The Effect of Weather on Respiratory and Cardiovascular Deaths in 12 U.S. Cities

of 5
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Information Report

Sheet Music


Views: 33 | Pages: 5

Extension: PDF | Download: 0

The Effect of Weather on Respiratory and Cardiovascular Deaths in 12 U.S. Cities
  Environmental Health Perspectives  • VOLUME 110 | NUMBER 9 | September 2002  859 The Effect of Weather on Respiratory and Cardiovascular Deaths in 12 U.S. Cities Alfésio L. F. Braga, 1, 2  Antonella Zanobetti, 1 and Joel Schwartz  1 1 Environmental Epidemiology Program, Harvard School of Public Health, Boston, Massachusetts, USA; 2 Environmental PediatricsProgram, University of Santo Amaro School of Medicine, and Laboratory of Experimental Air Pollution, Department of Pathology,University of São Paulo School of Medicine, São Paulo, Brazil  Weather is known to modulate health.Seasonal changes of temperature promotechanges in the daily number of respiratory and cardiovascular diseases (CVD) as well asin total and cause-specific mortality. Theseeffects are more prominent among elderly people and children ( 1 ). Although cold temperatures show greatereffects than do hot temperatures, other fac-tors such as respiratory epidemics, usually present during the winter, make unclear theprecise role of temperature on increased mor-bidity and mortality. On the other hand, heatand heat waves are associated with increasedmorbidity and mortality ( 2  ). Increases of heat-related illnesses have been reported dur-ing episodes of excessive temperature, espe-cially in mid-latitude cities ( 3,4  ). The effectof heat waves has gained more attentionbecause of the expected changes in mean tem-perature with the increase of greenhousegases. Because other factors contribute to theseasonal patterns in mortality, studies havebegun to focus on the short-term effects of  weather, controlling for season. In this regard,realization has been growing that weatherchanges might cause delayed effects and thatsome of the heat-related deaths might be very short-term displacements of the deaths of critically ill people, a phenomenon sometimesreferred to as harvesting.To address these issues, we have studied theeffect of temperature on mortality, focusing onits lag structure. Rather than look at simplemeans of, for example, the previous week’s or3 weeks’ temperature, we have allowed theeffect of weather to vary with the lag timebetween exposure and the related death, withlags up to 3 weeks. To reduce the noise thataccompanies estimating the effects of temper-ature on 21 different days, we applied a poly-nomial distributed lag model ( 5,6  ). In ourprevious study ( 7  ), we examined 12 U.S.cities and estimated the effect of mean daily temperature and relative humidity on each of the 21 days before the death on total deathsin each of the cities. We did meta-analysesstratifying the analyses in two groups: hot andcold cities. In cold cities, we found both highand low temperatures associated withincreased deaths. Although the cold effectpersisted for days, the effect of high tempera-tures was more immediate (day of and day before the death) and was twice as large as thecold effect. However, the hot temperatureeffect appears to involve primarily harvesting.In hot cities, neither hot nor cold tempera-tures had much effect on deaths. Moreover,the magnitude of the effect of hot tempera-ture varied with central air conditioning useand the variance of summer temperatures.These results agree with other studies thathave pointed out the impact of housing, airconditioning, and variability of mean temper-ature as important factors on heat-relatedhealth effects ( 2  ).Total mortality encompasses deaths froma wide variety of causes. Different diseasestates may show different sensitivities toextremes in temperature. Understanding these differences may help in understanding both the sensitive populations and the mech-anisms of action. In this study, we assessedthe lag structure between weather and respira-tory and CVD daily deaths in 12 U.S. cities,applying polynomial distributed lag models. Materials and Methods Data.  We extracted daily counts of deathscaused by pneumonia [ International Classification of Diseases, 9th Revision  (ICD-9), 480–487] ( 8  ),deaths caused by chronicobstructive pulmonary diseases (COPD)(ICD-9: 490–496),all CVD (ICD-9:390–429),and specifically myocardial infarc-tion (MI) (ICD-9: 410) in the metropolitancounties containing the cities of Atlanta,Georgia; Birmingham, Alabama; Canton,Ohio; Chicago, Illinois; Colorado Springs,Colorado; Detroit, Michigan; Houston,Texas; Minneapolis-St. Paul, Minnesota;New Haven, Connecticut; Pittsburgh,Pennsylvania; and Seattle and Spokane, Washington from National Center forHealth Statistics mortality tapes for the years1986 through 1993 ( 9  ). We combined data from Minneapolis and St. Paul and treatedthem as one city. We obtained daily weatherdata from the nearest airport station ( 10  ). Methods.  We modeled counts of daily deaths in a Poisson regression. Our modelsincluded two basic components. We exam-ined the effects of temperature and humidity allowing for nonlinear effects and for thoseeffects that persisted for up to 3 weeks. Wedescribe the methods for doing this below. We found 3 weeks to be more than sufficientto capture the effects on total deaths in our  Address correspondence to J. Schwartz,Environmental Epidemiology Program, Departmentof Environmental Health, Harvard School of PublicHealth, 665 Huntington Ave., Bldg. I, Room 1414,Boston, MA 02115 USA. Telephone: (617) 384-8752. Fax: (617) 384-8745. E-mail: jschwrtz@hsph.harvard.eduThis work was supported in part by NIEHS grantES 00002 and U.S. EPA Research Center AwardR827353. A.L.F.B. received personal grants from SãoPaulo State Research Support Foundation(98/130214) and University of Santo Amaro(UNISA).Received 17 August 2001; accepted 7 February 2002.  We carried out time-series analyses in 12 U.S. cities to estimate both the acute effects and thelagged influence of weather on respiratory and cardiovascular disease (CVD) deaths. We fit gener-alized additive Poisson regressions for each city using nonparametric smooth functions to controlfor long time trend, season, and barometric pressure. We also controlled for day of the week. Weestimated the effect and the lag structure of both temperature and humidity based on a distributed lag model. In cold cities, both high and low temperatures were associated with increased CVDdeaths. In general, the effect of cold temperatures persisted for days, whereas the effect of hightemperatures was restricted to the day of the death or the day before. For myocardial infarctions(MI), the effect of hot days was twice as large as the cold-day effect, whereas for all CVD deathsthe hot-day effect was five times smaller than the cold-day effect. The effect of hot days included some harvesting, because we observed a deficit of deaths a few days later, which we did not observe for the cold-day effect. In hot cities, neither hot nor cold temperatures had much effect onCVD or pneumonia deaths. However, for MI and chronic obstructive pulmonary disease deaths, we observed lagged effects of hot temperatures (lags 4–6 and lags 3 and 4, respectively). We saw no clear pattern for the effect of humidity. In hierarchical models, greater variance of summer and  winter temperature was associated with larger effects for hot and cold days, respectively, on respi-ratory deaths. Key words:  cardiovascular deaths, nonparametric smoothing, respiratory deaths,temperature, time series, weather. Environ Health Perspect 110:859–863 (2002).[Online18 July 2002]  Articles  Articles • Braga et al. 860 VOLUME 110 | NUMBER 9 | September 2002  • Environmental Health Perspectives previous study ( 7  ). We modeled the covari-ates we controlled for (season and trend, day of the week, and barometric pressure) by using nonparametric smoothing as describedbelow.In environmental epidemiologic studies, we expect the relationship between the out-come and some variables to be nonlinear. Thegeneralized additive model ( 11 ) fits smoothfunctions for these variables. We chose Loesssmoothes for our models ( 12  ).In this 8-year study, we used a smoothfunction of time to capture the basic long time trend represented by the expected six rises and falls in daily deaths over the periodbecause of seasonality ( 13  ). This approach hasbeen adopted systematically in environmentalepidemiologic studies of daily deaths( 6,14–16  ). Seasonal patterns can vary greatly among cities and for different causes of death. We chose a separate smoothing parameter ineach city and for each cause to both eliminateseasonal patterns in the residuals and reducethe residuals of the regression to “white noise”(i.e., remove serial correlation), as describedpreviously ( 17  ). In models with remaining ser-ial correlation from the residuals, we incorpo-rated autoregressive terms ( 18  ).The other covariates were barometricpressure on the same day and day of the week. To allow for city- and cause-specificdifferences, we chose the smoothing parame-ters for these covariates separately in eachlocation and for each cause to minimize Akaike’s information criterion ( 19  ). Distributed lag models. Distributed lag models have been used extensively in thesocial sciences ( 20  ), and their use in epidemi-ology was described by Pope and Schwartz( 21 ). Recently, this methodology has beenapplied to several studies estimating the dis-tributed lag between air pollution and healtheffects ( 6,15,22  ). The motivation for the dis-tributed lag model is the realization that tem-perature can affect deaths occurring notmerely on the same day but also on severalsubsequent days. Therefore, the converse isalso true: deaths today will depend on the“same-day” effect of today’s temperature, the“one-day lag” effect of yesterday’s tempera-ture, and so forth. Therefore, suppressing covariates and just focusing on temperaturefor the moment, the unconstrained Poissondistributed lag model assumesLog [ E  ( Y  t  )] = α + β 0  X  t  + … + β q   X  t–q  + ε t ,[1] where  X  t–q  is the temperature q  days beforethe deaths. In this study, we examined theeffect of temperature in the 12 cities ondeaths with latencies (lags) ranging fromzero to 20 days after the temperature event.Because the effects of temperature onmortality are usually nonlinear, with J-, U-,or V-shaped relations commonly reported, weused both a linear and a quadratic term fortemperature at each lag. Equation 1 can berecast asLog [E( Y  )] = α + ω 0  X  t  + … + ω q   X  t–q  + ω q  +1  X  t  2 + … + ω q+q   X  t–q  2 + ε t ,[2] where the ω i  are parameters.Because substantial correlation existsbetween temperatures on days close togetherand between temperature and its square, theabove regression will have a high degree of collinearity. This will produce unstable esti-mates of the individual ω i  and hence poorestimates of the shape of the distribution of the effect over lag.To gain more efficiency and more insightinto the shape of the distributed effect of thetemperature over time, constraining ω i  is use-ful. If this is done flexibly, substantial gains inreducing the noise of the unconstrained dis-tributed lag model can be obtained, withminimal bias ( 6  ). The most commonapproach is to constrain the shape of the vari-ation of the ω i   with lag number to fit somepolynomial function. We used separatefourth-degree polynomial constraints for thelinear and quadratic temperature terms,because that should be flexible enough toencompass any plausible pattern of delayedeffect over time. The result is a 10 degree-of-freedom surface of the effect of temperatureover the past 3 weeks on death from each spe-cific cause. We simultaneously included linearand quadratic terms for relative humidity upto 20 days before the death in the model,subject to similar constraints.The immediate effects of weather extremesmay represent harvesting—that is, deathsbrought forward by only a few days. To assessthis, we compared the estimated immediate(lag 0 and 1) effect of hot days with the sumof the estimated effect over 7 days.By fitting the same model in 12 differentlocations, for pneumonia, COPD, CVD, andMI deaths, and combining effect size esti-mates, by lag over the cities, we can estimatethe distribution of the effect of temperatureand humidity over time. To combine resultsacross cities, we used inverse variance weighted averages including a random vari-ance component to incorporate heterogeneity. We stratified analysis in two groups of cities: the hot cities (Atlanta, Houston, andBirmingham) and cold cities (Canton,Chicago, Colorado Springs, Detroit,Minneapolis, New Haven, Pittsburgh, andSpokane). As we observed in the total mortal-ity study ( 7  ), the differences in the temperatureranges between these two groups of cities pre-cluded a useful combination across all cities.In this hierarchical study (i.e., a study  with multiple levels of analysis), we first fitteda generalized additive Poisson regression foreach city and each outcome. In the secondstage of the analysis, we fitted an ecologicregression to investigate the role of the preva-lence of central air conditioning and the vari-ance of summer and winter temperature, thebackground mortality rate, percentage of population with a college degree, percentnonwhite, percent unemployed, percent liv-ing below the poverty level, city size, andmean age of the population on the estimatedeffect of hot days (24 hr mean of 30°C) andcold days (24 hr mean of –10°C) on cause-specific deaths. To do this, we regressed theestimated effect in each city at each of thosetemperatures against the above explanatory variables. We obtained prevalence of air con-ditioning from the American Housing Survey  Web site and the remaining demographicdata from the 1990 census. We used inversevariance weighting. Where heterogeneity remained, as assessed by a chi-square test, wefitted the regression including a random vari-ance component, estimated using a maxi-mum likelihood approach, following themethod of Berkey et al. ( 23  ). Results Table 1 presents the descriptive analysis of thevariables used in the study. The cities varied in Table 1. The populations and the descriptive analysis of the variables in the study in the 12 locations.1990 Temperature b  (°C)Humidity c  (%)Pressure d  CitiespopulationDeaths a 5%Mean95%5%Mean95%(mm Hg) Atlanta 1,642,533 36.2 3.3 17.1 28.3 41.0 67.0 93.0 736Birmingham 651,525 19.1 2.8 16.9 27.8 49.0 70.5 91.0 747Canton 367,585 9.9 –6.1 10.0 24.4 51.0 73.7 93.0 729Chicago 5,105,067 133.4 –7.2 10.1 25.6 50.0 70.8 92.0 744Colorado Springs 397,014 6.0 –6.1 9.5 22.8 25.0 51.0 84.0 610Detroit 2,111,687 59.7 –6.1 10.5 25.6 49.0 69.2 89.0 744Houston 2,818,199 47.0 7.2 20.3 30.0 54.0 75.0 92.0 760Minneapolis 1,518,196 32.3 –13.3 7.9 25.0 45.0 68.7 90.0 739New Haven 804,219 20.4 –6.1 10.7 25.0 43.0 66.8 92.0 760Pittsburgh 1,336,449 42.4 –5.0 11.2 25.0 48.0 69.3 90.0 732Seattle 1,507,319 29.3 2.8 11.4 20.6 52.0 77.0 93.0 752Spokane 361,364 8.7 –5.6 8.8 22.8 35.0 68.0 95.0 699 a  Daily mean. b  Daily mean temperature. c  Relative humidity. d  Barometric pressure.  Articles • Weather and respiratory and cardiac deaths Environmental Health Perspectives  • VOLUME 110 | NUMBER 9 | September 2002  861 size, although in 1990 seven cities of the study had more than one million inhabitants. Wedivided the cities in two groups according totheir meteorologic characteristics: hot (Atlanta,Birmingham, and Houston) and cold(Canton, Chicago, Colorado Springs, Detroit,Minneapolis, New Haven, Pittsburgh, andSpokane). Among the hot cities, Houston wasthe hottest and most humid; among the coldcities, Minneapolis was the coldest and had the widest range of temperatures. Seattle, locatedin the extreme northwest of the United States,had the narrowest range of temperatures of thecities in this study and rarely exhibited extremetemperatures.In the hot cities and in New Haven, tem-perature was positively associated withhumidity. Correlations between temperatureand barometric pressure were, in general,small and negative. We estimated the covariate-adjusted(including humidity) effects of temperature onrespiratory and CVD daily deaths by lag in the12 cities, using a standard range of tempera-tures. We then performed a meta-analysis of temperature effect for hot and cold cities. Wedid not include Seattle in this stratified analysisby temperature groups because its mild tem-perature range did not fit in either group.In cold cities (Figure 1), both high andlow temperatures were associated withincreased numbers of CVD deaths. In general,the effect of cold temperatures persisted fordays, whereas the effect of high temperatures was restricted to the day of the death orimmediately the day before. For MI deaths,the hot-day effect at lag zero was twice as largethe cold-day effect (6% and 3% increases indaily deaths, respectively), whereas for allCVD deaths it was five times smaller than thecold-day effect (1% and 5% increases in daily deaths, respectively). For MI deaths and hotdays we observed a harvesting effect: After 2days we found a 12% increase in deaths, which decreased to 4% when we looked at thecumulative effects up to 7 days. For CVDdeaths, we found a 3% increase after 2 daysthat decreased to –0.6% after 7 days. Also, only hot temperatures increasedCOPD deaths (25%); the cold effect waszero. Pneumonia deaths differed from theother causes of death in that the cold-day effect was larger, and the effect of hot temper-atures was stronger at lags 3–5 (an average of 15% increase).In hot cities (Figure 2), neither hot norcold temperatures had much effect on CVDor pneumonia deaths. However, for MI andCOPD deaths, we observed lagged effects of hot temperatures (lags 4–6, 4% increase, andlags 3 and 4, 6% increase, respectively).Similar to that observed in total mortality analysis ( 7  ),when we estimated the effect of humidity on respiratory and CVD daily deaths in each of the 12 cities, we observed noconsistent pattern, in terms of either lag struc-ture or differences between high and low humidity. Stratifying the cities by weathercharacteristics also did not suggest any patternfor humidity.In the meta-regressions, none of the pre-dictors significantly modified the effects of hot or cold days on CVD deaths (Table 2).However, for both COPD and pneumonia,the variance in summer temperature was asso-ciated with substantial increases in the effectof a hot day. The variance of winter tempera-ture was similarly associated with substantialincrease in the death rate on cold days.None of the demographic factors (back-ground mortality rate, percentage of popula-tion with a college degree, percent nonwhite,unemployment rate, percent below poverty level, city size, and mean age of the popula-tion) modified the effect of either cold or heat waves in our data (  p > 0.12). Discussion Temperature has been recognized as a physicalagent able to induce health effects ( 1,2,24  ).The rapid buildup of greenhouse gases isexpected to increase both mean temperatureand temperature variability around the world( 25  ). This has added urgency to the need tobetter understand the direct effects of suchchanges on daily death rates, and to betterunderstand the modifiers of those effects. Oneissue that has been extensively explored in thisfield is the shape of the relationship betweentemperature and deaths. U- and V-shapes havebeen reported in regions where both hot andcold temperatures have been associated withfatal events with similar magnitudes of effects, whereas J-shapes and even a linear shape havebeen reported in regions where the susceptibili-ties for extreme temperatures are not similar( 22  ). We have focused our attention onexploring the lag structure between tempera-ture and daily deaths using a systematicapproach to look at the delayed effects of  weather on mortality up to 3 weeks afterwards.In this study we looked at the temperatureeffect on cause-specific deaths in 12 U.S. cities. As observed in the total mortality study ( 7  ),hot and cold temperatures were associated withincreased deaths, and the shape of this relation-ship varied according to climatic characteristicsof the cities. However, we found sizable effectsof temperature on daily deaths just at lag 0. We found lagged effects of hot temperatures inhot cities and specifically for MI and COPD,and in cold cities for pneumonia.In cold cities, we found differences in termsof temperature effect on CVD. Although bothhot and cold temperatures affected MI andtotal CVD deaths, the relative impacts of theextreme temperatures were different. Cold pre-sented more homogeneous and persistenteffects on both outcomes, with no evidence of harvesting. Heat presented a much moreimportant effect on MI deaths than it did onCVD deaths. These effects were predominantly short-term mortality displacement. The patternobserved for temperature effects on CVDdeaths in cold cities is similar to those observedfor total deaths, probably because most of thetotal mortality is due to CVD deaths.Cold temperatures did not have mucheffect on respiratory mortality in cold cities.However, heat increased respiratory deaths. Figure 1. Overall effect of temperature (°C) on MI ( A ), CVD ( B  ), pneumonia ( C  ), and COPD ( D  ) daily deathsin the eight cold cities: Canton, Colorado Springs, Chicago, Detroit, Minneapolis, New Haven, Pittsburgh,and Spokane. In the z  -axis, a log relative risk of 0.1 represents a 10% increase in mortality.  2  0 1  6 1 2 8 4  0   L a g - d a y s  2  0 1  6 1 2 8 4  0 –  1   2   0   1   2   2   4    T   e  m   p e  r  a  t  u  r  e   –  1   2   0   1   2   2   4    T   e  m   p e  r  a  t  u  r  e    2  0 1  6 1 2 8 4  0 2  0 1  6 1 2 8 4  0 –  1   2   0   1   2   2   4    T   e  m   p e  r  a  t  u  r  e   –  1   2   0   1   2   2   4    T   e  m   p e  r  a  t  u  r  e––0.05–0.100.250–0.25–0.5–0.750.250.00–0.25    l  o  g   R  e   l  a   t   i  v  e  r   i  s   k   l  o  g   R  e   l  a   t   i  v  e  r   i  s   k   l  o  g   R  e   l  a   t   i  v  e  r   i  s   k   l  o  g   R  e   l  a   t   i  v  e  r   i  s   k ABCD   L a g - d a y s  L a g - d a y s  L a g - d a y s  Articles • Braga et al. 862 VOLUME 110 | NUMBER 9 | September 2002  • Environmental Health Perspectives For COPD, the heat effect was remarkableand acute (lag 0, 25-fold higher than the coldeffect), whereas we observed a lagged effectfor pneumonia.In hot cities, we found no relevant effectsof cold on both respiratory death and CVDdeaths. When we analyzed pneumonia, weobserved no association with temperature.The same behavior could be seen for CVD.However, for the relation between heat andboth MI and COPD deaths, we saw a patterndifferent from the total mortality results: Weobserved lagged effects for these two causes of death. Hence, even in hot cities, where peopleare more accustomed to hot temperatures andair conditioning is common ( 26  ), the effect of heat on health, leading to increased deaths, canovercome adaptive mechanisms.In our hierarchical model, we found thatthe variance of summer and winter tempera-ture was associated with substantial changesin the effects of hot and cold days on respira-tory but not CVD deaths. The substantialmortality increase in cities with more variabletemperature suggests that increased tempera-ture variability is the most relevant change inclimate for the direct effects of weather onrespiratory mortality.In many ways, the results of this study andour previous study of total mortality parallelthose of the Eurowinter study ( 27  ), whichassessed the association between daily deathsand temperature in the winter in eightEuropean regions. Daily deaths increased withfalling temperatures in all regions. However,the effect of a cold day was greater in warmerclimates than in colder climates. In our 12-city U.S. studies, the converse was true: Theeffect of hot days was worse in cities wherethey were less common. In the Eurowinterstudy, the effect of cold days was reduced by  warmer temperatures in the living room andmore hours per day of heat in the bedroom—that is, by greater use of space conditioning toreduce exposure to the cold weather. In ourstudy, greater use of central air conditioning  was associated with a reduced effect of hotdays for total and for cause-specific mortality,although the results were less significant forthe cause-specific mortality. Greater variability in either summer or winter temperatures, which might be expected to reduce protectivebehavior such as always wearing hats, wasassociated with increased effects of cold orheat waves. The overall message seems to bethat space conditioning and behavior can sub-stantially modify the adverse impacts of tem-perature extremes, but that this behavior ismore frequently found in the climates wherethose extremes are common. We found no association in the second-stage analysis with baseline mortality rates orsocial or demographic factors. However, a log-linear regression builds in interactions by design—that is, we estimated our tempera-ture effect as a relative or percentage changein each city. In cities with a higher baselinerate, a greater absolute effect is built in. Thesecond-stage regression therefore tests super-multiplicativity. This makes the failure tofind interactions with direct or indirect mark-ers of baseline risk understandable and theassociation with the temperature variancesmore impressive.In the present study and in the previousone ( 7  ), we have used mean temperature. Thebest indicator of the temperature effect onhealth is still debated ( 2  ). Further analysisusing different parameters (e.g., minimumtemperature and dew point temperature) areneeded to compare the results presented hereand elsewhere, and for finding the best instru-ment for estimating the health effect due toextreme weather exposure.In this cause-specific death study, we saw no consistent patterns for the relation of humidity to daily deaths by city. The com-bined city estimate reinforced this idea, show-ing no overall effect of humidity on totaldaily deaths. Using dew point temperaturecan give a more reasonable estimate of thehumidity effect on daily deaths and should bepursued in the future. Air pollution is a predictor of daily deaths. Effect modification was tested by Samet et al. ( 26  ) in a study of 20 years of data in Philadelphia. They stratified days into 20categories based on synoptic weather condi-tions and found no effect modification. Thisdoes not preclude the possibility that effect Figure 2. Overall effect of temperature (°C) on MI ( A ), CVD ( B  ), pneumonia ( C  ), and COPD ( D  ) daily deathsin the three hottest cities: Atlanta, Birmingham, and Houston. In the z-axis, a log relative risk of 0.1 repre-sents a 10% increase in mortality.  2  0 1  6 1 2 8 4  0   L a g - d a y s  2  0 1  6 1 2 8 4  0 –  1   2   0   1   2   2   4    T   e  m   p e  r  a  t  u  r  e   –  1   2   0   1   2   2   4    T   e  m   p e  r  a  t  u  r  e    2  0 1  6 1 2 8 4  0 2  0 1  6 1 2 8 4  0 –  1   2   0   1   2   2   4    T   e  m   p e  r  a  t  u  r  e   –  1   2    0    1   2    2    4     T   e  m   p  e  r  a  t  u  r  e   0.130.00––0.0500.00–0.06–0.12–0.180.000–0.120–0.240    l  o  g   R  e   l  a   t   i  v  e  r   i  s   k   l  o  g   R  e   l  a   t   i  v  e  r   i  s   k   l  o  g   R  e   l  a   t   i  v  e  r   i  s   k   l  o  g   R  e   l  a   t   i  v  e  r   i  s   k ABCD   L a g - d a y s  L a g - d a y s  L a g - d a y s Table 2. Percentage increase in cause-specific deaths at 30°C and at –10°C for the difference between the 90th and 10th percentiles in air conditioning, variance of summer temperature, and variance of winter temperature.Summer effectWinter effectPercent95% CIPercent95% CI CVDAir conditioning –1.15  – 14.72–14.60Variance summertime temperature 0.93  – 9.67–12.77Variance wintertime temperature 2.20  – 1.19–5.71MIAir conditioning –16.99  – 35.64–7.06Variance summertime temperature 15.67  – 7.54–44.71Variance wintertime temperature –3.63  – 11.62–5.08COPDAir conditioning –13.44  – 45.89–38.49Variance summertime temperature 42.76 4.54–94.94Variance wintertime temperature 25.86  – 1.12–60.20PneumoniaAir conditioning –8.31  – 30.79–21.47Variance summertime temperature 28.01 3.96–57.63Variance wintertime temperature 12.57 2.87–23.19  Articles • Weather and respiratory and cardiac deaths Environmental Health Perspectives  • VOLUME 110 | NUMBER 9 | September 2002  863 modification may be seen in other studies.However, the only air pollutant consistently associated with daily deaths in the U.S. is air-borne particles ( 28  ). Unfortunately, airborneparticles are measured only one day in six inmost U.S. cities. This would prevent us fromexamining the effect of multiple lags of  weather in our study. Hence we have chosennot to include it in our models.In summary, we found that temperatureis associated with increased daily cause-spe-cific deaths in both cold and hot cities. Incold cities, both heat and cold contributed with daily cause-specific deaths. In hot cities,only heat presented important effect on daily deaths, and its effect was smaller than thoseobserved in cold cities. In these cities, peopleseem to be more adapted to heat waves andalso are not exposed to very low temperatures.Therefore, we reinforce the concept thatanalysis of the impact of any climatic changeshould take into account regional weather dif-ferences and that further analysis using differ-ent weather indicators must be done. R EFERENCESAND N OTES 1.Kalkstein LS, Greene JS. An evaluation of climate/mortalityrelationship in large U.S. cities and the possible impacts ofa climate change. Environ Health Perspect 105:84–93 (1997).2.McGeehin MA, Mirabelli M. The potential impacts of cli-mate variability and change on temperature-related mor-bidity and mortality in the United States. Environ HealthPerspect 109(suppl 2):185–189 (2001).3.Semenza JC, McCullough JE, Flanders WD, McGeehin MA,Lumpkin JR. Excess hospital admissions during the July1995 heat wave in Chicago. N Engl J Med 16:269–277 (1999).4.Centers for Disease Control and Prevention. Heat-relateddeaths—Philadelphia and United States, 1993–1994.Available: [cited 17 July2001].5.Almon S. The distributed lag between capital appropria- tions and expenditures. Econometrica 33:178–196 (1965).6.Schwartz J. The distributed lag between air pollution anddaily deaths. Epidemiology 11:320–326 (2000).7.Braga ALF, Zanobetti A, Schwartz J. The time course ofweather related deaths: a tridimensional estimate in 12 UScities. Epidemiology 12:662–667 (2001).8.International Classification of Diseases, 9th Revision.Vol1. Commission on Professional and Hospital Activities.Ann Arbor, MI:Edward Brother, Inc., 1979.9.U.S. Department of Health and Human Services. PublicUse Data Tape Documentation. Hyattsville, MD:NationalCenter for Health Statistics, 1993.10.EarthInfo. NCDC Surface Airways. Boulder, CO:EarthInfo,Inc. Available: [cited 15 November 2000].11.Hastie T, Tibshirani R. Generalized Additive Models.London:Chapman and Hall, 1990.12.Cleveland WS, Devlin SJ. Robust locally-weighted regres-sion and smoothing scatterplotts. J Am Stat Assoc74:829–836 (1988).13.Schwartz J. Generalized additive models in epidemiology.In: International Biometric Society, Invited Papers. 17thInternational Biometric Conference, 8–12 August 1994,Hamilton, Ontario, Canada. Washington, DC:InternationalBiometric Society, 1994;55–80.14.Schwartz J. Assessing confounding, effect modification,and thresholds in the association between ambient parti-cles and daily deaths. Environ Health Perspect 108:563–568(2000).15.Braga ALF, Saldiva PHN, Pereira LAA, Menezes JJC,Conceição GMS, Lin CA, Zanobetti A, Schwartz J, DockeryDW. Health effects of air pollution exposure on childrenand adolescents in São Paulo, Brazil. Pediatr Pulmonol1:106–113 (2001).16.Conceição GMS, Miraglia SGEK, Kishi HS, Saldiva PHN,Singer JM. Air pollution and child mortality: a time-seriesstudy in São Paulo, Brazil. Environ Health Perspect109(suppl 3):347–350 (2001).17.Schwartz J. Air pollution and hospital admissions for heartdisease in eight US counties. Epidemiology 10:17–22 (1999).18.Brumback BA, Ryan LM, Schwartz J, Neas LM, Stark PC,Burge HA. Transitional regression models with application to environmental time series. J Am Stat Assoc 95(449):16–28(2000).19.Akaike H. Information theory and an extension of the max-imum likelihood principle. In: 2nd International Symposiumon Information Theory (Petrov BN, Csaki F, eds).Budapest:Akademiai Kiado, 1973;267–281.20.Judge GG, Griffiths WE, Hill RC, Lutkepohl H, Lee T-C. TheTheory and Practice of Econometrics. New York:JohnWiley and Sons, 1980.21.Pope CA III, Schwartz J. Time series for the analysis ofpulmonary health data. Am J Respir Crit Care Med154:S229–S233 (1996).22.Zanobetti A, Schwartz J, Dockery DW. Airborne particlesare a risk factor for hospital admissions for heart and lungdisease. Environ Health Perspect 108(11):1071–1077 (2000).23.Berkey CS, Hoaglin DC, Mossteller F, Colditz GA. A ran-dom-effects regression model for meta-analysis. Stat Med14:395–411 (1995).24.Houghton JT, Ding Y, Griggs DJ, Noguer M, van derLinden PJ, Xiaosu D, eds. Climate Change 2001: TheScientific Basis. Contribution of Working Group I to theThird Assessment Report of the Intergovernmental Panelon Climate Change (IPCC). Cambridge:CambridgeUniversity Press.25.McCarthy JJ, Canziani OF, Leary NA, Dokken DJ, WhiteKS, eds. Climate Change 2001: Impacts, Adaptation &Vulnerability Contribution of Working Group II to the ThirdAssessment Report of the Intergovernmental Panel onClimate Change (IPCC). Cambridge:Cambridge UniversityPress, 2001.26.Samet JM, Zeger SL, Kelsall JE, Xu J, Kalkstein LS. AirPollution, Weather, and Mortality in Philadelphia1973–1988. Health Effects Institute Special Report 1–29.Particulate Air Pollution and Daily Mortality: Analysis of the Effects of Weather and Multiple Air Pollutants. FinalReport. Cambridge, MA:Health Effects Institute 1997. 27.The Eurowinter Groups. Cold exposure and winter mortal-ity from ischaemic heart disease, cerebrovascular dis-ease, respiratory disease, and all cause in warm and coldregions in Europe. Lancet 349:1341–1346 (1997).28.Samet JM, Dominici F, Curriero F, Coursac I, Zeger SL.Particulate air pollution and mortality: findings from 20U.S. cities. N Engl J Med 343:1742–1757 (2000).
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!