions compared with the normative population. However, the AN group displayed a consistently lower performance than the matched HC group, which performed above normative means. The BN group performed at a level between that of the AN and HC groups. Copyright © 2013 John Wiley & Sons, Ltd and Eating Disorders Association. Keywords anorexia nervosa; bulimia nervosa; neuropsychology; intelligence *Correspondence Siri Weider, Department of Psychology, the Norwegian University of Science and Technology (NTNU), Dragvoll, 7491 Trondheim, Norway. Tel.: 0047 952 61 916; Fax: 0047 73 59 19 20. Email: [email protected] Published online 24 September 2013 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/erv.2260

Introduction Eating disorders (EDs) are characterized by disturbed eating behavior, as well as an intense preoccupation with body shape and size. The diagnostic classifications of EDs currently include the following diagnoses: anorexia nervosa (AN), bulimia nervosa (BN) and eating disorders not otherwise specified (American Psychiatric Association, 2000). In Norway, the prevalence of EDs has been reported to be 0.2–0.4% for AN and 1.2% for BN (Statens Helsetilsyn, 2000). These numbers are fairly consistent with those from other parts of the world (Hoek & van Hoeken, 2003; Preti et al., 2009; Smink, van Hoeken, & Hoek, 2012). Eating disorders most likely occur as a result of genetic and environmental factors interacting in a developmental and sociocultural context. Increasing evidence has shown neurobiological alterations in patients with EDs, including both genetic vulnerability (Bulik, 2005; Bulik, Sullivan, Wade, & Kendler, 2000; Bulik & Tozzi, 2004) and specific personality traits. In addition, several studies using brain imaging techniques (computed tomography and magnetic resonance imaging) have detected structural changes in the brains of patients with AN (Titova, Hjorth, Schioth, & Brooks, 2013) and to a lesser extent also in patients with BN (Krieg, Lauer, & Pirke, 1989).

Early studies on cognitive function in EDs was mainly performed on patients with AN and focused on providing a broad description of achievements in a variety of cognitive domains (Steinglass & Glasofer, 2011). The findings from these studies yielded highly inconsistent conclusions, which were likely caused to some extent by the lack of healthy control (HC) groups (Lena, Fiocco, & Leyenaar, 2004) and the use of a wide variety of tests (some of which have not been validated) and their diverging interpretations (Duchesne et al., 2004; Tchanturia, Campbell, Morris, & Treasure, 2005). In recent years, the focus has moved to the study of domain-specific areas of cognition where deficits have been shown. Currently, strong evidence suggests that patients with AN display impaired central coherence, as well as difficulties related to executive functions such as set shifting and inhibition (Holliday, Tchanturia, Landau, Collier, & Treasure, 2005; Lopez, Tchanturia, Stahl, & Treasure, 2009; Roberts, Tchanturia, Stahl, Southgate, & Treasure, 2007; Tchanturia et al., 2012). In contrast, deficits shown in patients with BN do not seem to represent a specific cognitive profile (Van den Eynde et al., 2011). With the release of the Wechsler Adult Intelligence Scale – Third Edition (WAIS-III), it has been argued that the Wechsler tests have evolved from pure intelligence tests to more comprehensive test batteries for measuring a range of cognitive

Eur. Eat. Disorders Rev. 22 (2014) 15–24 © 2013 John Wiley & Sons, Ltd and Eating Disorders Association.

15

Intellectual Function in AN and BN Patients

S. Weider et al.

functions (Egeland, Bosnes, & Johansen, 2009), which makes the test a valuable instrument for investigating the cognition of patients with EDs. Only a few studies have specifically addressed the issue of intelligence in patients with EDs. However, a number of studies have included intelligence screening results as a demographic variable. These studies have mainly been conducted on patients with AN. There has been a long-standing assumption that AN is a disorder that most commonly occurs in groups of higher socioeconomic status, and early studies revealed significantly elevated intelligence in these groups (Dally, 1969). More recently, several studies have found generally high intellectual function in patients with AN (Blanz, Detzner, Lay, Rose, & Schmidt, 1997; Connan et al., 2006). Other authors have reported average intellectual function in patients with AN, measured both in clinical groups (Galderisi et al., 2003; Ranseen & Humphries, 1992) and in populationbased studies (Gillberg, Gillberg, Rastam, & Johansson, 1996), whereas Koyama et al. (2012) found that patients with restrictive AN and a very low body mass index (BMI; 12.8) showed significantly reduced full-scale intelligence (FIQ) scores. A meta-analysis on intellectual functioning in AN examined whether patients with AN have higher intelligence than the general population (Lopez, Stahl, & Tchanturia, 2010). The authors concluded that different methods generate somewhat divergent results. In studies in which Wechsler’s tests were used, the average FIQ was shown to vary between 96.5 and 117.6. In studies using the National Adult Reading Test (NART) as a measure of intelligence, patients with AN had consistently been reported to have higher than normal intellectual function. In BN patients, an intelligence level within or above the normative range has been found in studies using Wechsler’s tests (Jones, Duncan, Brouwers, & Mirsky, 1991; Kim, Kim, & Kim, 2010; Kim, Lim, & Treasure, 2011; Marsh et al., 2009) and the NART (Bosanac et al., 2007; Davies, Swan, Schmidt, & Tchanturia, 2012; Harrison, Sullivan, Tchanturia, & Treasure, 2010; Kenyon et al., 2012; Southgate, Tchanturia, & Treasure, 2008; Tchanturia et al., 2011). Considering the high heritability of intelligence (Neisser et al., 1996), a large recent cohort study conducted by Kothari, Solmi, Treasure, and Micali (2013) on neuropsychological functioning in children of mothers with lifetime ED is interesting. In this study, children of mothers with lifetime AN showed significantly better FIQ and performance intelligence, in addition to better performance on the perceptual organization index than children of unaffected mothers. Better verbal intelligence (VIQ) than performance intelligence (PIQ) is reported in patients with AN (McCormick et al., 2008). A potential explanation for this finding is that VIQ is more resistant to cognitive impairment or that this group displayed a better premorbid VIQ than PIQ. This result is supported by findings from a study on patients with AN in whom asymmetric cognitive function (exhibiting better function on tasks believed to depend more on the left hemisphere than on the right hemisphere) was found (Maxwell, Tucker, & Townes, 1984). An opposite asymmetric cognitive profile, with better performance regarding PIQ than VIQ, was described in patients with BN, measured with a German intelligence test, the Prüfsystem für Schul-und Bildungsberatung (Blanz et al., 1997). Previous studies on intellectual function associated with EDs have generally only reported the FIQ. To our knowledge, only 16

two previous studies have explored the FIQ, VIQ and PIQ in both AN and BN patients (Jones, Duncan, Brouwers, & Mirsky, 1991; Ranseen & Humphries, 1992), and no studies have examined intellectual function at the index or subtest level in both groups. The aim of the present study was to explore the cognitive profiles of patients diagnosed with AN and BN compared with an HC group matched for sex, age and education. To highlight the relative strengths and weaknesses of the groups, we used the various global measures, indexes, subtests and the General Ability Index (GAI) from the WAIS-III and controlled for factors that might affect cognitive function, namely sex, age, education, psychotropic medication, BMI, nadir (lowest ever) BMI and depression.

Methods Participants Patients were recruited consecutively from two special ED units at Levanger Hospital (RKSF). These units provide healthcare coverage to Central Norway in addition to receiving patients from throughout the country through the Free Hospital Choice Norway program. These units treat patients of both sexes, 16 years or older, with all types of EDs. As Danielsen & Ro (2012) describe, indications for admission were earlier unsuccessful outpatient treatment and severe illness requiring inpatient care. Patients with different types of EDs of varying severities were admitted to the units. Thorough information about the study was given to the therapists treating these patients. Consequently, 110 patients were consecutively referred for inclusion between September 2008 and April 2010 and between August 2011 and February 2013. Fifteen of these patients suffered from eating disorders not otherwise specified and were excluded from this analysis. Further exclusion criteria included confirmed brain damage (n = 1; cerebral infarction), psychosis, diabetes (n = 1), neurological disease and neuropsychiatric disorders such as attention deficit hyperactivity disorder (n = 3), Tourette’s syndrome, autism spectrum disorder or chronic fatigue syndrome (n = 2), leaving 88 referred candidates. Seven women did not want to participate in the study after receiving detailed information. Hence, 81 of 88 (92.0%) patients participated. Among the patients who did not want to participate, four were diagnosed with AN. These patients displayed a higher average age (42.0 vs. 28.1 years) and a lower average BMI (15.4 vs. 16.3) than the patients with AN who participated in the study. The other three patients who chose not to participate were diagnosed with BN, and these patients did not differ from patients with BN who participated in the study The patients were diagnosed according to the criteria of the Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, Text Revision (American Psychiatric Association, 2000). The AN group consisted of 39 female patients and two male patients with an average BMI of 16.3 kg/m2 (SD = 1.9) (normal range 18.5–24.9). Among these subjects, 34 patients met all of the criteria for AN. Seven patients were characterized as suffering from subthreshold AN and met all but one of the diagnostic criteria for AN as a result of recent weight gain during their current hospitalization (BMI 17.6–18.5). The BN group consisted of 38 female patients and two male patients with an average BMI of 21.7 (SD = 3.9), who met all of the diagnostic criteria for

Eur. Eat. Disorders Rev. 22 (2014) 15–24 © 2013 John Wiley & Sons, Ltd and Eating Disorders Association.

S. Weider et al.

BN. Five of the patients with BN had formerly had an AN diagnosis. They all had normal weight at the time of assessment (BMI 18.5–22.0), and the period with AN was more than 1 year ago. None of them had been hospitalized for AN. All of the patients were undergoing a treatment regime that involved admission to full-time care. Efforts were made to recruit the patients as close to the time of admission as possible. However, all patients first underwent an introductory week and were at the earliest recruited at the end of the first week of ordinary admittance (at the end of the second week) to avoid the effect of starvation. In patients with severe AN, the testing, in some instances, was postponed until the patient was more nutritionally stable. In some cases, however, it was not possible to include the patient until a while after admission (≤4 weeks) or during the first week after the patient had been discharged from the units. Among the 41 AN patients, 37 (90.2%) were inpatients, one received day treatment and three were outpatients. Among the 40 BN patients, 32 (80.0%) were inpatients, one received day treatment and seven were outpatients. All patients were medically stable at the time of assessment, ate ordinary food and attended at least four meals per day. Patients with BMI < 17 usually received a fifth meal and were given a nutritional supplement in addition. The nutritional status of the outpatients was, however, more uncertain. The patients were maintained on their regular medication. In the AN group, the following medications were used: antidepressants (n = 26), antipsychotics (n = 15), anxiolytics (n = 7), hypnotics (n = 7), thyroid hormones (n = 3) and antiepileptics (n = 3), either as mood stabilizers or as a treatment for peripheral neuropathic pain. In the BN group, the following medications were used: antidepressants (n = 27), hypnotics (n = 7), antipsychotics (n = 6), anxiolytics (n = 3), thyroid hormones (n = 2) and antiepileptics as mood stabilizers (n = 1). For all patients, the ED was their primary diagnosis. However, several patients displayed comorbid disorders. In the AN group, the following comorbid diagnoses were present: depression (n = 17), post-traumatic stress disorder (PTSD) (n = 5), bipolar II disorder (n = 2), obsessive–compulsive disorder (OCD) (n = 3) and generalized anxiety disorder (GAD) (n = 3). Three of the patients with AN suffered from hypothyroidism. In the bulimia group, the following comorbid diagnoses were present: depression (n = 10), PTSD (n = 4), GAD (n = 2), OCD (n = 1) and bipolar disorder (n = 2). In this group, two patients were treated for hypothyroidism, one was treated for hypertension and one patient had celiac disease. The control group consisted of 40 healthy individuals that were recruited from the following educational facilities, located in both urban and rural areas: adult education participants at Ole Vig Upper Secondary School in Stjørdal, Sør-Trøndelag University College, Nord-Trøndelag University College, the Norwegian University of Science and Technology and the Folkeuniversitetet Adult Education Association. Three HCs were also recruited by word of mouth. The exclusion criteria for this group were the same as those for the patient groups, with the following additional exclusion criteria: a life history of EDs, currently being on a diet, having a known psychiatric diagnosis and displaying a BMI of 26 kg/m2. Five participants were excluded as a result of having a known psychiatric diagnosis (n = 2) or a BMI of >26 kg/m2 (n = 3). The matching was carried out individually,

Intellectual Function in AN and BN Patients

creating triplets consisting of one patient with AN, one with BN and one HC that were similar with respect to age (±3 years), sex and education. This was possible because the patient groups were similar regarding age and education. However, the age span and educational level of the patients in the AN group seemed to vary more than that of the BN group. As the HC group was recruited to resemble also the cases at the extremes, the HCs were in some cases a closer match to the AN patients than to the BN patients. The choice of including a healthy comparison group matched according to education was based on the hypothesis that patients with AN might have higher than average level of education, perhaps as a result of higher socioeconomic status. A sex-matched control group was necessitated by the knowledge of a skewed sex distribution in this population. The impact of sex on performance on earlier WAIS tests has been established (Kaufman, McLean, & Reynolds, 1988), and research seems to find better verbal functioning in women than in men and better visuospatial abilities in men than in women (Neisser et al., 1996). Measurements Clinical assessment Patients were interviewed using the Mini International Neuropsychiatric Interview (Sheehan & Lecrubier, 1994). This is a semistructured interview for the assessment of Axis I symptom disorders based on the Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, Text Revision (American Psychiatric Association, 2000). These interviews confirmed the ED diagnosis and identified comorbid psychopathologies. Beck depression inventory-II. The Beck Depression Inventory (BDI) is a self-reporting questionnaire that includes 21 questions, using a four-point grading scale from 0 to 3, measuring the degree of depression (Beck, Steer and Brown, 1996) over the past 2 weeks. Points are summed, and depressive symptoms may be scored as minimal (0–13 points), mild (14–19 points), moderate (20–28 points) or severe (29–63 points). Cognitive assessment The patients completed a comprehensive battery of neuropsychological tests. In the current study, only results from the general intelligence test were analysed. Wechsler adult intelligence scale – third edition. The WAIS-III (Wechsler, 1997) includes a number of subtests that measure various aspects of cognition. This intelligence scale was made available in Norwegian in 2003, following American norms, after thorough translation and adaption (Wechsler, Nyman, & Nordvik, 2003). The test provides measures of FIQ, VIQ and PIQ. In addition, the test quantifies achievements through four different indexes: the Verbal Comprehension Index (VCI), Perceptual Organization Index (POI), Working Memory Index (WMI) and Processing Speed Index (PSI). Scores in the normative sample for the subtests is defined as M = 10, SD = 3, whereas the IQ scores and indexes are defined as M = 100, SD = 15. In addition to these measures, the GAI was calculated according to the procedures and norms described by Tulsky, Saklofske, Wilkins, and Weiss (2001). The GAI is a composite score composed of

Eur. Eat. Disorders Rev. 22 (2014) 15–24 © 2013 John Wiley & Sons, Ltd and Eating Disorders Association.

17

Intellectual Function in AN and BN Patients

S. Weider et al.

the subtests included in the VCI and POI. The GAI may be a better measure of intelligence for this group, as processing speed and working memory are considered to be more vulnerable to neuropsychological damage (Tulsky, Saklofske, Wilkins and Weiss, 2001). The GAI is also assumed to be a better index of general ability and to be more resilient to cognitive impairments than the FIQ, and the GAI has been incorporated into the WAIS-IV. Ethics The study was approved by the Regional Committee for Medical and Health Ethics of Central Norway (reference 4.2007.2229). All participants gave written informed consent. Statistical analysis The data were analysed using SPSS 19.0 (IBM Corporation, Armonk, NY, USA). All tests were two-tailed, and the level of statistical significance was set at p < .05. The demographic and clinical characteristics of the AN, BN and HC groups (Table 1) were compared via one-way, between-group analysis of variance with a post hoc Sidak test. A general linear model was employed to analyse the differences between the three groups in the global measures of the WAIS-III, including the FIQ, VIQ and PIQ as well as the GAI and the four other indexes. Potential confounding factors (age, sex, years of education, psychotropic medication, BMI, nadir BMI and BDI-II) were each adjusted for separately. The normality of the residuals was checked through visual inspection of Q–Q plots. At the index and subtest levels, the test scores were analysed via analysis of variance with a post hoc Sidak test. Within-subject contrasts of relative weaknesses and strengths were performed using paired samples t-tests.

Results Demographic and clinical characteristics As shown in Table 1, no significant differences were found between the groups regarding age, years of education, the father’s or mother’s education level, years with an ED or years of treatment. As expected, the AN group exhibited a significantly lower BMI than the BN and HC groups. In addition, the nadir BMI

was significantly lower in the AN group than the BN group and the HC group. The nadir BMI was also lower in the BN group than in the HC group. Furthermore, a significant difference was observed in the degree of depression as measured by the BDI-II: both the AN and BN groups scored significantly higher than the HC group, and the AN group scored significantly higher than the BN group, indicating higher levels of depression. Statistics on level of education for each sex and age group was obtained from Statistics Norway (2013). On the basis of these statistics, we computed the expected percentage within the education levels ( 115) above the normative mean. It is worth noting that the control group consistently scored better than the normative sample. The test scores from the global IQ measures and the indexes obtained on the basis of the WAIS-III, adjusted for sex, age and education, are displayed in Table 2. The AN group displayed a significantly lower score than the HC group on all IQ and index measures. The BN group presented a higher score than the AN group on the PIQ, POI and PSI. The BN group scored below the HC group on VIQ and VCI. At the index level, the analyses revealed that the groups’ profiles showed similar patterns of strengths and weaknesses. All three groups performed best in the POI, followed by the VCI, PSI and WMI. However, the differences between the groups varied substantially. Five patients in the BN group had an AN diagnoses >1 year ago. These patients performed better than the AN group on the PSI but did otherwise not differ from either patient group.

Table 1 Demographic and clinical characteristics of groups with anorexia nervosa and bulimia nervosa compared with healthy controls AN (n = 41) Mean (SD) Age BMI Nadir BMI Number of years of education Number of years with an ED Father’s education Mother’s education BDI-II

28.07 16.25 13.68 13.49 12.28 12.83 12.85 30.20

(10.223) (1.901)* (2.420)* (2.204) (9.692)‡ (3.528) (2.954) (13.841)*

BN (n = 40) Mean (SD) 27.55 21.72 16.93 12.65 10.80 12.58 12.30 23.63

(8.629) (3.935) (2.933)† (1.929) (7.028)‡ (2.791) (3.156) (12.834)†

HC (n = 40) Mean (SD)

F

p-value

27.85 (10.225) 22.51 (1.833) 20.73 (1.667) 13.03 (1.915) N/A 12.65 (3.060) 13.05 (2.791) 5.53 (6.771)

0.029 63.355 99.913 1.746 37.539 0.070 0.686 48.936

0.971

Intellectual function in patients with anorexia nervosa and bulimia nervosa.

This study aimed to examine cognitive function in individuals with anorexia nervosa (AN) and bulimia nervosa (BN) on the basis of IQ measures, indexes...
328KB Sizes 0 Downloads 0 Views