Neurovegetative Symptoms in Multiple Sclerosis - CiteSeerX

Archives of Clinical Neuropsychology, Vol. 15, No. 5, pp. 387–398, 2000 Copyright © 2000 National Academy of Neuropsychology Printed in the USA. All rights reserved 0887-6177/00 $–see front matter

PII S0887-6177(99)00030-X

Neurovegetative Symptoms in Multiple Sclerosis: Relationship to Depressed Mood, Fatigue, and Physical Disability John J. Randolph, Peter A. Arnett, Christopher I. Higginson, and William D. Voss Washington State University

Some authors have suggested that when evaluating depression in multiple sclerosis (MS) patients, neurovegetative symptoms should be discounted and/or not considered, given the ostensibly high overlap between symptoms of MS (e.g., sleep disturbance, fatigue) and neurovegetative symptoms of depression. A further assertion is that inclusion of items assessing neurovegetative symptoms may artificially inflate overall depression scores and that mood scales may provide more accurate indices of depression in MS patients. The current study investigated the possibility that some neurovegetative symptoms may be specifically related to MS patients’ depressed mood and are not simply indicators of physical disability and/or fatigue. Seventy-six clinically definite MS patients in the northwestern United States were administered two depression inventories and measures of physical disability and fatigue as part of a larger study. Results revealed that one neurovegetative symptom—disinterest in sex—was uniquely associated with depressed mood, and other neurovegetative symptoms were associated with both depression and fatigue but not physical disability. The present findings suggest that certain neurovegetative symptoms are differentially associated with depression, fatigue, and physical disability in MS. Routinely discounting all neurovegetative symptoms when assessing depression in MS patients may thus be unwarranted. © 2000 National Academy of Neuropsychology. Published by Elsevier Science Ltd Keywords: neurovegetative symptoms, multiple sclerosis, depression

Multiple sclerosis (MS) is a neurological illness characterized by various somatic and emotional symptoms, such as physical disability, fatigue, and depression. Depression is a particularly common form of psychopathology in MS patients, with lifetime prevalence rates ranging from 27 to 54% across studies (Minden, Orav, & Reich, 1987; Minden & Schiffer, 1991; Rao, Leo, Bernardin, & Unverzagt, 1991). MS patients demonstrate a higher prevalence of depression even when compared to patients with other chronic medical or neurological conditions with similar levels of physical disability (e.g., muscular dystrophy) (Schiffer & Babigian, 1984; Whitlock & Siskind, 1980). Depression in MS patients is generally assessed using self-report questionnaires, such as the Beck Depression Inventory (BDI; Beck & Steer, 1987) or the Zung Self-rating Address correspondence to: Peter A. Arnett, PhD, Department of Psychology, College of the Liberal Arts, The Pennsylvania State University, Bruce V. Moore Bldg. Room 522, University Park, PA 16802-3104. E-mail: [email protected]

387

388

J. J. Randolph et al.

Depression Scale (SDS; Zung, 1965). One inherent difficulty in using measures such as the BDI or SDS is that such measures have items that assess neurovegetative symptoms of depression. Given the overlap between neurovegetative symptoms of depression (e.g., sleep difficulty) and somatic symptoms commonly associated with MS-related physical disability (e.g., restlessness, fatigue), scores of depression on self-report measures may be artificially inflated, leading to inaccurate portrayal of depressive symptomatology. Some authors (e.g., Huber & Rao, 1993; Nyenhuis et al., 1995) have suggested that when evaluating depression in MS patients, neurovegetative symptoms should be discounted and/or not considered, given the ostensibly high overlap between symptoms of MS and neurovegetative symptoms of depression. A further assertion is that mood scales, such as the one on the Chicago Multiscale Depression Inventory (Nyenhuis et al., 1995) may provide a more accurate characterization of depression in MS patients than scales containing neurovegetative items. One important consideration in evaluating depression in MS patients not commonly addressed in the literature is that certain neurovegetative symptoms may be specifically related to MS patients’ depression and are not simply indicators of physical disability or fatigue. Although researchers’ (e.g., Nyenhuis et al., 1995) assertions against the use of neurovegetative symptoms in MS are sensible, it is also possible that some of these symptoms are more associated with depression than other symptoms. Thus, routinely failing to include any neurovegetative items in assessment of depression in MS may not be warranted and may lead to an incomplete and possibly misleading description of depression in MS. The current study investigated the possibility that some neurovegetative symptoms may be specifically related to MS patients’ depression and are not simply indicators of physical disability and/or fatigue. In particular, we assessed whether various neurovegetative symptoms were differentially associated with physical disability, fatigue, and depression in MS patients.

METHOD Participants and Procedure Seventy-nine Caucasian MS patients were recruited from outpatient clinics and the Spokane, WA chapter of the National MS Society. Each MS participant was diagnosed as having definite or probable MS based on Poser et al. (1983) criteria by a board-certified neurologist who also determined disease course based on recently published criteria (Lublin & Reingold, 1996); scores from the Expanded Disability Status Scale (Kurtzke, 1983) were also rated. Participants were excluded from the present study if they: (a) had a history of drug or alcohol abuse or nervous system disorder other than MS, (b) had severe motor or visual impairment that might interfere with cognitive testing, or (c) could not be evaluated on an outpatient basis. Following the testing, it was subsequently learned that one patient had suffered a severe allergic reaction in childhood that resulted in him being comatose for a long period of time; another tested patient refused or was unable to complete a significant proportion of the test battery. Therefore, these two participants were excluded from data analyses. Additionally, one patient (a depressed MS participant) refused to complete one questionnaire and therefore could not be included in the data analyses. Thus, the MS patients described below included 76 tested participants (see Table 1 for demographic information). Additional description of patient characteristics and recruitment can be found in Arnett et al. (1999).

Neurovegetative Symptoms

389

TABLE 1 Summary of Participant Characteristics Variable

n CMDI Mood Scale BDI Age (years) Education (years) Symptom duration (years) Diagnosis duration (years) Sex, n (% Female) Diagnostic category Clinically Definite, n (%) Laboratory Definite, n (%) Clinically Probable, n (%) Laboratory Probable, n (%) Clinical course Relapsing-remitting, n (%) Secondary progressive, n (%) Primary progressive, n (%) Progressive relapsing, n (%) Medicationsa Antidepressant, n (%) Interferon-based, n (%) Beta-blocker, n (%) Antispasmodic, n (%) Benzodiazepine, n (%) Immunosuppressant, n (%) Anticholinergic, n (%)

76 21.88 10.67 46.49 (8.21) 14.88 (2.37) 14.33 (9.96) 7.82 (7.84) 61 (80%) 68 (89%) 4 (5%) 3 (5%) 1 (1%) 48 (63%) 19 (25%) 7 (9%) 2 (3%) 28 (37%) 17 (22%) 13 (17%) 11 (14%) 9 (12%) 9 (12%) 6 (8%)

Note. Unless otherwise specified, values represent means (standard deviation). CMDI ⫽ Chicago Multiscale Depression Inventory; BDI ⫽ Beck Depression Inventory. a 25 patients were using at least two medications at the time of data collection.

After providing informed consent, participants were administered a semi-structured psychosocial interview followed by a neuropsychological test battery. No participant was experiencing an exacerbation of MS symptoms at the time of testing. In return for their participation, participants received a brief written evaluation of their overall intellectual functioning, attention, memory, and mood. The treatment of participants was in accordance with the ethical standards of the American Psychological Association and the state of Washington.

Measures Expanded Disability Status Scale (EDSS; Kurtzke, 1983). The EDSS is a scale that evaluates physical and neurological disability in MS patients. Various functional systems are assessed by the EDSS, including pyramidal, cerebellar, brain stem, and sensory systems. Ratings are made on a 0 to 10 scale, ranging from no physical disability/disturbance in functional systems (0) to extreme functional system disturbance (e.g., inability to communicate or eat rated as a 9.5). Reasonable psychometric data for the EDSS are published elsewhere (Kurtzke, 1983). Higher EDSS scores are indicative of greater disability.

390


Fatigue Impact Scale (FIS; Fisk, Pontefract, Ritvo, Archibald, & Murray, 1994) . The FIS is a 40-item self-report measure assessing symptoms and sequelae of fatigue. Participants rate items on a 0 to 4 scale, from “no problem” to “extreme problem.” Sample items include, “Because of my fatigue I am less motivated to do anything that requires physical effort” and “Because of my fatigue I have to reduce my workload or responsibilities.” Higher FIS scores represent more severe fatigue.

Beck Depression Inventory (BDI; Beck & Steer, 1987). The BDI is a 21-item self-report inventory that assesses depressive symptoms over the past week. Items are rated on a 0 to 3 scale, from neutral feelings about a statement (e.g., “I have not noticed any recent change in my interest in sex,” indicated by a “0”) to depressotypic feelings (e.g., “I have lost interest in sex completely,” indicated by a “3”). Only items assessing neurovegetative symptoms of depression (i.e., sleep disturbance, tiredness, appetite change, decision-making difficulty, disinterest in sex) were used in data analyses; these items were considered individually in relation to Chicago Multiscale Depression Inventory, EDSS, and FIS scores. The BDI has consistently been found to be highly reliable and to demonstrate high convergent validity with independent ratings of depression severity with various clinical populations (Beck, Steer, & Garbin, 1988). Higher BDI scores represent greater depression.

Chicago Multiscale Depression Inventory (CMDI; Nyenhuis et al., 1995). The CMDI is a 42-item measure of depression developed specifically to assess depression in MS and other medical populations. The CMDI contains three subscales (14 items each) that respectively assess mood, evaluative, and neurovegetative symptoms of depression. Only the mood scale will be used in the present study in order to derive a relatively “pure” index of depressed mood independent of neurovegetative symptoms. Participants rate how well items describe their own experiences (e.g., “sad”; “easily awakened”; “rejected”) over the past week on a 1 to 5 scale, from “not at all” to “extremely.” Reliability and validity data for the CMDI attest to its usefulness as a measure of depression with MS patients; in validation studies, coefficient alpha of the three scales has ranged from .91 to .98 (Nyenhuis et al., 1995) and significant correlations have been found between all scales and other well-used depression inventories (Nyenhuis, Luchetta, Yamamoto, Terrien, & Garron, 1996). Items were scored such that higher scores indicate more depressed mood.

Data Analytic Strategy For data analyses, neurovegetative symptoms from the BDI were correlated with CMDI-assessed depressed mood. Two measures of depression were used in this way in order to compare neurovegetative symptoms on a commonly used measure of depression (BDI) with a scale designed specifically to assess mood independent of neurovegetative and evaluative symptoms of depression (CMDI) and to reduce effects of multicollinearity. We first correlated BDI items with the CMDI-mood scale and the EDSS to determine the relations among neurovegetative items, depressed mood, and physical/ neurological disability. Stepwise regression analyses were then conducted to determine the relative contribution of each neurovegetative item to depressed mood. An additional correlational analysis was then performed correlating the neurovegetative items with the


391

CMDI-mood scale and the Fatigue Impact Scale to assess the relationships among neurovegetative symptoms, depressed mood, and fatigue. A final data analytic procedure applied in the present study was structural equation modeling (SEM) using the EQS computer program (Bentler, 1995). When investigating theoretical connections between various independent and dependent variables, structural modeling has several advantages over other statistical procedures (e.g., multiple regression; see Baron & Kenny, 1986). For example, with SEM, all relationships among variables are explicitly stated through description of specific paths connecting certain variables; these relationships are also specified graphically, allowing for an elegant presentation of study hypotheses and findings. Further, because the present research relies on questionnaires that assess psychological variables, it is likely that measurement error is a methodological issue of concern. Importantly, SEM incorporates biases such as measurement error directly into estimated models, allowing for data to be analyzed and interpreted with more confidence. Using SEM, we estimated a post-hoc model based on correlational analyses in order to conceptualize more clearly the possible relations between neurovegetative symptoms and depressed mood, physical disability, and fatigue. To examine the fit of the model, four fit indexes were used. First, the chi-square statistic tests the departure of the specified model’s estimated covariance matrix from the actual sample covariance matrix. Given that the chi-square is known to be distorted with small or large sample sizes, some authors have suggested that the chi-square to degrees of freedom ratio is a more appropriate way to use the chi-square statistic to assess model fit (Bollen, 1989). A model is considered acceptable if this ratio is less than 2.0 (Ullman, 1996). Both the chi-square statistic and the chi-square/degrees of freedom ratio are reported below. Two other fit indexes, the Comparative Fit Index (CFI) and Incremental Fit Index (IFI) have been found to reflect model fit well regardless of sample size and have less sampling variance than other fit indexes (Bentler, 1990; Bollen, 1989). A fourth index, the LISREL Goodness-of-Fit Index (GFI), is based on arbitrary distribution theory and is computed according to the size of weighted residuals compared to the size of the weighted input data (Bentler, 1995). This index is commonly used in another structural modeling computer program (LISREL; Joreskog & Sorbom, 1988). CFI, IFI, and GFI values range from 0 to 1; values greater than .9 represent a good model fit (Bentler, 1995).

RESULTS Preliminary Analyses Of note is the number of female versus male participants in our study; our female:male ratio is approximately 4:1. This ratio is not unusual for studies in the MS literature (cf. Rao et al., 1991); however, considering that the sex distribution for MS patients in the general population is thought to approximate a 2:1 female:male ratio (Hogancamp, Rodriguez, & Weinshenker, 1997), it is possible that our results are not entirely representative of MS patients as a whole. Therefore, we conducted t-test analyses to determine whether male and female patients differed on study variables. Results revealed no significant differences on any study variable (all p’s ⬎ .35). We also conducted t-test analyses to determine whether patients with relapsingremitting versus progressive courses (collapsed across primary and secondary progressive courses) differed on study variables. Analyses revealed only one significant difference between groups; patients with a progressive course had significantly higher scores

392

J. J. Randolph et al. TABLE 2 Intercorrelations Among BDI Neurovegetative Items, CMDI-Mood Subscale, and EDSS Scores

1. Decision-making difficulty 2. Sleep disturbance 3. Tiredness 4. Appetite change 5. Disinterest in sex

CMDI-Mood

EDSS

.32** .41*** .28** .22 .38***

.11 .02 .21 .18 ⫺.05

BDI ⫽ Beck Depression Inventory; CMDI ⫽ Chicago Multiscale Depression Inventory; EDSS ⫽ Expanded Disability Status Scale. n ⫽ 76 for all variables. **p ⬍ .01; ***p ⬍ .001.

on the EDSS, t(72) ⫽ ⫺4.53, p ⬍ .001. However, given that the EDSS was not a critical variable in the tested structural equation model, we do not view this t-test finding as particularly noteworthy in the context of other findings. Correlational and Regression Analyses We first conducted correlational analyses to determine whether neurovegetative symptoms were differentially related to CMDI-Mood versus EDSS scores. 1 Results revealed that four of five neurovegetative items were significantly (all ps ⬍ .01) associated with CMDI-Mood but not with the EDSS (Table 2). In order to determine the relative contribution of each neurovegetative symptom to CMDI-Mood, we next conducted a stepwise regression analysis using CMDI-Mood as the criterion variable and all five neurovegetative items as predictors.2 As displayed in Table 3, three neurovegetative items were significant (all ps ⬍ .05) independent predictors of CMDI-Mood scores: sleep disturbance, disinterest in sex, and decision-making difficulty. We additionally correlated neurovegetative items, CMDI-Mood, and FIS scores to determine the relationships among neurovegetative symptoms, depressed mood, and fatigue (Table 4). Results revealed that all neurovegetative items were associated with FIS scores (all ps ⬍ .01), with decision making difficulty and tiredness being significantly [t(74) ⫽ 3.52, p ⬍ .001, two-tailed; t(74) ⫽ 2.74, p ⬍ .01, two-tailed, respectively] more associated with the FIS than CMDI-Mood. One neurovegetative item, disinterest in sex, was more highly correlated (though not significantly so) with CMDI-Mood than the FIS. To examine the possibility that this neurovegetative item was associated with depressed mood after controlling for fatigue, we conducted a hierarchical regression analysis using disinterest in sex as the criterion variable and entering the FIS and CMDI-Mood in two

1Given the risk of familywise error as a result of conducting multiple analyses, we set the criterion for significance for all zero-order correlational analyses at p ⬍ .01. 2We used stepwise regression to explore the relationship between depressed mood and each of the neurovegetative symptoms. Using this procedure, our subject-to-variable ratio was approximately 15:1. As a result, our ratio falls short of some authors’ recommendations that any multiple regression procedure, particularly a stepwise regression analysis, should have a minimum subject-to-variable ratio of 20:1 to minimize the risk of statistical artifacts (Hays, 1994). Given the exploratory nature of this analysis (we are unaware of other studies that have examined the relationships among specific neurovegetative symptoms and depressed mood, and thus did not have any a priori predictions for these relationships), it will be important for future studies to replicate our findings using a larger sample.


393

TABLE 3 Stepwise Regression Analysis for BDINeurovegetative Items Predicting CMDI-Mood Variable

Step 1 Sleep disturbance Step 2 Disinterest in sex Step 3 Decision-making difficulty

B

SE B

␤

3.54

.92

.41***

2.06

.90

.26*

1.97

.98

.22*

BDI ⫽ Beck Depression Inventory; CMDI ⫽ Chicago Multiscale Depression Inventory. n ⫽ 76 for all variables. R2 ⫽ .17 for Step 1; ∆R2 ⫽ .06 for Step 2; ∆R2 ⫽ .04 for Step 3 (ps ⬍ .05). *p ⬍ .05; ***p ⬍ .001.

steps. Results indicated that CMDI-Mood was associated with disinterest in sex after entering the FIS in the first step (p ⬍ .01; ⌬R2 ⫽ .09), suggesting that depressed mood is associated with disinterest in sex after accounting for fatigue.

Structural Equation Modeling Analyses In order to conceptualize the relationships among neurovegetative symptoms, depressed mood, physical disability, and fatigue more clearly, we constructed a post-hoc structural equation model based upon our correlational findings. Further, in constructing the model we noted that EQS essentially tests a specified structural model by conducting a series of stepwise regression analyses simultaneously. As with regression analyses, the variance in each independent variable in a structural equation model is often accounted for by the first or second tested dependent variable in the specified model. Considering this logic, we constructed the post-hoc model by predicting that each independent variable in the model (i.e., depressed mood, fatigue, and physical disability) would only be associated with neurovegetative symptoms in the model when the zeroorder correlations (computed earlier) between the independent variable and neurovegetative symptoms were robust. For the model, we proposed that depressed mood would be uniquely associated with disinterest in sex compared to physical disability and fatigue, given that disinterest in sex was not associated with physical disability and was associated with depressed mood after controlling for fatigue in earlier analyses. Fatigue was predicted to be uniquely associated with decision-making difficulty; depressed mood and fatigue were both hypothesized to be associated with sleep disturbance and appetite change. Finally, depressed mood, fatigue, and physical disability were all proposed to be related to tiredness, given that tiredness was the only neurovegetative symptom with which all three predictors correlated.3 An initial test of the post-hoc model revealed a moderate fit to the data, ␹2/df ratio ⫽ 2.10 [␹2 (17, N ⫽ 76) ⫽ 35.74, p ⬍ .05], CFI ⫽ .86, IFI ⫽ .87, GFI ⫽ .90. In an attempt to develop a better-fitting model, we made additional modifications based on

3Although the correlation between tiredness and fatigue was more robust than between tiredness and depression or physical disability, we opted to include paths from fatigue, mood, and physical disability to tiredness since tiredness was the only neurovegetative symptom associated with all three upstream variables.

394

J. J. Randolph et al. TABLE 4 Intercorrelations Among BDI Neurovegetative Items, CMDI-Mood, and Fatigue

1. Decision-making difficulty 2. Sleep disturbance 3. Tiredness 4. Appetite change 5. Disinterest in sex

CMDI-Mood

FIS

.32** .41*** .28** .22 .38***

.64*** .39*** .55*** .27** .25

CMDI ⫽ Chicago Multiscale Depression Inventory; FIS ⫽ Fatigue Impact Scale. n ⫽ 76 for all variables. **p ⬍ .01; ***p ⬍ .001.

EQS Wald and LaGrange modification tests.4 Results revealed that the modified posthoc model (Figure 1) provided a better fit to the data, with two of four indexes (including the ␹2/df ratio) indicating a good fit and two indexes nearly reaching the .90 criterion for good model fit, ␹2/df ratio ⫽ 1.84 [␹2 (20, N ⫽ 76) ⫽ 36.76, p ⬍ .05], CFI ⫽ .88, IFI ⫽ .89, GFI ⫽ .90. Further, all paths in the revised post-hoc model were significant at or beyond the p ⬍ .05 level, with depressed mood being associated with disinterest in sex and sleep disturbance, and fatigue associated with decision-making difficulty, sleep disturbance, tiredness, and appetite change.

DISCUSSION The present study sought to examine and clarify the relationship between neurovegetative symptoms and depressed mood, fatigue, and physical disability in a sample of MS patients. We found that four of five neurovegetative symptoms of depression were associated with depressed mood but not physical disability, suggesting that some neurovegetative symptoms are reliably associated with depressed mood in MS patients and are not merely symptoms that overlap with physical/neurological sequelae of MS. We also found that all neurovegetative symptoms were associated with both fatigue and depressed mood, although one symptom—disinterest in sex—was associated with depressed mood after controlling for fatigue. Two other symptoms—decision-making difficulty and tiredness—were more associated with fatigue than depressed mood. Further, our revised post-hoc structural equation model of neurovegetative symptoms bolstered correlational findings and supported differential associations among neurovegetative symptoms and various categories of emotional and physical sequelae of MS. Perhaps the most notable finding within the structural model was the relationship between depressed mood and two neurovegetative symptoms: disinterest in sex and sleep disturbance. In particular, disinterest in sex was associated with depressed mood independent of both fatigue and physical disability, suggesting that assessment of this neurovegetative symptom may be diagnostically valuable in assessment of depression in MS and should not be considered as a symptom that merely overlaps with somatic symptoms of MS.

4EQS Wald and LaGrange tests indicate paths to delete from and to add to a model, respectively, to improve model fit.


395

FIGURE 1. Structural equation model of neurovegetative symptoms in multiple sclerosis. Standardized path coefficients and significance levels are shown. *p ⬍ .05; **p ⬍ .01; ***p ⬍ .001.

Although others (e.g, Nyenhuis et al., 1995) have argued that neurovegetative symptoms as a class should not be considered as depressive symptoms in MS patients, the current findings suggest that such a strategy may be misguided. Indeed, by not considering certain neurovegetative symptoms, valuable diagnostic information may be neglected when assessing affective disorders in MS patients. Relatedly, findings from the present study have implications for clinicians and researchers working with MS patients. In particular, when assessing for depression in MS, it may be relatively more important to consider the neurovegetative symptom of decreased interest in sex than other neurovegetative symptoms; this symptom appears to be uniquely associated with depressed mood compared with symptoms of physical disability and fatigue. Other neurovegetative symptoms may be associated with both depression and fatigue (e.g., sleep disturbance) or only with fatigue (e.g., decision-making difficulty, tiredness).

396


We found that fatigue correlated similarly to, or better than, depressed mood with most neurovegetative symptoms. Further, depressed mood and fatigue were robustly correlated in the present study. It is also important to note that fatigue and physical disability were moderately correlated [r (74) ⫽ .49, p ⬍ .001], whereas mood and physical disability were not correlated [r (74) ⫽ .12, p ⬎ .10]. Consistent with most research using the EDSS as a measure of physical disability in MS (Fischer et al., 1994), we did not find depressed mood to be related to physical disability. It is possible that the high prevalence of depression in MS patients, particularly compared to patients with other chronic disabling disorders (see Schiffer & Babigian, 1984), is due to other life stressors MS patients experience. Additional factors that might contribute to depression in MS include cognitive impairment and awareness of such impairment, uncertainty of the disease process, decreased social support, social isolation, and, as our data suggested, fatigue. Considering these data, perhaps it is sensible to consider depression and fatigue as overlapping but somewhat independent constructs rather than as two distinct classes of symptoms associated with MS. Future studies are needed to delineate the complex web of interrelated variables that contribute to depression in MS. One limitation of our study is that the tested structural equation model of neurovegetative symptoms was constructed and estimated after considering correlational findings rather than a priori; however, based on previous research and suppositions of researchers that all neurovegetative symptoms of depression should be discounted when assessing depression in MS, little rationale existed for which particular symptoms to consider as being specifically related to depression versus related to physical disability and fatigue. Future studies might test the present model a priori to determine whether it would generalize to other MS samples as well as to different neurological populations. Relatedly, although the revised post-hoc model was considered to be the most statistically optimal representation of the data in the present study, it is by nature exploratory. A revised model using SEM represents an efficient statistical solution; this solution may or may not be an accurate representation of how variables interact in “real life.” Thus, the generalizability and applied relevance of the revised post-hoc model is dependent on future research that is able to examine and replicate the present findings. Further, although the model in the present study implies that depressed mood and fatigue lead to various neurovegetative symptoms, it is possible that other configurations of variables also provide a reasonable fit to the data. Future research might amend the present study’s post-hoc model by examining possible mediating or moderating variables within such a model, particularly by examining the proposed relationships longitudinally. Another limitation of our study relates to SEM and sample size. Although guidelines for appropriate sample size when using structural equation modeling are not firmly established (Tanaka, Panter, Winborne, & Huba, 1990), an accepted “rule of thumb” is 10 subjects per path to be estimated (Bentler, 1995). Thus, our sample is slightly below this standard. However, it is important to note that: (a) we couch our results tentatively, and acknowledge that the ultimate utility of our model depends on cross-validation and replication with other MS samples; (b) our sample is larger than some studies in the MS literature (e.g., Beatty, Goodkin, Monson, & Beatty, 1989; Maurelli et al., 1992), and similar in size to others (e.g., Klonoff, Clark, Oger, Paty, & Li, 1991; Rao et al., 1991); (c) other recent studies have tested structural equation models with subject-estimated path ratios similar to ours (e.g., Kafer & Hunter, 1997); and (d) our use of SEM in the present study represents one of the few attempts to apply this powerful statistical technique common in other literatures to neuropsychological/neurological phenomena. Thus, we feel our use of SEM is appropriate in the present study. It is also important to note that neurovegetative symptoms may be associated with


397

conditions other than depressed mood, fatigue, and physical disability. For example, decision-making ability in MS patients may also be affected by executive functioning impairment above and beyond depressed mood and fatigue. Indeed, impaired executive functions of planning, self-direction, and conceptual reasoning are all likely to be associated with difficulty in making decisions; these executive functions have been found to be impaired in up to one third of MS patients (Arnett et al., 1997; Rao et al., 1991). Executive impairment may also impact patients’ ability to self-monitor and accurately report symptoms they are experiencing. Relatedly, it is possible that self-report data collected from MS patients such as those in the present study are not valid due to poor insight and more general executive impairment (Arnett et al., 1997). However, in a separate report using the same sample as that in the present study (Arnett, Higginson, Voss, Bender, et al., in press), significant other ratings of patients’ depression (using the BDI) was nearly identical to patients’ self-reported depression, suggesting that patients are capable of reporting their symptoms accurately. Additional studies might consider the role of variables such as executive functioning impairment on reported neurovegetative symptoms. A final caveat is that, although our study suggests that it may be misguided to discount neurovegetative symptoms in evaluating depression in MS, there may be certain situations where examining homogeneous clusters of depressive symptoms is warranted. For example, Arnett et al. (1999) recently reported that when depressed and nondepressed MS groups were defined using extreme scores on the CMDI-Mood subscale, the depressed group was significantly more impaired on several attentional capacity-demanding tasks compared with the nondepressed group. Although the depressed mood MS group also showed significantly more neurovegetative depressive symptoms compared with the nondepressed group, neurovegetative symptoms were uncorrelated to performance on the capacity-demanding attentional tasks. In summary, our study demonstrates that neurovegetative symptoms in MS are differentially associated with depressed mood, fatigue, and physical disability. We found that: (a) nearly all neurovegetative symptoms in MS are associated with depressed mood and not physical disability; (b) one neurovegetative symptom, disinterest in sex, appears to be uniquely associated with depressed mood in MS; and (c) fatigue, like depressed mood, is associated with various neurovegetative symptoms. We suggest that researchers and clinicians alike consider the present findings as a caution against minimizing the importance of all neurovegetative symptoms when assessing depression in MS.

REFERENCES Arnett, P. A., Higginson, C. I., Voss, W., Bender, W. I., & Wurst, J. M. (in press). Depression in multiple sclerosis: Relationship to working memory capacity. Neuropsychology. Arnett, P. A., Higginson, C. I., Voss, W., Wright, B., Bender, W. I., & Wurst, J. M. (1999). Depression in multiple sclerosis: Relationship to capacity-demanding memory and attentional functioning. Neuropsychology, 13, 434–446. Arnett, P. A., Rao, S. M., Grafman, J., Bernardin, L., Luchetta, T., Binder, J. R., & Lobeck, L. (1997). Executive functions in multiple sclerosis: An analysis of temporal ordering, semantic encoding, and planning abilities. Neuropsychology, 11, 535–544. Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182. Beatty, W. W., Goodkin, D. E., Monson, N., & Beatty, P. (1989). Cognitive disturbances in patients with relapsing-remitting multiple sclerosis. Archives of Neurology, 46, 1113–1119.

398


Beck, A. T., & Steer, R. A. (1987). BDI: Beck Depression Inventory manual. New York: Psychological Corporation. Beck, A. T., Steer, R. A., & Garbin, M. G. (1988). Psychometric properties of the Beck Depression Inventory: Twenty-five years of evaluation. Clinical Psychology Review, 8, 77–100. Bentler, P. M. (1990). Comparative fit indexes in structural models. Psychological Bulletin, 107, 238–246. Bentler, P. M. (1995). EQS: Structural equations program manual. Encino, CA: Multivariate Software, Inc. Bollen, K. A. (1989). Structural equations with latent variables. New York: John Wiley and Sons. Fischer, J. S., Foley, F. W., Aikens, J. E., Ericson, G. D., Rao, S. M., & Shindell, S. (1994). What do we really know about cognitive dysfunction, affective disorders, and stress in multiple sclerosis? A practitioner’s guide. Journal of Neurological Rehabilitation, 8, 151–164. Fisk, J. D., Pontefract, A., Ritvo, P. G., Archibald, C. J., & Murray, T. J. (1994). The impact of fatigue on patients with multiple sclerosis. Canadian Journal of Neurological Science, 21, 9–14. Hays, W. L. (1994). Statistics. Fort Worth, TX: Harcourt Brace College Publishers. Hogancamp, W. E., Rodriguez, M., & Weinshenker, B. G. (1997). The epidemiology of multiple sclerosis. Mayo Clinic Proceedings, 72, 871–878. Huber, S. J., & Rao, S. M. (1993). Depression in multiple sclerosis. In S. E. Starkstein & R. G. Robinson (Eds.), Depression in neurologic disease (pp. 84–96). Baltimore: Johns Hopkins University Press. Joreskog, K. G., & Sorbom, D. (1988). Lisrel 7, A guide to the program and applications. Chicago: SPSS. Kafer, K. L., & Hunter, M. (1997). On testing the face validity of planning/problem-solving tasks in a normal population. Journal of the International Neuropsychological Society, 3, 108–119. Klonoff, H., Clark, C., Oger, J., Paty, D., & Li, D. (1991). Neuropsychological performance in patients with mild multiple sclerosis. Journal of Nervous and Mental Disease, 179, 127–131. Kurtzke, J. F. (1983). Rating neurological impairment in multiple sclerosis: An Expanded Disability Status Scale (EDSS). Neurology, 33, 1444–1452. Lublin, F. D., & Reingold, S. C. (1996). Defining the clinical course of multiple sclerosis: Results of an international survey. Neurology, 46, 907–911. Maurelli, M., Marchioni, E., Cerretano, R., Bosone, D., Bergamaschi, R., Citterio, A., Martelli, A., Sibilla, L., & Savoldi, F. (1992). Neuropsychological assessment in MS: Clinical, neurophysiological and neuroradiological relationships. Acta Neurologica Scandinavica, 86, 124–128. Minden, S. L., Orav, J., & Reich, P. (1987). Depression in multiple sclerosis. General Hospital Psychiatry, 9, 426–434. Minden, S. L., & Schiffer, R. B. (1991). Depression and mood disorders in multiple sclerosis. Neuropsychiatry, Neuropsychology, and Behavioral Neurology, 4, 62–77. Nyenhuis, D. L., Luchetta, T., Yamamoto, C., Terrien, A., & Garron, D. C. (1996). The standardization and continued validation of the Multiscale Depression Inventory. Presented at the 1996 meeting of the International Neuropsychological Society, Chicago, IL. Nyenhuis, D. L., Rao, S. M., Zajecka, J. M., Luchetta, T., Bernardin, L., & Garron, D. C. (1995). Mood disturbance versus other symptoms of depression in multiple sclerosis. Journal of the International Neuropsychological Society, 1, 291–296. Poser, C. M., Paty, D. W., Scheinberg, L., McDonald, W. I., Davis, F. A., Ebers, G. C., Johnson, K. P., Sibley, W. A., Silberberg, D. H., & Tourtellotte, W. W. (1983). New diagnostic criteria for multiple sclerosis: Guidelines for research protocols. Annals of Neurology, 13, 227–231. Rao, S. M., Leo, G. J., Bernardin, L., & Unverzagt, F. (1991). Cognitive dysfunction in multiple sclerosis. I. Frequency, patterns, and predictions. Neurology, 41, 685–691. Schiffer, R. B., & Babigian, H. M. (1984). Behavioral disorders in multiple sclerosis, temporal lobe epilepsy, and amyotropic lateral sclerosis: An epidemiological study. Archives of Neurology, 41, 1067–1069. Tanaka, J. S., Panter, A. T., Winborne, W. C., & Huba, G. J. (1990). Theory testing in personality and social psychology with structural equation models. In C. Hendrick & M. S. Clark (Eds.), Review of personality and social psychology: Research methods in personality and social psychology (pp. 217–242). Newbury Park: Sage. Ullman, J. B. (1996). Structural equation modeling. In B. G. Tabachnick & L. S. Fidell (Eds.), Using multivariate statistics (pp. 709–811). New York: Harper Collins College Publisher. Whitlock, F. A., & Siskind, M. M. (1980). Depression as a major symptom of multiple sclerosis. Journal of Neurology, Neurosurgery, and Psychiatry, 43, 861–865. Zung, W. K. (1965). A self-rating depression scale. Archives of General Psychiatry, 12, 63–70.