HYPERSENSITIVITY to low levels of common chemicals is a ubiquitous public health problem (1); however, the prevalence of chemical hypersensitivity in the U.S. population has not been previously established. Chemical hypersensitivity--often called multiple chemical sensitivity (MCS) (2)--is also referred to as toxicant-induced loss of tolerance (3) or environmental illness. (4) It is typically acknowledged to be a condition characterized by acute reactions that occur after exposure to even low levels of common chemical products such as fragrances, household cleaners, fresh paints, newsprint, pesticides, and other products that contain petrochemicals. (5) MCS can produce a wide range of symptoms, and individuals with the hypersensitivity can encounter great difficulty functioning in normal working and living environments. (6) Although a limited number of epidemiological studies have investigated the regional prevalence of chemical hypersensitivity in the U.S., its national prevalence is speculative. The National Academy of Sciences estimated that up to 15% of the U.S. population experiences some degree of hypersensitivity to common chemicals. (7) The
California Department of Health Services found a 15.9% prevalence of chemical hypersensitivity among Californians, (8) whereas similar studies in the Atlanta, Georgia, metropolitan area and in New Mexico found prevalences of 12.6% (9) and 16%, (10) respectively. These 3 studies used accepted epidemiological methodologies with randomly derived samples. Other published prevalence studies that have used methodologies such as anecdotal evidence, nonrandomly or self-selected subjects, or a broader definition of chemical hypersensitivity have produced varying findings. For instance, a study that relied exclusively on conversations with medical personnel in clinical settings estimated that 2% to 10% of the population experiences hypersensitivity. (11) Two subsequent studies in Arizona that used self-selected subjects--young adult college students and elderly persons--determined that approximately 15% of the younger group (12) and more than 37% of the elderly group (13) reported a hypersensitivity to chemicals. A population survey in rural North Carolina, using a broader definition of MCS that did not distinguish between
adverse reactions to harsh chemical odors and a true hypersensitivity, found a rate of 33%. (14)
Research suggests that MCS manifests as a 2-step process. (15) The first step, initiation, is when the hypersensitivity first develops. Initiation can result from an acute exposure to a particular toxic agent; a chronic exposure to one or more toxic substances, even at low levels; or a combination of those mechanisms. (3) The second step, triggering, is characterized by adverse physical reactions that occur after an exposure. Once the hypersensitivity has developed, symptoms can be triggered by low-level exposure to a wide range of toxic substances--usually much more diverse than the initiating substance(s).
Other researchers have hypothesized that MCS is linked to psychological factors and is an idiopathic disorder, which is either somatized (16) or a conditioned response. (17) These studies, however, have been criticized for methodological weaknesses, such as lack of preonset data or biased case selection. (18,19)
Method
We used a population-based telephone survey to investigate the prevalence of chemical hypersensitivity, as well as its etiology, symptomatology, diagnoses, and lifestyle effects, in the United States. To ensure a valid national sample, cases were obtained by random digital dialing of telephone numbers derived from the continental United States and regionally weighted for representativeness. To achieve a 95% confidence level and a [+ or -] 3% margin of error (i.e., generally accepted levels for a random population study), the sample would need to contain 1067 cases, based on the standard formula for determining population characteristics: proportion n = [proportion x (1 - proportion) x [Z.sup.2]]/[[(confidence interval).sup.2]], where n is the sample size, the proportion is 0.5, and Z is the statistical Z-score that corresponds to the confidence level. (20) Our study ultimately yielded 1054 valid cases, which was within accepted bounds (0.02%) for the desired margin of error ([+ or -] 3%) at a 95% confidence level.
A 22-item questionnaire was used as the survey instrument. A longer questionnaire was developed initially, but a pretest found that it caused excessive respondent fatigue, forcing several respondents to terminate their participation prematurely. The original questionnaire, therefore, was shortened to maximize participation. A pretest of the 22-item questionnaire revealed no significant problems, and the measurement instrument was judged acceptable. External validity was promoted by constructing the questionnaire to conform to recommendations of the Federal Interagency Workgroup on Multiple Chemical Sensitivities, (21) thus allowing comparisons with subsequent studies. The questions replicated ones in the regional studies and followed recommendations from other studies, to facilitate validity evaluations. (22) Data were collected on symptoms that characterize MCS. Questions were initially asked to determine if the respondents had negative reactions to scented products, air fresheners, or natural allergens, and if a medical professional had ever diagnosed them with MCS. Additionally, a key question on hypersensitivity, which replicated the wording in the earlier regional study, was used to determine the prevalence of chemical hypersensitivity. Respondents who gave a positive reply to the key question were asked a series of additional questions to determine the magnitude and nature of their sensitivity. Finally, demographic data were collected from all respondents.
To ensure reliability and to control for potential seasonal variations, the questionnaire was administered to 4 different seasonal cohorts. Each randomly dialed phone number was repeatedly called until answered by an adult willing to participate. Calling a number was discontinued after 4 unsuccessful efforts. Children aged younger than 14 yr who answered were not used but were asked to put an adult on the phone. College students who had been specifically trained for this project administered the questionnaire. Data for the 4 cohorts were obtained in spring 2002, summer 2002, fall 2002, and winter/spring 2003, respectively. The number of cases in each seasonal cohort varied somewhat, but the rates of a positive response to the key hypersensitivity question in each cohort (10.16%, 10.36%, 11.4%, and 16%, respectively) were all within the desired margin of error. After the data from each cohort were assessed and deemed statistically congruent with data from the other cohorts, the data from all 4 cohorts were aggregated to produce cumulative averages.
Statistical analysis. Data were aggregated and analyzed using the computer program Statistical Package for Social Science (SPSS) Windows Version 11.5 (SPSS, Inc., Chicago, IL) to obtain cross tabulations of variables, and statistical tests of data reliability. The reliability of the questionnaire was evaluated by using statistical measures of internal consistency. Measurements of Cronbach's coefficient of alpha and other measures of inter-item correlation were used on a cluster of related questions to determine consistency levels of the responses. A subsequent item analysis was conducted to further evaluate the findings and promote the integrity of the study. For the Cronbach and other inter-item analyses, questions about behavior modifications were clustered. An additional cluster consisting of questions about behavior modifications and dietary changes also was used. The Cronbach's alpha measurements indicated that respondents displayed different levels of behavior modification.
Results
The initial question asked of all respondents to our questionnaire (n = 1054) was whether they had ever been medically diagnosed with asthma; 14.1% (n = 148) of the sample answered "yes" (Table 1). A subsequent question asked all respondents whether they suffered from allergies to natural substances, such as dust, mold, grass, pollen, and animal dander, and 36.4% (n = 384) of the sample responded "yes." The next question asked about "air fresheners," and 17.6% (n = 185) of respondents reported getting headaches, breathing difficulties, or other health problems from them. All respondents also were asked whether being next to someone who was wearing a scented product was irritating or appealing; 31.1% (n = 329) said "irritating," 46.3% (n = 487) said "appealing," and 21.7% (n = 229) said "not sure."
The key question on chemical hypersensitivity replicated the wording used in the California and Atlanta studies and asked respondents the following: "Compared with other people, do you consider yourself to be allergic or unusually sensitive to everyday chemicals like those in household cleaning products, paints, perfumes, detergents, insect sprays, and things like that?" Among those responding (n = 1046), 11.2% (n = 117) answered "yes," with an additional 2.6% (n = 27) saying "not sure" (Table 1 ). All respondents were then asked whether they had ever been diagnosed by a medical professional as having MCS, and 2.5% (n = 26) answered "yes."
Other studies of chemically hypersensitive individuals have reported a variety of symptoms and reaction dynamics. (23) Thus, the respondents from our study who answered "yes" to the key hypersensitivity question (n = 117) or declined (n = 3) were asked follow-up questions to assess the characteristics and magnitude of their symptoms. The first follow-up question asked about the severity of the hypersensitivity. The percentage of hypersensitive respondents who characterized their symptoms as "severe" was 23.6% (n = 29), with 43.1% (n = 53) saying "moderately severe," 26% (n = 32) saying "mild," and 6.5% (n = 8) reporting that they did not know. The hypersensitive respondents also were asked whether they had developed sensitivities to certain foods since the emergence of their sensitivity to chemicals, and 18.5% (n = 22) said "yes." There were 2 mismarked questionnaires.
Individuals with hypersensitivity were then asked about any lifestyle changes they made following the onset of their hypersensitivity. Respondents were asked whether their hypersensitivity made it difficult for them to shop in stores or be in public places such as restaurants, houses of worship, or theaters. This question received a 39.5% (n = 47) positive response (Table 2). A subsequent question found that 72.7% (n = 85) of hypersensitive respondents had changed their household and personal care products. In addition, 13.2% (n = 16) reported losing their job, and 6.7% (n = 8) had moved from their home because of their hypersensitivity.
When we asked respondents with hypersensitivity to comment on the duration of their symptoms, 5.2% (n = 6) reported less than 1 yr, 25% (n = 29) reported 1 to 5 yr, 20.7% (n = 24) were in the 5 to 10 yr range, 29.3% (n = 34) reported 10 to 20 yr, and 19.8% (n = 23) responded "more than 20 yr." When asked how old they were when the hypersensitivity originally developed, 28.2% (n = 33) said under the age of 20; 28.2% (n = 33) said age 20-35; 21.4% (n = 25) said age 36-50; and 14.5% (n = 17) said over the age of 50, with 7.7% (n = 9) answering they were unable to recall or didn't know.
We also investigated the etiology of the chemical hypersensitivity. Respondents (n = 117) were asked whether they either knew or had a good idea what caused their hypersensitivity; 29.1% (n = 34) reported knowing the cause, 9.4% (n = 11) suspected the cause but could not be certain, 58.1% (n = 68) could not identify the cause, and 3.4% (n = 4) declined to answer this question. Those who could identify or suspect a potential cause (n = 65) were asked what they knew or suspected to be the main cause of their sensitivity to chemicals: 18.5% (n = 12) reported that an exposure to solvents or cleaners initiated their hypersensitivity, 4.6% (n = 3) reported an exposure to pesticides or herbicides, 10.8% (n = 7) reported an exposure to paint or building materials such as carpet or furnishings, 49.2% (n = 32) reported some other cause, and 16.9% (n = 11) declined to answer.
Respondents with hypersensitivity were also asked whether they had experienced any serious emotional problems prior to the emergence of their hypersensitivity, and whether any emotional problems had developed since its onset (Table 3). Of the hypersensitive respondents, 10.3% reported having emotional problems before the hypersensitivity emerged, and 19.5% reported these types of problems only after the hypersensitivity developed. It was assumed that people who had emotional problems before hypersensitivity emerged also had them after its emergence.
Demographic questions were asked of all individuals in the sample, and a cross-tabulation with the hypersensitivity was subsequently conducted. The age composition of the entire sample was as follows: 8.1% were under 20 yr of age, 26.4% were 20-35, 28% reported being 36-50, 37.5% were over 50 yr of age, and 2.1% did not know or declined to answer. Men constituted 38.4% and women 61.1% of the sample. The educational levels in the sample were as follows: 10.2% reported having less than a high school education, 28.6% were high school graduates, 26.3% had attended some college, 32.8% were college graduates and higher, and 2.1% answered "not sure." The ethnicity/race composition of the sample was 2% Asian or Pacific Islander, 5.7% Hispanic or of Latin origin, 11.1% black or African American, 76.4% Caucasian or European, and 4.8% other or declining to answer.
Cross-tabulation of the demographic characteristics of individuals who gave a positive response to the hypersensitivity question (n = 117) with the demographics of the entire sample (n = 1046) produced the characteristics of the hypersensitive (Table 4). The hypersensitive respondents were 13.8% men and 86.2% women; 5.5% were under 20 yr of age, 18.2% were 20-35 yr, 25.5% were 36-50 yr, and 50.9% were 50+ yr of age. Among hypersensitive respondents, 10.5% had less than a high school education, 27.2% were high school graduates, 26.3% had some college, 33.3% were college graduates or higher, and 2.6% declined to answer. Ethnically/ racially, the hypersensitive respondents were 0% Asian/ Pacific Islander, 5.3% of Hispanic/Latin origin, 13.2% black/African American, 78.9% Caucasian, and 2.6% other or declining to answer.
Discussion
The 11.2% national prevalence of chemical hypersensitivity we found in this study is statistically congruent, given the margin of error of [+ or -] 3%, with the regional prevalences of 12.6% and 15.9% found in the Atlanta and California studies, respectively, which used the same key question on hypersensitivity. The 31.1 % of our sample that found scented products on other people to be irritating also supports the prevalence rate of 33% found in the North Carolina study that used a broader definition of chemical hypersensitivity.
Although the cross-tabulation of demographic variables and hypersensitivity indicates that a disproportionate number of women and older individuals report chemical hypersensitivity, these findings are less significant when the gender (38.9% male and 61.1% female) and age (36.8% over 50) bias in the entire sample is considered. Additionally, although older people are more likely to report hypersensitivity, only 14.5% of the hypersensitive respondents said that their symptoms first appeared after age 50, whereas 56.4% reported that their symptoms first developed prior to age 35. This finding indicates that chemical hypersensitivity affects a wide age range and is not limited to older individuals. A comparison of the demographic characteristics of the entire sample with those of the hypersensitive shows close conformance among the percentages in the groups, indicating that chemical hypersensitivity is widely distributed throughout the general American population, cutting across racial/ethnic, age, and gender groupings.
An internal consistency analysis, which used a clustering of the behavior variables difficulty shopping, changing personal care products, loss of employment, and the necessity to change housing, produced a Cronbach alpha coefficient of 0.6189. When the same cluster was used with the additional variable of dietary change, the alpha coefficient rose to 0.6622. These results suggest that respondents with chemical hypersensitivity have taken varying actions to cope with their symptoms. Although 72.7% have changed their personal care products, for example, only 6.7% have found it feasible to move from their home.
The results from our question on prehypersensitivity emotional problems have etiological implications. That only 10.3% of the hypersensitive respondents had a history of emotional problems prior to the onset of their symptoms--and that hypersensitivity cuts across all demographic categories--weakens the notion that MCS is somatized. Although some respondents might have been reluctant to admit previous emotional problems, 19.5% reported experiencing problems after the symptoms appeared, suggesting a reasonable level of respondent veracity. These findings, therefore, suggest that chemical hypersensitivity is inconsistent with psychogenic disorders.
Although population studies are credible research tools, they have certain limitations. Randomly dialed numbers are an effective method of obtaining a valid national sample; however, there is no method to control who will voluntarily participate. Consequently, a sample could potentially contain a demographic bias. Also, incorrectly marked or unmarked responses on the questionnaires create somewhat different totals for different questions. Regardless of these limitations, however, the findings of this study provide valuable data for understanding the prevalence and nature of chemical hypersensitivity.
Conclusion
The findings of this study provide additional evidence that hypersensitivity to low levels of common chemicals is a widespread phenomenon in the American population. It is broadly distributed across racial/ethnic, age, and educational groups. Of the 11.2% of respondents who experienced chemical hypersensitivity, 66.7% described their symptoms as either severe or moderately severe, indicating that chemical hypersensitivity poses a significant public health problem. Consequently, the dynamics, etiology, and symptomatology of MCS deserve substantially more research and attention.
Submitted for publication September 7, 2004; accepted for publication December 1, 2004.
Requests for reprints should be sent to Stanley M. Caress, State University of West Georgia, Pafford Building #121, Carrollton, GA 30118.
E-mail: scaress@westga.edu
References
(1.) Sarnet JM, Davis DL. Introduction. In: Mitchell F (Ed). Multiple Chemical Sensitivities: A Scientific Overview. Atlanta, GA: U.S. Department of Health and Human Services, Agency for Toxic Substances and Disease Registry, 1995; pp 1-3.
(2.) Cullen MR, Pace PE, Redlich CA. The experience of the Yale occupational and environmental medicine clinics with multiple chemical sensitivities, 1986-1991. Toxicol Ind Health 1992; 8(4):15-9.
(3.) Miller C, Ashford N, Dory R, et al. Empirical approaches for the investigation of toxicant-induced loss of tolerance. Environ Health Perspect 1997; 105(suppl 2):515-9.
(4.) Levin AS, Byer VS. Environmental illness: a disorder of immune regulation. In: Cullen M, editors. Workers with Multiple Chemical Sensitivities. Philadelphia: Hanley & Belfus, 1987; pp 669-82.
(5.) Ashford N, Miller C. Chemical Exposures: Low Levels and High Stakes, 2nd ed. New York: Van Nostrand Reinhold, 1998; pp 19-33.
(6.) Gibson PR, Elms AN, Ruding LA. Perceived treatment efficacy for conventional and alternative therapies reported by persons with multiple chemical sensitivities. Environ Health Perspect 2003; 111 (12):1498-1504.
(7.) National Research Council. Indoor Pollutants. Washington, DC: National Academy Press, 1981.
(8.) Kreutzer R, Neutra RR, Lashuay N. Prevalence of people reporting sensitivities to chemicals in a population-based survey. Am J Epidemiol 1999; 150(1):1-12.
(9.) Caress SM, Steinemann AC. Prevalence of multiple chemical sensitivities: a population-based study in the southeastern United States. Am J Public Health 2004; 94(5): 746-7.
(10.) Voorhees R. Results of Analyses of Multiple Chemical Sensitivities Questions. 1997 New Mexico Behavioral Risk Factor Surveillance Systems. Santa Fe, NM: New Mexico Department of Health, 1999; p 25.
(11.) Mooser SB. The epidemiology of multiple chemical sensitivities (MCS). Occup Med 1987; 2(4):663-81.
(12.) Bell IR, Schwartz GE, Peterson JM, et al. Self-reported illness from chemical odors in young adults without clinical syndromes or occupational exposures. Arch Environ Health 1993; 48(1):6-13.
(13.) Bell IR, Walsh ME, Goss A, et al. Cognitive dysfunctions and disabilities in geriatric veterans with self-reported intolerance to environmental chemicals. J Chron Fatig Synd 1997; 2:5-42.
(14.) Meggs WJ, Dunn KA, Bloch RM, et al. Prevalence and nature of allergy and chemical sensitivity in a general population. Arch Environ Health 1996; 51(4):275-82.
(15.) Miller C. Chemical sensitivity: history and phenomenology. Toxicol Ind Health 1994; 10(4/5):253-76.
(16.) Black D, Rathe A, Goldstein R. Environmental illness: a controlled study of 26 subjects with 20th century disease. JAMA 1990; 264:3166-70.
(17.) Siegel S, Kreutzer R. Pavlovian conditioning and multiple chemical sensitivity. Environ Health Perspect 1997; 105: 521-6.
(18.) Davidoff AL, Fogarty L. Psychogenic origins of multiple chemical sensitivities syndrome: a critical review of the research literature. Arch Environ Health 1994; 49:316-25.
(19.) Davidoff AL, Fogarty L, Keyl PM. Psychiatric inferences from data on psychological/psychiatric symptoms in multiple chemical sensitivities syndrome. Arch Environ Med 2000; 55(31):163-75.
(20.) O'Sullivan E, Rassel GR. Research Methods for Public Administrators. White Plains, NY: Longman Publishers, 1995; p 135.
(21.) Interagency Workgroup on Multiple Chemical Sensitivities. A Report on Multiple Chemical Sensitivity (MCS) [Predecisional draft]. Atlanta, GA: U.S. Centers for Disease Control and Prevention, Agency for Toxic Substances and Disease Registry, 1998; pp 26-31.
(22.) Fiedler N, Maccia C, Kipen H. Evaluation of chemically sensitive patients. J Occup Environ Med 1992; 34: 529-38.
(23.) Miller CS, Mitzel HC. Chemical sensitivity attributed to pesticide exposure versus remodeling. Arch Environ Health 1995; 50(2):119-29.
STANLEY M. CARESS
State University of West Georgia
Carrollton, Georgia
ANNE C. STEINEMANN
University of Washington
Seattle, Washington
COPYRIGHT 2004 Heldref Publications
COPYRIGHT 2005 Gale Group