| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
Department of Psychology (W.W., S.D., I.V.D., P.E., O.V.D.B.), and Faculty of Medicine (B.N., H.V., K.V.D.W.), University of Leuven, Leuven, Belgium.
Address reprints requests to: Omer Van den Bergh, PhD, Department of Psychology, Tiensestraat 102, B-3000 Leuven, Belgium. Email: omer.vandenbergh{at}psy.kuleuven.ac.be
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXACKNOWLEDGMENTSREFERENCES |
|---|
METHOD: One group received prior information framing the study in the context of the rapidly increasing chemical pollution of our environment. Another group received no prior information. Conditional odor stimuli (CS) were diluted ammonia (foul-smelling) and niaouli (neutral-positive smelling); the unconditional stimulus (UCS) was 10% CO2-enriched air. Each subject breathed one odor mixed with CO2 and a control odor mixed with air in 80-sec breathing trials. The type of odor mixed with CO2 was counterbalanced across participants. Next, the same breathing trials were administered without CO2. Breathing behavior was measured during each trial; subjective symptoms were assessed after each trial.
RESULTS: Only participants who had been given warnings about environmental pollution reported more symptoms to the odor that had previously been associated with CO2, compared with the control odor. This was so for both the foul- and the pleasant-smelling odor. Symptom learning did not occur in the group that did not receive warnings. The elevated symptom level could not be accounted for by altered respiratory behavior, nor by experimental demand effects.
CONCLUSIONS: Raising environmental awareness through warnings about chemical pollution facilitates learning of subjective health symptoms in response to chemical substances.
Key Words: associative learning, • odors, • subjective symptoms, • multiple chemical sensitivity, • environmental pollution, • psychosomatic medicine.
Abbreviations: AN(C)OVA = analysis of (co)variance;; CO2 = carbon dioxide;; CS = conditioned stimulus;; CR = conditioned response;; ES = effect size;; f = breathing frequency;; FETCO2 = end-tidal fractional concentration of CO2;; M = mean;; MCS = multiple chemical sensitivity;; NA = negative affectivity;; PANAS = positive and negative affect scale;; SD = standard deviation;; TI = inspiratory time;; TE = expiratory time;; UCR = unconditioned response;; UCS = unconditioned stimulus;; VE = minute ventilation;; VT = tidal volume;
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXACKNOWLEDGMENTSREFERENCES |
|---|
Wessely et al. (4) consider MCS as another version of the so-called functional syndromes, (eg, chronic fatigue syndrome, fibromyalgia, etc.) because patients show considerable symptom overlap, share the same nonsymptom characteristics (eg, predominance of women, elevated psychiatric co-morbidity) and respond to the same therapies. The authors assume that these syndromes have a common ground and that the different diagnostic categories reflect the specialty of the consulted physician. Before that, however, symptom perception processes and patient’s causal attributions of symptoms to cues may determine which particular physician will be consulted. This perspective results in two distinct questions concerning MCS : 1) what causes the symptoms; and 2) why are they attributed to chemicals in the environment? The present study will focus on this latter question.
Because MCS often starts with a toxic exposure (5), this aversive experience can be construed in terms of associative (Pavlovian) learning: The toxic exposure can be regarded as an unconditional stimulus (UCS), inducing symptoms and the (odorous) context as the conditional stimulus (CS). However, we believe that any symptom episode (UCS) without a clear explanation (eg, stress-induced hyperventilation, chronic fatigue syndrome, fybromyalgia, etc.; cf. the first question above) may be perceived in contingency with or attributed to the presence of environmental cues (CS; cf. the second question). Regardless of the cause of the symptoms, it is hypothesized that subsequent confrontations with the conditional stimulus (CS) alone may trigger anticipatory processes at several levels of functioning subserving the experience of subjective symptoms.
To test this learning account of MCS, our group has carried out several experiments in which symptoms were induced using inhalations of air enriched with 7.5% CO2 (UCS) (see 6 for a review). Fast breathing, smothering sensations, chest tightness, feelings of choking, pounding heart, sweating, hot flushes, lump in throat, headache, and anxious feelings typically start after about 20 seconds, are moderate in intensity and disappear quickly after a 2-minute trial. A harmless odor was mixed with the CO2-enriched air and served as CS (CS + trial). Control trials contained another odor mixed with regular air (CS trial). After a number of trials of each type (acquisition phase), the participant received both odors mixed with regular air (test phase).
Our studies showed that the odor previously associated with CO2 induced an elevated level of (mainly respiratory) symptoms compared with the control odor (7, 8). Acquired symptoms persisted after 1 week and spread to new odors (9), and presenting the odor a number of times without CO2 readily eliminated the symptoms (10). Interestingly, symptom learning was more pronounced in normal participants scoring high for negative affectivity (or neuroticism) (9, 11) and in psychosomatic patients (8). One of the most interesting findings was that symptoms were only acquired in response to foul smelling as CS (eg, ammonia or butyric acid) and not to pleasant ones (eg, fresh-smelling niaouli), whereas in real life, patients with MCS typically also report symptoms in response to fragrances. The present study wanted to elucidate this discrepancy.
An important difference between acquiring symptoms in real life compared with in the laboratory is social context. It is well known in human learning psychology that culturally transmitted information, existing beliefs, and emotions elicited by the CS may bias the perception of contingency between a CS and a UCS (12, 13). For example, one can imagine that concern about environmental pollution and about the ubiquitous presence of hazardous chemicals may induce beliefs that facilitate associating harmless chemicals with symptoms. The present experiment was designed to test this hypothesis. Before the experiment we gave half of the participants a leaflet containing information similar to that found on websites and other media about environmental pollution and a description of a patient with MCS. The remaining half did not receive this leaflet. Both groups were then tested in a set up with ammonia and niaouli as CS. We expected symptom learning in response to both the foul smelling and pleasant odor used as CS in participants provided with prior information.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXACKNOWLEDGMENTSREFERENCES |
|---|
Participants
Thirty-two participants were engaged in the experiment (27 women; age 18–30 years.). Six participants were psychology freshmen receiving course credit, 26 volunteer students were paid 500 BEF ($12). The study was part of a project that was approved by the Ethical Committee of our department in accordance with the Helsinki Declaration of 1975.
Materials and Apparatus
Odors.
Ammonia (NH3) and niaouli (a volatile oil, containing 65% eucalyptus) were used as CS. Ammonia was dissolved in water in a 0.8% concentration.
Subjective measures.
Before the experiment, a brief health inquiry was administered to exclude participants with past or present diseases (eg, asthma, cardiac problems, etc.) and the trait version of the Positive and Negative Affect Scale (PANAS, 14), assessing negative affectivity. Subjective symptoms were measured after each trial with a 16-item symptom checklist (subsets were arousal, respiratory, cardiac/warmth, tingling, and unclassified symptoms) and 5 "dummy" symptoms (usually not reported during CO2 inhalation). All items had a 5-point answering scale reflecting the intensity of the sensation. The list started with an extra item measuring intensity and (un)pleasantness of the inhalation trial as a whole and ended with an extra item measuring experienced intensity and (un)pleasantness of the odor perceived during the inhalation trial. All these questionnaires were administered by computer. After the experiment, the participants were asked to indicate their level of worry experienced during the different stages of the experiment and the credibility of the given information.
Physiological measures.
Subjects breathed through a mask, enclosing mouth and nose, which was connected to a pneumotachograph (Fleish no. 2, Epalinges, Switzerland) and an infrared CO2 monitor (Poet II, Criticare, Waukesha, WI). The following primary parameters were extracted per breathing cycle: inspiratory and expiratory time (TI, TE), inspiratory and expiratory volume (VI, VE), drive and end-tidal fractional CO2 (FETCO2) (see 9 for technical details). Odors were vaporized using a Devilbiss 646 nebulizer and fed into a tube attached to the mask. The air mixture consisted of 21% O2 and 79% N2. The CO2 mixture consisted of 10% CO2, 21% O2, and 69% N2. Compared with previous studies, we raised the concentration of CO2 in the breathing air from 7.5% to 10%, but we shortened the trial duration from 120 to 80 seconds. In addition, we applied only one rather than three acquisition trials. This was done to increase the ecological validity of this laboratory model for MCS: a stronger UCS given once may mimic a single, accidental exposure more closely.
Procedure
Half of the participants (information condition) were given a leaflet in the waiting room, describing "the widespread chemical pollution of our environment" as a potential cause for MCS and a MCS case description (see Appendix). The other half were given no information and were directly led to the experimental room. There, all subjects were given a brief health inquiry and the PANAS. Next, it was explained 1) that the purpose of the study was to test breathing responses to different air mixtures; 2) that low intensity transient complaints could occur; 3) and that they could stop the experiment at any moment. Written informed consent was then collected. In the information condition the purpose of the study was further explained as part of an investigation into the mechanisms of MCS, implying a test of the reaction of normal persons to the substances that had been tested previously on persons with this disease. In the "no information" condition, no references to MCS or other illnesses were made. Subsequently, participants were informed that two different gas mixtures were to be inhaled and that symptoms of moderate intensity could occur that would disappear after a few minutes. In the information condition, it was repeated that the gas mixtures had been tested previously on MCS patients and that the specific aim of this study was to find out more about the reaction of normal subjects. To avoid differences in pre-existing beliefs and associations (ammonia is commonly used in household products and niaouli contains eucalyptus which is also widely known), the two mixtures were fictitiously labeled as "hydrogen nitride" for the foul smelling ammonia and "nialin citrate" for the neutral to fresh-smelling niaouli. These new labels were also meant to increase credibility of the given information. Next, instructions were given to handle the mask properly.
The experiment consisted of an acquisition and a test phase, each starting with a baseline trial implying breathing regular room air through the system without odor. Next, half of the participants within each information condition received sequentially ammonia mixed with CO2 (CS+ trial) and niaouli mixed with room air (CS- trial). The other half received the reversed combination: niaouli mixed with CO2 (CS+ trial) and ammonia mixed with room air (CS- trial). The order of the CS+ and CS- trial was randomized across participants. In the test phase, the same trials were administered in the same order without the CO2. During each trial, a sentence on the screen displayed the label of the substance that was being administered ("hydrogen nitride" or "nialin citrate") and—in the information condition—whether the substance was generally well or poorly tolerated by MCS patients for the CS- and CS+ trial, respectively. A 10-minute pause was inserted between acquisition and test.
Lastl, a few postexperimental questions were answered, the participants were thanked and debriefed about the substances they had inhaled, and explicitly asked not to tell potential participants about the specific details of the experiment.
Design and Analysis
The 32 participants were randomly assigned to one of four between-subject conditions (Information: yes/no; CS+ odor type: ammonia CS+/niaouli CS+; N = 8 per condition). A 2 (Information; between subjects) x 2 (CS+ type; between subjects) x 2 (Conditioning: CS+ vs CS-; within subjects) ANOVA was run on the total symptom score per trial (sum of the 16 items) in test phase. Student’s t tests were used to test a priori differences between specified conditions directly (15) and standardized effect size (ES) (16) and power were calculated. Because our data were not normally distributed, we first performed a reciprocal transformation (17). However, in accordance with the general idea that the F and t-statistics are robust against normality violations (15, 17) , tests on nontransformed data yielded basically the same results. Also the nonparametric alternative to the t tests (Wilcoxon matched-pairs test) did not alter the conclusions. Therefore, we report the analyses on the nontransformed data to allow easy interpretation of the symptom scores and straightforward comparisons with data from our previous studies using this paradigm. Significance level was set a priori at 0.05 and Bonferroni sequential correction for multiple tests was applied (18).
Effects on subjective symptoms were further explored by looking at breathing behavior in corresponding cells of the design. Means per trial for FETCO2 (%), TI and TE (sec), VI and VE (ml) and drive were analyzed.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXACKNOWLEDGMENTSREFERENCES |
|---|
Acquisition Phase
More symptoms were reported during trials containing CO2 (M = 38.3; SD = 12.3) than during trials with regular air (M = 22.6; SD = 8.1) and higher FETCO2 was measured (M = 7.5; SD = 0.8 vs M = 4.6; SD = 0.8 for air trial). The effect of CO2 inhalation on symptoms did not interact with CS+ odor type or Information. There was no difference in symptom level reported to inhaling either odor mixed with room air (CS-).
Test Phase
Overall, more symptoms were reported to the CS+ odor (conditioning, F(1,28) = 6.5; p < .05). This effect interacted with Information (F(1, 28) = 8.7; p < .01) and with CS+ odor type (F(1,28) = 5.2; p < .05), replicating selective conditioning to ammonia as CS+ (7–9, 11).
Testing the specific expectation, we found a learning effect for both types of odors as CS+ with prior information (Conditioning with ammonia as CS+ : t(7) = -3.6; p < .01; ES = 1.4 (large), power = 0.9; with niaouli as CS+, t(7) = -3.0; p < .02; ES = 1.0 (large), power = 0.7; see Fig. 1). However, learning of symptoms did not occur to either odor in the condition without prior information (Conditioning for ammonia CS+ : t(7) = -0.9; ns; ES = 0.3 (small), power = 0.1; for niaouli CS+, t(7) = 0.9; ns; ES = -0.3 (small), power = 0.1).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXACKNOWLEDGMENTSREFERENCES |
|---|
A potential methodological problem with experimentally induced beliefs is that they may create demand effects, that is, the participant may deliberately attempt to comply to his/her understanding of the experimenter’s expectations. However, this explanation is quite unlikely here: First, there was no learning effect for "dummy" symptoms. Second, the learning effects were mostly pronounced for the respiratory, the cardiac and the unclassified symptom subset. Both findings replicate the pattern of results in our previous studies without belief induction. Compliance because of belief induction would not predict different effects for the subsets and certainly not the specific pattern of results found earlier.
Which processes mediated the effect of the information manipulation on symptom learning? Conveying information similar to that found on MCS websites and in other media to mimic real life individual differences in beliefs about potential hazards of chemical pollution may have several effects. Davey (12) described cognitive and emotional variables that may modulate associative learning, either by influencing outcome expectancy or by modulating the evaluation of the acquired response. The information we conveyed may have affected both processes. For example, the message "nialin citrate is poorly tolerated by MCS patients" may have affected outcome expectancy, while the description of a disabled MCS patient may have turned the evaluation of the symptom episode more negatively. In any case, ratings of the affective valence of the CS+ odors were more negative with information than without information (-1.18 vs -0.31, t(30) = -2.33; p < .05) and this was more so for niaouli. Apparently, the affective quality of the odors is a critical feature for symptom learning, which reinforces conclusions reached in our previous studies (9, 10). However, the present data add that negative affective valence may not only imply the sensory-affective quality of the odor, but also its meaning. In other words, a nice smelling fragrance may become a cue for learned symptoms, if one beliefs that the chemical composition is detrimental for one’s health. Other explanations for the effect of information, such as more intense odor perceptions or more worrying were not corroborated by the data. Unlike in previous studies, subjects receiving niaouli as CS+ without information had more symptoms overall (Fig. 1). Currently, we have no explanation for this finding.
The present data may be relevant for prevalence rates of MCS. Although it is difficult to give reliable estimates, the impression exists that rates are higher in the U.S. than in Europe. Conservative estimates in the U.S. ranged between 3% to 5% (19). However, the 1995 California Behavior Risk Factor Survey showed 6.3% of doctor-diagnosed "environmental illness" or MCS among the general population and 15.9% reported being allergic or unusually sensitive to everyday chemicals (20). In Europe, important differences exist among countries as regards the specific triggers (eg, pesticides, wood preservatives, new carpet installations) that are held responsible for the same symptoms (21). An explanation may be that differences in media coverage may have induced differences in awareness, chronic concerns or even irrational fears (22) about chemical pollution and different a priori beliefs about causal connections between (undetectable) quantities of toxins in the environment and symptoms (23–25).
The present findings may also be relevant for real life conditions where stress-related symptoms occur in the context of environments full of salient smells or wherein chemical substances have been released accidentally (Ground Zero, NY City) or intentionally (eg, anthrax in the aftermath of the September 11 terrorist attacks). It is likely that fears that go along with salient beliefs about (innocuous) stimuli being dangerously contaminated facilitate associations between such stimuli and stress-related symptoms.
To summarize, reading a message inducing beliefs about the potential harm of chemical substances in our environment facilitated learning of symptoms in response to odorous chemicals: less trials were required and learning occurred to pleasant odors as well. This suggests that warnings and campaigns against environmental pollution, while having important beneficial effects for the environment, may inadvertently facilitate acquiring symptoms to chemicals in the environment and promote the spreading of MCS, mass sociogenic illnesses, and the like.
|
APPENDIX |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXACKNOWLEDGMENTSREFERENCES |
|---|
What Is Multiple Chemical Sensitivity or MCS?
MCS is a syndrome that is not yet very known in Belgium. In the U.S., however, it is widely spread; approximately 3% to 5% of the population suffers from it. Common complaints are light-headedness, dizziness, pounding heart, pain in the chest, difficulty concentrating, fatigue, coughing, lack of appetite, etc. But also headaches, peculiar odor perception, tingling sensations in hands or face and other complaints occur. The symptoms often seem to be evoked by exposure to chemical substances. Therefore, many lay people and specialists think they are caused by increasing environmental pollution in modern society. Indeed, you probably have heard about ever more PCBs and dioxins being around, and about food additives serving to enhance color and taste or to conserve the food. In addition, cleaning products are extensively used in and outside the home and exhausts and fumes from cars, factories, and garbage ovens pollute the air. Think further about the ever increasing use of pesticides and insecticides in agriculture, about new types of plastics and polymers in tools and other user goods (even in toys !!), and about all kinds of new materials used to insulate our houses (eg, polyurthanes, and the like).
Experts believe that, because of this "overdose" of chemical products in our environment, some people may develop a hypersensitivity to the extent that they eventually cannot tolerate exposures to substances any more that are generally considered harmless, such as perfumes. They are often unable to walk in the city, because the air that is only mildly contaminated by car exhausts is making them sick.
The Story of Eric
Eric has been complaining for some time about not feeling well when he is exposed to chemical products. When he enters a room where such products have been used recently, he becomes light-headed, has problems with his balance, and has difficulty breathing. He is about to faint and feels dysphoric, looses his vigour and strength. This condition may last for hours and afterward, he often has a headache for a couple of days and prefers to stay home. He lives in a rural community. He likes to work in the garden and regains his strength this way. "Nature heals itself," he says.
The problems started 5 years ago. Eric was working as an employee in a company that produces silicones. Regularly, he had to enter the production units and the warehouses for control and advice. All the employees from the warehouses had a medical check-up regularly, Eric as well. There were never serious problems reported by anyone except by Eric.
The complaints developed gradually in Eric’s case. First they were tolerable, but they slowly became more disturbing. After a while, he was avoiding any odor that could trigger complaints. He avoided entering the productions units and the warehouses and he called in sick very often. Finally, his problem and avoidance behavior escalated such that he got fired.
Now he is unemployed, but his problem has not improved. Meantime, all kinds of products that—according to him—had a chemical odor have been removed from the house: paint, thinner, white spirit, some types of soap, ethyl alcohol, several household cleaning products, and even perfumes; he considered them all poisonous. His avoidance behavior has become so bad that when the house is being cleaned, he has to leave.
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXACKNOWLEDGMENTSREFERENCES |
|---|
Received for publication February 18, 2002.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXACKNOWLEDGMENTSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  K Osterberg, R Persson, B Karlson, F C. Eek, and P Orbaek Personality, mental distress, and subjective health complaints among persons with environmental annoyance Human and Experimental Toxicology, March 1, 2007; 26(3): 231 - 241. [Abstract] [PDF]
|
|
|
|
 |
|
 M. Lopez-Grancha, C. Sanchez-Amate, M. Navarro, F. Carvajal, F. Sanchez-Santed, and I. Cubero Lateral Parabrachial Lesions Disrupt Paraoxon-Induced Conditioned Flavor Avoidance Toxicol. Sci., May 1, 2006; 91(1): 210 - 217. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K Grill and S O Hansson Epistemic paternalism in public health J. Med. Ethics, November 1, 2005; 31(11): 648 - 653. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Greenhalgh and S. Wessely 'Health for me': a sociocultural analysis of healthism in the middle classes Br. Med. Bull., June 1, 2004; 69(1): 197 - 213. [Full Text] [PDF]
|
|
|
|
 |
|
 W. Evans Bibliography Public Understanding of Science, October 1, 2003; 12(4): 445 - 446. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
