Profile of patients with chemical injury and sensitivity.

Patients reporting sensitivity to multiple chemicals at levels usually tolerated by the healthy population were administered standardized questionnaires to evaluate their symptoms and the exposures that aggravated these symptoms. Many patients were referred for medical tests. It is thought that patients with chemical sensitivity have organ abnormalities involving the liver, nervous system (brain, including limbic, peripheral, autonomic), immune system, and porphyrin metabolism, probably reflecting chemical injury to these systems. Laboratory results are not consistent with a psychologic origin of chemical sensitivity. Substantial overlap between chemical sensitivity, fibromyalgia, and chronic fatigue syndrome exists: the latter two conditions often involve chemical sensitivity and may even be the same disorder. Other disorders commonly seen in chemical sensitivity patients include headache (often migraine), chronic fatigue, musculoskeletal aching, chronic respiratory inflammation (rhinitis, sinusitis, laryngitis, asthma), attention deficit, and hyperactivity (affected younger children). Less common disorders include tremor, seizures, and mitral valve prolapse. Patients with these overlapping disorders should be evaluated for chemical sensitivity and excluded from control groups in future research. Agents whose exposures are associated with symptoms and suspected of causing onset of chemical sensitivity with chronic illness include gasoline, kerosene, natural gas, pesticides (especially chlordane and chlorpyrifos), solvents, new carpet and other renovation materials, adhesives/glues, fiberglass, carbonless copy paper, fabric softener, formaldehyde and glutaraldehyde, carpet shampoos (lauryl sulfate) and other cleaning agents, isocyanates, combustion products (poorly vented gas heaters, overheated batteries), and medications (dinitrochlorobenzene for warts, intranasally packed neosynephrine, prolonged antibiotics, and general anesthesia with petrochemicals). Multiple mechanisms of chemical injury that magnify response to exposures in chemically sensitive patients can include neurogenic inflammation (respiratory, gastrointestinal, genitourinary), kindling and time-dependent sensitization (neurologic), impaired porphyrin metabolism (multiple organs), and immune activation.


Introduction
The study of medicine "begins with the bodily functions. Porphyrin disturbances patient, continues with the patient and of various types following chemical and ends ... with the patient," according to heavy metal exposures were reported by William Osler (1). Exposure to chemicals, several authors in a special conference on particularly petrochemicals and combus-chemically induced porphyrinopathies tion products, has been associated in the sponsored by the New York Academy of literature with a variety of alterations in Sciences (2). Abnormally elevated levels of urinary coproporphyrins were reported in several papers, but other fecal and urinary porphyrins could be increased as well. Specific types of heavy metal and petrochemical exposures seem to cause specific patterns of porphyrin disturbance in rats (3). Chlorinated benzenes can induce porphyria in rats (4), and small exposures (such as swallowed mouthwash) can aggravate congenital porphyria (5). Immune disturbances following chemical exposure have been reported by several authors. In the following immune references, patients were studied only after probable causal exposure and the results were compared to those of laboratory normals. Impaired mitogenesis has been noted after exposure to chlordane (6) and isocyanates (7). A higher number of helper cells have been described in workers exposed to solvents (8) and isocyanates (7), and in persons consuming chlorinated solvents in drinking water (9). A greater number of immune complexes are reported with vinyl chloride exposure, and scleroderma has been noted after aromatic or chlorinated solvent exposure (10). Immune activation with higher levels of TAI/CD26 has been reported with isocyanates (7), formaldehyde (11), formaldehyde with aliphatic amines (12), silicone (13), chlordane and chlorpyrifos (14), and sick-building exposures (15).
Neurologic effects long have been associated with exposure to petrochemicals. Solvent exposure is associated with autonomic dysfunction (25), neurocognitive impairment (26,27), vestibular abnormalities (28), and impaired hearing (29). In a study of the effects of solvents on 15 Environmental Health Perspectives -Vol 105, Supplement 2 * March 1997 industrial painters (30), impairment was observed on a variety of neuropsychological measures. The Halstead-Reitan Battery was administered and impairment was found on the Impairment Index, Trails A, Digit Symbol, Seashore Rhythm and Speech Sounds-Perception Tests. In addition, subjects reported personality change and decreased memory.
Neurological abnormalities in a group of organophosphate-exposed subjects were described using the Halstead-Reitan Battery (31). Visual retention, memory dysfunction, and constructional deficits were reported (32).
Studies on chlorinated hydrocarbon solvents have also shown adverse effects. Trichloroethylene in low concentrations and for relatively brief times can lead to significant and prolonged impairment (33). Psychomotor speed and memory were two of the areas most affected, with the memory impairments characterized by storage and retrieval difficulties.
Neuropsychological changes have been found among community residents living in a supposedly benign environment such as areas adjacent to a wood-treating plant (34). Of the 34 subjects tested, more than 40% had sensory impairments, 86% had motor or psychomotor speed problems, and 72% had concentration difficulties. Disturbed autonomic function has been reported with chemical sensitivity (35), and abnormal neurocognitive function with chemical sensitivity/cacosmia (36).
Chemical sensitivity has been reported in the literature following exposure to chlordane (14), chlorpyrifos (37), pesticides (38), formaldehyde (39), tight or sick buildings (15), and organophosphates and solvents (40). These chemicals are not structurally related, although all may form free radicals and cause tissue damage.
A wide range of neurologic abnormalities [single photon emission computed tomography (SPECT), nerve conduction, electro encephalograms (EEGs), evoked potentials, neurocognitive] have been reported in chemically sensitive patients (41)(42)(43). Since depression and mood swings can occur with solvent/hydrocarbon exposures and with porphyrin disturbances, these are not in themselves evidence of a psychologic etiology of MCS (40). These and other studies strongly suggest that chemical sensitivity is a physiologic not a psychologic disorder (35,36,44).
A 1994 study found that 67% of patients with fibromyalgia and the same percentage with chronic fatigue syndrome (CFS) reported that their symptoms worsened on exposure to gas, paint, or solvent fumes, and 46 to 64% of these groups reported sensitivities to three other categories of common chemical exposures (45). Fibromyalgia patients have shown reduced current perception threshold (46) and T-cell changes (47). Chronic fatigue syndrome patients often meet criteria for fibromyalgia (48) and have impaired NK function, increased TA1/CD26, altered CD8, impaired mitogenesis (49,50), and vestibular and other neurologic abnormalities (51). Both chemical sensitivity (52) and chronic fatigue syndrome (53) have been postulated to involve limbic encephalopathy.
We describe medical findings on history, physical exam, and laboratory testing for patients who, in the wake of chemical exposure, developed chronic illness with multi-systemic symptoms exacerbated by exposures to multiple different chemicals at levels usually tolerated by most healthy members of the general population and previously tolerated by the patient.

Methods
Ziem's medical practice has been evaluating and caring for patients with MCS accompanied by chronic illness for many years, assisting them with environmental controls to reduce exposure and addressing clinical issues related to their exposure. Typically, an initial evaluation requires 1.5 to 2 hr for medical and exposure history, physical exam, evaluation of environmental aggravating factors, recommendations for environmental controls, and otherwise addressing clinical problems. Follow-up evaluations usually require 1 to 1.5 hr. This time-intensive approach to the evaluation of chemically injured patients contrasts with what we consider to be a too-cursory assessment prevalent among many industrial and academic professionals. We do not believe it is possible to evaluate and manage these disorders adequately in the 10 to 30 min usually allotted to such patients.
In the late 1980s, Ziem introduced a standardized questionnaire for initial patient assessment to evaluate the types of symptoms present, possible organ systems involved, and the types of chemical exposures perceived by these patients to be aggravating their symptoms. The full questionnaire is 16 pages long with 46 detailed questions. Reported here are the responses to questions 14 and 6 (Appendices 1 and 2) from 91 chemically sensitive patients whose first visit occurred after the introduction of the questionnaire. These two questions originally were developed by Dr. Ann Davidoff of the Johns Hopkins School of Hygiene and Public Health as part of a research project on chemical sensitivity (54).
Patients responding to the questionnaire came from many states from the eastern region of the United States, with most coming from the mid-Atlantic area. Most were referred by physicians not specializing in occupational medicine or toxic substances. Some were referred by patient support groups, friends, or relatives, and a few were legal referrals (although we have no data on the actual number in each category). Most but not all patients were well before the onset of their MCS. Former psychiatric diagnoses and treatment were unusual rather than typical.
In following these patients over time, Ziem's clinical impression is that patient responses to these two questions seem to correlate well with the clinical severity of their symptoms: sicker patients appear to have more frequent symptoms and to be affected by more exposures. However, it was not possible for this presentation to confirm this clinical impression by comparing response rates to exposures with clinical outcomes. This would be useful for future clinical studies. Symptom frequency can also be studied as an indicator of clinical severity.
These patients filled out the questionnaires at or shortly before their first visit. Some had been using some environmental controls; others had not. Many but not all patients were aware of chemical sensitivity as a potential problem for them. To assess the frequency of 51 different symptoms commonly associated with MCS, patients were asked in question 14 to describe whether these symptoms occur daily to almost daily, several times a week, once a week, several times per month, once per month or less, rarely, if ever, or not sure. In the analysis below, daily or almost daily responses are combined with several times a week or more responses to obtain a total figure for frequent symptoms.
All patients reporting increased sensitivity to chemicals accompanied by chronic illness were included in this analysis. They typically had frequent recurring or chronic symptoms in more than one organ system for over 6 months before completing the questionnaire. For some items in question 14, responses were only available for 89 or 90 patients: the bar graph in Figure 1 shows the number of patients (n) answering for each question. Outcomes are reported as percentages of the total answering each  item. The bar for each symptom displays separately symptoms occurring daily to several times a week (combined totals from the first two columns from Appendix 1) and once a week to several times a month (combined totals from columns 3 and 4 from Appendix 1). These percents are then added to show the total percent with symptoms several times a month or more (this percent figure is shown to the right of first bar for each symptom 57% of patients had he. eral times a week and weekly to several times of 75% of patients exp several times a month or Question 6 of the q from the Davidoff stud conducted at Johns Ho1 parts. In the first part, p: they thought they would be sick following a 20-min or a 4-hr carefully defined exposure to a variety of common petrochemicals and irritants ( Figure 2). In the second part of the questionnaire, they were asked if their symptoms would be exacerbated by a different set of briefer exposures ( Figure 3).
Laboratory testing was recommended for most patients, more so in recent years as knowledge of MCS-related abnormalities grew. A few patients have had EEG, quantitative electroencephalogram, nerve conduction studies, SPECT scans, etc., but these data are not reviewed here because of -__________________ the small numbers involved [and because these abnormalities in MCS patients have been reported elsewhere (42)]. Also, for all testing, financial constraints were often RMa Em operative: disabled patients without adequate insurance were less able to afford testing. As a result, it is possible that tested patients were less severely affected than untested patients.
Immune testing was recommended for  laboratory also included tests of T-and B-cell function, which were not requested of the AAL. Beginning in 1995, MCS patients with neurological and/or cutaneous symptoms suggesting porphyrin disorders have been screened for the various biomarkers of those conditions that may be detected in blood, urine, and stool. Samples were collected locally by independent laboratories and sent for analysis (via overnight delivery) to the Mayo Medical Laboratories in Rochester, Minnesota. Mayo was selected both because of its excellent reputation and because it offers more porphyrin-related tests than any other commercial laboratory in the United States. A full panel includes tests for five of the porphyrin-related enzymes involved in the heme synthesis pathway [aminolevulenic acid dehydratase (ALA-D), porphobilinogen deaminase, also known as uroporphyrinogen I synthase, uroporphyrinogen III cosynthase, uroporphyrinogen decarboxylase, and coproporphyrinogen oxidase], 6-aminolevulinic acid in urine, porphobilinogen in urine, quantitative urinary porphyrins, and fractionated fecal porphyrins. Results are presented for six MCS patients from Ziem's practice who have undergone all these tests and eight others who have undergone most of them (Table 1). Data on the three other patients screened are not included, as they each had undergone only one or two of the recommended tests, although two patients showed some abnormalities-one with decreased coproporhyrinogen oxidase (but no urine or fecal testing) and one with increased urinary coproporphyrins (but no blood or stool testing). In terms of both symptoms and history, the patients tested for porphyrin disorders were fairly representative of all chemically sensitive patients in this medical practice and none were known to have any other disorder associated in the literature with abnormal porphyrin metabolism (such as lead poisoning, hepatitis C, or liver cancer).
Patients diagnosed with MCS and describing difficulty with thinking, concentration, and/or memory were referred for neurocognitive testing. Those who live near Baltimore, Maryland, were referred to Dr. James McTamney, a clinical psychologist. As of September 1995, 13 patients had completed testing with Dr. McTamney. Patients in other geographic areas sometimes went elsewhere for testing, but because testing approaches and the tests used were different, these patients' data are not analyzed here. Dr. McTamney's evaluation included the Wechsler Adult Intelligence Scale-Revised (WAIS-R) ( Table 2) and the Halstead-Reitan Battery ( Table 3). The Halstead-Reitan Neuropsychological Battery is generally conceded to be a reliable psychological means of identifying patients with brain damage. It is an individually administered battery composed of the following tests: the Category test, the Tactile Performance test, the Seashore Rhythm test, the Speech Sounds Perception test, the Finger Oscillation test, and Trails A and B.
The Category test is a complex test of new problem solving, judgment, abstract reasoning, concept formation, mental flexibility, and mental efficiency. It requires a number of higher order functions such as the ability to note similarities and differences among stimuli and to formulate hypotheses regarding the principle that determines the correct answer. It is also a test of learning ability utilizing nonverbal  Trails A and B do not contribute to the Halstead Impairment index but are considered a part of the Halstead-Reitan Battery. On the Trails A, the person is required to connect 24 numbered circles distributed in a random pattern while being timed for the performance. On Trails B, the person is required to connect circles numbered 1 through 13 and letters A through L alternating from number to letter in sequence.
The overall impairment index is calculated as the number of subtests within the impaired range divided by the overall number of subtests taken.
WAIS-R is a scale of an individually administered composite of tests in battery format. For all but the most severely impaired adults, a WAIS-R battery constitutes a substantial portion of the test framework of the neuropsychological examination. Eleven different subtests make up the WAIS-R battery. Six of the subtests are classified as verbal tests: Information, Comprehension, Arithmetic, Similarities, Digit Span, and Vocabulary. Five are termed performance tests and measure nonverbal/visuospatial factors. The 100 tests are Digit Symbol, Picture Completion, Block Design, Picture Arrangement and Object Assembly. We invite independent scientific review ed of our actual clinical data to further understanding of these patients' chemical injury and chemical sensitivity.

Results
material. Finally, there is a memory component to the test that requires the person to remember which principle was correct in determining the answer to an individual item. This requires a longer recall of previously learned correct responses. The Tactile Performance test is a complex test requiring the individual to sustain adequate strength and speed of movement. It also requires the tactile perception and the ability to form a visual map of the board. The person, while blindfolded, must place ten blocks with the dominant hand on the first trial and then ten blocks with the nondominant hand on the second trial. On the third trial, the individual is allowed to use both hands together. After the trials are completed, the blindfold is removed and the person asked to draw the blocks approximating their size, shape, and relationships to one another while the blocks are out of sight.
The Seashore Rhythm test requires the individual to listen to 30 pairs of rhythmic beats from a tape recording and to select the proper response on an answer sheet as to whether the two tones in each pair are the same or different. The test requires the individual to discriminate between different patterns of nonverbal sounds while maintaining attention and concentration throughout the test.
On the Speech Sounds Perception test, the individual has a sheet of paper on which 60 groups of four nonsense words are listed. The individual must underline the correct response after hearing it on the audio tape recording.
On the Finger Oscillation test, the individual is required to tap as rapidly as possible with the index finger on a small lever attached to a mechanical counter. The person is given five consecutive 10-sec trials with the preferred hand and then five consecutive trials with the nonpreferred hand. The scores in this test are the average number of taps in a 10-sec period for each hand. The cutoff point indicating impairment would be less than 50 taps per hand on average. The vast majority of these patients reported some symptoms occurring daily to almost daily ( Figure 1). Symptoms that could reflect respiratory responses to irritant exposures were frequent (several times a week or more): patients indicated nasal symptoms (60%), sinus discomfort (48%), throat discomfort (53%), weak voice/ hoarseness (44%), chest tightness (42%), and wheezing (25%). The higher percentages of patients reporting nose and throat symptoms suggest that irritant effects may be greatest in the upper respiratory tractareas that first encounter respiratory irritants (nose, throat, etc.)-and less as these irritants move down the respiratory tract. This is consistent with partial irritant removal as the inhaled air moves down the respiratory tract.
(49%), and reduced bladder control exposure.) Symptoms suggesting inflamand outdoor petrochemical air pollutants (27%). The latter suggests possible involvemation such as swollen glands, muscle dis-(new carpets, pesticides, paint, scented ment of the autonomic nervous system, as comfort/spasm, and joint discomfort also products, etc.); half or more of the patients do some other frequently reported sympwere commonly reported to be frequent by reported symptoms when exposed to other toms such as flushing skin (39%), rapid 21, 49, and 52% of patients, respectively. irritants such as detergents or chlorine. In pulse (24%), palpitations (25%), reduced Frequent abdominal discomfort was expe-this group of patients, the vast majority cold tolerance (31 %), and reduced heat rienced by 40%, consistent with abnormal-who reported symptoms with 4-hr expotolerance (24%). Based on patient exams ities in porphyrin metabolism, as described sures also experienced symptoms with only and Holter monitoring, ectopic heartbeats below. Headache and unusual fatigue were 20-min exposure. This has implications for appear more frequently during exposures frequent in 57 and 69% of patients, reasonable accommodation: requiring daily and in sicker patients, but these data have respectively. The 40% reporting frequent commuting, work, or even brief meetings not been analyzed in the aggregate. Mitral unusual thirst was an unexpected finding in newly carpeted, remodeled or recently valve prolapse, a probable autonomic disor-and may indicate endocrine changes in this pesticide-treated areas could aggravate der, appeared to be unusually common in group. The finding of significant muscu-symptoms in a significant proportion of our chemically sensitive patients, possibly loskeletal aching and fatigue among MCS persons with chemical sensitivity. It also 10% or more, but this too has not been patients is interesting in light of the overlap suggests that many current products and quantified in the aggregate. Murmurs in symptomatology and clinical findings practices involving chemicals (remodeling, seemed to subside with clinical improve-between MCS, fibromyalgia syndrome, deaning, repairs, pesticide application, etc.) ment in some cases. and chronic fatigue syndrome (45). present frequent problems for these patients.
Symptoms of sensory organs were also In response to question 6 ( Figure 2), Ziem first encountered cellular immune reported to be frequent by a significant more than two-thirds of the patients abnormalities while evaluating a cluster of percentage of MCS patients: changes in reported symptoms with exposure to eight chlordane-exposed patients with hearing (32%), visual changes (48%), and combustion products such as passive ciga-chemical sensitivity in the late 1980s. She ringing ears (36%). (Clinically, visual rette smoke and vehicle exhaust and to has checked for similar immune abnormalichanges usually involve blurred vision with many other frequently encountered indoor ties in other chemically exposed patients Environmental Health Perspectives * Vol 105, Supplement 2 * March 1997 with chemical sensitivity, using the AAL for testing on the first 68 patients (Table 4). Compared to lab normals, the majority of those tested by AAL had increased total CD26 (also known as TA1) cells, which is considered a marker of immune activation (55). Seventeen (25%) had increased CD4 (helper T lymphocytes). Only five had increased CD8 (suppresser T lymphocytes). NK (CD57) evaluations were done on only 15 patients but were reduced in number in 5. B lymphocyte (CD 14) values are available for 57 patients; these were reduced in 18 (32%) and increased in 1 patient. Judgments of abnormal and normal ranges were made by the laboratory, but the raw data and the laboratory's normal reference ranges are provided for independent review.
Autoantibodies were evaluated and found to be present in increased titer in some cases: antimyelin (nervous system) autoantibodies and antismooth muscle (liver) in about half the patients, with antiparietal (stomach), antibrush border (kidney), antimitochondrial, and antinuclear in only a small number of patients. AAL considered myelin antibodies abnormal if titers were over 1.8; other autoantibodies were considered abnormal if titer levels were 1:20 or greater. Exact titer levels were not reported (by AAL) but may have provided additional information. Ziems's clinical impression is that as patients improved, some titers fell within the laboratory's normal range, but longitudinal data have not been quantified in aggregate. Titer levels facilitate comparison of changes in clinical status with increase or decline in titer levels. The striking proportion with autoantibodies against liver and nervous system tissue (about half the total patients tested for each) is consistent with neurologic effects and liver involvement, which is consistent with porphyrin disturbance. Table 4 also lists the location, situation, and chemical(s) involved in each patient's presumed causal exposure when this could be determined. An exposure was considered causal for this purpose if: there were no symptoms of multiple chemical sensitivity accompanied by chronic illness before the exposure; significant symptoms occurred during or shortly after the exposure (within days); symptoms improved away from the exposure on at least two occasions; and symptoms recurred on reexposure on at least one occasion. For any particular products associated with causal exposures, the chemicals involved were identified whenever possible, usually from material safety data sheets.
Over half the patients listed in Table 4 were considered occupational in probable causal exposure: 23 in the private sector (OCP in Table 4) and 14 in government or the public sector (OCG). Many publicsector exposures appeared to involve tight buildings. Public sector reluctance to tackle the chemical sensitivity problem might be affected by a substantial number of compensation cases filed by public sector employees. School exposures accounted for eight cases, all involving pesticides, and primarily affected students. Home exposures involved over a fourth of the total cases seen by Ziem. Nearly a third of all cases were felt by Ziem to be caused by pesticide exposure.
After Simon raised questions about the reliability ofAAL testing in April 1994-he said the results of split samples secretly submitted to the laboratory showed that "the reliability on most of their measures is no better than chance" (56)-Ziem arranged for subsequent immune function testing to be performed by the ISL. In this presentation of ISL data, both normal and abnormal results are shown but only for those particular tests done on more than half the patients ( Table 5). The only ISL data excluded after analysis are those on MY, GP, SG, and AS neural autoantibodies, each of which were tested in less than five patients, although some other parameters were not analyzed for this paper.
Autoantibody levels were normal for most antigens, but 7 of 21 showed increased levels compared to laboratory normals for both rheumatoid factor and total immune complex. Patients tested through ISL typically had normal numbers of total helper and suppresser cells (CD4, CD8) and normal CD26 (TAI) cells. Results from the two laboratories may have differed because of different normal ranges, different patients being tested, and/or techniques; this merits further study. However, abnormalities in ISL tests did include higher than expected levels of chemical antibodies (11 of 24 patients), typically involving benzene, which is most likely encountered as a metabolite of aromatic petrochemicals. Reduced T-cell function was common (9 of 21 patients tested), with 7 of the 9 being abnormal using both concanavalin A (conA) and phylohemagglutinin (PHA), an indication of intratest consistancy. Impairment of B-cell function was seen in five patients, of whom four were abnormal in all three tests (pokeweed mitogen, lipopolysaccharide, Staphalococcus aureus). This consistency across tests increases the likelihood of their validity. While the measure of total NK cells was typically normal, function was impaired in 14 of 22 tested patients.
Immune indicators are likely to change significantly over time following exposure, with initial activation leading to a cascade of other immune changes and later resolution of activation. Thus, single immune measurements in a population with varying time intervals following various causal and significant aggravating exposures are likely to obscure certain abnormalities. The optimal way to study immune changes is to observe patients shortly after casual exposure and to follow them through time. The time interval to maximum immune activation and other immune responses and to return to preexposure levels, if ever, at this time is not known for chemically sensitive patients. Further, different immune measures are likely to peak at different time intervals following onset of illness.
For a patient whose chemical sensitivity was thought to be caused by occupational exposure to carbonless copy paper, we compared immune measurements before the patient left and after she returned to her job, which required sitting near and working with large quantities of carbonless copy paper (Table 6). These data suggest that immune measures pre-and postchallenge testing are unlikely to show major changes. The patient had been away from exposure for many months and showed significant clinical improvement. Chemical sensitivity has been seen in a hundred patients with substantial occupational exposure to carbonless copy paper (E Panitz, personal communication).
In 1994, a few physicians began testing for porphyrin disturbances in chemically injured patients previously diagnosed with MCS. Like patients with some types of congenital porphyria, chemically sensitive patients frequently describe intolerance to small amounts of alcohol and many synthetic (petrochemical) medications, ingestion of which provokes acute outbreaks of neurological and psychological symptoms such as imbalance, tremor, abdominal pain, and mood swings as well as symptoms or signs of autonomic involvement such as rapid heart rate and increased blood pressure. Patients also sometimes exhibit cutaneous symptoms of photosensitivity such as skin rashes and other lesions.
Intrigued by informal reports of porphyrin abnormalities being found in over 70% of MCS patients being tested (57), Ziem  Donnay, developed a protocol for diagnosing disorders of porphyrin metabolism in chemically sensitive patients to try to control for the many variables that may affect testing for porphyrins and related enzymes (58). The protocol was developed in consultation with the porphyria laboratory of the Mayo Medical Laboratories and is based on Mayo's 1995 test catalog and its 1994 interpretive handbook. The protocol includes a one-page patient testing questionnaire (Appendix 3), most of which has been incorporated into the Health and Environmental History Questionnaire. It was surprising to learn from this questionnaire and verbal interviewing that even dark brown or red urine (not attributable to blood) was relatively common in MCS patients during more severe episodes of illness.
It appears that most of the patients questioned experienced significant porphyrialike neurological reactions (including sharp abdominal pain) when exposed to such well-known porphyrogenic stressors as alcoholic beverages (even a single drink was enough to trigger symptoms), usual therapeutic doses of many synthetic pharmaceuticals, and fasting or skipping meals. Most of those with skin symptoms also reported adverse reactions to sunlight exposure. It was found that because of problems with memory, concentration, and general fatigue, patients often experienced difficulty understanding and following the necessary instructions. A separate patient testing questionnaire provided a double check on such compliance (Appendix 3). Since many biomarkers of neurological and neurocutaneous porphryinopathy commonly are known to be abnormal only during periods of acute attack, patients usually were advised to wait for testing until their symptoms were "worse than usual." Unfortunately, many went for testing without waiting for an exacerbation of their illness. Of the two biomarkers in urine associated with acute attacks of neurological congenital porphyrias, porphobilinogen was normal in 10 of 11 patients tested, and 8-aminolevulinic acid was normal in 7 of 8 tested. Despite lacking these indicators of acute attack, surprisingly 12 of the 14 nevertheless tested positive for a variety of other porphyrin-related abnormalities (defined here as any result outside Mayo's range of normal).
Of the 13 patients tested for 5 different urinary porphyrins, 8 had at least one abnormality. Four patients had only elevated coproporphyrins, 1 had only elevated pentacarboxylporphyrins, 2 had elevations in both of these, and 1 had elevated coproporphyrins and uroporphryins. Of the 11 patients with stool evaluations, 3 showed abnormalities: 1 had only elevated coproporhyrins, 1 had only elevated protoporphrins, and the third patient had elevated copro-, uro-and protoporphyrins. Those with elevated fecal coproporphyrins also had elevated urinary coproporhyrins. One patient submitted a stool specimen that Mayo felt was too small to represent a 24-hr sample, so this result was not included.
All patients were given blood tests for the activity of porphyrin-related enzymes. The most common enzyme deficiency was in the activity of coproporhyrinogen oxidase (CpgO), which was below normal in 9 of the 14 patients tested. CpgO was evaluated in reticulocytes in conjunction with a reticulocyte count which, because of lab confusion was done in most but not all patients. When completed, the reticulocyte counts were normal. The activity of both ALA-D and porphobilinogen deaminase (PbgD) was below normal in 7 of the 14 patients. Uroporphyrinogen decarboxylase and uroporphyrinogen III co-synthase activity was tested in 12 and 14 patients, respectively, and found to be normal in all cases. Only two patients had no enzyme deficiencies; four had one deficiency each, four had two each, and four had three each (ALA-D, PbgD, and CpgO). The enzymes affected here occur in cytosol and mitochondria (59), and the presence of disturbed enzymes in both these compartments suggests chemical injury to both mitochondrial and cytosol cellular areas.
Despite lack of a control group, it is difficult not to conclude that porphyrin disturbances are present in a substantial percentage of chemically sensitive patients. The multiple enzyme deficiencies found in many of these MCS patients (6 of 14) are not characteristically associated with any of the known types of congenital porphyria, which usually are marked only by a single enzyme deficiency or with secondary coproporphyrinuria due to other unrelated conditions, usually an isolated abnormality, unaccompanied by enzyme deficiencies. (Because of testing techniques at Mayo, when porphobilinogen deaminase is reduced, ALA-D can also appear to be reduced, so the presence of these two abnormalities together cannot be presumed to involve two enzymes.) Given also that none of the patients tested had known family histories of active or latent porphyria, the fact that such relatively rare abnormalities and porphyrialike symptoms were found in 12 of 14 patients strongly suggests that those with MCS suffer from an environmentally acquired rather than an inherited disorder of porphyrin metabolism. (A genetic predisposition cannot be ruled out conclusively without further testing the patients and family members, but statistically such a finding is highly unlikely, since congenital porphyrias are so rare.) Although the porphyrin disturbances and enzyme deficiencies found in these MCS patients appear to be milder than in the acute types of congenital porphyria that primarily affect heme synthesis in the liver, other organs clearly are affected in cases of toxic exposure and chemical injury (e.g., irritant effects on the respiratory system, petrochemical neurotoxicity, immune dysfunction). These factors together with the involvement of multiple enzyme deficiencies may account for the somewhat more diverse symptomatology seen in MCS patients compared to those with strictly congenital porphyria. This certainly is the case with other toxically acquired porphyrinopathies such as the well-recognized forms caused by overexposure to lead, dioxin, and hexachlorobenzene, all of which, as neurotoxins, affect more than just heme synthesis and cause chemical injury unrelated to porphyrin abnormalities. Given all the above, it is not unreasonable to suspect that the different patterns of porphyrinopathy evident in MCS patients may be caused and/or exacerbated by these patients' own unique overexposures to specific toxic chemicals.
On the WAIS-R neuropsychological evaluation done by McTamney (Table 2), 6 of the 13 MCS patients tested showed a significant difference of 15 points between the verbal and performance phases of the test, all scoring higher on the verbal portion. Only one patient scored in the superior range for IQ with no apparent difficulties.
The Halstead-Reitan battery showed a number of significant scores indicating impairment (Table 3). On the Category test, 53% of patients scored in the impaired range. On the Tactile Performance test, 60% scored in the impaired range on the total time required to complete the three trials. None of the patients had problems with the dominant hand, while 30% were slower with the nondominant hand despite the previous trial with the dominant hand. Twenty-three percent of patients scored in the impaired range on the memory phase of the Tactile Performance test, while 84% scored in this range on the Localization Scale.    other chemical antibodies were tested but only those shown increased.
The Rhythm Test showed 53% in the impaired range, whereas the Speech Sounds Perception Test showed 46% in the impaired range. On the Finger Oscillation Test, 53% scored in the impaired range with the dominant hand and 61% for the nondominant hand. Trails A had 53% in the impaired range, whereas Trails B showed 61%. The overall Impairment Index scores showed 53% impairment. The most common neurologic changes on the physical exam were reduced vibratory perception (hands) and abnormal Romberg (suggesting vestibular neuropathy).

Discussion
Chemically sensitive patients from a medical practice, who often have experienced different initiating exposures, may have different patterns of porphyrin disturbance, as was found in this study. This is the pattern to be expected from the literature on acquired porphyrin disturbance. Future research on porphyrin abnormalities should compare results by type of original exposure and by type and level of current symptoms, with patients tested consistently both more common diagnoses whose which he asserts is "no better than chance during both acute and nonacute phases. symptoms have been found to overlap on most of their measures," but he also Discrepancies are found between our those of multiple chemical serjitivity (and admits that these critical data on test reproimmune testing results and those of Simon of each other) so as to be almost indistinducibility were not published in his paper et al. (60). Simon found significantly lower guishable in as many as 67% of all cases (56). Given these problems, and the TAI/CD26 cell activation among MCS (45,61). Musculoskeletal controls certainly reports of immune abnormalities here and patients compared to that of a group of would be expected to include patients with by other authors, it appears that the issue of controls from a musculoskeletal clinic, inflammation of joints, muscles, and/or cellular immune disturbance in chemically whereas our results from AAL and the literconnective tissue in whom cellular immune sensitive patients needs further study. ature show higher values compared to changes of inflammation also could be pre-Immune response probably varies greatly those of normal reference ranges following sent. In addition, these controls are likely with time and appears to differ with different chemical exposure. It is possible a coding to be taking medication for pain, a frequent types of exposures. Immune studies should error occurred in the Simon study; raw reason for clinic visits in such patients. evaluate exposed individuals over time, data should be made available to indepen-Patients taking medication for pain are not beginning as soon as feasible after causal dent researchers. Ziem personally coded an appropriate control group for neurocogexposures and following for several years. immune data for Table 4, but it could also nitive test comparisons, which also were Major immune changes may not be seen folbe independently reviewed. Since both reported in this study. We believe the clinilowing acceptable challenge doses; the studies used AAL, this extent of discrepancy cal records of Simon's control patients patient whose immune measures increased is unexpected.
should be independently reviewed to evalu-mildly to moderately with return to work-Other confounding problems exist with ate for fibromyalgia, chronic fatigue, chem-place exposure ( Table 6) had serious exacerthe Simon data, which is often cited ical sensitivity, inflammatory processes, bation of many clinical symptoms lasting because it uses a control group. However, and use of pain medication. several months. This level of exposure with the control group is problematic in the Both the control and study groups' raw accompanying symptoms would not be Simon study because it consists of patients data for the TAI /CD26 parameter also acceptable as a clinical chamber challenge. seen in a musculoskeletal clinic who were should be independently reviewed. Simon Failure to follow chemically sensitive not screened for MCS, chronic fatigue attributes his study's unusual immune findpatients over time leads to lack of undersyndrome, or fibromyalgia. The latter are ings to the laboratory's poor reliability, standing of responses to exposures and of removal from exposure. We are following more than 300 chemically sensitive patients and have observed both significant clinical improvement with reduced exposures and clinical exacerbation on reexposure (including increased abnormalities on exam and laboratory testing). This is a clinical impression, although it is not quantified in aggregate for this paper. It appears that to date the professionals who have published studies suggesting a psychological origin of chemical sensitivity do not follow these patients over time, do not remove them from exposure and observe responses and then return them to exposure and observe responses. Also, these researchers are not physicians treating patients for the disorder, and therefore are not able to observe the ongoing clinical course of the illness.
We reported onset of chemical sensitivity shortly following exposure to a wide variety of petrochemicals, combustion products, and irritants (Table 4). In some cases these were discrete events such as leaks, spills, or other acute exposure events. Often, especially in occupational settings, exposures were chronic. Typically, in these situations the illness began as a more limited sickbuilding syndrome, which with further exposure developed into chronic illness with associated chemical sensitivity. This suggests that exposure controls in the early phase may prevent the more disabling phase, and that sick-building syndrome is a more selflimited form of chemical sensitivity.
The combined abnormalities of the immune, respiratory, porphyrin, and nervous systems discussed here are incompatible with a psychologic etiology for chemical sensitivity. A systematic review of ten recent studies purporting to show a psychologic origin for chemical sensitivity revealed serious methodologic flaws in all studies including the Simon study, sufficient for the reviewing authors to conclude that none of the studies had the methodologic strength to determine that chemical sensitivity was psychologically induced (40).
Vasospasm appears to be a problem in chemically sensitive patients. Most of our patients noted frequent headaches that often were consistent with migraine and diagnosed as such by other evaluating physicians. Migraine is known to be triggered by chemical odors (62) and involves abnormal vasospasm. Cerebral vasospasm with reduced cerebral blood flow has been reported with encephalopathy following solvent exposure (63) and in patients with chemical sensitivity (42,64,65). One of our patients developed new onset of Raynaud's phenomenon (a vasospastic disorder), which we observed to be triggered by doubleblinded finger contact with carbonless copy paper (the probable cause of her chemical sensitivity). Only a few patients developed new onset of high blood pressure (another vasospastic response), which is seen primarily following exposure (their blood pressure readings usually remained normal between exposures). These vasospastic responses may involve abnormal autonomic function, which has been observed to occur with chemical sensitivity (35).
Diagnoses of fibromyalgia and chronic fatigue syndrome are common among our MCS patients (currently 75 and 85%, respectively). We recommend further research on the incidence of chemical sensitivity in these groups using the types of screening questionnaires developed by Kipen, Bell, and Davidoff. If these disorders are essentially the same, much of the research already done for fibromyalgia and chronic fatigue syndrome may also apply to MCS; this could reduce research costs and time delays. Increased sensitivity to chemicals also occurs with migraine, asthma, and other disorders. However, fibromyalgia syndrome, chronic fatigue syndrome, and porphyrin disorders are multiple-system disorders involving chemical sensitivity.
Over one-third of our patients described tremor several times or more a month ( Figure 1). Several patients were told by other physicians that they were developing Parkinson's disease during more serious periods of their illness only to have the diagnosis rescinded when they improved after environmental controls were put in place to reduce exposures. Parkinson's disease has been reported following pesticide exposure (66).
Nearly half our patients reported increased symptoms during or after swimming in a chlorinated pool (Figure 3), which has been associated with increased chloroform levels (67). Patients also often reported reduced symptoms during and after showering with, compared to without, an activated charcoal filter that helps remove chloroform. Longer showers and/or those with greater flow rates release more chloroform and other chlorinated products. Water filters that reduce chlorine therefore appear to be a reasonable control for MCS.
In addition to filters for chlorinated water, other reasonable means exist to control exposures that aggravate MCS symptoms. Aggravation from vehicle exhaust ( Figure 2) can be reduced by using an auto filter device that provides activated charcoal filtration and by avoiding ozonating devices which generate irritants. Symptom exacerbation from pesticides, cleaning products, building materials, etc. can be reduced by using less toxic products. Reasonable accommodation at the work place, at home (apartments, condominiums, etc.) and at school and public areas could be available to those with this debilitating condition by requesting that less toxic products and procedures be used. Further helpful suggestions are discussed in our Environmental Control Plan (68).
In our 10-year experience with chemically sensitive patients, no patients lost their sensitivity to chemicals. No patients were able to go to problem environments (new carpets, recent pesticides, etc.) consistently without deterioration of their conditions. Thus, MCS should be considered a permanent condition. Some patients were able to continue working if their employers provided nontoxic accommodations for their conditions. More than half the patients continued to be too sick to work under the prevailing conditions at the work place. We believe this proportion could be reduced if society pays more attention to use of nontoxic products.
The large and growing epidemic of chemical sensitivity in the United States is partly because of inadequate exposure limits for chemicals. These exposure limits have been shown to lack scientific merit (69) and to have been seriously influenced by vested interests (70)(71)(72)(73). Exposures that cause MCS at home or the work place sometimes reflect relatively widespread practices or products, some of which may be within legal limits. Many aggravating exposures are probably below legal limits but often do not provide an adequate safety margins even for the healthy population (69). The widespread indoor use of pesticides in U.S. schools, homes, work places, and public buildings is in striking contrast to practices in Germany and Scandinavia. Major policy changes with regard to chemical product formulation and use will be necessary to reduce future cases of permanent chemical sensitivity as well as to reduce disability for currently affected patients.

Summary
Most of our patients with symptoms of MCS appear to have developed chronic illness following exposure to petrochemicals, combustion products, and other irritants. Symptoms and signs in patients Environmental Health Perspectives * Vol 105, Supplement 2 * March 1997 reporting chemical sensitivity suggest involvement of the immune, respiratory, limbic, and other nervous systems (central, peripheral, autonomic) as well as impaired porphyrin metabolism. This suggests that multiple mechanisms of chemical injury probably are involved, all of which can intensify response to an exposure. Immune activation, neurogenic inflammation (74), kindling (75), and/or time-dependent sensitization (76) can amplify the body's response to chemical exposure in the immune system, respiratory system, and nervous system. It is possible that impaired porphyrin metabolism reduces the amount of heme available for cytochrome P450, part of the liver's major detoxification system for foreign chemicals, which could result in intensified symptoms for a wide range of exposures.
Research strategies are needed that allow evaluation of multiple sites of chemical injury and multiple mechanisms of injuryalmost all studies to date focus on only one type of chemical injury.
Immune activation also could lead to increased response to foreign substances, such as conventional allergens. We note in our patients an apparent high rate of new onset of allergies to mold, dander, etc., following onset of chemical sensitivity. Kipen reports what appears to be a significant level of sensitivity to chemicals among asthmatics (77). Asthma also is increased with higher levels of indoor volatile organic compounds (VOCs), formaldehyde and/or limonene (78). Persons with other allergic diatheses also should be studied for sensitivity to chemicals. If indeed chemical exposure via immune activation contributes significantly to the high and increasing rates of allergies, we may have the means to counter this alarming trend.
Studies of chemically injured populations should compare MCS patients with specific and identifiable initial exposures to MCS patients who cannot identify any specific triggering exposure. Other patient groups that should be studied for MCS include those with recurring migraines, chronic sinusitis or rhinitis, degenerative neurologic diseases (such as acute types of congenital porphyria), autoimmune disorders (such as multiple sclerosis, autoimmune hepatitis, rheumatoid arthritis, and lupus), hyperactive children, and patients with attention deficit disorder. Chemically exposed groups also merit study, especially those that have avoided further occupational exposures following work with pesticides, solvents, etc. We believe such studies will find levels of chemical sensitivity in these subpopulations that are considerably greater than currently recognized.
14. For the symptoms and health problems listed below, if you have had the problem in the last year, CIRCLE the number that best describes how often the symptom occurs, or circle 7 if not sure.