The Fatigue Severity Scale
Application to Patients With Multiple Sclerosis
and Systemic Lupus Erythematosus
Lauren B. Krupp, MD; Nicholas G. LaRocca, PhD; Joanne Muir-Nash, RN; Alfred D. Steinberg, MD
\s=b\ Fatigue is a prominent disabling
symptom in a variety of medical and neu-
rologic disorders. To facilitate research in
this area, we developed a fatigue severity
scale, subjected it to tests of internal con-
sistency and validity, and used it to com-
pare fatigue in two chronic conditions:
systemic lupus erythematosus and multi-
ple sclerosis. Administration of the fatigue
severity scale to 25 patients with multiple
sclerosis, 29 patients with systemic lupus
erythematosus, and 20 healthy adults re-
vealed that the fatigue severity scale was
internally consistent, correlated well with
visual analogue measures, clearly differ-
entiated controls from patients, and could
detect clinically predicted changes in fa-
tigue over time. Fatigue had a greater de-
leterious impact on daily living in patients
with multiple sclerosis and systemic lupus
erythematosus compared with controls.
The results further showed that fatigue
was largely independent of self-reported
depressive symptoms and that several
characteristics could differentiate fatigue
that accompanies multiple sclerosis from
fatigue that accompanies systemic lupus
erythematosus. This study demonstrates
(1) the clinical and research applications
of a scale that measures fatigue severity
and (2) helps to identify features that dis-
tinguish fatigue between two chronic med-
ical disorders.
(Arch Neurol. 1989;46:1121-1123)
Accepted for publication April 20,1989.
From the Departments of Neurology, State
University of New York at Stony Brook (Dr
Krupp) and Albert Einstein College of Medicine,
Bronx, NY (Dr LaRocca); and the National Insti-
tute of Arthritis and Musculoskeletal and Skin
Diseases, Bethesda, Md (Ms Muir-Nash and Dr
Steinberg).
Read in part before the 40th annual meeting of
the American Academy of Neurology, Cincinnati,
Ohio, April 19,1988.
Reprint requests to Department of Neurology,
School of Medicine, Health Sciences Center, State
University of New York at Stony Brook, Stony
Brook, NY 11794-8121 (Dr Krupp).
Downloaded From: http://archneur.jamanetwork.com/ by a University of Manitoba User on 06/05/2015
T^atigue, long recognized as a major
symptom in neurology and in med¬
icine, has been notoriously difficult to
define or study as a distinct entity. Al¬
though it is nonspecific and highly
subjective, fatigue is a prominent com¬
plaint in many disorders. In both mul¬
tiple sclerosis (MS) and systemic lupus
erythematosus (SLE), for example, it
may be the presenting symptom or a
chronic and disabling problem.13 Re¬
cently, a chronic fatigue syndrome has
been described in which fatigue is the
overwhelming feature.4 Whereas in¬
terest in the role of fatigue in cognitive
dysfunction and in distinguishing it
from somatization disorders or affec¬
tive disorders has grown,5 our under¬
standing has been hampered by the
lack of suitable techniques for its mea¬
surement. To facilitate research and
patient treatment, we have developed
a fatigue severity scale (FSS) that as¬
sesses disabling fatigue across two
different clinical disorders. Use of this
scale has helped elucidate the relation¬
ship between fatigue and depressive
symptoms and has identified features
of fatigue that might be characteristic
of specific diseases.
PATIENTS, MATERIALS, AND METHODS
The samples consisted of 25 patients with
chronic progressive MS with Expanded Dis¬
ability Status Scale scores ranging from 3.0
to 6.5, 29 patients with SLE, and 20 normal
healthy adults (NHAs). Patients chosen for
this study were outpatients of the National
Institute of Arthritis and Musculoskeletal
and Skin Diseases of the National Insti¬
tutes of Health, Bethesda, Md, or the Re¬
search and Training Center for Multiple
Sclerosis at the Albert Einstein College of
Medicine, Bronx, NY. Criteria for entry into
the study included a diagnosis of definite
MS as defined by the Schumacher et al' cri¬
teria or SLE as defined by the American
Rheumatology Association.7 Patients hav¬
ing other chronic medical conditions in ad¬
dition to MS or SLE were excluded. Healthy
controls were selected from volunteers un¬
familiar with this study.
Subjects were chosen consecutively from
clinic outpatients, were interviewed by one
of us (L.B.K. or J.M.N.), and were adminis¬
tered a 28-item fatigue questionnaire. Sub¬
jects were asked to read each statement of
the questionnaire and choose the number
from 1 to 7 that best described their degree
of agreement with each statement: 1 indi¬
cates strongly disagree and 7 indicates
strongly agree. Using factor analysis, item
analysis, and theoretical considerations, 9
items from the fatigue questionnaire were
selected that identified features of fatigue
common to the MS and SLE patient groups.
This 9-item scale was then employed as the
FSS. It had a high degree of internal con¬
sistency as measured by Cronbach's alpha,
an estimate of the reliability of a scale
based on intercorrelation of the individual
items of a multi-item scale. Scores obtained
by healthy controls on the FSS were com¬
pared with scores from the patient groups.
In addition to the FSS, fatigue was as¬
sessed with a 100-mm visual analogue scale
(VAS). Patients were asked to indicate on a
line the point that best described their fa¬
tigue, and then this distance was scored
from 0 to 1.0. Depressive symptoms were
assessed with the Center for Epidemiologie
Studies Depression (CES-D) Scale. This is a
self-report, 20-item screening question¬
naire in which the score ranges from 0 to 60.
A score of 16 or greater is indicative of
clinical depression.8
To assess the ability of the FSS to detect
clinical changes in fatigue over time, the
FSS was administered to a second nonran-
dom sample of patients in whom changes in
fatigue were clinically predicted. Two pa¬
tients with MS were given the FSS before
and after drug therapy for fatigue. Six pa¬
tients with clinical histories consistent wih
the diagnosis of Lyme disease and with im¬
munologie evidence of reactivity to Borrelia
burgdorferi were administered the FSS be¬
fore and after a 3-week course of antibiotic
therapy.
RESULTS
Patient Characteristics
All subjects were interviewed in per¬
son regarding their fatigue. Patient
and control characteristics are shown
in Table 1. Scores above the CES-D
cutoff for depression were more com-
 Table 1.—Characteristics of Study Samples* Table 4.—Sensitivity of Fatigue
Severity Scores (FSS) to Clinical
SLE, MS, NHAs,
Mean ± SD Mean ± SD Mean ± SD Changes and Measures of Its
Age, y 35.6 ± 8.9 44.8+10 39.7 ± 9 6.48 <.01t Consistency Over Time
CES-D 16.0 ± 10.9 14.1 ± 11 6.75 ± 7.2 5.46 <.01 Time 1, Time 2, Weeks,
VAS 0.58 ± 0.31 0.58 ± 0.20 0.16 ± 0.18 <.001§
16.38 Experi- Mean Mean Mean
*SLE indicates systemic lupus erythematosus; MS, multiple sclerosis; NHAs, normal healthy adults; CES-D, ment ±SD ±SD ±SD
Center for Epidemiologie Studies Depression Scale scored from 0 to 60 where 16 or greater correlates with de¬ 1*
pression; and VAS, visual analogue scale of fatigue scored 0 to 1.0. (n 8)
=
5.7±0.8 3.6±1.2t 16.9±9.3
tAnalysis of variance for SLE, MS, and NHA; the following contrasts were significant: SLE vs MS.
tAnalysis of variance for SLE, MS, and NHA: the following contrasts were significant: SLE, vs NHA and MS (n=11) 5.1±1.3 5.2±1.1§ 10.0±11.3
vs NHA. *
§Analysis of variance for SLE, MS, and NHA: the following contrasts were significant: SLE vs NHA and MS experiment, eight subjects (six patients
In the first
vs NHA. with Lyme disease and two patients with multiple
sclerosis [MS]) were given the FSS at two time points
separated by a mean of 16.9 weeks. All subjects re¬
ceived medication intended to reduce their fatigue.
tr(7) 2.16;P< .01, Student's two-tailed paired
=

Table 2—Fatigue Severity Scale (FSS)' t test.

tin the second experiment, 11 subjects (five pa¬
Statement tients with systemic lupus erythematosus and six pa¬
My motivation is lower when I am fatigued.
1. tients with MS) were studied for fatigue in situations
2. Exercisebrings on my fatigue. in which no change in fatigue was clinically antici¬
3. I am easily fatigued.
pated and medication for fatigue was not adminis¬
tered.
4. Fatigue interferes with my physical functioning. §Paired r test differences were not significant. /
5. Fatigue causes frequent problems for me. (10) .08; correlation coefficient was r .84,
= =

6. My fatigue prevents sustained physical functioning. p< .01.

7. Fatigue interferes with carrying out certain duties and responsibilities.
8. Fatigue is among my three most disabling symptoms.
9. Fatigue interferes with my work, family, or social life. nificant correlation of FSS with VAS
'Patients are instructed to choose a number from 1 to 7 that indicates their degree of agreement with each scores (r =
.68; < .001).
statement where 1 indicates strongly disagree and 7, strongly agree.
Sensitivity, Reliability, and Internal
Consistency of the FSS
FSS in MS, SLE, and NHAs
Table 3.—Comparison of Fatigue To assess the potential utility of the
Severity Scale (FSS) Across The 9 items selected from the 28- FSS for clinical research and patient
Samples* item fatigue questionnaire to form the treatment, the ability of the scale to
FSS are shown in Table 2. These items detect clinically appropriate and pre¬
SLE MS NHA were selected on the basis of their dicted changes in fatigue was exam¬
FSS, ability to identify common features of ined. The FSS was administered to six
mean±SD 4.7±1.5 4.8±1.3 2.3±0.7
Cronbach's
fatigue in both MS and SLE. The patients with Lyme disease before and
.89 .88 fatigue scores for patient and control after antibiotic therapy (Table 4). Two
alpha .81
Correlation groups along with their respective patients with MS were treated with
with VAS .81t .47 .50 Cronbach's alphas are shown in Table pemoline, a central nervous system
*SLE indicates
systemic lupus erythematosus; 3. The mean FSS for patients with MS stimulant (Table 4). In all patients,
MS, multiple sclerosis; NHA, normal healthy adults; was 4.8 and was not significantly dif¬ clinical improvement in fatigue was
and VAS, visual analogue scale. ferent from the mean FSS for patients associated with reductions in FSS. The
\P< .001.
*P< .05.
with SLE of 4.7. In contrast, even when magnitude of change on the FSS varied
adjusted for differences in age and de¬ for individual patients, however, rang¬
pression, patient scores were signifi¬ ing from 3.5 to 0.5 points.
mon among the
patients with MS and cantly higher than the scores for nor¬ The individual Cronbach's alphas
SLE than the NHA controls. Forty- mal subjects (P < .001). for the subpopulations of subjects
eight percent (12/25) of the patients Discriminant function analysis was (shown in Table 3) were .89 for SLE, .81
with MS, 39% (11/28) of the patients used to evaluate the ability of the nine for MS, and .88 for NHAs. The FSS had
with SLE, but only 15% (3/20) of items to distinguish patients from con¬ a high internal consistency as demon¬
healthy controls had CES-D scores of trols. Separate discriminant functions strated by a Cronbach's alpha of .88 for
16 or greater corresponding to clinical were calculated for subjects with MS the entire sample. Test-retest reliabil¬
depression. As shown in Table 1, the and SLE and NHAs. These discrimi¬ ity of the scale over time was examined
mean CES-D score for patients with nant functions correctly classified 98% in five patients with SLE and six pa¬
MS was 14.1 and for patients with SLE (MS vs NHA) and 90% (SLE vs NHA) tients with MS in whom there was no
it was 16.0 vs 6.75 for the NHA con¬ of the subjects, respectively. Five per¬ clinical reason to expect changes in
trols. cent (1/20) of the controls had a score their fatigue state. The patients were
On the VAS for fatigue, patients of 4 or higher in comparison with 60% tested at two time points separated by
with SLE and MS had identical mean of patients with SLE and 91% of pa¬ 5 to 33 weeks. No significant changes in
values of 0.58. These scores were sub¬ tients with MS. Thus, the nine-item FSS scores were noted (Table 4).
stantially higher than the VAS for the scale clearly distinguished patients
Relationship Between FSS and
NHA controls, 0.16. After correcting from controls. To further validate the Depression
for age and depression scores, the dif¬ FSS, a correlation was sought between
ference between the VAS of controls FSS scores and VAS scores. For the Having established the scale's inter¬
and patients was still significant. entire sample, there was a highly sig- nal consistency and reliability, we used

Downloaded From: http://archneur.jamanetwork.com/ by a University of Manitoba User on 06/05/2015


Table 5.—Percentages of Patients With SLE and MS With Positive
on
Itemst
Heat exacerbates fatigue
Cool temperatures improve
fatigue
Fatigue predates other disease
symptoms
Fatigue is the most disabling
_
disease symptom
Fatigue makes other disease
symptoms worse
*
SLE indicates systemic lupus erythematosus; MS, multiple sclerosis.
tltems were selected from the 28-item fatigue questionnaire administered to all subjects (see text).
^Percentages of patients responding with a "5," "6," or "7" on questionnaire where 1 indicates strongly dis¬
agree and 7, strongly agree.
the FSS to analyze the relationship
between fatigue and depression. The
CES-D was used as a measure of de¬
pressive symptomatology. The corre¬
lations between FSS and CES-D were
weak. Among patients with MS, the
correlation coefficient was .26 and
among controls it was .20; neither
reached statistical significance.
Among the patients with SLE, how¬
ever, there was a statistically signifi¬
cant correlation between fatigue and
depression (r =
.46; < .05).
Characteristics of Fatigue Distinguishing
MS From SLE
Analysis of the responses on the 28-
item questionnaire provided addi¬
tional information on qualities of fa¬
tigue that distinguished patients with
MS from patients with SLE. Five items
that distinguished MS from SLE are
shown in Table 5. None of the five items
were part of the nine-item FSS. Cer¬
tain fatigue characteristics were
present more often in patients with
MS than in patients with SLE. Among
patients with MS, fatigue was exacer¬
bated by heat, improved by cool tem¬
peratures, and made other disease
symptoms worse. In contrast, fatigue
was more likely to predate other symp¬
toms of SLE and was more frequently
reported as the most disabling prob¬
lem in patients with SLE.
COMMENT
Among patients with MS and SLE,
fatigue is a common and frequently
disabling symptom. On a nine-item fa¬
tigue scale, patients with MS and SLE
had similar degrees of fatigue severity
with scores that were significantly
higher than the scores for healthy
controls. The FSS could distinguish
MS or SLE from the NHA controls
with a high degree of accuracy. Hence,
Downloaded From: http://archneur.jamanetwork.com/ by a University of Manitoba User on 06/05/2015
Responses to Items
Fatigue Questionnare*
SLE MS
37.9* 92.0
37.9 84.0 10.0 .0016
64.3 8.7
.0002
58.6 26.1 .0389
35.7 76.2 6.4 .0117
the scale was successful in identifying
features of fatigue that are specific to
the medically ill.
Previous studies of fatigue in medi¬
cally ill populations have relied on
self-report visual analogue measures
in which subjects choose a number
from 1 to 5 that reflects their degree of
fatigue.910 Although a useful approxi¬
mation of a quantitative approach,
this method fails to assess the impact
of fatigue on daily functioning and is
vulnerable to impulsive answers. In
contrast, the nine items of the FSS re¬
quire somewhat more reflection. More¬
over, a scale consisting of multiple
items is inherently more reliable than
a single indicator. The reliability of the
FSS was suggested by its high Cron¬
bach's alpha and stability of scores
over time. The FSS clearly distin¬
guished patients from controls. Fur¬
thermore the scale proved to be a sen¬
sitive method for detecting clinically
predicted improvements in fatigue in
all of the eight subjects tested.
Analyzing fatigue in these patients
revealed that fatigue severity was
largely independent of depressive
symptoms common among patients
with MS and SLE. High CES-D scores
were common among patients with MS
and SLE. Nevertheless, the severity of
fatigue did not have a statistically sig¬
nificant correlation with CES-D scores
in patients with MS or in NHAs. In the
patients with SLE, there was a positive
relationship between fatigue and de¬
pression. These data suggest that fa¬
tigue and depression are separate, al¬
though overlapping, entities in both
chronic illnesses as well as in healthy
adults. They do not support the notion
that fatigue can be largely explained
as simply a manifestation of depres¬
sion. The results also suggest that
there are qualitative features of fa¬
tigue that distinguish MS from SLE.
As has been reported previously1-2 heat
made fatigue accompanying MS worse
while cool temperatures helped. This
was not as true for patients with SLE.
Fatigue that accompanies multiple
sclerosis also made other MS symp¬
toms worse while patients with SLE
did not report such an effect. However,
patients with SLE were more likely to
experience fatigue as their first symp¬
tom of disease and more frequently
stated that fatigue was one of their
most disabling symptoms.
In summary, an FSS has been tested
in samples of patients with MS and
SLE and in NHA controls. The FSS
should prove useful in the evaluation
of fatigue in both clinical research
studies and surveys. Its brevity and
simple self-report format make it a
cost-effective alternative to more elab¬
orate methods. As a brief self-report of
fatigue severity, the FSS has ac¬
ceptable internal consistency, stability
over time, and sensitivity to clinical
changes. While the FSS consists of
characteristics common to the two
conditions tested (MS and SLE), it is
also accurate in distinguishing pa¬
tients with either of these conditions
from normal controls.
This investigation was supported in part by
grant H133B80018 from the National Institute on
Disability and Rehabilitation Research, Bethes-
da, Md, and by grant RDP 12-D-70287-2-01 from
the Social Security Administration, Washington,
DC.
Joseph Schwartz, PhD, made helpful comments
and Irene London gave assistance in preparing
the manuscript.
References
1. Freal JG, Kraft GH, Coryell JK. Symptom-
atic fatigue in multiple sclerosis. Arch Phys Med
Rehabil. 1984;5:135-137.
2. Krupp LB, Alvarez LA, LaRocca NG, Schein-
berg LC. Fatigue in multiple sclerosis. Arch Neu-
rol. 1988;45:435-437.
3. Steinberg AD. Systemic lupus erythemato-
sus. In: Wyngaarden JB, Smith LH, eds. Cecil
Textbook of Medicine. Philadelphia, Pa: WB
Saunders Co; 1988:211-218.
4. Holmes GP, Kaplan JE, Gantz NM, et al.
Chronic fatigue syndrome: a working case defini-
tion. Ann Intern Med. 1988;108:387-389.
5. Manu P, Lane TJ, Mathews DA. The fre-
quency of the chronic fatigue syndrome in pa-
tients with symptoms of persistent fatigue. Ann
Intern Med. 1988;109:534-556.
6. Schumacher GA, Beebe G, Kibler RF, et al.
Problems of experimental trials of therapy in
multiple sclerosis. Ann NY Acad Sci. 1965;
122:552-568.
7. Tan EM, Cohen AS, Fries JF, et al. The 1982
revised criteria for the classification of systemic
lupus erythematosus. Arthritis Rheum. 1982;
25:1271-1277.
8. Radloff LS. CES-D scale: a self-report de-
pression scale for research in the general popula-
tion. Psychol Meas 1977;1:385-401.
9. Cardenas DD, Rutner NG. The problem of
fatigue in dialysis patients. Nephron. 1982;30:
336-340.
10. Murray TS. Amantadine therapy for fa-
tigue in multiple sclerosis. J Neurol Sci. 1985;
12:251-254.