Results of the investigation of three apparent clusters of chronic fatigue syndrome (CFS) occurring in three well-defined populations are reported to illustrate pitfalls to be avoided in the investigation of future clusters. We conducted multidisciplinary investigations of three clusters of CFS in very different, well-defined populations: members of a symphony orchestra, teachers at an elementary school, and members of a convent.
The first cluster was brought to our attention because of the diagnosis of non-Hodgkins lymphoma in one of the members of tie orchestra. The second cluster was brought to our attention by a patient who knew of our interest in CFS clusters. Since both clusters occurred before the development of a consensus regarded the definition of CFS, a case of CFS was defined by a physicians diagnosis. Through a review of their medical records, we found that five numbers of the orchestra met the current Centers for Disease Control and Prevention (CDC) research case definition of CFS.
The residual three cases met Group 2 criteria according to the latest
CDC definition. In the school, where review of teachers
medical records was not possible, we used questionnaire data to determine that three teachers met the diagnostic criteria for
CFS, and four did not. Of the four cases who did not meet the criteria, one was later diagnosed as having rheumatic fever; the
other three met Group 2 criteria of the recent CDC research case definition of CFS. The third cluster was in a convent where we made a more intensive effort to define each case by making detailed clinical evaluations of severely affected individuals,
However, we report the findings for this group to illustrate how individuals
in an apparent CFS cluster tend to mistakenly label
other diseases as CFS. This cluster was brought to our attention by a colleague who treated two of the nuns and knew of our
interest in CFS clusters.
ln the third study, only assays of viral immunity were done, whereas
in the first two clusters laboratory studies included as-
says of cellular immunity (natural killer [NK] cell activity, T cell assays, mitogen stimulation, etc.) and viral immunity (antibod-
ies to Epstein-Barr virus [EBV], human herpesvirus type 6 [HHV-6], cytomegalovirus, and human T cell leukemia virus
I/II). For reasons of similarity in age, race, and socioeconomic status as well as convenience, we decided to compare the members of the orchestra with CFS with the unaffected members of the orchestra to establish a control group. We did not anticipate that we would uncover similar laboratory findings for both affected and unaffected cohort members.
Our first investigation focused on eight aflected members of the 67-member orchestra. Laboratory findings for many of the unaffected members of the orchestra were found to be similar to those for members with CFS (elevated EBV antibody titers and a low level of NK cell activity). In fact, the only laboratory finding that was significantly different in the two groups was lower NK cell activity in members with CFS. This difference occurred despite the fact that very low levels of NK activity were found for several of the unaffected members. NK cell activity was measured in terms of lytic units. The mean NK cell activity among members with CFS was 50.8 lytic units (LU) vs. 134.5 LU in noncases and 235 LU in laboratory controls. Results from the study of this cluster suggest that comparisons between individuals with CFS and their close contacts is difficult because unaffected individuals may show similar patterns of laboratory abnormalities compatible with perturbations of immune function. This finding suggests transmission of an infectious agent or exposure to a common source of a noninfectious causal agent that might cause similar laboratory abnormalities in symptomatic and asymptomatic individuals. These data also demonstrate that laboratory tests cannot be used to definitively diagnose CFS.
The second cluster of CFS was an outbreak among teachers in an elementary
school. In this outbreak, seven of 38 teachers had CFS (one was found to
have rheumatic fever). We were unable to study the students. We found a
similar picture to that observed with the orchestra members. Laboratory
findings for most of the teachers with CFS and many without CFS were similar
(elevated EBV and HHV-6 antibody titers and low NK cell activity). No tests
results were significantly different between the two groups. For example,
the mean level of NK cell activity was 50.6 LU for teachers with CFS and
64.7 LU for
teachers without CFS (normal laboratory control value, 170 LU). These findings suggest again that these laboratory tests are of little or no diagnostic value.
The third cluster occurred among a group of nuns living in the same convent. Seven of the 26 nuns in the group were initially suspected of having CFS. Some of the diagnoses of CFS were proposed by the nuns rather than by their physicians. Two of the seven nuns reporting CFS were found to have borderline diagnoses of systemic lupus erythematosus (SLE), and both had family histories of SLE. Both were severely affected and met the criteria sufficiently for a diagnosis of CFS had SLE not been present. Although detailed evaluations were performed on several individuals, especially in regard to attempts to diagnose autoimmune disease, the only laboratory studies done for the entire group were assays for viral immunity. The results of these studies were similar for individuals with and without CFS. This investigation demonstrates that when there appears to be clustering of a disease such as CFS, members of the group tend to attribute the signs and symptoms of other diseases to CFS. This observation underscores the need for extensive medical evaluation of individuals within putative clusters of CFS to rule out other mimicking disorders.
In summary these investigations suggest that in clusters of CFS in discrete
populations the definition of cases of CFS is often quite difficult. In
the two clusters in which we used abnormal laboratory results putatively
associated with CFS as potential diagnostic criteria, we could not clearly
define cases from noncases. Whether or not this was because of exposure
to a common source of a noninfectious causal agent or to an infectious
etiology that causes disorders of immune function cannot be determined
from these investigations. However, these findings suggest that the use
of comparable control groups outside the cluster cohort would be preferable
in future investigations. Another important problem is that when CFS clusters
occur, people with other chronic diseases tend to classify themselves as
Reprints or correspondence: Dr. Seymour Grufterman, Department of emily Medicine and Clinical Epidemiology, M-200 Scaife Hall, University dPittsburgh School of Medicine. Pittsburgh, Pennsylvania 15261
'Keywords: grufferman cluster immune transmissable infectious cause pathology immune natural killer cells ebv teachers nuns