A Report on Multiple Chemical Sensitivity (MCS)

The Interagency Workgroup on
Multiple Chemical Sensitivity

August 24, 1998

Predecisional Draft

Prevalence Studies

Descriptive Epidemiologic Studies

Summary Comments on Epidemiology

Table of Contents

II. Epidemiologic Considerations

Investigators gathering data about MCS often develop their own case definitions and, in almost all investigations, they rely on symptoms reported by the study participants. Likewise, data concerning the prevalence of MCS are limited and must be reviewed in light of any potential reporting errors and other errors that lead to misclassification, selection, and other related biases. Some investigators express the prevalence of self-reported, physician-diagnosed MCS, whereas other investigators report the prevalence of persons who report feeling ill from exposure to chemicals, or being sensitive to chemicals, but do not profess to have MCS.

As described in the following sections, epidemiologic investigations of MCS can be grouped into two kinds of study, depending on the populations comprising the studies. Epidemiologic studies of the prevalence of MCS in the general population constitute one kind of study. The other kind of study uses methods from descriptive epidemiology and has focused on groups who have indicated their hypersensitivity to chemicals.

Top of Page.

Prevalence Studies

On the basis of conversations with clinicians, Mooser (1987) suggested that 210 percent of persons in the general population have substantive disruption of their lives because of MCS. Cullen et al. (1992) later contested that this range was too high.

The U.S. Environmental Protection Agency (EPA) reported that about one-third of persons working in sealed buildings felt especially sensitive to one or more common chemical exposures (USEPA, 1991).

Bell and colleagues investigated the prevalence of chemical sensitivity in two groups of subjects, young adult college students and elderly persons. They found that 15 percent of a sample of college students reported feeling moderately or severely ill after exposure to at least four of five common substances (i.e., pesticides, paint, perfume, car exhaust, and new carpet) (Bell et al., 1993a). Twenty-two percent of college students (Bell et al., 1993b) reported feeling moderately or severely ill after exposure to at least three of the five substances. The same investigators (Bell et al., 1996a) found that the proportion of college students who report feeling ill or sensitive to environmental chemicals depends on how they are queried about such symptoms. In this study, 9.7 percent reported illness, sometimes or more often, from 6 to 10 environmental chemicals (new carpet, newsprint, disinfectant, paint, natural gas, perfume, tar, pesticide, car exhaust, tobacco smoke). Twenty-eight percent considered themselves to be "especially sensitive to certain chemicals." Only 0.2 percent reported physician-diagnosed MCS.

Bell et al. (1993c) reported that 17 percent of a group of retired elderly persons participating in a longitudinal study of osteoporosis reported feeling moderately or severely ill after exposure to at least four of five common substances (pesticides, paint, perfume, car exhaust, and new carpet). In subsequent reports 34 percent of the same study group of elderly persons (Bell et al., 1994) and 37 percent of elderly veterans (Bell et al., 1997a) considered themselves "especially sensitive to certain chemicals." Overall, 4 percent of the participants in the study of community elderly (Bell et al., 1994) reported physician-diagnosed chemical sensitivity.

Baldwin et al. (1997) found that 22.7 percent of a subset of urban employed persons participating in a study investigating relationships between indoor and outdoor air contaminants on respiratory health reported feeling moderately or severely ill after exposure to at least three of five substances (pesticides, paint, perfume, car exhaust, and new carpet) .

Meggs et al. (1996) conducted a random telephone survey to determine the self-reported prevalence of allergy and sensitivity to chemicals in a rural population in North Carolina. The survey defined allergy as "becoming sick from exposure to natural things (e.g., pollen, dust, grass, trees, cats, dogs, mold, feathers, foods)." Sensitivity to chemicals was defined as "becoming sick after smelling chemical odors (e.g., perfume, pesticides, fresh paint, cigarette smoke, new carpets, car exhaust)." Of 1,027 persons who participated in the survey, 35 percent reported allergies and 33 percent reported sensitivity to chemicals. Thirty-five percent of those who were sensitive to chemicals reported that symptoms occurred at least once each week. Symptoms of sensitivity to chemicals that occurred daily were reported by 3.9 percent of the total population; allergic symptoms that occurred daily were reported by 5.3 percent of the total population. The investigators cautioned, "This study was not designed to determine the prevalence of specific syndromes for which chemical sensitivity has been described (e.g., asthma and rhinitis, solvent-exposed workers, office workers in >sick buildings', people poisoned by organophosphate pesticides, individuals with the MCS syndrome). These results, therefore, cannot be used to draw conclusions about the prevalence of these specific conditions."

In 1993, the California Department of Health Services (CDHS) received a MCS-relevant grant from ATSDR to assemble a multidisciplinary panel of experts. Their task was to advise the agency on development of questionnaires, a battery of medical examinations, laboratory tests, and recommendations for epidemiologic study designs for implementation in the event of a community chemical spill. Four questionnaires were developed: (1) a set of 12 general population-screening questions, (2) a set of household-screening questions for use in a post-spill situation, (3) a "long-form" questionnaire for in-depth population-based research, and (4) a follow-up questionnaire. The 12 general population-screening questions were pilot tested by being included on the 1995 California Behavioral Risk Factors Surveillance (BRFS) telephone survey of 4,000 randomly selected California residents. According to Kreutzer and Neutra, the results showed that 16 percent of survey participants reported sensitivities to everyday chemicals, and 6 percent claimed to have been diagnosed with multiple chemical sensitivities by a doctor (personal communication, 1996). CDHS has developed a plan to further evaluate these incidence and prevalence data through in-depth interviews with the BRFS survey respondents who reported chemical sensitivities and a smaller number of respondents not reporting sensitivities.

In summary, studies report the prevalence of feeling ill after exposure to chemicals or being sensitive to chemicals, but not necessarily having MCS, ranges from 15 to 37 percent. However, because the relationship between cacosmia (i.e., a heightened sensitivity to odors) and the later development of MCS is unknown, the relevance of most of these studies to MCS is unclear. Only three studies have reported the prevalence of self-reported physician-diagnosed MCS. The prevalence of self-reported physician-diagnosed MCS ranges from published values of 0.2 percent in college students (Bell et al., 1996a) to 4.0 percent in elderly persons (Bell et al., 1994) and an unpublished value of 6 percent among randomly selected California residents (Kreutzer and Neutra, personal communication, 1996).

Top of Page.

Descriptive Epidemiologic Studies

In 1989, the National Foundation for the Chemically Hypersensitive, an MCS patients advocacy group, surveyed 6,800 persons claiming to be hypersensitive and found that 80 percent knew "when, where, with what, and how they were made ill." Pesticides were described as the initiating factor by 60 percent of this group (48 percent of all participants) (cited in Ashford and Miller, 1991). However, these results may not be applicable to the overall U.S. population.

Kipen et al. (1995) modified a questionnaire designed to identify environmental exposures in patients' living situations and tested it for use in population studies to assess the presence or absence of chemical sensitivity. The questionnaire asked whether exposure to various substances caused symptoms (defined as some discomfort or bothersome change). The questionnaire was administered (1) to 41 patients receiving health care at a general medicine clinic (medical clinic patients) and (2) to persons visiting an environmental and occupational health center, including 436 persons referred for routine prescribed surveillance or baseline examinations because of their employment (surveillance patients), 39 persons who met the investigators' criteria for MCS (MCS patients), 43 persons with a diagnosis of asthma or airway hyperreactivity but not MCS (occupational clinic patients with asthma), and 137 patients with a wide range of occupational and environmental health diagnoses (occupational clinic referrals). The MCS patients reported that significantly more substances elicited symptoms than did other groups. Patients with asthma reported significantly more substances elicited symptoms than the surveillance, occupational medicine referral, and medical clinic patient groups. Four percent of surveillance patients, 15 percent of occupational clinic referrals, 20 percent of medical clinic patients, 54 percent of occupational clinic patients with asthma, and 69 percent of MCS patients were identified as reporting symptoms with exposure to 23 or more substances. Although the investigators did not know if some of the patients in the medical clinic and surveillance groups would have qualified as having MCS, they point out that this would have improved both specificity and predictive values.

In a case series of 100 consecutive patients admitted to an environmental ecology (primarily MCS) unit at a Texas hospital, the median age of the patients was 40 years; 56 percent of these patients began having their symptoms before 30 years of age, 77 percent were female, and more than half had 4 years of college (as cited in Spyker, 1995). A self-assessment survey of initiating factors was also conducted. Twelve percent of the study population considered the initiating factor to have been an occupational exposure, and 11 percent, a new environment (i.e., home, job, or college); 58 percent could not identify an initiating factor (as cited in Spyker, 1995). Rea (1995) recently expanded this case series to a review of 30,000 patients seen over a 20-year period. In this case series, females outnumbered males at a ratio of 7:1. The predominant age range for developing chemical sensitivity was 3050 years, although MCS had been diagnosed at Rea's clinic in both infants and older patients. Rea also noted that, in this case series, males had a higher death rate than females. He also stated that although the rate of end organ failure was the same for males and females, the organs predominantly affected differed by sex. No data were presented to support these findings (Rea, 1995).

Miller and Mitzel (1995) reviewed the history of symptoms in 112 persons already diagnosed as having MCS: 83 percent had onset of symptoms after age 30 years and had a predominance of cognitive symptoms; 81 percent had been working full time when exposed, but only 12 percent were fully employed at the time of the survey. A majority stated that they had quit their jobs or changed jobs or careers because of their illness. About 40 percent had consulted 10 or more medical practitioners.

To evaluate workplace conditions possibly associated with MCS, Lax and Henneberger (1995) investigated MCS patients examined by an occupational health clinic. Computerized records of 605 patients were reviewed to identify those individuals who met a MCS case definition similar to the Cullen definition. Thirty-nine of the 605 patients (6.4 percent) met the criteria for MCS diagnosis; four were excluded because they had non-occupational-related onsets of MCS. The remaining 35 MCS patients were compared with 557 patients who had no indication of MCS. The 35 MCS patients were interviewed by telephone to survey health status and exposure histories. Findings showed 80 percent of the MCS patients and only 25 percent of the other patients were female. Also, the two groups of patients were very different with respect to the industries in which they were employed. For example, 54 percent of the non-MCS patients worked in industries considered to have a greater potential for hazardous exposures than other occupational settings; only 26 percent of the MCS patients were employed in the more hazardous industries. Notwithstanding the small number of MCS patients and limitations in exposure history surveys, Lax and Henneberger concluded, "Commonalities in exposure and symptoms suggest that Multiple Chemical Sensitivities represents a distinct diagnostic category."

Davidoff and Keyl (1996) compared four diverse MCS study groups with one another and with a matched general population group with respect to self-reported health and mental health history and status variables. Three of the MCS groups were similar in that they met the investigators' criteria for multiple chemical sensitivities syndrome, but they were unique in having developed this syndrome in response to different sensitizing exposures (i.e., organic solvents, organophosphate pesticides, and sick building syndrome). The fourth MCS group consisted of 10 workers from a potato processing plant who had developed conditions that resembled MCS after they were exposed by accident to both chlorine dioxide gas and chloroform. With respect to health status, ratings of "fair" or "poor" health and changes in health were more characteristic of the four MCS groups than for the general population group. The study suggested that the MCS groups were not significantly different from one another in health and illness status. While members of all groups reported changed tolerances to substances, MCS group members were more likely to attribute illness to chemical exposures on a daily basis and to place recovery from the ill effects at more than 12 hours after exposure. Frequency of members reporting three or more chronic childhood health conditions, which the investigators suggested was evidence of longstanding somatic complaints or complaints of sickliness in childhood, was seen relatively often in the MCS groups, but was observed rarely in the general population group. Negative affect, a psychiatric index score used in this study, was consistently higher in MCS groups than in the general population group. The investigators recognized possible limitations of the study, such as group selection methods.

Most studies show that a preponderance of patients with MCS are females, 30-50 years of age, with an above-average socioeconomic status (Cullen, 1992). Sparks et al. (1994) cited four studies documenting a preponderance of females reporting MCS symptoms, and other investigators also have published a similar finding. The reasons for this apparent preponderance are unknown.

A higher prevalence of a variety of medical conditions has been observed among cacosmic persons identified by the Cacosmia Screening Index and the relatives of these persons (Bell et al., 1994; Bell et al., 1995b; Bell et al., 1996; Baldwin et al., 1997). However, only one study (Bell et al., 1995a) compared personal and family medical and psychiatric histories of persons who have MCS with a control group. In this study, healthy cacosmics and healthy noncacosmics served as comparison groups. The prevalence was highest in the MCS group for self-reported physician-diagnosed rhinitis, chronic bronchitis, migraine headache, irritable bowel, arthritis, chronic fatigue syndrome (CFS), hypoadrenocortical function, candidiasis, ovarian cysts, menstrual cycle irregularities, painful menses, chronic pelvic pain, heavy menstrual bleeding, depression, anxiety, and panic disorder. The MCS group also reported the highest prevalence of rhinitis and diabetes mellitus among family members. The prevalence of family histories of depression, anxiety, panic disorder, and substance abuse were not significantly different between the groups.

Top of Page.

Summary Comments on Epidemiology

A particularly important question about MCS is its prevalence in the general population. Only three studies have reported the prevalence of self-reported physician-diagnosed MCS. The prevalence of self-reported physician-diagnosed MCS ranges from published values of 0.2 percent in college students (Bell et al., 1996a) to 4.0 percent in elderly persons (Bell et al., 1994) and an unpublished value of 6 percent among randomly selected California residents (Kreutzer and Neutra, personal communication, 1996).

A review of the current knowledge about epidemiology of MCS reveals that the only consensus descriptive or demographic data are the age range and sex of MCS patients. In addition, it is difficult to compare the results of studies on MCS conducted by different investigators. There is no consensus definition of MCS among the investigators, and they usually do not provide detailed descriptions of the methods used to document symptom complaints, identify environmental triggers that produced symptoms, or specify criteria used for inclusion or exclusion of subjects. The use of standardized questionnaires is one way to improve comparison of results across MCS investigations (CDHS, 1996). Other ways to improve the design and conduct of epidemiologic studies are discussed in Section IV.

Top of Page.

Return to Table of Contents

Return to Committee Reports Page