Noise From Civilian Aircraft in the Vicinity of Airports - Implications for Human Health - Noise, Stress and Cardiovascular Disease

3. Discussion and Conclusions

3.1 Stress

There is evidence that acute noise exposure can cause temporary elevations in heart rate, as well as increases in peripheral vasoconstriction and blood pressure (WHO, 1999; Passchier-Vermeer, 1993; HCN, 1994; Berglund and Lindvall, 1995; IEH, 1997). The evidence for effects of chronic noise exposure, however, is not consistent across studies. Studies of laboratory animals do suggest persistent elevation of blood pressure, but human laboratory studies are less consistent (Stansfeld and Haines, 1997).

Noise can be one of many environmental stressors. It does not elicit a unique stress response. The stress response is an adaptation or coping mechanism that occurs when the brain perceives experiences or challenges as threats. It is associated with secretion of the stress hormones, such as epinephrine, norepinephrine and cortisol, and changes in heart rate and blood pressure level. Normally, these return to baseline levels when the individual adapts or the experience(s) end (McEwen, 1998). These physiological changes are widely accepted as 'biomarkers' of stress (Frankenhauser, 1986; Scheuch, 1986) and represent a generalized response to any non-specific stressor, such as noise.

Two factors largely govern individual stress responses:

1. the magnitude of the perceived threat or challenge; and
2. the individual's general state of physical health, which largely depends on genetic factors and one's developmental history, experiences, and behavioural and lifestyle choices (Lazarus and Folkman, 1984; McEwen, 1998).

It has been hypothesized that stress hormone levels and blood pressure may remain elevated as a result of frequent or excessive stress in susceptible individuals. The source of such stress can range from daily hassles to traumatic life events (McEwen, 1993, 1998; Rosmond, et al., 1998).

If the release of the stress hormones is sustained or excessive, the functional integrity of many organs and tissues can be compromised in susceptible individuals (Chrousos and Gold, 1998). Sustained release of cortisol has been associated with elevated blood pressure, depression, osteoporosis, immunosuppression, insulin resistance, visceral obesity, and the excessive stimulation of the amygdala, the fear center in the brain (Chrousos and Gold, 1998; McEwen, 1998; Tsigos and Chrousos, 1996; Friedman, et al., 1996). High levels of cortisol can also damage neurons in the hippocampus, an integral part of a negative feedback system that is responsible for returning cortisol levels to normal (McEwen, 1998).

Chronic stress can also have adverse effects on health if the behavioural response to perceived challenges or threats leads to harmful behaviours such as social isolation, aggression, and resorting to the excessive consumption of alcohol, tobacco, food and drugs (McEwen, 1998).

3.1.1 Stress - Conclusions

Noise can act as a short term stressor and has the potential, in susceptible individuals, to cause chronic physiological effects such as elevated blood pressure and stress hormone levels. There is evidence to suggest that physiological effects arising from chronic stress, as well as harmful behaviours, may exacerbate a variety of mental and physical adverse health effects such as cardiovascular disease, depression, osteoporosis, susceptibility to infections and diabetes (via insulin resistance).

3.2 Stress Related Physiological Effects in Children

Several epidemiological studies have examined whether stress related physiological effects in children were associated with exposure to aircraft noise. The endpoints studied were resting blood pressure and stress hormone levels. The detailed reviews and conclusions are provided in sections 3.2.1 to 3.2.4 below. The overall findings can be summarized as follows. The recent Munich airport study (Evans, et al., 1995; Evans, et al., 1998; Hygge, et al., 1998), particularly because of its longitudinal design, has provided the strongest evidence for an association between aircraft noise and physiological effects, especially an increase in epinephrine and norepinephrine (catecholamine) levels. However, there are too few studies to provide conclusive evidence of a cause and effect relationship between aircraft noise and physiological effects. Also, for the few studies that have been done: (i) the characterization of the noise exposure was sometimes difficult to interpret; (ii) associations were not consistently found; and (iii) there was a lack of controls for potentially important confounders. These findings cast doubt as to whether some factor other than aircraft noise was responsible for the observed differences between exposed and control populations.

3.2.1 Noise-induced Blood Pressure Effects

The systolic and diastolic blood pressures in children living in high noise areas around the Los Angeles (Cohen, et al., 1980; Cohen, et al., 1981), Munich (Evans, et al., 1995; Evans, et al., 1998; Hygge, et al., 1998,) and Sydney airports (Morrell, et al., 1998) were compared to those of children living in low aircraft noise areas.

The Los Angeles airport study used a matched group design, where matching was statistically successful for grade level and socioeconomic status. There were 262 subjects, 142 in the noisy area and 120 in the quiet control area. The study was longitudinal in design but only initial findings and 1year follow-up results were published. Additional statistical controls were applied using regression techniques for the confounding factors of racial distribution, known to have a significant effect on blood pressure, mobility (amount of time lived in the area prior to the endpoint measurements), and pon-derosity (ratio of weight to height). This study utilized audiometric screening and controlled environments for the blood pressure measurements. The noise was described as yielding peak sound level readings in the school of 95 dBA in an air corridor with over 300 flights per day. Noise levels were not stated for the control group.

A statistically significant increase of 3 mm Hg in both systolic and diastolic blood pressure was observed in the noise-exposed group initially (Cohen, et al., 1980). The probability was less than 0.03 that this increase occurred by chance. Closer examination of the increase indicated that a statistically significant effect occurred for African-Americans in the study, but not for Caucasians. The authors also reported that a rise in systolic blood pressure for Caucasian noise-exposed school children disappeared as length of enrollment increased. The probability that the rise occurred by chance was less than 0.07. The LA study was unable to find a statistically significant association between blood pressure and aircraft noise at the 1 year follow-up. The authors ascribed this to relocation of susceptible individuals from the study area, but this was not verified (Cohen, et al., 1981).

The authors reported that the Los Angeles study suggested a link between aircraft noise and increases in blood pressure in chronically exposed schoolchildren. However, the suggested link is weakened by the inconsistency of the results as described above.

The Sydney study, which was cross-sectional in design, showed no effect of aircraft noise on blood pressure. Systolic and diastolic blood pressure levels were measured for 1,230 Year 3 schoolchildren from a random sample of primary schools within a 20 km radius of the Sydney airport. Response rates for the study were about 80% of schools approached and 40% of children in Year 3 from the participating schools. The authors stated that this was adequate because the outcome was a physical measurement. The accuracy of the blood pressure measurements was reported as ± 2 mm Hg.

Aircraft noise exposure was reported as monthly energy averaged noise levels accurate to single Australian Noise Energy Index (ANEI) units. They were geocoded to individual school and residential addresses of each participant. A level was assigned to each survey participant. The levels ranged from 15 to 45 ANEI.

Multiple linear regression was used to determine, simultaneously, the magnitude and statistical significance of the effect of aircraft noise and potentially confounding variables. The potential confounding factors included body size, child activity levels, use of salt on food, family history of high blood pressure, whether the child ate breakfast before school, ambient temperature, rail and road traffic noise. A correction for cluster sampling was made in the statistical analysis. All data were obtained between March 11, 1994 and May 6, 1995. The new runway at the Sydney airport opened in the middle of the study, October, 1994.

The study found that blood pressure was not associated with noise exposure. Diastolic blood pressure decreased with time after the opening of the new runway, and systolic blood pressure decreased if the house was insulated. No association was found with road or rail noise. Statistically significant confounders were weight, pulse rate, not eating before school (systolic), using salt on food (diastolic), non-English speaking background (systolic).

The authors noted the potential difficulty of finding an effect because blood pressure is normally highly variable, both between and within individuals. The estimation of the statistical power of the study is not stated in the 1998 Sydney conference paper. Therefore the possibility of a Type II error being committed is not addressed. (A Type II error occurs if the study finds no statistically significant association between outcome and exposure when, in fact, an association exists.) Any possibility that aircraft noise has an effect on childhood blood pressure can only be confirmed or disproved with longitudinal follow-up.

In the recent Munich airport study, there were two experimental groups, each with a less exposed control group, matched for sociodemographic characteristics. The first experimental group was exposed to the noise of the old Munich airport. The second experimental group was not initially exposed to aircraft noise, only after the opening of the new Munich airport (in a new location). The study was longitudinal because there were three testing times during the span of two years (wave 1: occurred 6 months prior to the change over of airports; wave 2: one year later; and wave 3: two years after wave 1).

In the first experimental group around the old Munich airport (Evans, et al., 1995), a 3 mm increase in systolic blood pressure was found to be associated with aircraft noise. The authors concluded that the result was statistically significant because their analysis indicated that the probability (p-value) was less than 0.08 that the observed increase was due to chance. (Most scientists and statisticians would consider a p-value less than 0.05 to indicate statistical significance. Some scientists and statisticians describe p-values less than a number between 0.06 and 0.10 as being indicative of marginal or borderline significance. This latter description is somewhat controversial and some epidemiologists would state that p-values this large would likely be due to chance.) As socioeconomic status can be a confounding factor for blood pressure, it was important that the authors of the article showed that households in the noise-exposed and control areas did not differ in socioeconomic status. However, there was insufficient detail in the reporting of the statistical analysis to assess its validity. For example, for the endpoints of interest, the standard deviations were not reported, so that applicability of the t-test could not be verified.

For the second experimental group around the new Munich airport, over the 3 waves of the study, the increase in systolic and diastolic blood pressures for the noise affected community was 3.4 mm greater than for its matched control group (Evans, et al., 1998; Hygge, et al., 1998). Repeated measures statistics indicated that the probability was less than 0.05 that the difference in systolic blood pressure could be due to chance. (Most scientists and statisticians would consider it unlikely that the observed difference was due to chance). The rise in systolic blood pressure associated with aircraft noise was small compared to normal physiological variations in either population and was essentially the same as the difference in blood pressure level between the two populations at the beginning of the study. The observed rise in average diastolic blood pressure was assessed to have a probability of less than 0.06 of occurring by chance.

The exposure data makes it somewhat difficult to interpret the observed associations. The values for the 24 hour time-averaged sound levels, in A-weighted decibels (dBA), at the new airport in the noise-exposed and control groups (Hygge, private communication) are given in Table 1 below.

Table 1.
Time-averaged sound levels (24 hr.), L_eq (dBA) new airport

Subject Group	Wave 1	Wave 2	Wave 3
Noise	53	66	62
Control	53	61	55

These data were obtained only outside the school that the children attended and only during the 24 hour periods in which the children underwent the physiological and psychological tests used in the study. Therefore it is difficult to tell how representative these sound levels were of the chronic exposure of the children. This difficulty is increased by the fairly large variations that were found in the time-averaged sound levels. For example, at Wave 2, the exposure level for the control neighbourhood was essentially the same as for the noise neighbourhood at Wave 3. This appears to weaken support for the hypothesis that aircraft noise significantly elevates blood pressure among children.

Except for socioeconomic status and ponderosity, the ratio of weight to height, the Munich study did not appear to control for confounding factors which have a bearing on blood pressure in childhood and adolescence . These factors include: differences in diet, such as salt intake (Elliott, 1991), body mass index, height, weight (the correlation coefficient for age 10 is about 0.4 for weight as a predictor of blood pressure (De Swiet, et al., 1992)), levels of physical activity and age (De Swiet, et al., 1992; Law, et al., 1993; Task Force, 1987). As a result, the study's conclusions of an association between chronic noise exposure and increased blood pressure may not be valid.

Even if the epidemiological studies had reliably demonstrated an effect of chronic aircraft noise exposure on blood pressure, the observed elevations were probably not clinically significant. Throughout the Munich study, blood pressures measured in the noise-impacted and control groups were both around the 50^{h t} percentile range of a U.S. and U.K. population, according to the standards developed by the Second Task Force on Blood Pressure Control in Children (Task Force, 1987).

The only cause for clinical concern would be if the observed elevation in children could lead to elevated blood pressure in adulthood. This stems from evidence that suggests that a lower blood pressure will be associated with a lower risk for cardiovascular disease (MacMahon, et al., 1990). Conversely, any increase in blood pressure could be considered as representative of a higher risk. Although, there is some evidence that blood pressure in children can be correlated with blood pressure later in adulthood (Ingelfinger, 1994), the correlation is weak for 10 year old children (De Swiet, et al., 1992), the age group in the Munich study. Therefore, it is unlikely that the observed elevation in blood pressure in children would lead to raised blood pressure in adulthood and a subsequent increased risk of cardiovascular disease.

3.2.1.1 Noise Induced Blood Pressure Effects -Conclusions

There were inconsistent findings between and within studies as to whether observed differences in blood pressure between controls and noise-exposed groups were due to chance. In addition, characterization of the noise exposure was difficult to interpret in the Munich study, casting some doubt as to whether observed differences were due to noise exposure. Furthermore, lack of control for some potentially confounding factors in the Munich study further weakens support for the hypothesis that noise exposure alone was responsible for the observed differences in blood pressure.

The differences in blood pressure between control and exposed populations of 3 mm Hg would not be clinically significant in the subject population even if they had been reliably demonstrated by the epidemiological studies reviewed here.

3.2.2 Noise-induced Stress Hormone Effects

The Munich airport study is the only one around civilian airports to test for stress hormone levels in children. Measurements were made of the resting levels of the catecholamine (epinephrine and norepinephrine) and cortisol stress hormones (Evans, et al., 1998; Hygge, et al., 1998). The results showed evidence of elevated catecholamines but no change in cortisol associated with aircraft noise.

The design of the study and noise exposure values have been described above in Section 3.2.1. The results for the cate-cholamines are shown in Tables 2 and 3 across the 3 waves of measurement.

Table 2.
Changes in epinephrine levels

	Epinephrine ng/hr	Epinephrine ng/hr	Epinephrine ng/hr
Subject Group	Wave One	Wave Two	Wave Three
Aircraft Noise-impacted	229.2	328.1	341.9
Quiet Community	251.8	280.9	246.2

At Wave One, before the opening of the new airport, the levels of both catecholamines in the Quiet Community were higher than those in the community that was to be noise-affected by the new airport. However, at Waves Two and Three, the levels of both catecholamines in the Noisy Community increased much more than in the Quiet Community. The authors concluded that these results indicated a statistically significant association of catecholamine level with aircraft noise.

Table 3.
Changes in norepinephrine levels

Subject Group	Wave One Norepinephrine (ng/hr)	Wave Two Norepinephrine (ng/hr)	Wave Three Norepinephrine (ng/hr)
Aircraft Noise-impacted	610.7	1,228.5	1,556.3
Quiet Community	660.0	879.7	950.7

As noted above in Section 3.2.1 the difficulties concerning the measured noise levels reduces the confidence that the observed stress hormone response arises from aircraft noise, as opposed to other factors associated with the development of the airport.

Also, although all groups were from the third and fourth grade at the start of the study, some age confounding cannot be ruled out. There is a fairly strong effect of age on urinary epinephrine output. For adults and children over the age of 10 years, the upper limit of normal for urinary epinephrine output is about 20 µg/day while for children under 10 years it is 14 µg/day (Behrman, et al., 1987). For norepinephrine, this limit is about 100 µg/day for adults, 80 µg/day limit for children over 10 years of age and 65 µg/day for those under 10 years of age (Behrman, et al., 1987). Therefore, over a two year study involving 9-11 year old children there would be a naturally occurring increase in epinephrine and norepineph-rine output as the children's catecholamine output approaches adult values. The differences might be more pronounced if the children in the 2 groups were not matched for age.

It has been hypothesized that chronic and excessive elevations of the catecholamines can have adverse impacts on the cardiovascular and immune systems later in life. In the Munich study, the average excretion rates of epinephrine and norepinephrine for both Noisy and Quiet Communities were well within the normal limits of, on average, an output of 20.8 - 833 ng/h for epinephrine and 625 - 3333 ng/h for norepinephrine. (Normal values for 24h urinary output of epinephrine and norepinephrine for children greater than 10 years old are listed as 0.5 - 20 µg/d and 15 - 80 µg/d respectively (Behrman, et al., 1987).)

3.2.2.1 Noise-induced Stress Hormone Effects - Conclusions

The results showed evidence of elevated catecholamines between control and exposed populations, but difficulties in interpreting the noise exposure and potential confounding factors due to age cast some doubt on how much of the observed difference in catecholamine levels was due to noise. Furthermore, the lack of a corroborating change in cortisol levels does not support the conclusion that any observed changes in cate-cholamines was a sign of chronic stress. Independent longitudinal studies would be needed to assess whether chronic exposure to aircraft noise leads to a chronic increase in stress hormone levels.