- Email: [email protected]
Influence of weather on osteoarthritics
Sm. Sci. Med. Vol. 23, No. 6, pp. 549-554, 1986
0277-9536/86 $3.00 + 0.00
Pergamon Journals Ltd
Printed in Great Britain
INFLUENCE OF WEATHER
ON OSTEOARTHRITICS
JOYCEM. LABORDE,’WILLIAMA. DANDO’ and MARJORIEJ. POWERS~ ‘College of Nursing, 2Department of Geography, University of North Dakota, Grand Forks, ND 58201 and )College of Nursing, University of Illinois Health Sciences Center, Chicago, Illinois, U.S.A. Abstract-This exploratory study examined the effects of selected weather variables on pain and pain-related stress in osteoarthritic subjects. Urban and rural dwelling arthritics who perceived that weather made their symptoms worse and those who did not were surveyed. Some persons with osteoarthritis in urban Chicago were more weather sensitive than their rural counterparts in Grand Forks, North Dakota. Multiple regression analysis showed that precipitation affected degree of pain for urban subjects who identified weather as a pain-generating factor; barometric pressure, relative humidity and sunshine were significant factors influencing pain-related stress. Wind speed correlated with pain and pain-related stress; relative humidity and precipitation correlated with pain-related stress for urban subjects who did not perceive weather as a problem. Specific weather variables were not identified as affecting rural subjects’ pain. However, temperature and barometric pressure affected degree of pain-related stress in rural subjects who perceived weather as a problem. Subtle differences between Chicago urban and Grand Forks rural climates are reflected in arthritic subjects’ degree of pain and their perception of pain-related stress. Key words-osteoarthritis,
pain, climate, symptomatology
INTRODUCrION Climate and the degree to which osteoarthritics complain about their symptoms have been inextricably linked (Table 1). Arthritics living in colder regions (55”N) have been shown to have more complaints about their illness than those living in warmer areas (18”N) even though differences in pathologic joint involvement were not found [l]. It was proposed that joint sensitivity to meteorologic elements is not sufficient to give rise to spontaneous pain for arthritics who live in warmer climates, whereas, when tissue temperature falls in arthritics living in colder regions, spontaneous pain occurs [2]. According to Hill [3], approx. 80-90% of persons with arthritis are weather sensitive; conditions affecting these persons are cold climate and falling barometric pressure combined with increased humidity. Pain occurs because diseased tissues retain fluid (joint swelling), thereby causing intracellular pressure to become higher than ambient pressure. Increased pain was reported by hospitalized arthritics when decreased barometric pressure, decreased ambient pressure, and other weather changes that indicated approaching storm systems occurred [4]. Utilizing a controlled environment, Hollander [S, 61 noted that 5 out of 24 patients with arthritis had increased stiffness after exposure to dropping temperatures, and 10 others (7 rheumatoid and 3 osteoarthritics) reported increased stiffness after exposure to decreased barometric pressure and rising humidity. Moreover, Svarcova [7] found that most of the 130 arthritics he interviewed experienced pain at least 12 hr or more before a weather change occurred. Stress is usually a concomitant of pain. Weisenfeld and Hallin [8] found that 13 out of 30 partially acclimated and non-acclimated rats with peripheral nerve damage that produced pain exhibited an increase in stress-related behavior upon exposure to a
cold (4°C) environment. A stressful environment has also been found to affect pain in humans. Hospitalized medical surgical patients scoring high on a stress assessment scale reported more pain than lowscoring patients with similar complaints (91. The individual’s perception of pain also affects the degree of stress related to noxious stimulation [lo]. For some arthritics environmental factors such as weather conditions may influence how pain and its resultant stress is perceived. This exploratory study examined the effects of selected weather variables on pain and pain-related stress in urban and rural subjects who perceived that weather made their arthritic symptoms worse and those who did not perceive weather as a problem. Chicago, Illinois was chosen as the urban setting and Grand Forks, North Dakota was chosen as representative of a rural area [1 11. Chicago’s constructed environment has created a totally different climatic realm from that which existed there before the city’s founding-it has a classic urban climate. The site’s natural vegetation has been replaced with more than 228 square miles of concrete, asphalt, and glass; over three million urbanites are housed and work in structures, ranging from single to multi-storied complexes of extreme vertical extent, that have replaced the largely horizontal surface; vast amounts of energy are imported and combusted in one of the nation’s largest manufacturing and transportation centers; and Chicago’s atmosphere is polluted by industrial and transportation by-products [12, 131. Grand Forks encompasses less than 9 square miles, a fraction of Chicago’s built-up area; houses fewer than 45,000 people; and does not possess an urban profile nor any basic industry. It is primarily a small university town with virtually pollution free air and no statistically discemable urban climate [14]. Subjects from the urban Chicago area were from a senior center, a university hospital and a private hospital.
1. Climatolotical
*P < 0.05.
Barometric pressure Relative humidity Sunshine (minutes)
-0.36* R = 0.36
B
0.37. -0.36* -0.35’ R = 0.50
stress (Z’)
Pain (Z’)
Pain-related
Precipitation
Variables
Pain-related 0.52t 0.37’ -0.31’ R = 0.55
*P < 0.05; tP < 0.01.
Wind speed Relative humidity Precipitation
stress (Z’)
Pain (Z’)
Wind speed
Variables
(A) Urban Meteorological parameters (a) Temperature (b) Barometric pressure (c) Relative humidity (d) Wind speed (e) Precipitation (f) Sunshine (minutes) Group 2 Group 1 (n = 37) (n = 37) Those who perceived weather as Those who did not perceive making their symptoms worse weather as a problem
Table
-
J
.
I-J
‘P < 0.05.
Temperature Barometric pressure
effects
B
0.45. -0.33’ R = 0.62
stress (Z’)
No significant
Pain (Z’)
Pain-related
Variables
Pain (Z’)
effects
effects
stress (Z’) No significant
Pain-related
No significant
Variables
(B) Rural Meteorological parameters (a) Temperature (b) Barometric pressure (c) Relative humidity (d) Wind speed (e) Precipitation (f) Sunshine (minutes) Group 3 Group 4 (n = 26) (n = 26) Those who perceived weather as Those who did not perceive making their symptoms worse weather as a problem
effects on urban and rural osteoarthritics
B
Weather and arthritis Subjects from the rural Grand Forks area were from a senior center and an independent living facility. Weather conditions on arthritic subjects’ interview days were considered as environmental stimuli and subjects’ responses to these stimuli were assessed through self-reports of degree of arthritic pain and subjective evaluation of pain-related stress. METHOD
Urban sample
In a nursing intervention study of 135 urban osteoarthritic community residents, 66 (48.9%) reported that weather affected their symptoms [15]. From this group, 37 subjects who perceived that weather made their symptoms worse (group 1) were matched on age, sex and location with 37 subjects from the same study who did not perceive weather as a problem (group 2). The resultant sample consisted of 62 females and 12 males; ages ranged from 50 to 89 years (M = 70.22, SD = 7.80). Except for overrepresentation of females (84%), Caucasians (75%), and retirees (62%), the sample was heterogeneous on other background variables. The average duration of osteoarthritic joint disease was 12.77 years (SD = 11.27). Thirty-eight percent (n = 14) of the subjects in group 1 reported having pain every day; 62% (n = 23) reported their pain occurred two or three times a week. Six subjects used assistive devices for ambulation most of the time; two used assistive devices occasionally. Two persons had undergone surgical joint replacement procedures and four were under consideration for surgery related to their arthritis. In group 2, 70% (n = 26) reported having pain on a daily basis and 30% (n = 11) stated that their pain varied but occurred two or three times a week. Eleven subjects used assistive devices most of the time and two were under consideration for surgical joint replacement procedures. Although the extent of arthritic joint involvement was not significantly different between both groups of subjects (x2[l] = O.Ol), frequency of pain occurrence differed appreciably (x2[1] = 7.36, P < 0.01) when Chi-square tests for independent samples were used. Rural sample
Twenty-six elderly osteoarthritic community residents who perceived that weather made their symptoms worse (group 3) were selected from a group of 60 rural residents who had participated in a nursing intervention study. These subjects were matched on age, sex and location with 26 osteoarthritic subjects who did not perceive weather as a problem who participated in the same study. All persons identified by chart review or who had affirmed by self-report they had osteoarthritis and suffered pain were eligible. The resultant rural sample was comprised of 50 females and 2 males; ages ranged from 64 to 93 years (M = 81.36, SD = 7.78). All were Caucasian and retirees. The average duration of osteoarthritic joint disease was 14.59 years (SD = 16.70). Of those who perceived that weather made their symptoms worse (group 3), 77% (n = 20) reported having pain every day; 23% (n = 6) reported their pain occurred intermittently several times a week.
551
Eight persons used assistive devices most of the time, two used assistive devices for ambulation purposes occasionally, and one had undergone a surgical joint replacement procedure. In group 4 (those who did not perceive weather as a problem), 65% (n = 17) reported daily pain and 35% (r~ = 9) reported intermittent pain during the course of a week. Two subjects used asssistive devices occasionally. No significant differences between groups for extent of joint involvement (x2[1] = 1.46) or frequency of pain (x2[1] = 0.84) were found. However, frequency of assistive device use differed significantly for those who perceived weather as making theirsymptomsworse(X2[1] = 4.60, P < 0.05). Measurement
Interview schedules developed for the overall study contained 16 items related to osteoarthritis [15]. However, only the following items will be discussed: perceived weather sensitivity, pain, duration of osteoarthritis, age, and pain-related stress. Perceived weather sensitivity was elicited by posing the following question from the McGill Pain Questionnaire [16]: “What kind of things increase your pain?” Responses such as cold, damp, high humidity and changing weather conditions indicated weather sensitivity; responses that excluded weather-related factors indicated that weather was not perceived as a problem. Using the pain intensity rating scale from the McGill Pain Questionnaire, subjects were asked to indicate which word best described their arthritic pain: 1 ‘= mild, 2 = discomforting, 3 = distressing, 4 = horrible, 5 = excruciating. Duration of osteoarthritis in terms of years was quanitified from raw data. Likewise, years of age were also quantified. In order to differentiate pain-related stress from overall life stress, a rating on general life stress was obtained first. Using a self-anchoring ladder scale technique, with 1 representing the least stress experienced in life and 10 the most, subjects were asked to indicate which point on the scale represented their present amount of stress. Subjects were then asked to indicate on the same scale the amount of stress associated specifically with the pain of arthritis. Because the McGill Pain Questionnaire has been psychometrically evaluated in previous studies [16. 171, validity of the pain intensity rating scale has already been documented. To establish reliability for the original 16 interview items however, a pilot test was administered to eight female retirement home residents (M age = 82.40 years, SD = 6.97). This resulted in a Cronbach’s alpha of 0.80 which indicates support for the homogeneity of these items. At each site, subjects were interviewed to ascertain their degree of pain, their perception of factors that increase pain and their amount of pain-related stress. Two weeks after baseline data had been obtained, subjects were contacted by phone and questions related to pain, factors that influenced pain and degree of pain-related stress were repeated. (Retest reliability for these questions resulted in a coefficient of 0.98.) Only data obtained via telephone interviews from September 5 to December 16, 198 1 for urban subjects and from October 6, 1983 to February 29, 1984 for subjects from the rural area were used for analysis.
552
JOYCE M. LABORDE et al. Table 2. Climatic changes associated with cities Change compared with rural areas
Atmosoheric element Radiation, total on horizontal surface Ultraviolet Sunshine duration Temperature (nights) Heating needs Relative humidities Rainfall Number of rainy days with small amounts Snowfall
Reduced Reduced Reduced Increased Reduced Reduced Increased Increased Reduced
Cloudiness Fog and low vlslblhty Wind speeds Contaminants: solids easeous
Increased Increased Reduced
12-20% l&30% 5-15% l-3% 10% 2-10% 5-10% 10% (amount depending on latitude) S-10% 5&100% l&30%
Increased Increased
1000% 50&2500%
Source: Lansberg H. E. Wearher andtlealfh, p. 88. Doubleday, New York, 1969; and Climate and urban planning. In Proceedings of the Symposium in Urban and Building Climatology, Technical note 108, p. 372. World Meteorological Organization, Geneva, 1970.
Weather data
Information on weather conditions correspondent to data collection schedules was obtained via printed reports from the National Weather Service for Chicago, Illinois, and Grand Forks, North Dakota. The Local Climatological Data (LCD) reports provided information on mean climatic temperatures, mean barometric pressures, three hourly humidity averages, wind speed, precipitation, and minutes of sunshine for each 24-hr period from the two weather stations. At the first station (latitude 41”59’N; longitude 87”54’W), where 74 urban residents were located, ground elevation was 658 ft above mean sea level. At the second weather station (latitude 47”56’N; longitude 97”05’W), where 52 rural residents resided, ground elevation was 830 ft. Meteorological variables as quantified on the LCD reports were not coded but were analyzed directly. Because solar radiation would be markedly reduced for urban as opposed to rural areas along with temperature, wind and humidity changes [18], symptoms of arthritis affected by weather in the urban and rural osteoarthritic subject may not be the same. To illustrate differences between urban and rural climates, Landsberg’s [18] chart is presented (Table 2). Although the existence of urban effects upon climate are accepted, estimates of the degree of climatic change due to urbanization have been questioned
[191. Nonetheless, climatic changes associated with cities are presented to provide a framework from which the two locales can be compared. To determine if weather-related variables affected degree of arthritic pain and stress related to pain in urban and rural subjects, stepwise multiple regression analyses were performed. In each instance, when predictor variables were deemed not statistically significant (i.e. > 0.05) the analysis was terminated. Moreover, even though weather variables are not independent, multicollinearity in the regression equations was not found. RESULTS
Urban sample
Summary statistics for the urban sample are shown in Table 3. As expected, there were no significant differences between subjects via paired t-tests. Likewise, we found that weather variables did not differ significantly when independent t-tests were used. When pain was used as the criterion variable for group 1, precipitation was a significant predictor. When pain-related stress was used as the criterion variable for group 1, statistically significant predictors were barometric pressure, relative humidity and sunshine (Table 1). In the regression analysis using pain as the criterion variable for group 2, wind speed
Table 3. Summary statistics for urban subjects who perceived weather as affecting their arthritis (group 1) and those who did not (group 2) Group 1 (n = 37)
Group 2 (n = 37)
M
SD
M
SD
t
Age
70.43
7.58
70.00
8.17
0.24
Duration of disease (yr) Pain Pain-related stress
14.73 2.70 5.46
12.98 1.31 2.57
10.81 2.51 5.97
9.02 1.19 2.96
1.51 0.65 -0.80
49.46 29.36 71.84 9.12 0.11 374.59
15.73 0.17 6.18 2.98 0.21 225.88
49.27 29.29 71.22 9.00 0.11 287.24
14.01 0.17 7.71 2.72 0.20 233.68
0.05 1.83 0.38 0.17 -0.06 1.07
Variables
Weather-related
Temperature Barometric pressure Relative humidity Wind speed Precipitation Sunshine (minutes)
553
Weather and arthritis Table
4. Summary statistics for rural subjects who perceived weather as affecting their arthritis (group 3) and those who did not (group 4) Group 4 (n = 26)
Group 3 (n = 26) Variables
Age Duration of disease (yr) Pain Pain-related stress Weather-related
Temperature Barometric pressure Relative humidity Wind speed Precipitation Sunshine (minutes)
SD
M
SD
1.89 19.36 0.98 2.21
81.12 11.31 2.13 4.73
7.81 13.12 1.22 2.41
0.23 1.43 -1.13 0.60
25.00
16.79
23.27
14.41 0.19 9.85 6.09 0.01 44.48
0.40 1.32 0.01 -0.47 0.14 0.29
30.01 78.92 7.61 0.03 586.84
as a predictor was identified. Regression of painrelated stress on weather variables for group 2 showed that wind speed, relative humidity and precipitation were statistically significant predictors of this variable. Rural sample
Summary statistics for patient and weather variables for the rural sample are shown in Table 4. We did not find significant differences between groups in terms of patient and weather variables via t-tests. When pain-related stress was regressed on weather variables for group 3, temperature and barometric pressure were statistically significant predictors (Table 1). However, regression of pain on weather variables for group 3 was non-statistically significant. Likewise, regression equations for pain and painrelated stress were non-statistically significant for group 4 subjects. DISCUSSION
As noted by Landsberg [18], “not all rheumatics have weather-related pain, but most of those who have chronic complaints react to atmospheric pulsations” (p. 104). Characteristics of urban and rural subjects differed somewhat. For instance, there were more males (16%) in the urban sample than there were for the rural group (4%). Moreover, proportionately fewer retirees (62%) and Caucasians (75%) appeared in the urban sample than among the rural subjects-100% of whom were retired Caucasians. In general, rural subjects averaged 10 years older than urban subjects and consequently reported having their disease longer-by about 2 years. Use of an assistive device for ambulation purposes was reported more frequently (33%) by rural subjects than by the urban sample (29%). Furthermore, only one person in the rural sample had undergone a surgical procedure for arthritis, whereas in the urban sample two reported having joint replacement surgeries and six were under consideration for this procedure. For urban subjects who perceived weather as increasing their symptoms (group l), precipitation was significantly related to pain. However, as a predictor of pain, precipitation was inversely related. Out of seven arthritics who implicated rain as making their symptoms worse, two indicated that their symptoms increased before it rained but decreased before the rain ceased. Therefore, it could be postulated that for SSM 23,&E!
t
M 81.61 17.88 2.38 5.12
0.23 12.07 6.40 0.07 44.58
29.93 78.88 8.42 0.03 583.27
some persons changing weather conditions preceding rain may be more bothersome than the presence of rain and might help explain the inverse relationship between precipitation and pain found in this study. There was a positive relationship between barometric pressure and an inverse relationship between relative humidity and pain-related stress for group 1. Relationships between these variables we thought reflected changing weather conditions. As pointed out by Hollander and Yeostros [20], it is the changing conditions rather than high humidity or low barometric pressure that are responsible for affecting the weather sensitive arthritic. Our study revealed an inverse relationship between sunshine and painrelated stress. This finding could reflect that as their pain-related stress decreased, the arthritics’ sense of well-being increased. Rosen [21] proposed that the most significant weather element that increases people’s judgement of ‘pleasantness’ is increased sunshine. Thus, with increased sunshine it could be speculated that the arthritics’ expectancy of pain occurrence concomitantly decreased, hence the inverse relationship. Interestingly, urban subjects who did not perceive weather as a problem (group 2) were also affected by weather variables. For instance, wind speed was the most significant predictor of pain and pain-related stress. The predominant wind direction for days during data collection was southeast. Although it could be expected that southeasterly winds would usher in warm weather to the Chicago area, this may not have occurred. According to Cox and Armington [22], air systems that affect decrements in the temperature of Chicago often pass the city to the west and south. Another factor to be considered may be wind turbulence. Wind action becomes complicated in large cities because tall buildings penetrate the boundary layer causing different air temperatures to mix, resulting in turbulence [23]. Therefore, wind speed related to colder temperature and increased turbulence may be factors affecting pain and painrelated stress in osteoarthritics who did not identify weather as a problem in terms of their disease. Relative humidity and precipitation were other meteorological variables that influenced pain-related stress for urban subjects who did not identify weather as a problem. As suggested by Lawrence [24], when tissue temperature falls resultant to decreased air temperature or diminished solar radiation, or possibly from the influence of increases in humidity in a
JOYCE M. LABORDE et al.
554
pain occurs in the arthritic. In turn, Howarth and Hoffman [25] found that anxiety increased considerably for 24 male subjects whenever it rained or snowed. Nonetheless, the inverse relationship between precipitation and painrelated stress, noted in this study, may indicate decrements in anxiety about pain when rain appeared. Therefore, it could be proposed that cognitive awareness of increasing relative humidity and precipitation may have affected increments or decrements in anxiety manifesting as stress related to the expectation that pain might occur. Meteorological variables did not affect degree of pain for rural arthritics who perceived that weather made their symptoms worse (group 3). However, weather variables such as temperature and decreasing barometric pressure were significant predictors of pain-related stress in this group, thus supporting Hill’s [3] postulation that temperature and falling barometric pressure may affect the weather sensitive arthritic. One explanation for the finding that painrelated stress rather than pain was influenced by these weather variables could be that even though rural arthritics were aware of unfavorable weather conditions, they exerted cognitive control over their degree of pain. By altering one’s awareness to pain and adopting a passive attitude, it is possible to decrease pain [26]. Indeed, if this occurred for group 3 it might help account for the finding that only their painrelated stress was affected by the weather. Further findings were that rural arthritics who did not perceive weather as a problem (group 4) appeared not to be affected by meteorological variables in terms of their pain or pain-related stress. The minimal impact of weather-related variables on rural osteoarthritics may also be related to the fact that they were more acclimated to changing weather conditions than their urban counterparts. In the rural sample 20 subjects (38.5%) reported farming as a past occupation as opposed to only 3 subjects (2%) in the urban group. It could therefore be assumed that rural subjects working and living in areas where weather change was expected became accustomed to these conditions. Conversely, urban subjects may not be accustomed to changing weather patterns because of Chicago’s location in relation to Lake Michigan which often produces unpredictable, highly unstable ‘lake effect’ weather conditions in the city area. Basic elements of weather and climate acted upon differently by human-modified controls in Chicago produced a distinct urban climate that was conducive to the generation of arthritic pain. The same basic elements synthesized dissimilarly and their biomedical impact was not as discernible by arthritics residing in Grand Forks. Thus differences in location may be a factor affecting this study’s findings along with differences in characteristics that were found between urban and rural groups. cold climate, spontaneous
Acknowledgements-This research was partially funded by awards from Sigma Xi, The Scientific Research Society, and AARP, Andrus Foundation made to Dr Joyce M. Laborde in 1980, and 1983, respectively. REFERENCES 1. Lawrence
J. S. and
Molyneux
M. Degenerative
joint
2.
3.
4. 5.
6. 7.
8
9. 10.
11.
12. 13. 14.
15.
16. 17.
18. 19.
20.
21.
22.
23. 24. 25.
26.
disease among populations in Wenslydale, England and Jamaica. J. Biomet. 12, 163-175, 1968. Lawrence J. S. Climate and arthritis. In Arthritis and Physical Medicine (Edited by Licht S.), pp. 429446. Waverley Press, Baltimore, 1969. Hill D. F. Climate and arthritis. In Arthritis and Allied Conditions: A Textbook of Rhuemalology (Edited by Hollander J. L. and McCarty D. C. Jr), 8th edn, pp. 256263. Lea & Febiger, Philadelphia, 1972. Renschler E. B. et al. Arthritic pain in relation to weather change. J. Am. med. Ass. G, 1995-2000, 1929. Hollander J. L. The controlled climatic chamber for study of neurological changes in human diseases. Trans. N.Y. Acad. Sci. 2, 167-172, 1961. Hollander J. L. Environment and musculoskeletal diseases. Archs Emiron. Hlth 6, 527-536, 1963. Svarcova J. Subjective difficulties of rheumatic patients associated with weather change. Fysiat. Reumar. Vest. 61, 277-285, 1983. Weisenfeld Z. and Hallin R. G. Stress-related nain behavior in rats with peripheral nerve damage. Pain 8, 279-284, 1980. Volicer h. J. Hospital stress and patient reports of pain and physical stat&. J. Hum. St&s 4, 28137, 1978. Duffy E. Activation. In Handbook of Psychopathophysiology (Edited by Greenfield N. S. and Sternbath R. A.), pp. 577-622. Holt, Rinehart & Winston, New York, 1972. Dando W. A. Introduction to North Dakota’s Weather and Climate. University of North Dakota Press, Grand Forks, 1981. The World Almanac and Book of Facts 1984. Newspaper Enterprise Association, New York, 1984. Lowry W. P. The climate of cities. Sci. Am. 217, 15-23, 1967. Bureau of Business and Economic Research, North Dakota Office of Intergovernmental Assistance. In Slatistical Abstract of North Dakota 1983,2nd edn. University of North Dakota Press, Grand Forks, 1983. Laborde J. M. and Powers M. J. Evaluation of educational interventions for osteoarthritics. Mull. Lin. Reg. Viewprs 12, 12-37, 1983. Melzack R. The McGill pain questionnaire: Major properties and scoring methods. Pain 1, 277-299, 1975. Brena S. F. et al. Chronic pain states: Their relationship to impairment and disability. Archs Phys. Med. Rehab. 60, 387-389, 1979. Landsberg H. Weather and Health. Doubleday, New York, 1969. Lowry W. P. Empirical estimation of urban effects on climate: A problem analysis. J. appl. Met. 16, 129-135, 1977. Hollander J. L. and Yeostros S. J. The effect of simultaneous variations of humidity and barometric pressure on arthritis. Bull. Am. meteorol. Sot. 44,489494, 1963. Rosen S. Weathering: How the Atmosphere Conditions Your Bodv. Your Mind. Your Moods--and Your Health. Evans, N&v York, 19j9. Cox H. J. and Armington J. H. The Weather and Climate of Chicago. University of Chicago Press, Chicago, 1914. Eppinger J. Defying the wind. Sci. Digest 90, 78-83, 1982. Lawrence J. S. Influence of weather and climate on rheumatic disease. Prog. Biomet. A. 1, II, 83-88, 1977. Howarth E. and Hoffman M. S. A multidimensional approach to the relationship between mood and weather. Br. J. Psychol. 75, 15-23, 1984. Benson J. et al. The relaxation response: Psychophysiologic aspects and clinical applications. In Psychosomatic Medicine: Current Trends and Clinical Applications (Edited by Lipowski Z. J. et al.), pp. 377-388. Oxford University Press, New York, 1977.
0277-9536/86 $3.00 + 0.00
Pergamon Journals Ltd
Printed in Great Britain
INFLUENCE OF WEATHER
ON OSTEOARTHRITICS
JOYCEM. LABORDE,’WILLIAMA. DANDO’ and MARJORIEJ. POWERS~ ‘College of Nursing, 2Department of Geography, University of North Dakota, Grand Forks, ND 58201 and )College of Nursing, University of Illinois Health Sciences Center, Chicago, Illinois, U.S.A. Abstract-This exploratory study examined the effects of selected weather variables on pain and pain-related stress in osteoarthritic subjects. Urban and rural dwelling arthritics who perceived that weather made their symptoms worse and those who did not were surveyed. Some persons with osteoarthritis in urban Chicago were more weather sensitive than their rural counterparts in Grand Forks, North Dakota. Multiple regression analysis showed that precipitation affected degree of pain for urban subjects who identified weather as a pain-generating factor; barometric pressure, relative humidity and sunshine were significant factors influencing pain-related stress. Wind speed correlated with pain and pain-related stress; relative humidity and precipitation correlated with pain-related stress for urban subjects who did not perceive weather as a problem. Specific weather variables were not identified as affecting rural subjects’ pain. However, temperature and barometric pressure affected degree of pain-related stress in rural subjects who perceived weather as a problem. Subtle differences between Chicago urban and Grand Forks rural climates are reflected in arthritic subjects’ degree of pain and their perception of pain-related stress. Key words-osteoarthritis,
pain, climate, symptomatology
INTRODUCrION Climate and the degree to which osteoarthritics complain about their symptoms have been inextricably linked (Table 1). Arthritics living in colder regions (55”N) have been shown to have more complaints about their illness than those living in warmer areas (18”N) even though differences in pathologic joint involvement were not found [l]. It was proposed that joint sensitivity to meteorologic elements is not sufficient to give rise to spontaneous pain for arthritics who live in warmer climates, whereas, when tissue temperature falls in arthritics living in colder regions, spontaneous pain occurs [2]. According to Hill [3], approx. 80-90% of persons with arthritis are weather sensitive; conditions affecting these persons are cold climate and falling barometric pressure combined with increased humidity. Pain occurs because diseased tissues retain fluid (joint swelling), thereby causing intracellular pressure to become higher than ambient pressure. Increased pain was reported by hospitalized arthritics when decreased barometric pressure, decreased ambient pressure, and other weather changes that indicated approaching storm systems occurred [4]. Utilizing a controlled environment, Hollander [S, 61 noted that 5 out of 24 patients with arthritis had increased stiffness after exposure to dropping temperatures, and 10 others (7 rheumatoid and 3 osteoarthritics) reported increased stiffness after exposure to decreased barometric pressure and rising humidity. Moreover, Svarcova [7] found that most of the 130 arthritics he interviewed experienced pain at least 12 hr or more before a weather change occurred. Stress is usually a concomitant of pain. Weisenfeld and Hallin [8] found that 13 out of 30 partially acclimated and non-acclimated rats with peripheral nerve damage that produced pain exhibited an increase in stress-related behavior upon exposure to a
cold (4°C) environment. A stressful environment has also been found to affect pain in humans. Hospitalized medical surgical patients scoring high on a stress assessment scale reported more pain than lowscoring patients with similar complaints (91. The individual’s perception of pain also affects the degree of stress related to noxious stimulation [lo]. For some arthritics environmental factors such as weather conditions may influence how pain and its resultant stress is perceived. This exploratory study examined the effects of selected weather variables on pain and pain-related stress in urban and rural subjects who perceived that weather made their arthritic symptoms worse and those who did not perceive weather as a problem. Chicago, Illinois was chosen as the urban setting and Grand Forks, North Dakota was chosen as representative of a rural area [1 11. Chicago’s constructed environment has created a totally different climatic realm from that which existed there before the city’s founding-it has a classic urban climate. The site’s natural vegetation has been replaced with more than 228 square miles of concrete, asphalt, and glass; over three million urbanites are housed and work in structures, ranging from single to multi-storied complexes of extreme vertical extent, that have replaced the largely horizontal surface; vast amounts of energy are imported and combusted in one of the nation’s largest manufacturing and transportation centers; and Chicago’s atmosphere is polluted by industrial and transportation by-products [12, 131. Grand Forks encompasses less than 9 square miles, a fraction of Chicago’s built-up area; houses fewer than 45,000 people; and does not possess an urban profile nor any basic industry. It is primarily a small university town with virtually pollution free air and no statistically discemable urban climate [14]. Subjects from the urban Chicago area were from a senior center, a university hospital and a private hospital.
1. Climatolotical
*P < 0.05.
Barometric pressure Relative humidity Sunshine (minutes)
-0.36* R = 0.36
B
0.37. -0.36* -0.35’ R = 0.50
stress (Z’)
Pain (Z’)
Pain-related
Precipitation
Variables
Pain-related 0.52t 0.37’ -0.31’ R = 0.55
*P < 0.05; tP < 0.01.
Wind speed Relative humidity Precipitation
stress (Z’)
Pain (Z’)
Wind speed
Variables
(A) Urban Meteorological parameters (a) Temperature (b) Barometric pressure (c) Relative humidity (d) Wind speed (e) Precipitation (f) Sunshine (minutes) Group 2 Group 1 (n = 37) (n = 37) Those who perceived weather as Those who did not perceive making their symptoms worse weather as a problem
Table
-
J
.
I-J
‘P < 0.05.
Temperature Barometric pressure
effects
B
0.45. -0.33’ R = 0.62
stress (Z’)
No significant
Pain (Z’)
Pain-related
Variables
Pain (Z’)
effects
effects
stress (Z’) No significant
Pain-related
No significant
Variables
(B) Rural Meteorological parameters (a) Temperature (b) Barometric pressure (c) Relative humidity (d) Wind speed (e) Precipitation (f) Sunshine (minutes) Group 3 Group 4 (n = 26) (n = 26) Those who perceived weather as Those who did not perceive making their symptoms worse weather as a problem
effects on urban and rural osteoarthritics
B
Weather and arthritis Subjects from the rural Grand Forks area were from a senior center and an independent living facility. Weather conditions on arthritic subjects’ interview days were considered as environmental stimuli and subjects’ responses to these stimuli were assessed through self-reports of degree of arthritic pain and subjective evaluation of pain-related stress. METHOD
Urban sample
In a nursing intervention study of 135 urban osteoarthritic community residents, 66 (48.9%) reported that weather affected their symptoms [15]. From this group, 37 subjects who perceived that weather made their symptoms worse (group 1) were matched on age, sex and location with 37 subjects from the same study who did not perceive weather as a problem (group 2). The resultant sample consisted of 62 females and 12 males; ages ranged from 50 to 89 years (M = 70.22, SD = 7.80). Except for overrepresentation of females (84%), Caucasians (75%), and retirees (62%), the sample was heterogeneous on other background variables. The average duration of osteoarthritic joint disease was 12.77 years (SD = 11.27). Thirty-eight percent (n = 14) of the subjects in group 1 reported having pain every day; 62% (n = 23) reported their pain occurred two or three times a week. Six subjects used assistive devices for ambulation most of the time; two used assistive devices occasionally. Two persons had undergone surgical joint replacement procedures and four were under consideration for surgery related to their arthritis. In group 2, 70% (n = 26) reported having pain on a daily basis and 30% (n = 11) stated that their pain varied but occurred two or three times a week. Eleven subjects used assistive devices most of the time and two were under consideration for surgical joint replacement procedures. Although the extent of arthritic joint involvement was not significantly different between both groups of subjects (x2[l] = O.Ol), frequency of pain occurrence differed appreciably (x2[1] = 7.36, P < 0.01) when Chi-square tests for independent samples were used. Rural sample
Twenty-six elderly osteoarthritic community residents who perceived that weather made their symptoms worse (group 3) were selected from a group of 60 rural residents who had participated in a nursing intervention study. These subjects were matched on age, sex and location with 26 osteoarthritic subjects who did not perceive weather as a problem who participated in the same study. All persons identified by chart review or who had affirmed by self-report they had osteoarthritis and suffered pain were eligible. The resultant rural sample was comprised of 50 females and 2 males; ages ranged from 64 to 93 years (M = 81.36, SD = 7.78). All were Caucasian and retirees. The average duration of osteoarthritic joint disease was 14.59 years (SD = 16.70). Of those who perceived that weather made their symptoms worse (group 3), 77% (n = 20) reported having pain every day; 23% (n = 6) reported their pain occurred intermittently several times a week.
551
Eight persons used assistive devices most of the time, two used assistive devices for ambulation purposes occasionally, and one had undergone a surgical joint replacement procedure. In group 4 (those who did not perceive weather as a problem), 65% (n = 17) reported daily pain and 35% (r~ = 9) reported intermittent pain during the course of a week. Two subjects used asssistive devices occasionally. No significant differences between groups for extent of joint involvement (x2[1] = 1.46) or frequency of pain (x2[1] = 0.84) were found. However, frequency of assistive device use differed significantly for those who perceived weather as making theirsymptomsworse(X2[1] = 4.60, P < 0.05). Measurement
Interview schedules developed for the overall study contained 16 items related to osteoarthritis [15]. However, only the following items will be discussed: perceived weather sensitivity, pain, duration of osteoarthritis, age, and pain-related stress. Perceived weather sensitivity was elicited by posing the following question from the McGill Pain Questionnaire [16]: “What kind of things increase your pain?” Responses such as cold, damp, high humidity and changing weather conditions indicated weather sensitivity; responses that excluded weather-related factors indicated that weather was not perceived as a problem. Using the pain intensity rating scale from the McGill Pain Questionnaire, subjects were asked to indicate which word best described their arthritic pain: 1 ‘= mild, 2 = discomforting, 3 = distressing, 4 = horrible, 5 = excruciating. Duration of osteoarthritis in terms of years was quanitified from raw data. Likewise, years of age were also quantified. In order to differentiate pain-related stress from overall life stress, a rating on general life stress was obtained first. Using a self-anchoring ladder scale technique, with 1 representing the least stress experienced in life and 10 the most, subjects were asked to indicate which point on the scale represented their present amount of stress. Subjects were then asked to indicate on the same scale the amount of stress associated specifically with the pain of arthritis. Because the McGill Pain Questionnaire has been psychometrically evaluated in previous studies [16. 171, validity of the pain intensity rating scale has already been documented. To establish reliability for the original 16 interview items however, a pilot test was administered to eight female retirement home residents (M age = 82.40 years, SD = 6.97). This resulted in a Cronbach’s alpha of 0.80 which indicates support for the homogeneity of these items. At each site, subjects were interviewed to ascertain their degree of pain, their perception of factors that increase pain and their amount of pain-related stress. Two weeks after baseline data had been obtained, subjects were contacted by phone and questions related to pain, factors that influenced pain and degree of pain-related stress were repeated. (Retest reliability for these questions resulted in a coefficient of 0.98.) Only data obtained via telephone interviews from September 5 to December 16, 198 1 for urban subjects and from October 6, 1983 to February 29, 1984 for subjects from the rural area were used for analysis.
552
JOYCE M. LABORDE et al. Table 2. Climatic changes associated with cities Change compared with rural areas
Atmosoheric element Radiation, total on horizontal surface Ultraviolet Sunshine duration Temperature (nights) Heating needs Relative humidities Rainfall Number of rainy days with small amounts Snowfall
Reduced Reduced Reduced Increased Reduced Reduced Increased Increased Reduced
Cloudiness Fog and low vlslblhty Wind speeds Contaminants: solids easeous
Increased Increased Reduced
12-20% l&30% 5-15% l-3% 10% 2-10% 5-10% 10% (amount depending on latitude) S-10% 5&100% l&30%
Increased Increased
1000% 50&2500%
Source: Lansberg H. E. Wearher andtlealfh, p. 88. Doubleday, New York, 1969; and Climate and urban planning. In Proceedings of the Symposium in Urban and Building Climatology, Technical note 108, p. 372. World Meteorological Organization, Geneva, 1970.
Weather data
Information on weather conditions correspondent to data collection schedules was obtained via printed reports from the National Weather Service for Chicago, Illinois, and Grand Forks, North Dakota. The Local Climatological Data (LCD) reports provided information on mean climatic temperatures, mean barometric pressures, three hourly humidity averages, wind speed, precipitation, and minutes of sunshine for each 24-hr period from the two weather stations. At the first station (latitude 41”59’N; longitude 87”54’W), where 74 urban residents were located, ground elevation was 658 ft above mean sea level. At the second weather station (latitude 47”56’N; longitude 97”05’W), where 52 rural residents resided, ground elevation was 830 ft. Meteorological variables as quantified on the LCD reports were not coded but were analyzed directly. Because solar radiation would be markedly reduced for urban as opposed to rural areas along with temperature, wind and humidity changes [18], symptoms of arthritis affected by weather in the urban and rural osteoarthritic subject may not be the same. To illustrate differences between urban and rural climates, Landsberg’s [18] chart is presented (Table 2). Although the existence of urban effects upon climate are accepted, estimates of the degree of climatic change due to urbanization have been questioned
[191. Nonetheless, climatic changes associated with cities are presented to provide a framework from which the two locales can be compared. To determine if weather-related variables affected degree of arthritic pain and stress related to pain in urban and rural subjects, stepwise multiple regression analyses were performed. In each instance, when predictor variables were deemed not statistically significant (i.e. > 0.05) the analysis was terminated. Moreover, even though weather variables are not independent, multicollinearity in the regression equations was not found. RESULTS
Urban sample
Summary statistics for the urban sample are shown in Table 3. As expected, there were no significant differences between subjects via paired t-tests. Likewise, we found that weather variables did not differ significantly when independent t-tests were used. When pain was used as the criterion variable for group 1, precipitation was a significant predictor. When pain-related stress was used as the criterion variable for group 1, statistically significant predictors were barometric pressure, relative humidity and sunshine (Table 1). In the regression analysis using pain as the criterion variable for group 2, wind speed
Table 3. Summary statistics for urban subjects who perceived weather as affecting their arthritis (group 1) and those who did not (group 2) Group 1 (n = 37)
Group 2 (n = 37)
M
SD
M
SD
t
Age
70.43
7.58
70.00
8.17
0.24
Duration of disease (yr) Pain Pain-related stress
14.73 2.70 5.46
12.98 1.31 2.57
10.81 2.51 5.97
9.02 1.19 2.96
1.51 0.65 -0.80
49.46 29.36 71.84 9.12 0.11 374.59
15.73 0.17 6.18 2.98 0.21 225.88
49.27 29.29 71.22 9.00 0.11 287.24
14.01 0.17 7.71 2.72 0.20 233.68
0.05 1.83 0.38 0.17 -0.06 1.07
Variables
Weather-related
Temperature Barometric pressure Relative humidity Wind speed Precipitation Sunshine (minutes)
553
Weather and arthritis Table
4. Summary statistics for rural subjects who perceived weather as affecting their arthritis (group 3) and those who did not (group 4) Group 4 (n = 26)
Group 3 (n = 26) Variables
Age Duration of disease (yr) Pain Pain-related stress Weather-related
Temperature Barometric pressure Relative humidity Wind speed Precipitation Sunshine (minutes)
SD
M
SD
1.89 19.36 0.98 2.21
81.12 11.31 2.13 4.73
7.81 13.12 1.22 2.41
0.23 1.43 -1.13 0.60
25.00
16.79
23.27
14.41 0.19 9.85 6.09 0.01 44.48
0.40 1.32 0.01 -0.47 0.14 0.29
30.01 78.92 7.61 0.03 586.84
as a predictor was identified. Regression of painrelated stress on weather variables for group 2 showed that wind speed, relative humidity and precipitation were statistically significant predictors of this variable. Rural sample
Summary statistics for patient and weather variables for the rural sample are shown in Table 4. We did not find significant differences between groups in terms of patient and weather variables via t-tests. When pain-related stress was regressed on weather variables for group 3, temperature and barometric pressure were statistically significant predictors (Table 1). However, regression of pain on weather variables for group 3 was non-statistically significant. Likewise, regression equations for pain and painrelated stress were non-statistically significant for group 4 subjects. DISCUSSION
As noted by Landsberg [18], “not all rheumatics have weather-related pain, but most of those who have chronic complaints react to atmospheric pulsations” (p. 104). Characteristics of urban and rural subjects differed somewhat. For instance, there were more males (16%) in the urban sample than there were for the rural group (4%). Moreover, proportionately fewer retirees (62%) and Caucasians (75%) appeared in the urban sample than among the rural subjects-100% of whom were retired Caucasians. In general, rural subjects averaged 10 years older than urban subjects and consequently reported having their disease longer-by about 2 years. Use of an assistive device for ambulation purposes was reported more frequently (33%) by rural subjects than by the urban sample (29%). Furthermore, only one person in the rural sample had undergone a surgical procedure for arthritis, whereas in the urban sample two reported having joint replacement surgeries and six were under consideration for this procedure. For urban subjects who perceived weather as increasing their symptoms (group l), precipitation was significantly related to pain. However, as a predictor of pain, precipitation was inversely related. Out of seven arthritics who implicated rain as making their symptoms worse, two indicated that their symptoms increased before it rained but decreased before the rain ceased. Therefore, it could be postulated that for SSM 23,&E!
t
M 81.61 17.88 2.38 5.12
0.23 12.07 6.40 0.07 44.58
29.93 78.88 8.42 0.03 583.27
some persons changing weather conditions preceding rain may be more bothersome than the presence of rain and might help explain the inverse relationship between precipitation and pain found in this study. There was a positive relationship between barometric pressure and an inverse relationship between relative humidity and pain-related stress for group 1. Relationships between these variables we thought reflected changing weather conditions. As pointed out by Hollander and Yeostros [20], it is the changing conditions rather than high humidity or low barometric pressure that are responsible for affecting the weather sensitive arthritic. Our study revealed an inverse relationship between sunshine and painrelated stress. This finding could reflect that as their pain-related stress decreased, the arthritics’ sense of well-being increased. Rosen [21] proposed that the most significant weather element that increases people’s judgement of ‘pleasantness’ is increased sunshine. Thus, with increased sunshine it could be speculated that the arthritics’ expectancy of pain occurrence concomitantly decreased, hence the inverse relationship. Interestingly, urban subjects who did not perceive weather as a problem (group 2) were also affected by weather variables. For instance, wind speed was the most significant predictor of pain and pain-related stress. The predominant wind direction for days during data collection was southeast. Although it could be expected that southeasterly winds would usher in warm weather to the Chicago area, this may not have occurred. According to Cox and Armington [22], air systems that affect decrements in the temperature of Chicago often pass the city to the west and south. Another factor to be considered may be wind turbulence. Wind action becomes complicated in large cities because tall buildings penetrate the boundary layer causing different air temperatures to mix, resulting in turbulence [23]. Therefore, wind speed related to colder temperature and increased turbulence may be factors affecting pain and painrelated stress in osteoarthritics who did not identify weather as a problem in terms of their disease. Relative humidity and precipitation were other meteorological variables that influenced pain-related stress for urban subjects who did not identify weather as a problem. As suggested by Lawrence [24], when tissue temperature falls resultant to decreased air temperature or diminished solar radiation, or possibly from the influence of increases in humidity in a
JOYCE M. LABORDE et al.
554
pain occurs in the arthritic. In turn, Howarth and Hoffman [25] found that anxiety increased considerably for 24 male subjects whenever it rained or snowed. Nonetheless, the inverse relationship between precipitation and painrelated stress, noted in this study, may indicate decrements in anxiety about pain when rain appeared. Therefore, it could be proposed that cognitive awareness of increasing relative humidity and precipitation may have affected increments or decrements in anxiety manifesting as stress related to the expectation that pain might occur. Meteorological variables did not affect degree of pain for rural arthritics who perceived that weather made their symptoms worse (group 3). However, weather variables such as temperature and decreasing barometric pressure were significant predictors of pain-related stress in this group, thus supporting Hill’s [3] postulation that temperature and falling barometric pressure may affect the weather sensitive arthritic. One explanation for the finding that painrelated stress rather than pain was influenced by these weather variables could be that even though rural arthritics were aware of unfavorable weather conditions, they exerted cognitive control over their degree of pain. By altering one’s awareness to pain and adopting a passive attitude, it is possible to decrease pain [26]. Indeed, if this occurred for group 3 it might help account for the finding that only their painrelated stress was affected by the weather. Further findings were that rural arthritics who did not perceive weather as a problem (group 4) appeared not to be affected by meteorological variables in terms of their pain or pain-related stress. The minimal impact of weather-related variables on rural osteoarthritics may also be related to the fact that they were more acclimated to changing weather conditions than their urban counterparts. In the rural sample 20 subjects (38.5%) reported farming as a past occupation as opposed to only 3 subjects (2%) in the urban group. It could therefore be assumed that rural subjects working and living in areas where weather change was expected became accustomed to these conditions. Conversely, urban subjects may not be accustomed to changing weather patterns because of Chicago’s location in relation to Lake Michigan which often produces unpredictable, highly unstable ‘lake effect’ weather conditions in the city area. Basic elements of weather and climate acted upon differently by human-modified controls in Chicago produced a distinct urban climate that was conducive to the generation of arthritic pain. The same basic elements synthesized dissimilarly and their biomedical impact was not as discernible by arthritics residing in Grand Forks. Thus differences in location may be a factor affecting this study’s findings along with differences in characteristics that were found between urban and rural groups. cold climate, spontaneous
Acknowledgements-This research was partially funded by awards from Sigma Xi, The Scientific Research Society, and AARP, Andrus Foundation made to Dr Joyce M. Laborde in 1980, and 1983, respectively. REFERENCES 1. Lawrence
J. S. and
Molyneux
M. Degenerative
joint
2.
3.
4. 5.
6. 7.
8
9. 10.
11.
12. 13. 14.
15.
16. 17.
18. 19.
20.
21.
22.
23. 24. 25.
26.
disease among populations in Wenslydale, England and Jamaica. J. Biomet. 12, 163-175, 1968. Lawrence J. S. Climate and arthritis. In Arthritis and Physical Medicine (Edited by Licht S.), pp. 429446. Waverley Press, Baltimore, 1969. Hill D. F. Climate and arthritis. In Arthritis and Allied Conditions: A Textbook of Rhuemalology (Edited by Hollander J. L. and McCarty D. C. Jr), 8th edn, pp. 256263. Lea & Febiger, Philadelphia, 1972. Renschler E. B. et al. Arthritic pain in relation to weather change. J. Am. med. Ass. G, 1995-2000, 1929. Hollander J. L. The controlled climatic chamber for study of neurological changes in human diseases. Trans. N.Y. Acad. Sci. 2, 167-172, 1961. Hollander J. L. Environment and musculoskeletal diseases. Archs Emiron. Hlth 6, 527-536, 1963. Svarcova J. Subjective difficulties of rheumatic patients associated with weather change. Fysiat. Reumar. Vest. 61, 277-285, 1983. Weisenfeld Z. and Hallin R. G. Stress-related nain behavior in rats with peripheral nerve damage. Pain 8, 279-284, 1980. Volicer h. J. Hospital stress and patient reports of pain and physical stat&. J. Hum. St&s 4, 28137, 1978. Duffy E. Activation. In Handbook of Psychopathophysiology (Edited by Greenfield N. S. and Sternbath R. A.), pp. 577-622. Holt, Rinehart & Winston, New York, 1972. Dando W. A. Introduction to North Dakota’s Weather and Climate. University of North Dakota Press, Grand Forks, 1981. The World Almanac and Book of Facts 1984. Newspaper Enterprise Association, New York, 1984. Lowry W. P. The climate of cities. Sci. Am. 217, 15-23, 1967. Bureau of Business and Economic Research, North Dakota Office of Intergovernmental Assistance. In Slatistical Abstract of North Dakota 1983,2nd edn. University of North Dakota Press, Grand Forks, 1983. Laborde J. M. and Powers M. J. Evaluation of educational interventions for osteoarthritics. Mull. Lin. Reg. Viewprs 12, 12-37, 1983. Melzack R. The McGill pain questionnaire: Major properties and scoring methods. Pain 1, 277-299, 1975. Brena S. F. et al. Chronic pain states: Their relationship to impairment and disability. Archs Phys. Med. Rehab. 60, 387-389, 1979. Landsberg H. Weather and Health. Doubleday, New York, 1969. Lowry W. P. Empirical estimation of urban effects on climate: A problem analysis. J. appl. Met. 16, 129-135, 1977. Hollander J. L. and Yeostros S. J. The effect of simultaneous variations of humidity and barometric pressure on arthritis. Bull. Am. meteorol. Sot. 44,489494, 1963. Rosen S. Weathering: How the Atmosphere Conditions Your Bodv. Your Mind. Your Moods--and Your Health. Evans, N&v York, 19j9. Cox H. J. and Armington J. H. The Weather and Climate of Chicago. University of Chicago Press, Chicago, 1914. Eppinger J. Defying the wind. Sci. Digest 90, 78-83, 1982. Lawrence J. S. Influence of weather and climate on rheumatic disease. Prog. Biomet. A. 1, II, 83-88, 1977. Howarth E. and Hoffman M. S. A multidimensional approach to the relationship between mood and weather. Br. J. Psychol. 75, 15-23, 1984. Benson J. et al. The relaxation response: Psychophysiologic aspects and clinical applications. In Psychosomatic Medicine: Current Trends and Clinical Applications (Edited by Lipowski Z. J. et al.), pp. 377-388. Oxford University Press, New York, 1977.