4. Results

Appendices A and B provide details of the designs of selected studies which collected information from respondents on the health-related impacts of wind turbines. Appendix A includes studies which sampled populations or representative samples of population for the purposes of estimating 1) the magnitude of an outcome, 2) the magnitude of differences between groups with respect to an outcome, or 3) the relationships between exposure and outcomes. This table presents only the initial studies reporting data collection, and does not include subsequent analyses of the same dataset.  Reviews are likewise not included.

Appendix B includes case-series and descriptive studies, the most basic of designs.

Campbell and Stanley¹³¹ comment that "such studies have such a total absence of control as to be of almost no scientific value".

A summary of evidence for health effects listed in Table 5 is presented in this section. Wherever possible, it is noted whether evidence is found in the studies of Appendix A or B.

 

Annoyance

The main studies addressing annoyance are included in Appendix A.

Three studies consistently demonstrated a dose-response relationship between wind turbine sound exposure and annoyance, two in Sweden^15,32  and one in the Netherlands^28,34.  In all three studies data were collected by questionnaire, and in all three the purpose of the survey was masked as a survey of general noise in the community. All three studies were postal surveys, with one hand delivered_¹⁵.  The survey questionnaire asked about several health outcomes: annoyance (indoors and outdoors), sleep interruption, chronic disease, diabetes, high blood pressure, cardiovascular disease, tinnitus, impaired hearing, headache, undue tiredness, feeling tense or stressed and feeling irritable. The data sets for these surveys have been subsequently re- analysed^27,72,73, and presented in meta-analyses^29,74,75.

Of all the health outcomes included in the questionnaire, only annoyance consistently demonstrated a dose-response relationship with sound exposure^{15,27-29,32,34,72-75}. The percentages of all respondents who were fairly or highly annoyed were 4% (of 754) in the 2007 study by Pedersen and Persson Waye³²  and 16% (of 341) in their 2004 study¹⁵, with a combined estimate of 8% (of 1095)29.  At exposures above 40 dBA, the two studies report percentages of 15%32 and 44%15  annoyed, although only 20 and 25 respondents respectively were exposed to such high levels, and therefore confidence intervals are large (CI95% 3%-38% and 24%-65%). In the third study, Pedersen, van den Berg, Bakker and Bouma reported 10% (of 706) respondents annoyed²⁸.  They found that annoyance was highest between 35-40 dBA (18%, CI95% 12%-24%) and 40-45 dBA (18%, CI95% 11%-28%). Above 45 dBA, 12% (CI95% 5%-23%) were annoyed²⁸.

Working from the same dataset, Bakker and colleagues reported increasing annoyance with sound exposure, with 2% (CI95% 1%-6%) of residents reporting being rather or very annoyed at

<30 dBA, 8% (CI95% 5%-13%) at 30-35 dBA, 20% (CI95% 14%-28%) at 36-40 dBA, 25% (CI95% 15%-38%) at 41-45 dBA, and 29% (CI95% 11%-52%) at >45 dBA72.  These levels represent annoyance while outside. Annoyance while indoors exhibited the same dose-response relationship in all three studies, but with fewer respondents reporting annoyance^15,28,72.  Using a combined dataset from all three studies, Janssen, Vos, Eisses and Pederson⁷⁵ found that at an Lden** of 45 dB at the facade of the dwelling, 12% of persons would be highly annoyed indoors, while 26% would be highly annoyed outdoors. The association was not strong, however. In the 2004 study of Pedersen and Persson Waye¹⁵, wind turbine noise exposure explained 13% of the variance in outdoor annoyance, while 20% of the variance was explained by wind turbine noise in Pedersen et al²⁷. Using a dose-response curve developed from the results of these studies, Verheijen, Jabben, Schteurs and Smith²⁶ estimated that approximately 1,500 of the 440,000 individuals exposed to wind turbine sound above 28 dBA outdoors in The Netherlands are severely annoyed by wind turbine noise indoors.

Shepherd and colleagues provide some support for these findings in a 2011 study in New Zealand⁵¹. The study used the tool "2010 Well-being and Neighbourhood Survey" and asked questions about health-related quality of life, annoyance, perception of living environment (amenity), and noise sensitivity.  Study subjects lived within 2km of with turbines, while the comparison group lived >8km from the turbines. The survey was masked; there were no questions about wind turbines, and distracter questions on air quality, traffic, neighbours and neighbourhood problems were included. Because respondents had to write in "wind turbines" if they were a source of annoyance, these responses were noted only in the study population; 59% (of 39) indicated wind turbines to be a source of annoyance, and the mean rating of (4.59 (of 5), SD 0.65) indicated severe annoyance. By contrast, the two groups reported similar annoyance with noise from traffic and neighbours.

In addition to the above findings, it has been demonstrated in laboratory experiments by Persson Waye and Öhrström that wind turbine sound is annoying⁷⁶, and it was confirmed by Lee et al. that the specifically modulated character of the sound is annoying¹⁹.

Wind turbines are perceived to be annoying at sound levels below that of other industrial or transportation noise sources. In the most recent and comprehensive study done to date (2011), using the combined dataset from Pedersen and Persson Waye ^15,32 and Pedersen et al²⁸, Janssen and colleagues found that indoor annoyance and high annoyance begin at lower noise levels and increase faster with wind turbines than with industrial noise sources, although comparisons are difficult to carry out because industrial noise annoyance is generally not assessed below 45 dBA⁷⁵. Further, both indoor annoyance and high annoyance with wind turbine noise begin at lower noise levels and increase at roughly comparable rates as annoyance with air, road and rail traffic noise. In the words of Janssen and colleagues⁷⁵, "the results suggest that, at relatively low exposure levels, wind turbine noise induces an annoyance response that is expected to occur only at much higher levels of transportation noise and other industrial noise sources". As the following section describes, annoyance is related to respondent variables as well as to attributes of noise.

 

** the mean of day, evening and night levels

Non-acoustic sources of annoyance with wind turbines

Pedersen and Persson Waye^15,32, Pedersen et al.²⁸ and subsequent analyses of these datasets found that attitudes about the visual impact of wind turbines on the landscape was correlated with annoyance^{15,28,29,32,74,75}, and elsewhere the visual impact was found to be a major annoyance^77-79. Pedersen and Persson Waye²⁹ found that visual impact had roughly the same effect on annoyance as sound exposure²⁹. In cross-sectional studies of approved and existing wind turbine sites in Scotland and Ireland, Warren, Lumsden, O'Dowd and Birnie found that complaints about visual appearance far outweighed complaints about noise, and that fears about noise prior to construction tended to be exaggerated⁸⁰.  A Swedish commission on wind turbines found that visual impact was the most troublesome factor⁷⁸. A postal survey of a random population sample in Sweden (n=547, 56% response rate) reported  by Ek found that far more respondents reported visual impact on the landscape than 'disturbing noise pollution' as a negative impact of turbines, but in this study respondents were drawn from the general population, not just those living near turbines⁸¹.

The shadow flicker produced by the visual of the shadows of the turbine blades moving against the ground has also been suggested by Knopper and Ollson as a possible source of annoyance⁸². In a grounded theory study, Pedersen et al. found that of fifteen interviewees, three considered "flickering light" and one "movement" to be the primary stimuli of concern, compared to ten who cited noise⁵⁰.  It should be noted that at the speed at which large modern wind turbines rotate (5-20 rpm)¹² the shadow flicker they produce is not likely to pose any risk of inducing seizures among the photosensitive population^83,84.

In addition to the visual impact of wind turbines, several other factors have been found to be associated with annoyance. Janssen et al. reported noise sensitivity to be associated with annoyance⁷⁵. General attitudes towards wind turbines were also found to be correlated with annoyance by several authors^{15,28,29,32,33,74}.  However, using logistic regression to predict noise annoyance, Pederson and Persson Waye found that attitudes to wind turbines and noise sensitivity made no significant contributions to a model that already contained noise exposure and attitudes to visual impact, and that attitude to visual impact had the strongest relationship with annoyance (OR=5.11, CI95% = 3.10,8.41)¹⁵.

The perception that a location is a suitable place to "restore myself and regain strength" demonstrated a relationship with outdoor annoyance with wind turbines^29,32. Other factors have been reported to contribute to the acceptance of wind farms or annoyance with them. These include the consultation process during the development of the wind farm⁸⁵ and perceived (in)ability to effect the development process⁵⁰, perceptions of the purpose of the land⁵⁰, experiences dealing with local authorities⁵⁰, lack of control over the noise source⁵⁰, and perceptions of intrusiveness^28,32,50. In their 2009 study, Pedersen et al. found a strong negative correlation between financial compensation and annoyance²⁸. Only 3 of 99 respondents reporting financial benefits reported being annoyed, even though those reporting financial gain were significantly more likely to live in higher exposure levels (76/99 above 40 dBA)²⁸. Bakker et al. found that those benefiting from wind turbines were four times less likely than those who did not receive compensation to report being rather or very annoyed (3% and 12% respectively, p<0.05) and no respondent who benefited reported indoor annoyance⁷².

 

Other relevant studies of annoyance as an adverse health effect

In a 2005 bulletin entitled 'Community Noise Annoyance,' Health Canada indicated that "there is concern about excessive community noise because people can find it highly annoying" and that over a period of time, high annoyance can affect quality of life⁸⁶. They note also that sleep quality and ability to concentrate can be affected. Health Canada indicates that noise starts to make people highly annoyed when the sound level outside the home is around 55 dBA, whereas the studies cited above found that high levels of annoyance seem to begin much lower with wind turbines, at about 40 dBA^28,75.  A Health Canada-commissioned study found roughly 8% (CI95% 7%-9%) of Canadians are extremely or very annoyed by noise sources outside their home, with the highest proportion of complaints relating to traffic noise⁸⁷. Wind turbines were not listed as a choice of noise source in this study.

Guidelines for community noise are published by the World Health Organization, in which noise annoyance is considered to be an adverse health effect and the relationship between noise and annoyance is considered to be complex, with many variables involved^88,89.  A more recent study by the WHO Regional Office for Europe states only that the WHO definition of health "implies that noise-induced annoyance may be considered an adverse effect on health" and that some experts believe that severe annoyance should be thus classified⁹⁰. Both documents are dealing with much higher exposure levels than the wind turbine studies, in which all exposure levels are below 55 dBA^88-90.

A review by Berglund, Hassmen and Job found that acoustic variables (sound level, exposure length, frequency) predicted up to one-third of the variance in noise annoyance, measured here by "being disturbed by noise"⁹¹; reinforcing the importance of other, individual factors⁹².  The same review reported the finding that long-term disturbance should be considered negative stress and that long-term stress could have an adverse health impact⁹¹. A meta-analysis by Fields of 136 studies in 1993, prior to the use of modern wind turbines, found no association between annoyance and demographic variables or financial benefit⁹³, in contrast to the more recent work⁷⁵.  Fields also found no association between annoyance and ambient noise levels, but did find an association between annoyance and personal attitudes (fear of danger from the noise source, noise prevention beliefs, general noise sensitivity, beliefs about the importance of the noise source, and annoyance with impacts of the source, other than noise)⁹³. It should also be noted that at equal intensity, noise sources with more low-frequency components receive more complaints than those with fewer low-frequency noise components⁹¹.

Sleep quality

The generic label 'sleep quality' is used here because some studies look at sleep disturbances, while others look at satisfaction with sleep. 'Diminished sleep quality' in this section refers to any undesirable sleep outcome.Pedersen and Persson Waye¹⁵ and Pedersen et al.²⁸ reported a dose-response relationship between wind turbine sound levels and sleep interruption (interruption of sleep by any noise source) but a subsequent analysis by Pedersen revealed low odds ratios in both studies (OR 1.12 (CI95% 1.03-1.22) and 1.03 (CI95% 1.00-1.07), respectively)⁷³. Upon further examination, the rates of sleep interruption were stable across sound level categories with increases only at the highest levels (40 dB in Pedersen and Persson Waye¹⁵ and 45 dB in Pedersen et al.^28,73). Pedersen and Persson Waye³² and Pedersen and Persson Waye²⁹, the latter using combined data from the 2004 and 2007 studies of Pedersen and Persson Waye^15,32, did not report the same relationship, although Pedersen⁷³ speculates this could be because Pedersen and Persson Waye³² studied more built-up areas.   Bakker et al.⁷², in subsequent analysis of Pedersen et al.²⁸, found that there was a significant difference in sleep disturbance (waking up at least once a month) between respondents exposed to < 30 dBA and those exposed to >45 dBA from wind turbines, although this includes disturbances attributed to other noise sources as well as wind turbines. Among 265 people who noticed wind turbines and indicated that they did not directly financially benefit from the wind turbines, wind turbine sound exposure was not found to predict sleep disturbance; this was found regardless of whether they lived in rural (quiet) or urban (noisy) areas, and when the two areas were combined in analysis⁷².  Five percent (CI95% 3%-6%) of respondents (6%, CI95% 4%-10% rural; 4%, CI95% 2%-6% urban) reported wind turbines as the sound source of their sleep disturbance, while 13% (CI95% 11%-16%) reported being disturbed by people or animals, and 15% (CI95% 13%-18%) by traffic or mechanical sources; 67% (CI95% 63%-70%) were not disturbed⁷².In Pedersen and Persson Waye^15,32 and Pedersen et al.²⁸, sleep interruption was strongly associated with both indoor and outdoor annoyance, after adjusting for sex, age, dBA exposure category, and economic benefit for Dutch respondents⁷³.  In a subsequent analysis of the data using structural equation modelling, Bakker et al. report that among respondents who noticed the sound of wind turbines, annoyance was the only factor to predict sleep disturbance⁷². They speculated that annoyance may increase sensitivity to environmental sounds, and that night-time disturbances are reactions to combined environmental noise levels, not just wind turbines.Shepherd et al. reported that residents living within 2km of wind turbines were significantly less satisfied with their sleep (p=0.006) than those who lived beyond 8km, which the authors suggest is in keeping with the findings from the above studies⁵¹.Nissenbaum et al, in an American study, also report supporting evidence of sleep disturbances related to distance from wind turbines. Surveys were sent to all residents living within 1.5km of two wind farms, who were then compared to a randomly selected control group of residents living between 3 and 7km from the same wind farms⁵². The survey was not blinded. The questionnaire included validated instruments measuring sleep disturbance (Pittsburgh Sleep Quality Index (PSQI)) and sleepiness (Epworth sleepiness scale (ESS)). Results suggested a dose response relationship between distance from the turbines and the scores on both sleep scales (mean PSQI 7.8 in the study group and 6.0 in controls, p= 0.046; mean ESS 7.8 in the study group and 5.7 in controls, p=0.032). However, both study and control groups had the mean PSQI scores above 5, indicating poor sleep quality, and the mean ESS scores were in the normal range (a score below 10). There were no significant differences between the two groups in the proportion of respondents with abnormal scores (PSQI >5 for 65.8% of the study group and 43.9% of controls (p=0.075); ESS >10 for 23.7% of the study group and 9.8% of controls (p=0.13)).  This study was limited by a small sample size.Phipps et al. report evidence for sleep disturbances at some distance from wind turbines^54,55.   A postal survey was sent to households at varying distances from a series of wind turbines of mixed age and size (N=614, 56% response rate).†† The addresses of respondents were checked against postal records to determine the distance to turbines, accurate within 0.5 km.   All respondents lived more than 2 km from the nearest turbine, with the majority (85%) living within 3.5 km. The survey was unblinded, and asked a series of questions concerning the visual impact, noise and general attitudes towards turbines, whether they had complained or considered complaining to authorities about the turbines, and opinions on future proposed developments. Seven percent of respondents (42/614) reported occasional sleep disturbance, 3% (n=21) reported frequent sleep disturbance and 1% (n=5) reported sleep disturbance 'most of the time', although distances from turbines are not reported for those disturbed. No dose-response relationship was examined.This section has, until now, discussed studies reported in Appendix A. However, sleep disturbances are commonly reported as adverse health outcomes in case series and case reports (Appendix B). Since 2009, Krogh, Gillis, Kouwen and Aramini⁵⁷, have conducted an on-going survey of Ontario residents living near wind turbines who believe that they are experiencing adverse health effects because of the turbines. The availability of the survey was advertised by word of mouth, fliers, community gatherings and the internet, and all those experiencing health effects they attributed to wind turbines were encouraged to fill out a survey. The survey permits multiple respondents from the same family. The survey instrument was developed by Harry⁵⁶ and consists of a list of symptoms: headaches, palpitations, excessive tiredness, stress, anxiety, tinnitus, hearing problems, sleep disturbance, migraines, depression and any other symptom; respondents are asked to check those that they have experienced in "excess...since living near a wind turbine". The survey also asks for the distance to the nearest turbine, length of time at home, whether they have sought help from a doctor, and whether their quality of life has been altered, as well as demographic information. Of 97 respondents, 71% reported disturbed sleep. Results are reported to be suggestive of a dose-response relationship, although not significant at p=.05. Harry, using the same questionnaire, surveyed 42 people from across the UK, including multiple members of families, experiencing adverse health effects that they attributed to wind turbines⁵⁶.  Of 42 respondents, 67% percent reported sleep disturbances at distances between .3 and 1.6 km from the nearest turbine. Dose-response was not investigated.  Both of these studies are subject to selection bias, as respondents are self-selected on the basis of adverse health events which they attribute to wind turbines.

In a similar case series by Pierpont, again subject to selection bias, sleep disturbance was among the most frequently reported symptoms and is among the 'core' symptoms of wind turbine syndrome⁵³.  Pierpont conducted telephone interviews with 23 adults and teenagers from 10 households in five countries, collecting medical case histories from them and for 15 children in the households. The households were selected for the severity of the symptoms experienced after the wind turbines were installed; to be included they must have moved or spent significant time away from the turbines in an effort to alleviate the symptoms. All households were located between 305m and 1.5km from the nearest turbine, all of which were recent installations (most in 2006 and 2007). Pierpont generated a baseline, or pre-exposure, medical status, and then compared the symptoms that had worsened or arisen after the installation of wind turbines that were "closely linked in time and space" to the turbines, including whether the symptoms disappeared when away from the turbines.  It should be noted that there were a significant number of respondents with serious re-existing medical conditions, with only 17 adult respondents, 18 or older, being recorded as 'good' pre-exposure health, although part of Pierpont's research design was to determine what pre-existing conditions created the most risk of wind turbine syndrome. Of the 36 respondents with data, ranging in age from 2-75, 89% suffered from sleep disturbances, all but three of whom reported improvements away from turbines. No dose response relationship was investigated.

 

†† Results are drawn largely from the cited presentation, because they are more up-to-date and correct some problems in the original report. The report was used only when details were lacking in the presentation. 

Psychological distress

The term 'psychological distress' is used here to indicate a cluster of symptoms, or sometimes a general measure of psychological distress, due to the range of questions asked in the various studies. Four symptoms defined as stress symptoms by Pedersen and Persson Waye: headaches, undue tiredness, feeling tense/stressed, and feeling irritable⁷³, are reported here individually under the heading of "psychological distress", along with reports of findings using validated general measures of psychological distress.

Of studies included in Appendix A, re-analysis of the data of Pedersen and Persson Waye^15,29,32 revealed no association between sound exposure and symptoms of psychological distress described above⁷³.  Bakker et al., using the data from Pedersen et al.²⁸, found psychological distress (measured using the 12-item General Health Questionnaire) to be correlated with sound exposure for all respondents who noticed wind turbine sound and in the sample of respondents from quiet areas who noticed wind turbine sound⁷².  However, including annoyance in the model, Bakker found that psychological distress was not related to sound levels, but was related to annoyance in subjects from quiet areas who noticed the sounds of the wind turbines⁷². Among those who did not notice the sounds of the wind turbines, there was no relationship found between exposure and either sleep disturbance or psychological distress.

On the other hand, annoyance has been found to be associated with psychological distress. In further analysis⁷³, the results of Pedersen and Persson Waye^15,32 and Pedersen et al.²⁸ showed annoyance with wind turbines to be associated with symptoms of psychological distress. In all three studies, feeling 'tense and stressed' and feeling 'irritable' were significantly, albeit weakly, associated with outdoor annoyance with wind turbines, although only Pedersen et al.²⁸ found an association between both of these and indoor annoyance.  'Headache' was associated with outdoor annoyance in data from Pedersen and Persson Waye¹⁵ and Pedersen et al.²⁸, and with indoor annoyance in Pedersen et al.²⁸. 'Undue tiredness' was associated with both indoor and outdoor annoyance in Pedersen and Persson Waye¹⁵. Pedersen considers findings significant only if they were consistently reported across all three surveys⁷³. Using the data from Pedersen et

al.²⁸, Pedersen et al.²⁷ found psychological distress to be associated with annoyance with noise due to both wind turbines and road traffic. Bakker et al. found that annoyance predicted psychological distress among those who noticed wind turbine sound⁷².

Shepherd and colleagues found significantly lower energy levels reported by residents living within 2km of wind turbines than by those living beyond 8km, but these groups reported similar scores on a Quality of Life measure which uses seven items from the Health Related Quality of Life (HRQOL) survey as a measure of psychological distress⁵¹. Nissenbaum et al.⁵² measured psychological distress using the mental component score of the SF36-v2. Those living within 2km of a wind turbines had a significantly lower mean score (indicating more distress) than those living 3-7km from the turbines (42 in the study group; 52.9 in controls, p=0.002).^‡‡ Neither study measured sound exposure from the wind turbines directly, but used distance from the turbines as a measure of exposure.

From Appendix B studies,  Krogh et al.⁵⁷ reported that 78% (of 97) respondents indicated excessive tiredness, 64% headaches, 71% stress, while 74% indicated 'distress' by checking stress and/or anxiety and/or depression. The authors suggest a dose-response relationship between exposure to wind turbines and excessive tiredness, as well as between exposure to wind turbines and headaches, but neither of these was found to be significant at p=.05.^§§ Harry reports that 71% (of 42) respondents reported excess tiredness, 64% headaches and 52% stress. Using stress and/or anxiety and/or depression as the definition of distress, 62% of the respondents in Harry can be calculated to have experienced distress⁵⁶.  Again, no dose response was investigated. Pierpont reported that of 34 respondents, of ages 6-55, 56% reported headaches of increased frequency, intensity or duration, which was found to be associated with pre-existing migraine disorder, among respondents aged 5 and older (p=0.004). Excluding those with pre- existing migraines, 9 of 17 respondents (all over 22 years of age) experienced headaches, with no correlation between any pre-existing medical conditions or other symptoms. Of 36 respondents, ages 2-64, 75% reported increased fatigue or lack of motivation, with all but six adult subjects recovering to baseline when away from turbines. Of 37 respondents, of ages 2-64, 76% reported increased feelings of irritability or anger⁵³.

 

‡‡ A score of 42 is in the lowest quartile of scores for the total US population, while a score of 52.9 is just above the median (52.82).(94) The results differ by age group and gender.

§§ Masotti and Hodges were unable to reproduce Krogh's results using the stated methods. Fisher's exact tests yield

p values of: .0535 instead of .0307 as reported for excessive tiredness; .1370 instead of .099 as reported for headaches. The regressions have reported p values of .1005 for excessive tiredness and .1837 for headaches.

Diabetes

Further analysis⁷³ of data from Pedersen and Persson Waye³² revealed a weak association between noise exposure and diabetes (OR 1.13, CI95% 1.00-1.27), but this was not found in data Pedersen and Persson Waye¹⁵ or Pedersen et al.²⁸. Data from Pedersen et al.²⁸ showed an association between diabetes and both outdoor annoyance (OR 1.70, CI95% 1.14-2.56) and indoor annoyance (OR 1.62, CI95% 1.10-2.40), but no such associations were found in data from the two studies of Pedersen and Persson Waye^15,32. Pedersen considers only those results common across all three studies to be significant⁷³. From studies included in Appendix B, Pierpont reported one subject who experienced increased difficulty with diabetes control⁵³.

 

Chronic Disease

In further analyses of data from Pedersen and Persson Waye^15,32 and Pedersen et al.²⁸ no association was found between noise exposure and chronic disease, nor between annoyance and chronic disease⁷³. Seven respondents reported by Pierpont experienced exacerbations of various chronic diseases⁵³.

 

Impaired hearing

The data of Pedersen and Persson Waye^15,32 and Pedersen et al.²⁸ revealed no association between noise exposure and impaired hearing, nor between annoyance and impaired hearing⁷³. In studies included in Appendix B, Harry⁵⁶ and Krogh et al.⁵⁷ include 'hearing problems' in their instruments, with 17% of 42 and 35% of 97, respectively, reporting the outcome.  Pierpont includes ear pressure and pain among the core symptoms of wind turbine syndrome, with 30% (of 34) subjects reporting, which was found to be correlated with tinnitus (p=0.008)⁵³. Wind turbines are not thought to pose any risk as sources of damage to the auditory system, because the levels of sound associated with them, even under worse-case conditions close to turbines, are not high enough to cause tissue damage^89,91.

 

High blood pressure, hypertension

Data from Pedersen and Persson Waye^15,32 and Pedersen et al.²⁸ show no association between noise exposure and high blood pressure, nor between annoyance and high blood pressure⁷³. Hypertension has been attributed to wind turbine noise in sporadic adverse events reports^59,60,62, and by Frey and Hadden in a narrative review⁹⁵. Two of Pierpont's participants with pre-existing hypertension experienced increased hypertension, including after the exposure period⁵³.

Cardiovascular disease

In analysis of data from Pedersen and Persson Waye^15,32 and Pedersen et al.²⁸, no association was found between noise exposure and cardiovascular disease, nor between annoyance and cardiovascular disease⁷³. The 'hallmark' of vibroacoustic disease (see below) is pericardial thickening, and vibroacoustic disease is considered to be systemic pathology with cardiovascular effects⁶⁹. No residents living near wind turbines have shown pericardial thickening⁶⁸.  Two of Pierpont's participants reported exacerbated pre-existing dysrhythmias⁵³.

 

Wind turbine syndrome

The term wind turbine syndrome was coined by Pierpont⁵³ for a cluster of symptoms commonly reported by participants, including sleep disturbance, headaches, tinnitus, ear pressure or pain, external auditory sensations, dizziness or vertigo, irritability, problems with concentration and memory, fatigue and visceral vibratory vestibular disturbance (see Table 5 for definition). Sleep disturbances, irritability, headaches and impaired hearing have already been discussed; tachycardia and tinnitus are discussed below.

In the work of Pierpont, external auditory canal sensations, experienced by 15% (of 34) subjects, aged 42-75, were described as "tickling, blowing, or undefined sensations" in the auditory canal. Dizziness or vertigo was reported by 59% (of 27) of subjects, aged ^12-64.

Memory and concentration problems were reported by 93% (of 30) of subjects, aged 5-64, including with school performance and daily tasks. What Pierpont calls visceral vibratory vestibular disturbance includes many components of symptoms reported elsewhere, including anxiety, irritability, sleep disturbances, and tachycardia, but associated with an internal feeling of pulsating or vibrating.  Pierpont alone describes this symptom, reported by 67% (of 21) respondents, ages 32-75.  The description of this syndrome was developed by Pierpont from her case series of respondents who believed they were experiencing health effects from wind turbines⁵³.

 

Tachycardia

Tachycardia is reported only in uncontrolled case series (Appendix B).  Heart palpitations ('palpitations') were included by Harry56 and Krogh et al.⁵⁷, with 12% of 42 and 34% of 97, respectively, reporting this symptom.  It has also been reported variously in adverse events reports in the media^58-60,62, and by Frey and Hadden in a narrative review⁹⁶. Tachycardia is among the symptoms associated with what Pierpont calls visceral vibratory vestibular disturbance (see above).

 

Tinnitus

Tinnitus is among the most controversial health effects in the literature. Tinnitus seems to mean different things to different authors, but is often roughly equated to 'ringing in the ears'⁸⁹. It is formally defined as "the sensation of sounds in the ears, head, or around the head in the absence of an external sound source" and "the perception of a sound in the absence of any external sound applied to the ear"⁹⁷.  It "may arise from abnormal activity at several different points in the auditory system, but the exact mechanisms are not understood, and tinnitus may occur in individuals with otherwise normal hearing" and can occur in both high and low frequencies^89,97.

Data from Pedersen and Persson Waye¹⁵ show a weak association between sound exposure and tinnitus (OR 1.25, CI95% 1.03-1.50)⁷³, but no such relationship was found in analysis of data from Pedersen and Persson Waye 32 and Pedersen et al.²⁸. No association in any of these three studies between annoyance and tinnitus was found.

Among studies included in Appendix B, Pierpont reported that 58% (of 24) subjects, ages 19-64, experienced worsened or new sensations of tinnitus, and this was correlated with pre- existing noise exposure (p=0.01), prior tinnitus (p=0.02), baseline hearing loss (p=0.04), and with ear pain (see above)53. Harry56 and Krogh et al.57 included it on their instrument, reported by 19% of 42 and 56% of 97 of respondents, respectively.

In a 2008 study by Pedersen, Moller and Waye of 21 randomly-selected cases of low- frequency noise complaints, six were found to be caused by low-frequency tinnitus, not external noise sources⁹⁶. Knopper and Ollson estimate that 50 million people in the United States experience tinnitus, which is associated with a number of demographic and psychological variables, including anxiety⁸².

 

Vibroacoustic disease (VAD)

The Portuguese investigators, Alves-Pereira and Castelo Branco, describe vibroacoustic disease as a whole body systemic pathology, characterized by the abnormal proliferation of extra-cellular matrices, where pericardial thickening (in absence of inflammation and diastolic dysfunction) is one of the main diagnostic criteria^70,98-100.  They identified this disease in studies of the effects of infrasound and low frequency noise in low frequency noise rich environments such as aviation and industry^70,98.  Prior to 2003, the researchers frequently indicated that vibroacoustic disease was caused by long term exposure to high levels (>90 dB) of low frequency noise (< 500 Hz)^98,102-104.  After 2003, the 90 dB risk threshold appears to have been replaced by the term "excessive exposure" ^{98,99,105-110}.

With respect to wind turbines, vibroacoustic disease was reported in a 2007 publication about two unrelated cases of residential exposure to infrasound (< 20 Hz)^68,69.  In one family living near a grain elevator (the suggested source of the infrasound), the researchers diagnosed VAD primarily based upon thickening of cardiovascular structures indicated in an echocardiogram. The second residence had four wind turbines located at distances of 322 to 643 meters and had slightly higher levels of infrasound exposure. The authors reported that no symptoms of vibroacoustic disease were present because the wind turbines had not been operating long enough, but would develop if they stayed at their residence, based on the experience of the family near the grain elevator.

Subsequent study results contradict claims made by Castelo Branco and Alves-Pereira that VAD is associated with both respiratory pathology and cardiovascular diseases, and that commercial pilots and crew are at high risk because they work in low frequency noise rich environments^99,112.  Three studies with sample sizes ranging from 2,678 to 44,000 pilots and crew found expected or lower than expected mortality rates from cardiovascular and respiratory diseases^112-115. Castelo Branco, Alves-Pereira and colleagues concluded that all Vieques Island residents exposed to sonic booms (130 dB, 30Hz) from USA military testing had VAD, diagnosed from thickened pericardium¹¹⁶.  A subsequent study to confirm these findings, addressing methodological issues such as the sampling frame and lack of blinding, was conducted by the Centers for Disease Control and Prevention and the Agency for Toxic Substances and Disease Registry and found no evidence of abnormal pericardial thickening¹¹¹.

Further arguments against a causal relationship are the lack of evidence for Hill's criteria of consistency, temporality and a dose-response relationship. Castelo Branco and Alves-Pereira indicated that cars, trains and subways routinely expose people to 92-100 dB levels⁹⁸. This suggests we should observe high rates of VAD among people whose employment places them in these environments. In addition, studies by Hepburn and Howe Gastmeier Chapnick Engineering illustrate that living in windy areas^25,30 can expose people to higher levels of infrasound than the new VAD risk threshold⁶⁰.

Pierpont argues that infrasound and low frequency noise below the hearing threshold could be the cause of disturbances in our vestibular organs, which control our senses of balance, motion and position; this would explain many of the symptoms associated with wind turbine syndrome⁵³.  This hypothesis has been the subject of much debate, focusing on whether wind turbines actually create levels of infrasound and low frequency noise high enough to produce this outcome. While no published studies demonstrate that long-term exposure to low levels of infrasound is harmful, Salt and colleagues have recently argued that infrasound from wind turbines could be effecting our ears, even below the 'hearing' threshold, due to stimulation of the outer hair cells, based on laboratory studies on guinea pigs^4,6. Bolin, Bluhm, Eriksson and Nilsson¹¹⁷ note that Salt and Hullar⁴ cite two studies as evidence that wind turbines produce high enough levels of infrasound for stimulation, but both use measurements no further than 20m from the wind turbines and thus are not typical of residential exposure.  However, Salt and Kaltenbach⁶ also cite van den Berg¹¹, whose measurements at 750m are used to demonstrate that wind turbines could stimulate the outer hair cells at roughly 8Hz-80Hz while still being below the hearing threshold.  Measurements by Hepburn at an Alberta wind farm demonstrate that at 1000m, in low winds the wind turbines would not produce the infrasonic levels necessary to stimulate these organs, but in high winds they could just meet the threshold at certain frequencies²⁵.  However, these measurements also revealed that in low winds, infrasonic levels at a receptor located at 1000m were higher when the wind turbines were off, and were only slightly lower (~5 dB) when off in high winds. Recent studies by Møller and Pedersen⁵, and O'Neal et al.¹⁸, as well as reviews by Knopper and Olsen⁸² and Bolin et al.¹¹⁷ have concluded that infrasound is not likely to be the culprit.  It has recently been suggested that Pierpont has not identified a new disease, but rather has put a new label on a well-known set of symptoms known from decades of work on 'sick' buildings and other industrial noise sources that output modulated infra- and low-frequency noise¹¹⁹.

 

Annoyance as a mediator (pathway variable) in the relationship between noise, psychological distress and sleep disturbance

Bakker et al.⁷² analysed the relationships among the variables wind turbine sound, annoyance, sleep disturbance and psychological distress, using structural equation modeling. The model they tested is presented in Figure 1. The analysis took into account whether respondents noticed the wind turbine noise, the general level of noise in the areas in which they lived, the age and sex of the respondents, and whether they experienced financial gain from the turbines.

 

Figure 1: Model tested by Bakker, Pederson, van den Berg, Stewart, Lok and Bouma

 [graphic]

They found partial support for the model among respondents who noticed the sound of wind turbines; no support was found among respondents who did not notice the sound. Sound levels were associated with annoyance (labeled (1) in the diagram) among non-financially benefiting respondents.  Annoyance was found to be associated with sleep disturbance (4 in the diagram) and psychological distress (5). The model was stronger in quiet areas. In the model, no association was found between sound levels and sleep disturbance (2) or psychological distress (3), or between sleep disturbance and psychological distress (6). This is in spite of the observation that analysis using binary logistic regression showed that those exposed to high levels (45dB and higher) of wind turbine sound were significantly more likely to be disturbed in their sleep than those exposed to low levels (30dB and lower).

The study of Bakker et al. supports annoyance as a mediator, or pathway variable, in a causal pathway between wind turbine noise and other adverse health effects. However, wind turbine annoyance is complicated, and cannot be attributed solely to wind turbine sound, as the results of our current review demonstrates. While Bakker and colleagues found no dose-response relationship between sleep disturbances and sound, their paper and results of studies in this review support a potential sound exposure threshold above which sleep disturbances exist.

It should be noted that Pedersen cautions that there may not exist a causal pathway between noise and stress mediated by annoyance.  This author suggests an alternative cognitive stress theory, in which stress develops in relation to negatively judged environmental stressors, with annoyance arising due to uncontrollable noise adding to that stress⁷³.

In their review of the literature, Knopper and Ollson find support for this hypothesis, and conclude that it is most likely "change in the environment" that finds wind turbines associated with negative health outcomes, and "not a turbine-specific variable like infrasound"82.