|
Health effects of predatory beneficial mites and wasps in greenhouses 7 Relation between exposure and health effects
7.1 IntroductionA considerable part of the study was measurements of health effects of the exposure. In the BIOGART study the persons were, at the annual examinations asked about a series of symptoms mainly from the respiratory tract. For every group of symptoms the first question was whether the person had experienced the symptom in question the previous year. The next questions were about the frequency, type, and relation to factors and activities, both at work and outside work (Larsen & Bælum, 2002). The analyses follow the model of disease shown in figure 1.1, which shows the relation between exposure, allergy and symptoms. During the four annual investigations the persons registered a series of symptoms, mainly from eyes, nose, and lungs (Larsen & Bælum, 2002). The symptoms were grouped in categories. Three categories from the lower airways; cough, chest tightness, and wheeze. From the upper airways; symptoms of always stuffed nose, itching nose or pharynx, and running nose. Questions about itching eyes and skin rash were also included. For each category there was a general question about symptoms at work, symptoms after a holiday, and specific additional symptoms to the main symptom. Finally the persons were asked about asthma within the last three months. In a follow-up study two principally different methods of estimating health effects are possible. As in a cross sectional study the prevalence of symptoms in relation to the concomitant measures of exposure can be estimated at each session. The other possibility is to estimate incident symptoms, i.e. newly developed symptoms between the examinations. Both methods have advantages and disadvantages. Therefore both types of analyses have been done. 7.2 Analysis of the symptom score7.2.1 Analytical methods7.2.1.1 Analysis of prevalencesAnalyses were made on two subsets of the observations, the first sample for each person (n=579) and the last sample restricted to either run 2 or run 3 (n=365). In the latter group those who had left the trade but anyhow were examined (n=27) were excluded, as no recent exposure estimates could be obtained. The latter sample corresponded with those who had measures of specific IgE against the four beneficial animals and Tetranychus urticae. Therefore, a third group of analyses were made, the relation between the prevalence of symptoms and sensitization against the four beneficial animals and T. urticae. In order to study the occurrence of asthma more specifically, an index adopted from the European study of asthma (ECRHS) was applied (Pekkanen et al., 2005). This index consists of eight symptoms of which seven were included in the present interview. For each set of data a logistic regression was made. Firstly, analyses were made only with the exposure variable (0= not exposed, 1= exposed in greenhouse, 2= applied within the last year). Secondly, the additional factors, sex, atopy (one or more positive standard skin prick tests at inclusion into the study) and smoking habits (been smoking within the last 1 year) were included. The results are presented as odds ratios with 95% confidence intervals. Due to the type of distribution of the asthma index a negative binomial regression analysis was made. 7.2.1.2 Analysis of incidenceThe incident cases for five main symptoms: cough, chest tightness, wheeze, asthma, and running nose were calculated. An incident case was defined, if a symptom occurred which had not been reported in the previous run. The time of occurrence was set to the day of examination in the actual run. However, if the symptom was reported in any of the questionnaires returned every three months between the examinations, the date of occurrence was set to the day of the first report in these questionnaires. The incidences were then analysed by a Cox-regression proportional hazard analysis including the exposure variables as well as sex, atopy, and actual smoking habit. The asthmascore variable was analysed by a cross sectional time series analysis based on the negative binomial distribution. 7.2.1.3 Lung functionThe values of FEV1, FVC, and FEV1/FVC in the first and the last measurement for each person were selected and the decline per year was calculated. These declines were anlysed by a linear regression including exposure variables as the averages over the period, height, smoking habits, and atopy as independent variables. According to bronchial hyperreactivity very few observations showed a positive value of PD20 (Larsen & Bælum, 2002). Therefore a slope value for the change in FEV1 in response to the increasing doses of histamine was calculated by the formula slope = log((x0-x)/(x0*dose)+1) adopted by (Miller et al., 2002). In order to normalize the distributions the values were log transformed before the linear regression analysis. 7.3 ResultsThe tables 7-1 to 7-5 and the figures 7.1 to 7.5 containing the results are situated at the end of this chapter. Aphidius colemani Figure 7.1 shows the prevalences in the first sample in relation to exposure to Aphidius colemani. In table 7-1 odds-ratios between the exposure estimates and the prevalence of the symptoms corrected for sex, status of atopy, and smoking habits are shown. Odds-ratios significantly different from 1 are shown in bold. If no odds-ratio is shown, one of the groups had no observations. In the initial samples a lower frequency of cough at work, chest tightness at work and itching eyes at work was seen in those exposed to A. colemani in the greenhouses, while skin rash was more frequent in those directly applying the wasp. In the last sample no significant differences between groups were seen, but characteristically no lung or nose symptoms were seen in those directly applying the species. A more consistent pattern was seen in correlation with IgE, where increased chest tightness at work, stuffed nose, itching nose or pharynx and skin rash was seen in those sensitized to A. colemani. Those exposed to A. colemani showed a higher incidence of chest tightness (IRR = 1.82 (1.07-3.10)) and running nose (IRR=2.10 (1.03-4.23)). Amblyseius cucumeris Figure 7.2 shows the prevalences in the first sample. In table 7-2 odds-ratios between the exposure estimates and the prevalence of the symptoms corrected for sex, status of atopy, and smoking habits are shown. Odds-ratios significantly different from 1, are shown in bold. If no odds-ratio is shown, one of the groups had no observations. No significant differences between the exposed groups were seen in the first sample, except a lower prevalence of skin rash at work in the exposed groups. In the last sample the same pattern was seen, especially no wheeze or asthma was seen in those applying Amblyseius cucumeris. Coughing at work and chest tightness at work were marginally more frequent in those sensitized to A. cucumeris. Wheeze at work was significantly increased while the general question “Wheeze” tended to be higher in the sensitized group. Phytoseiulus persimilis Figure 7.3 shows the prevalences in the first sample acoording to exposure to Phytoseiulus persimilis. In table 7-3 the odds-ratios between the exposure estimates and the prevalence of the symptoms corrected for sex, status of atopy, and smoking habits are shown. Odds-ratios significantly different from 1 are shown in bold. In the first measurement only one person applied P. persimilis himself, why figure 7.3 only has two columns. No relation between exposure and symptoms was seen. On the other hand, those sensitized to P. persimilis had significantly more chest tightness at start of work and wheeze than those not sensitized. Exposure to P. persimilis showed no effect on the incidence of symptoms. Hypoaspis miles Figure 7.4 shows the prevalences in the first sample in relation to exposure to Hypoaspis miles. In table 7-4 odds-ratios between the exposure estimates and the prevalence of the symptoms corrected for sex, status of atopy, and smoking habits are shown. Odds-ratios significantly different from 1 are shown in bold. No relation between the estimates of exposure in either the first or the last sample was seen, except a lower prevalence of wheeze in those exposed in the first sample. Those sensitized to H. miles had more cough and chest tightness at work than those not sensitized, while a number of other lower airways symptoms showed the same pattern all though not reaching significance. Exposure to H. miles showed no effect on the incidence of symptoms. Tetranychus urticae Figure 7.5 shows the relation between sensitization to Tetranychus urticae and symptoms. The odds-ratios corrected for sex, atopy, and smoking habits are shown in table 7-5. Significantly more chest tightness and wheeze at start at work, was seen in those sensitized to T. urticae. No specific pattern in the other symptoms was seen. 7.4 Asthma score, lung function, and bronchial reactivityThe asthma score (sum of symptoms from 0 to 7) was highly skewed, 5.2 % of the persons had more than one symptom, more among females than males and this prevalence decreased from run 0 to run 3 (7.5 % to 2.5 %). Table 7-6 shows the analyses of the change in asthma score in relation to the different exposures. This composite score did not show any relation with the exposure estimates, if any tendency, this was negative. A clear relation with atopy and smoking habits shows that relevant factors influence on the score. In the lung function values only one significant effect and no general tendencies were seen. Again smoking habit and atopy showed the expected effect, an indication of the validity of the measurements. 7.5 Cases with possible symptoms related to sensitizationTo look for the relation between sensitization and development of symptoms the persons developing sensitization during the study were scrutinized. Five persons with suspected symptoms related to sensitization are mentioned here: Female working with poinsetta exposed to A. colemani, A. cucumeris, and P. persimilis. She develops positive HR against A. colemani and concomitantly report itching nose and eyes as well as skin rash. IgE against A. colemani increased gradually from 0.08 to 0.17 OD. She has a positive prick test for birch, grass, dust mites and Alternaria. Male (no. 6 in table 6-2) exposed to A. cucumeris develops positive HR against A. cucumeris and transient weak reactions to P. persimilis and A. colemani. IgEs were, except against A. colemani in a single measurement, very low. He had a single nose symptom before sensitization but develops more as well as skin rash and eye symptoms. He had a positive prick test for birch, grass, mug worth, and dog. Female working with Campanula exposed to A. cucumeris, and H. miles, and from run 1 A. colemani. She develops in run 2 HR titre level 3 against P. persimilis and A. colemani disappearing in run 3. IgEs were negative. She develops cough, chest tightness, wheeze and nose symptoms in relation to work and general annoyance. In run 3 both sensitization and symptoms disappear. Female (no. 10 in table 6-2) working with Campanula applying A. cucumeris and H. miles. She develops a gradual increase in HR against A. cucumeris and P. persimilis although not exposed to the latter. IgEs against A. cucumeris were below 0.01 OD and the others negative. She develops nose and eye symptoms in run 2 but no report in run 3. In the study she had negative prick tests but was later in 2003 tested having several positive tests (cat, the two house dust mites, and Lepitoglyphus destructor) and prick test using a live A. cucumeris gave a very strong allergic reaction. The eye and nose symptoms were clearly related to the handling of A. cucumeris. Male (no. 16 in table 6-2) working in a large firm with several plants. They were using both A. colemani, A. cucumeris, P. persimilis, and H. miles. He had in all three runs positive HR against A. cucumeris, developed a titre of 2 against P. persimilis, and a titre of 1 against A. colemani in the first two measurements but not in the last measurement. IgEs were negative. He had eye and nose symptoms in all the runs. He had a well known seasonal rhinitis and positive prick test against birch. 7.6 DiscussionA large number of symptoms correlated with the five different exposures in three different settings gives a very large number of possible statistical tests and thereby a considerable probability of mass significance. Therefore the interpretation of the relation between symptoms and exposure rely on the pattern and directions of the significances or tendencies to effect. In the present chapter it is seen, that only spurious effects are seen in relation to the exposure estimates, especially a very low prevalence of symptoms in those actually applying the beneficial animals. On the other hand, a number of significant relations between sensitization and especially lower airway symptoms as chest tightness and wheeze were seen for all the three mites, while the wasp, A. colemani showed effects on the nose and skin rash. This wasp also showed, as the only one of the tested animals, exposure relation with the incidence data. The pattern was the same for T. urticae, a well known allergen. In the statistical analysis no relation was seen between sensitization to mites and symptoms from the upper air ways, a finding seen in both the Dutch study of Amblyseius species and the Swedish study of P. persimilis (Kronqvist et al., 2005; Groenewoud et al., 2002a). In the cases suspected of developing reactions against the mites a parallel development of sensitization and symptoms seem to be the case. It was also seen that the persons suspected of developing a clinical allergic disease were atopics with mostly several allergies to the most frequently occurring allergens. Besides, when reacting to one of the beneficial species there was a tendency to reaction to one of the others, too. The expected relations between both upper and lower air way symptoms, especially, chest tightness, wheeze, and the composite asthma score were seen with status of atopy, but correcting for this did not change the effects of beneficial mites. When looking at the composite score of asthma and lung function parameters no effect was seen in any of the exposure variables. A two or three years observation period for measuring a decline in FEV1 or FVC is very short although the large number of observations gives a high precision. On the other hand bronchial hyperreactivity is generally a sensitive measure of an asthmatic reaction. Therefore the possible effect of the predatory animals on inflammatory lung diseases, if apparent, is limited. The use of general epidemiological methods as used in the present study for allergic effects give some problems, as only a fraction of the population is in risk of developing the diseases. This risk group can only to some extend be identified and the studies will always be hampered by a lack of power, unless very large populations are used. Using a logistic regression in this setting may be a problem, when adjusting for a strong factor such as atopy. Therefore a control analysis using stratification and a Mantel-Hentzell test was done. This did not change the results. Therefore case findings can give additional information although the etiology often is hard to prove. In this present study a number of suspected cases were found. Characteristics were that only upper airway symptoms were seen, and the sensitization to the predator might be part of a general development of allergy to several factors. Whether asthma will develop at a later stage is uncertain. A very low frequency of bronchial reactivity can be a sign of a selection in the material, as persons with reactive airways leave the trade because of work related symptoms. 7.7 ConclusionsIn case reports a development of concomitant sensitization and development of upper airway and eye symptoms were seen. A number of relations between exposure estimates and symptoms were seen, both upper and lower airways symptoms as well as skin rash. On the other hand no indication between exposure to any of the predators and asthma was seen, neither in the composite asthma score nor in the lung function tests. Figure 7.1. Prevalence of the symptoms for each of the exposure groups of Aphidius colemani in the last sample of persons participating in run 2 or run 3 and still employed in the greenhouses (n=338).
Table 7-1. The odds-ratios corrected for sex, atopy, and smoking habits for the symptoms in the first sample (n_579), the last sample (n=338), and in relation to the specific IgE (n=365). Figure 7.2. Prevalence of the symptoms for persons with and without positive IgE (>0.05 OD) to Amblyseius cucumeris in the last sample of the persons (n=365).
Table 7-2. The odds-ratios corrected for sex, atopy, and smoking habits for the symptoms in the first sample (n_579), the last sample (n=338), and in relation to the specific IgE (n=365). Bold numbers denote odds-ratios significantly different from 1. Figure 7.3. Prevalence of the symptoms for persons with and without positive IgE (>0.05 OD) to Phytoseiulus persimilis in the last sample of the persons (n=365).
Table 7-3. The odds-ratios corrected for sex, atopy, and smoking habits for the symptoms in the first sample (n_579), the last sample (n=338), and in relation to The specific IgE (n=365). Bold numbers denote odds-ratios significantly different from 1. Figure 7.4. Prevalence of the symptoms for persons with and without positive IgE (>0.05 OD) to Hypoaspis miles in the last sample of the persons (n=365).
Table 7-4. The odds-ratios corrected for sex, atopy, and smoking habits for the symptoms in the first sample (n_579), the last sample (n=338), and in relation to specific IgE (n=365). Bold numbers denote odds-ratios significantly different from 1. Figure 7.5. Prevalence of the symptoms for persons with and without positive IgE (>0.05 OD) to Tetranychus urticae in the last sample of the persons (n=365).
Table 7-5. The odds-ratios corrected for sex, atopy, and smoking habits for the symptoms in relation to the specific IgE (n=365). Bold numbers denote odds-ratios significantly different from 1.
Table 7-6. The incidence rate ratios and their 95% confidence intervals for the five main symptoms in relation to exposure to the four species. The results are corrected for, sex, smoking habits, and atopy. Bold numbers denote values significantly different from 1.
Table 7-7. The relation between asthmascore and the exposure variables. The values are the estimates of the negative binomial regression on the score (0-7) corrected for sex, smoking habits, and atopy. *) denotes no or only single observations. IRR Incidence rate ratio, RD risk difference.
Table 7-8. The values of lung function and bronchial hyperreactivity in relation to the exposure values. The estimates corrected for sex, height, atopy status, and smoking habits are shown. The persons applying the products are included in the exposed group. Bold values are signifant effects (p<0.05)
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||