Fatigue in immune-mediated polyneuropathies

Participants.

A group of 113 patients with a stable clinical condition (83 GBS, 22 CIDP, 8 MGUSP) were recruited from the Rotterdam immune-mediated polyneuropathy databank and the Dutch GBS study group (table 1). A stable clinical condition was required to obtain the highest reliability for the selected fatigue scale. The selected patients still had residual signs or symptoms resulting from illness, representing a broad range of disability (functional grading score [f-score] of at least 1; see also assessment scales6). None of the patients have been treated with steroids or another immune-suppressive agent in the 6 months before the start of the study. Nine CIDP patients required interval treatment, ranging from weeks to months, with intravenous immunoglobulins. With this therapy, their clinical condition was stable for more than 6 months. The patients with GBS and CIDP met the international criteria for their illness.1,2 The diagnosis of MGUSP was established after excluding all possible underlying causes for the gammopathy and polyneuropathy.7 Six patients with MGUSP had an associated demyelinating polyneuropathy with minor concurrent axonal damage in three. An axonal polyneuropathy was diagnosed in the remaining two patients with MGUSP. Patients were excluded from participation if there was any concomitant disease or use of medication that might cause chronic fatigue. All selected patients declared to have experienced only mild and transient fatigue before their illness.

Controls, stratified for gender and age, were recruited from hospital personnel, companions (relatives, friends) of patients visiting our outpatient clinic, and volunteers unfamiliar with the study (see table 1). The controls declared themselves to be healthy, free from any chronic medical condition, and were not taking medication that could contribute to fatigue.

Assessment scales.

The Fatigue Severity Scale (FSS), a brief and simple self-assessed questionnaire, was selected and translated into Dutch before its use.8 The FSS is a nine-item questionnaire with answers ranging from 1 (strongly disagree) to 7 (strongly agree) for each inquiry. The mean score of the 9 inquiries ranges from 1 (no signs of fatigue) to 7 (most disabling fatigue). Internal consistency, reliability, validity, and sensitivity of the FSS have been established in patients with systemic lupus erythematosus, MS, and Lyme disease but never in patients with immune-mediated polyneuropathies.8

The Medical Research Council (MRC) sumscore and functional grading scale (f-score) were applied in the patient group to measure the physical condition of the patients at the impairment and disability levels.6 The MRC sumscore ranges from 0 (paralysis) to 60 (normal strength). For each muscle group, a standardized joint/limb position as well as the point at which counter-force is administered was predefined and taken when assessing muscle strength. The f-score of the patients included in this study ranges from 1 (minor neurologic symptoms or signs and able to run) to 4 (bed or chair bound). Good reliability and validity have been demonstrated for these two scales.6

The INCAT sensory sumscore ranges from 0 (normal sensation) to 20 (most severe sensory deficit) and is composed by the summation of the following sensation qualities: pinprick arm grade (range 0 to 4) + vibration arm grade (range 0 to 4) + pinprick leg grade (range 0 to 4) + vibration leg grade (range 0 to 4) + 2-point discrimination grade (range 0 to 4). The vibration examination was performed using the recently validated graduated Rydel-Seiffer tuning fork, and the obtained measures were compared with the reported normative threshold values.9 Pinprick and vibration sense examination took place from distal to proximal, and only the highest extension of dysfunction of the most affected arm and leg was recorded for both qualities. The sites of examination with corresponding grades were defined as followed: normal (grade 0) or disturbed (grade 1) pinprick or vibration sense at the dorsum distal interphalangeal joint of the index finger or hallux; and abnormal sense at the ulnar styloid process or medial malleolus (grade 2), at the medial humerus epicondyle or patella (grade 3), and at acromioclavicular joint or anterior superior iliac spine (grade 4). For the 2-point discrimination quality, a sliding esthesiometer was used where the exact measurable distance in millimeters could be read on the instrument. This instrument was assessed in a “static” manner at the ventral side, distal phalanx of the index finger, and the corresponding grades were arbitrarily chosen (grade 0: ≤4 mm; grade 1: 5 to 9 mm; grade 2: 10 to 14 mm; grade 3: 15 to 19 mm; grade 4: ≥20 mm).

The medical outcomes study 36-item short form health survey (SF-36), a generic health status questionnaire, consists of 36 items assigned to the domains of physical functioning (10 items), role limitations—physical (4), role limitations—emotional (3), social functioning (2), body pain (2), mental health (5), vitality (4), general health perception (5), and change in health which is scored separately.10 The numbers of response categories per item range from two to six. Each domain has a scoring range from 0 to 100. A high score indicates better health or less body pain. The Dutch version we used was developed as a part of the International Quality of Life Assessment Project, which aims to translate, validate, and determine norms for the SF-36 in a range of languages and cultural settings.11

Test procedure.

All participants gave informed consent for the study. Participants were lucid and competent to answer the questionnaires to the best of their ability. All individuals received brief instructions on how to fill in the fatigue and health status forms. These questionnaires were answered in random order. Patients were examined at our outpatient clinic, and all scales were completed at study enrollment. The FSS was mailed 4 to 6 months later to the patients for a second assessment. The patients also were asked to report whether their clinical condition had improved, gotten worse, or remained the same since completing the first questionnaire. Patients whose health had remained the same were included in the FSS test/retest reliability analyses. The FSS was completed once by all healthy controls. SF-36 was self-administered by 59 controls, taking into account the age and sex distribution of the patients. The study took place between March 1997 and August 1998 and was part of a more comprehensive outcome assessment research in patients with immune-mediated polyneuropathies on behalf of the European Inflammatory Neuropathy Cause and Treatment Group.

Statistics.

“Severe” fatigue was arbitrarily defined as a score ≥FSS 95th percentile in healthy individuals. The FSS scores obtained in the patients and in healthy controls were compared using the Mann-Whitney U test and univariate/multivariate quantile regression analyses.12 The median FSS values were compared within the categories (healthy controls, GBS, CIDP, and MGUSP patients) using design variables (0 to 1) representing each category. The design variables were defined in an appropriate manner, making it possible to test (by the likelihood ratio test) the hypothesis that the median fatigue values were equal for the diagnoses of CIDP and MGUSP but unequal to the median fatigue values in GBS patients and healthy controls, separately. Box plots, with the upper and lower adjacent values defined according to Tukey, were applied to visualize some of these observations (also see the legend to figure 1 for explanation of the various components of a box plot).13

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Box plots demonstrating the obtained Fatigue Severity Scale (FSS) scores in the healthy controls versus patient group (A) related to diagnosis (B) and sex (C). The patients were more fatigued than the healthy controls (Mann-Whitney U test, whole patient group versus healthy controls: p < 0.0001). All diagnostic categories were more fatigued than the healthy controls (likelihood ratio test combined with quantile regression analysis: p < 0.0001). The GBS patients had a higher median fatigue value compared with the other subgroups and healthy controls (likelihood ratio test combined with quantile regression analysis: p < 0.0001 for the comparisons GBS versus CIDP + MGUSP, and GBS versus healthy controls). Women were more tired in both groups, but this was significant only in the patient group (Mann-Whitney U test; in healthy controls: p = 0.28; in patient group: p = 0.02). HC = healthy controls; HM = healthy males; HF = healthy females; MP = male patients; FP = female patients. The various components of the box plots can be explained as follows. The median (50th percentile) FSS value corresponds to the horizontal line within the box. The box extends from the first quartile (25th percentile = x^[25]) to the third quartile (75th percentile = x^[75]), the so-called interquartile range (IQR). The vertical lines emerging from the box extend to the upper and lower adjacent values. The upper adjacent value is defined as the largest FSS value less than or equal to x^[75] + 1.5 × IQR. The lower adjacent value is defined as the smallest FSS value greater than or equal to x^[25] − 1.5 × IQR. Measured FSS values more extreme than the adjacent values are referred to as “outside values” and are individually plotted.13

The mean SF-36 subscale values were compared within the three diagnostic categories and between the whole patient group and healthy controls for each domain. One-way analysis of variance (ANOVA) with corrections according to Bonferroni multiple-comparison tests were applied.

The Cronbach’s alpha coefficient was estimated for the FSS in the patients and healthy participants.14 The test/retest reliability for the recruited FSS values in the polyneuropathy patients was quantified by estimation of the intraclass correlation coefficient using a one-way random effects analysis of variance model. The correlation between the FSS and other scales was analyzed using Spearman rank-correlation test. All analyses were performed using Stata 5.0 for Windows 95 (Stata Statistical Software, Stata Corporation, release 5.0, College Station, TX 1997). A value of p ≤ 0.05 was considered to be statistically significant.

Characteristics of participants.

The characteristics of all participants are presented in table 1. One hundred thirteen patients and 113 age- and sex-matched healthy individuals were enrolled. The CIDP patients were relatively younger (mean age 45.0 [SD 18.0] years) compared with the GBS (mean age 55.7 [SD 13.6] years) and the MGUSP patients (mean age 66.1 [SD 8.1] years). One hundred seven of the 113 patients (95%) completed the second FSS assessment. One hundred five patients reported a stable clinical condition, whereas two reported an improved clinical condition at the second FSS assessment. These two were excluded from reliability analyses. Two patients died during the interval between the two measures, and another was hospitalized because of myocardial infarction. The remaining three patients did not return the second FSS assessment for various reasons. Fifty-nine healthy volunteers (29 women, 30 men; mean age 53.5 [SD 13.6] years) completed the SF-36 questionnaire.

FSS comparison between patients and controls.

At entry, the median FSS value in the patient group was slightly higher (6.1) (corresponding mean value: 5.6 [SD 1.4]) compared with the median value at the second assessment (5.8) (corresponding mean value: 5.5 [SD 1.4]). Unless otherwise stated, further comparisons were made using only the FSS values obtained at entry. The median FSS value in the patient group was higher compared with the result in healthy individuals (median fatigue value 2.9; corresponding mean value: 2.9 [SD 1.1]) (figure 1A; Mann-Whitney U test: p < 0.0001). By the likelihood ratio test, it was demonstrated that the median FSS values were equal for the CIDP and MGUSP patients but unequal to the median fatigue values in healthy controls and GBS patients. This test also revealed that the GBS patients had a higher median fatigue value (6.2) compared with the other subgroups and healthy controls (CIDP 5.6, MGUSP 5.3, healthy controls 2.9; figure 1B; quantile regression analyses combined with likelihood ratio test; p < 0.0001 for the comparisons GBS versus CIDP + MGUSP, and GBS versus healthy controls).

The FSS 95th percentile in the healthy controls was 5.0. Severe fatigue, arbitrarily defined as a FSS score more than the FSS 95th percentile in the healthy controls, was demonstrated in 90 of 113 patients (80%). Remarkably, of the patients with a completely recovered general strength (MRC sumscore 60) relatively more patients reported a fatigue score that fell within the severe fatigue range (19 of 22, 86%). Severe fatigue also was reported in 17 of the 21 (81%) patients with normal sensation (sensory sumscore 0). Six of the eight patients (75%) with both normal general strength and sensation experienced severe fatigue. Eighty percent of the patients scored 5 or more on the 7-point scale for question 8 (“Fatigue is among my three most disabling symptoms”) compared with 12% (13/113) of the healthy controls.

FSS correlation with clinical parameters.

General strength, as measured by the MRC sumscore, the f-score values, and sensory sumscore, showed no significant association with the FSS values (Spearman rank correlation test: r = 0.07 to 0.13; p ≥ 0.18). Fatigue score was not significantly related to age and duration of symptoms in the patient group. Fatigue in women was consistently higher than in men, but this was significant only in the patient group (figure 1C; Mann-Whitney U test in healthy controls: p = 0.28; in patient group, p = 0.02).

SF-36 evaluation and its association with FSS in patients.

The SF-36 values showed no substantial difference between the three diagnostic categories (GBS, CIDP, and MGUSP) within each domain (ANOVA with Bonferroni tests: all p values ≥ 0.46). A separation was made between the GBS (acute subgroup; n = 83) and the CIDP and MGUSP patients (chronic subgroup; n = 30) based on these results combined with the FSS likelihood ratio test findings (also see FSS Comparison Between Patients and Controls in the Results section). For all SF-36 domains, both subgroups scored notably lower values compared with the scores obtained in healthy controls (figure 2; ANOVA with Bonferroni tests: p < 0.0001 for all domains, except for the dimension “role functioning—emotional” in the chronic subgroup versus controls: p = 0.03).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Short Form-36 (SF-36) health status in patients with immune-mediated polyneuropathies and healthy controls. Both patient subgroups demonstrated a lower score for all SF-36 domains compared with the scores of the healthy controls (p < 0.0001 for all domains, except “role functioning—emotional” in the CIDP/MGUSP patients versus controls: p = 0.03). No statistically significant difference was obtained at all domain levels between the GBS and the CIDP/MGUSP patients (all p values ≥ 0.27). Analyses were performed using ANOVA with corrections according to the Bonferroni multiple-comparison tests.

The correlations between the FSS and SF-36 domains for the two subgroups are presented in table 2. In both subgroups, FSS demonstrated the highest significant inverted association with the domains “general health perception” and “vitality.” Fatigue scores in GBS patients were significantly associated with the socioemotional domains but not with the physical subscales. Conversely, fatigue values in the chronic subgroup were significantly related to the physical (but not the socioemotional) dimensions (table 2, boxes).

Internal consistency, reliability, and validity evaluation of the FSS.

A good internal consistency for the FSS was found at both assessments for the whole patient group (α = 0.93 and 0.95, respectively) and in healthy controls (α = 0.84). The FSS demonstrated good test/retest reliability (R = 0.86) and construct validity when correlated with the SF-36 vitality domain (r = −0.56, p < 0.0001).