III. Theories of
Causation and Mechanisms
The proposed theories of causation of MCS can only be
summarized in this report. Although these theories can be grouped into three broad
categories (immunologic, neurologic, and psychological), there are many variations. Some
theories are interrelated, and each theory is still being considered and debated within
the scientific community. For example, Miller et al. (1997) proposed a theory of
"toxicant-induced loss of tolerance" (TILT), which suggests that acute or
chronic chemical exposures might cause certain susceptible persons to lose their tolerance
for previously tolerated chemicals, drugs, and foods. Subsequently, even minute quantities
of these and other substances may trigger symptoms. They argue that TILT may prove to be a
new theory of disease causation parallel to the germ, immune, and cancer theories.
Discussions regarding the mechanisms of MCS can be divided into two distinct
categories. Some persons assert that MCS symptoms are psychologically based or have strong
psychological components. Others who accept a physiological basis for MCS may concur that
psychological factors are present, but contend that they are only a component of the
condition or are the natural result of coping with an intractable chronic condition.
The following sections summarize key literature on immune system, neurologic, and
psychological mechanisms postulated to be associated with MCS.
Top of Page.
Disorders of the immune system have been suggested as
causing or contributing to MCS (Rea et al., 1992; Levin and Byers, 1987; Thrasher et al.,
1990; Heuser et al., 1992; Ross, 1992; Levin and Byers, 1992). Some practitioners employ
tests for immune sensitization and/or laboratory determinations of immune parameters in
diagnosing MCS, and some use therapeutic regimens directed at correcting putative immune
deficiencies (Rea et al., 1992; Heuser et al., 1992; Ross 1992; Levin and Byers, 1992).
However, these investigators generally agree that MCS differs from disorders known to be
associated with overt immunopathology (i.e., allergies, immune deficiencies, and
autoimmune diseases) and suggest that MCS is not mediated solely by known immune
mechanisms (Ziem, 1992). Instead, MCS is argued to be associated with a more general form
of "immune dysregulation" which leads to the MCS symptom complex, perhaps
through interaction of immune mediators with the neuroendocrine systems (Meggs, 1992;
Levin and Byers, 1992).
The theories and reports of immune involvement in MCS presented by some environmental
medicine practitioners have not been accepted by most physicians and researchers (Terr,
1987; Albright and Goldstein, 1992), but some do acknowledge that allergic or
immunotoxicologic reactions could be contributing factors in at least a subset of MCS
patients (Selner and Staudenmayer, 1992; Albright and Goldstein, 1992; Meggs, 1992).
Because immune responsiveness and inflammation are closely related, hypotheses relating
inflammation to MCS (which are discussed in the next section) are likely to overlap with
immunologic considerations (Meggs, 1992).
Evidence for or against immune involvement in MCS depends to some extent on laboratory
measurements of certain immune biomarkers (defined by the National Research Council [NRC]
 as "[i]ndicators of events in biological systems or samples"). Some
studies have reported that results of immune biomarkers lie outside "normal"
ranges in many MCS patients (Heuser et al., 1992; reviewed in Meggs, 1992). Immune
abnormalities reported to be associated with MCS include alterations in the distribution
of peripheral blood lymphocyte subsets, increases in the proportion of activated T-cells
in circulation, and abnormal serum antibodies to tissue antigens and chemical-protein
conjugates (Rea et al., 1992; Thrasher et al., 1990; Heuser et al., 1992; Levin and Byers,
1992). In contrast, studies conducted by other researchers have not detected abnormal
immune test results in MCS patients (Terr 1986; Simon et al., 1993).
The role of the immune system in MCS is difficult to assess from many of the published
reports because the laboratory methods are inadequately documented or, in some cases,
clearly deficient. Results reported for lymphocyte phenotypes illustrate these
methodologic concerns. For example, Levin and Byers (1987) cited results originally
reported in court records but gave no methodologic information. Similarly, an article by
Terr (1986), provides no description of laboratory methods. Another example is the use of
an inadequate method, peripheral blood lymphocyte phenotypes determined by manual
fluorescent microscopy on separated cells, to assess the role of the immune system in MCS
(Thrasher et al., 1990; Simon et al., 1993). This technique is considered to be
unacceptable for clinical use (Kidd and Vogt, 1989; CDC, 1992). Comparable methodologic
uncertainties attend reports on immune tests for autoantibodies, antibodies to
chemical-protein conjugates, and cellular function assays in MCS patients (Vogt 1991; Vogt
and Margolick , 1994). Fully controlled studies with appropriate quality assurance are
needed to verify the suggested changes in immune system markers postulated to be
associated with MCS.
In addition to limitations in laboratory analyses, another major difficulty with
interpreting data on immune functions and MCS is the lack of sound epidemiologic methods
in most of the published reports, which are limited to individual cases or small numbers
of individuals identified in clinical settings. While such reports have value in
suggesting directions for rigorous investigations with sound study design, they must be
interpreted cautiously because they do not account for the numerous confounding variables
(including age, sex, smoking, diurnal and seasonal variation, and stress) that can
influence immune parameters (Vineis et al., 1993).
The importance of both laboratory methods and epidemiologic design has been underscored
by one particular study of MCS and immunologic tests (Simon et al., 1993). The study used
careful epidemiologic design to determine the usefulness of certain immunologic tests in
discriminating MCS patients from matched controls. The results showed clearly that the
immunologic tests, as selected and performed by a laboratory specializing in tests for
MCS, were of no use for identifying MCS patients. Subsequent to the publication of this
study, one of its investigators, Simon, released unpublished results from 10 sets of split
(i.e., duplicate) samples (20 specimens), to which the specialty laboratory had been
blinded, showing that results across samples had not been replicated (Friedman et al.,
The sound epidemiologic design of this study was important for raising questions about
the reliability of results from the specialty laboratory and reports previously published
by the laboratory (Thrasher et al., 1990). Moreover, because the same laboratory was used
by at least one physician who claimed to have found immunologic tests useful for
diagnosing MCS (Friedman et al., 1994), the results of the Simon et al. (1993) study cast
doubt on some accounts that suggest immune system involvement in MCS. However, because of
the laboratory deficiency, the study provides inadequate information about immune effects
in MCS patients.
Clarification of the role of the immune system in MCS may be forthcoming from an
ongoing multicenter study that is comparing results on a comprehensive panel of
immunologic biomarkers between MCS patients and matched controls using rigorous
inter-laboratory quality assurance (personal communication, Joseph B. Margolick, Johns
Top of Page.
Inflammation has been suggested as being causally related
to MCS as a result of the initiation of mediators released from cell membranes by the
action of free radicals produced from toxic chemical exposures (Sparks et al., 1994).
Bascom (1992) has suggested that exposure to low-level irritants may result in chronic
respiratory health effects and that "[d]ifferential susceptibility exists to
illnesses resulting from chronic exposure to irritant mixtures." She suggests this
occurs through several mechanisms, primarily induction of inflammation through irritation
of the upper airway epithelium.
The role of respiratory tract inflammation in MCS has also been hypothesized to
resemble the changes seen in other conditions that include hyperreactivity of the airways.
It has been suggested that a single acute, high-dose induction exposure to a chemical is
followed by a chronic intolerance to low levels of chemicals. This two-stage process has
been observed in reactive airways dysfunction syndrome, in which a high-dose exposure to
airway irritants is followed by chronic asthma with bronchial hyperactivity (Brooks et
al., 1985). Meggs has theorized that patients who develop rhinitis after a single
high-dose exposure can be said to have reactive upper-airways dysfunction syndrome (Meggs,
1995). He suggests that chemical sensitivity may be a symptom of airway inflammation. In
support of this hypothesis, Meggs gives several examples of studies where chemical
sensitivity was associated with upper airway disease for which the examined health
outcomes did not require rhinolaryngoscopic examinations (e.g., Doty et al., 1988;
Chester, 1991). Fiber-optic rhinoscopy has been used to detect nasal inflammation in a
chemically sensitive population, and nasal biopsies have indicated chronic inflammation
and a cobblestone appearance of the pharynx and tongue accompanied by mucosal injection
(Meggs and Cleveland, 1993).
Meggs (1995) reported airway inflammation that he noted on rhinolaryngoscopic
examinations; he hypothesized that inflammation in the airways can produce many of the
extra-airway symptoms seen in MCS. He argued that patients with allergies have
extra-airway symptoms such as nausea, fatigue, mental confusion, and myalgia at sites
other than the site of inoculation with antigen. He speculated that the mechanism by which
chemicals cause airway inflammation is a chemical interaction with chemoreceptors on
sensory nerves, leading to release of substance P and other mediators of neurogenic
inflammation. Leznoff, however, reported that five patients he examined who complained of
throat-related symptoms (choking, cough, dysphonia, and "swollen glands")
following challenge with a chemical showed no visible changes in the throat or larynx
using fiber-optic laryngoscopy and no change in recorded phonograms (Leznoff, 1992). It
should be noted that in neither of the two previous studies where the investigators were
blinded (i.e., unaware) to the MCS status of the patients.
In studies that have been useful in elucidating inflammogenic pathways, inflammatory
mediators (including cytokines and neuropeptides) have been quantified in serum and in
nasal biopsies, scrapings, and washings of persons with well-defined allergic and
nonallergic inflammatory reactions (Straight and Vogt, 1997). There is no convincing
evidence that such mediators are involved with MCS (Salvaggio, 1992), although the
hypothesis has not been adequately tested. The analytical methods for measuring these
mediators require close attention, because many of the assays yield highly variable
results between different sources and even between different reagent lots from the same
Top of Page.
Neurologic Mechanisms Including
Altered Sense of Smell
Of the neurophysiologic models that have been advanced to
explain MCS-related clinical phenomena and to provide possible mechanisms for the
condition, the olfactory-limbic and neural sensitization model developed and refined by
Bell and colleagues is the one most completely explicated (Bell et al., 1992; Sparks et
al., 1994). In a description of this model, Bell et al. (1997b) proposed that neural
stimulation is the underlying mechanism for the disorder. Neural stimulation is defined as
the "[p]rogressive amplification of responsivity by the passage of time and repeated,
intermittent exposures" and can be initiated by a single exposure to a chemical or by
multiple low-level exposures. Several forms of sensitization are proposed, including
"limbic kindling," a phenomenon described in animal research in which exposures
to excitants, such as electricity or chemicals, result in abnormal electrical activity in
the brain and seizures or seizure-like phenomena. Other forms of sensitization include
time-dependent sensitization of neurochemical, immunologic, endocrinologic, and behavioral
responses. According to Bell's model, these forms of sensitization directly involve limbic
and mesolimbic systems in the brain. Because these brain systems include structures that
are known to be associated with emotion and cognition, Bell concludes that the cognitive
and mood symptoms associated with MCS are related to the involvement of these brain
regions through sensitization. She feels that sensitization is distinct from other
possible mechanisms associated with MCS symptomatology (e.g., conditioning and
habituation), but suggests that these distinct mechanisms might be integrated to better
To test their model of MCS, Bell and colleagues have conducted a number of studies
comparing persons who are chemically intolerant and chemically tolerant (based on the
cacosmia screening index). In studies of college students, Bell has tied the phenomenon of
chemical intolerance to limbic system function through demonstration of associations
between higher reports of psychological distress and drug use (Bell et al., 1996a) and
higher rates of personal histories of anxiety and depression and family histories of
substance abuse (Bell et al., 1995b) among students who report chemical intolerance in
comparison with those without such intolerance. In the latter study, she also reported
that the chemically intolerant students scored higher on a measure of "limbic system
symptoms" than did the chemically tolerant students. As support of the limbic system
hypothesis, Bell and colleagues have also reported lower scores on a memory test among
chemically intolerant persons in a sample of Veterans Administration patients (Bell et
al., 1997a) and slowed reaction times on a divided attention task performed by chemically
intolerant retired adults (Bell et al., 1996b). In a series of studies, Bell et al.
(1996c, 1996d, 1997d) reported changes in endorphin levels, blood pressure, and
wakefulness in chemically intolerant persons relative to controls. All of these studies
represent attempts to explore the hypothesized model. It should be noted that the
experiments have been carried out primarily among persons who had not received a diagnosis
Animal models of sensitization have been proposed and initiated (Sorg et al., 1994;
Sorg, 1995; Bell et al., 1997c). Gilbert (1995) reported that rats with chronic, low-dose
exposure to lindane developed electrical changes in the brain and seizure-like symptoms,
while those with a single-dose exposure did not. He tied these results to the concept of
chemical kindling. Animal models have also been used to test hypotheses that responsivity
to chemicals might have a genetic component (Wang et al., 1993). In addition, the use of a
specific breed of rats (Flinders Sensitive Line rats) that are highly sensitive to the
organophosphate diisopropylfluorophosphate and which have increased cholinergic receptors
and behavioral changes resembling those seen in human depression has been suggested for
studies in animal models of MCS (Overstreet et al., 1996).
In a study of odor responsivity among persons diagnosed with MCS, Fiedler et al. (1995)
tested 31 subjects to assess odor detection thresholds to rose-scented alcohol and an
unpleasant-smelling pyradine; no differences were found between the MCS subjects,
controls, and asthma patients.
Quantitative electroencephalography (EEG), brain electrical activity mapping (BEAM),
positron emission tomography (PET), and single photon emission computed tomography (SPECT)
have been used as neurologic correlates of MCS. Mayberg (1994) reviewed the studies that
used these methods and concluded that, although patterns seen in some studies have shown
abnormalities that might be related to MCS (particularly with SPECT), these studies are
deficient in standardization of techniques, replication of results using testable
hypotheses, and use of appropriate control groups.
Top of Page.
Psychiatric factors have been seen as the cause of
MCS, an effect of having MCS, a predisposing factor in the development of
MCS, and a co-morbid occurrence with MCS. Some investigators believe that MCS is a
somatoform reaction (i.e., physical symptoms not explained by objective clinical
findings), if not a frank psychiatric condition. For example, one investigator believes
the symptom complex of MCS resembles the DSM-III-R description for panic disorder (Kurt,
1995), and others have suggested that MCS is an "odor-triggered panic attack"
(Shusterman and Dager, 1991).
Black et al. (1990) conducted a study on the emotional profile of persons identified as
having "environmental illness." Persons who were included had illnesses
diagnosed as chronic yeast disease, environmental allergy syndrome, 20th century disease,
and the multiple chemical hypersensitivity syndrome. No physical or laboratory
examinations were included. Significantly more study subjects than controls met lifetime
criteria for a major mental disorder, suggesting that patients with a diagnosis of
environmental illness may have one or more commonly recognized psychiatric disorders that
could explain some or all of their symptoms. Psychiatric diagnoses were recommended for
consideration as an explanation for patients with multiple ill-defined symptoms in the
absence of clinical or laboratory findings. This suggestion was supported by an earlier
study (Stewart and Raskin, 1985) of patients with "20th century disease."
Another study (Simon et al., 1993) compared 41 patients who had chemical sensitivity with
34 control patients who had chronic musculoskeletal injuries to examine the role of
psychological and other factors in MCS. Psychological evaluation included standardized
measures of anxiety, depression, and somatization. Patients with chemical sensitivity
reported a greater prevalence of current anxiety or depressive disorder, but this
difference did not appear to precede the onset of chemical sensitivity. The investigators
concluded that psychological symptoms, while not necessarily a cause, are a central
component of chemical sensitivity. Davidoff and Fogarty (1994) have pointed out
methodologic problems, such as patient selection, in these and other reports.
Fiedler and colleagues have examined neuropsychological test performance as a marker of
central nervous system (CNS) dysfunction in patients in whom MCS has been diagnosed or who
have related problems. In 1992, Fiedler et al. summarized data on 11 patients who met
Cullen's criteria for diagnosis of MCS. On the basis of these data, the investigators
concluded that there were neuropsychological findings suggestive of CNS involvement in
MCS. This conclusion was questioned in a later study (Fiedler et al., 1996). In this
study, Fiedler and colleagues compared neuropsychological and psychiatric function among
MCS patients who met and those who did not meet Cullen's criteria of MCS (labeled MCS and
CS, respectively), patients who had chronic fatigue syndrome (CFS), and healthy controls.
Standardized measures of psychiatric and neuropsychological function did not distinguish
the MCS and CS groups from the CFS group. The prevalence of current Axis I Psychiatric
Diagnosis was higher in the MCS, CS, and CFS groups than in controls. Seventy-four percent
of MCS, 38 percent of CS, and 61 percent of CFS patients did not meet criteria for any
current Axis I Psychiatric Diagnosis. Neuropsychological test results did not account for
the level of impairment implied by the patients' symptom reports.
It has been suggested that MCS is an example of a conditioned response (Siegel and
Kreutzer, 1997). In classical conditioning, a neutral conditioned stimulus is paired with
an unconditioned stimulus. The unconditioned stimulus reflexively elicits some response,
termed the unconditioned response. Initially, the conditioned stimulus does not evoke a
response. However, as a result of conditioned stimulus-unconditioned stimulus pairings,
the conditioned stimulus becomes associated with the unconditioned stimulus. As a result,
the previously neutral stimulus elicits a new response, termed the conditioned response.
Once established, generalization may occur during which the conditioned response is
elicited by stimuli other than the conditioned stimulus. Typically, the greater the
similarity between the novel stimulus and the conditioned stimulus used during
acquisition, the greater the strength of the generalized conditioned response. Davidoff
(1992) examined three models of MCS, including the classical conditioning model. She
listed predictions derived from this model, including those that incitants will have
somewhat similar odors and that responses will be "stereotyped and reflex-like"
and will occur predictably with certain odors. This model also does not require prior
psychopathology, because emotional responses can be reflexively induced. Davidoff listed
data consistent with the classical conditioning model, including reports of the absence of
a psychiatric history predating the condition and reports that odor awareness is often
salient in MCS. She also listed data inconsistent with the classical conditioning model,
including MCS patients' reporting that odors and response patterns to incitants vary
A 1993 exposure chamber study that was designed to investigate odor thresholds and the
ability of MCS patients to determine the presence of chemicals included double-blind
provocation challenges to 20 patients (Staudenmayer et al., 1993). The investigators used
an olfactory masker and a variety of chemicals (i.e., one chemical per patient); each
patient received five–10 challenges. All patients believed that they were reactive or hypersensitive to
low-level exposure to multiple chemicals. Clean air challenges that contained the
olfactory masker were used as placebo controls. As a group, the patients did not show a
"reliable response pattern across a series of challenges." The patients as a
group showed 33.3 percent sensitivity, 64.7 percent specificity, and 52.4 percent
efficiency in responding to stressors. The investigators concluded that such testing helps
differentiate toxicologic mechanisms from psychological mechanisms such as stress
psychophysiology and learned sensitivity. However, the results of this study may have been
influenced by the choice of placebos used in the experiment, the use of masking, and the
outcome measures that were used. In an earlier report, the same investigators suggested
that, in any environmentally related case, an evaluation should be made of psychological
motivation and evidence of psychological symptoms, including repressed childhood trauma
(Selner and Staudenmayer, 1992).
In summary, the suggestion that psychiatric disorders are the basis of MCS has
complicated communication between those who believe that, if present, psychiatric symptoms
are a secondary accompaniment to a chronic disease process and those who believe that MCS
is primarily the symptomatic manifestation of a psychiatric disorder. The thread of this
debate runs throughout the discussion of MCS. It continues despite recent evidence that
many disorders and syndromes that are considered to be psychiatric (e.g., panic disorder
and post-traumatic stress disorder) are accompanied by measurable changes in brain
function as assessed by techniques such as functional magnetic resonance imaging and
single photon emission computed tomography (Dager and Steen, 1992).
Although causal psychological mechanisms in MCS remain uncertain, the data suggest that
psychological factors should be carefully evaluated in the diagnosis and treatment of
patients who have MCS. The workgroup finds the need for carefully designed studies to
evaluate both the primary and secondary psychological factors in MCS.
Top of Page.
Sick Building Syndrome
Although not discussed in detail in this report, the syndrome known as Sick Building
Syndrome (SBS) has been linked to MCS as an initiating factor. Persons with SBS experience
symptoms that include eye, nose, and throat irritation; headaches; cough; difficult
breathing; fatigue; dizziness; and difficulty in concentrating. These symptoms are
temporally related to being in a particular building. The cause of SBS is unknown, but it
is often thought to result from poor building ventilation causing a buildup of vapors from
sources that include building materials, furnishings, and office equipment. SBS has also
been linked to contamination of indoor spaces or ventilation systems by biologic
organisms. Occasionally, some persons with SBS report that they later develop MCS. One
published study describes the clinical follow-up of 20 persons whose work-related illnesses
were considered related to a "sick building" (Welch and Sokas, 1992). Over time,
three of the 20 persons had ongoing symptoms consistent with Cullen's definition of MCS.
Several MCS patients have been reported with neurologic and/or cutaneous symptoms
suggestive of porphyrin disorders (Ziem and McTamney, 1995). These patients reported
symptoms such as "dark brown or red urine," skin sensitivity to sunlight
exposure, and sharp abdominal pain. Porphyrin disturbances were reported in a substantial
percentage of the patients. Others have pointed out that derangements of porphyria
metabolism do not result in symptoms reported by MCS patients, and that laboratory
evaluations of confirmed porphyria-related conditions generally do not resemble those seen
in MCS patients (Hahn and Bonkovsky, 1997; Washington State, 1995, 1996b; Gots, 1996).
Other conditions putatively linked to MCS include systemic lupus erythematosus,
chronic fatigue syndrome, fibromyalgia, scleroderma, and multiple sclerosis (NRC, 1992c).
Buchwald and Garrity (1994) compared 30 adults with CFS, 30 with
fibromyalgia, and 30
with MCS to evaluate the similarities between these three conditions. The majority of the
persons in each group were female. The mean age (40.8–44.0 years) and mean
educational level (14.7–14.9 years) of the three groups were similar. Approximately 80 percent of both
the fibromyalgia and MCS groups met the major criteria of the Centers for Disease Control
and Prevention's (CDC) CFS criteria (Holmes et al., 1988), and both groups also frequently
reported the minor CFS symptom criteria. Persons with MCS most frequently reported adverse
effects after exposure to pollution; perfume; and gas, paint, and solvent fumes. However,
5367 percent of
the CFS group and 47–67 percent of the fibromyalgia group also reported adverse effects with exposure
to these substances. Persons in all three groups were infrequently employed full time (1323
percent) and often were receiving
disability (3057 percent).
The mean number of visits per person to medical providers during the preceding year was
22.1 for persons with CFS, 39.7 with fibromyalgia, and 23.3 with MCS. The investigators
concluded that the data, though limited, suggest that these illnesses may be similar, if
not identical, conditions. They noted that the diagnosis assigned to an individual with
one of these conditions may depend more on the chief complaint and the medical specialty
of physicians making the diagnosis than on the actual illness process.
Fiedler et al. (1996) compared 23 persons who had MCS, 13 who had chemical sensitivity
(CS), 18 who had CFS, and 18 healthy controls. Individuals with MCS met the full criteria
for MCS proposed by Cullen, including an initial identifiable environmental exposure. The
individuals with CS met the same criteria for MCS with the exception of a clear onset.
Psychiatric and neuropsychological evaluation demonstrated more similarities than
differences between the CFS group and MCS and CS groups. In comparison with the control
group, the CFS group reported twice as many substances, on average, caused symptoms.
However, 30 percent of the CFS group reported that no substance causing illness, and 39
percent reported more than 20 substances. The investigators suggested that investigators
may want to consider stratifying individuals with CFS by chemical sensitivities in future
studies that evaluate differences between CFS and chemical sensitivities.
In 1994, a conceptual framework and guidelines were proposed for a comprehensive,
systematic, and integrated approach to the evaluation, classification, and study of
persons with CFS and other fatiguing illnesses (Fukuda et al., 1994). The guidelines
specifically stated that MCS and other conditions, including fibromyalgia, that are
defined primarily by symptoms that cannot be confirmed by diagnostic laboratory tests, do
not exclude an individual from the diagnosis of CFS.
Top of Page.
Clinical Ecology Approach
Members of the American Academy of Environmental Medicine (AAEM) support the
application of a comprehensive model of environmental medicine (see Section I). A number
of definitions and descriptive terms have been developed from this model that are not well
known or understood by other clinicians or scientists. The terms were developed in part
because physicians practicing clinical ecology believe that these terms and descriptors
better define their patients' conditions than do other terms. A basic understanding of
these terms gives perspective to the clinical ecology approach to MCS and attendant
theories of causation and mechanisms.
The following definitions are taken verbatim from An Overview of the Philosophy of
the American Academy of Environmental Medicine (AAEM, 1992):
is the sum total at any one time of all an individual's exposures to
specific environmental stressors to which he is individually susceptible.