Abstract
Background
While several studies have examined associations between temperature and cardiovascular-disease-related mortality, fewer have investigated the association between temperature and the development of acute myocardial infarction (MI). Moreover, little is known about who is most susceptible to the effects of temperature.
Methods
We analyzed data from the Worcester Heart Attack Study, a community-wide investigation of acute MI in residents of the Worcester (MA) metropolitan area. We used a case-crossover approach to examine the association of apparent temperature with acute MI occurrence and with all-cause in-hospital and post-discharge mortality. We examined effect modification by sociodemographic characteristics, medical history, clinical complications, and physical environment.
Results
A decrease in an interquartile range (IQR) in apparent temperature was associated with an increased risk of acute MI on the same day (hazard ratio=1.15 [95% confidence interval= 1.01–1.31]). Extreme cold during the 2 days prior was associated with an increased risk of acute MI (1.36 [1.07–1.74]). Extreme heat during the two days prior was also associated with an increased risk of mortality (1.44 [1.06–1.96]). Persons living in areas with greater poverty were more susceptible to heat.
Conclusions
Exposure to cold increased the risk of acute MI, and exposure to heat increased the risk of dying after an acute MI. Local area vulnerability should be accounted for as cities prepare to adapt to weather fluctuations as a result of climate change.
Human activity is expected to result in a global increase in temperature, as well as differential changes by season and location. Additionally, the frequency of extreme temperature episodes is projected to rise.1 Many studies have linked increased mortality to changes in or extremes of temperature.2-5
While the association between temperature and cardiovascular mortality is important,6 fewer studies have investigated the association between the incidence of cardiovascular disease (CVD) and ambient temperature. Many of the studies that have examined the association between temperature and myocardial infarction (MI) have used MI mortality as an outcome rather than non-fatal events. Moreover, few of these studies have adjusted for potentially important confounders, including air pollution.7
The Interagency Working Group on Climate Change and Health recently highlighted vulnerability to the human health effects of climate change as a research priority.8 While there is some evidence that sociodemographic characteristics, including advanced age, race, and lower educational attainment, can increase the risk of mortality due to temperature,6,9 little is known about susceptibility in the relationship between temperature and occurrence of MI.7
We examined the association between temperature and occurrence of acute MI, as well as subsequent mortality. In addition to the main effects of temperature, we examined effect modification of the association between temperature and acute MI, as well as subsequent mortality, by individual and area-level characteristics.
Methods
The data for this investigation are from the Worcester Heart Attack Study, an ongoing community-wide investigation examining changes over time in the incidence and case-fatality rates of confirmed episodes of acute MI in residents of the greater Worcester area who were hospitalized with acute MI at all area medical centers. The details of this study have been described previously.10-12 In brief, during the 5 study years of 1995, 1997, 1999, 2001 and 2003, the medical records of the 11 acute care general hospitals serving residents of the Worcester metropolitan area were searched for patients with a possible discharge diagnosis of acute MI. The records were individually reviewed and validated according to diagnostic criteria described previously.10 These criteria included a suggestive clinical history, increased serum biomarker levels above each hospital's normal range, and serial electrocardiographic findings indicative of acute MI. At least 2 of these 3 criteria were necessary for study inclusion. The present investigation was limited to patients aged 25 years and older who were hospitalized with independently confirmed acute MI. Patients' residential addresses at the time of acute MI were collected from information contained in hospital medical records and geocoded. Date of death was ascertained for any patient who died after hospital admission for acute MI. This included patients who died in the hospital as well as those who were discharged from the participating medical centers. Long-term survival status was ascertained through the end of calendar year 2005 by reviewing records for additional hospitalizations and by searching state and national death certificates for residents of greater Worcester. The study was approved by the Committee for the Protection of Human Subjects at the University of Massachusetts Medical School and the Human Subjects Committee at the Harvard School of Public Health.
Exposure
We used daily mean apparent temperature, calculated from air and dew-point temperature,2 as the exposure metric in our analysis because it is a combined index of temperature and humidity that captures the physiologic experience better than temperature alone, and has been used in other studies assessing morbidity and mortality.13,14 Air and dew-point temperatures at the Worcester Airport were obtained from the National Climatic Data Center. While we used only one temperature measurement site for our exposure, the intraclass correlation between ambient temperatures during the study period at Worcester Regional Airport and Boston Logan Airport, which is 80 km away, was 0.934. This suggests that day-to-day variation in temperature at Worcester Regional Airport is a good surrogate for day-to-day variation at the participants' residential locations within the Worcester metropolitan area. The median distance between subjects' homes and the Worcester Regional Airport was 9.3 kilometers (standard deviation = 7.0 km, range 1.3 to 35.4 km).
Covariates and effect modifiers
Ozone and PM2.5 measurements were obtained from a stationary monitoring source located in Boston, MA. Absolute humidity was calculated from the dew-point temperature (www.gorhamschaffler.com/humidity_formulas.htm).
A number of individual- and area-level characteristics were available to examine effect measure modification (Table 1). Individual-level data were abstracted from hospital medical records for each patient and included demographic characteristics, medical history (angina, diabetes, hypertension, family history of coronary heart disease), smoking status (current vs. non), acute MI order (initial vs. prior), acute MI type (Q-wave vs. non–Q-wave), and hospital complications of acute MI (heart failure, atrial fibrillation, and cardiogenic shock).
Table 1. Individual and area-level characteristics of patients in the Worcester Heart Attack Study cohort.
| AMI Occurrence Analysis (n = 4,765) | Mortality Analysis (n = 2,427) | |
|---|---|---|
| Age (years); mean (SD) | 70.3 (13.9) | 76.7 (11.4) |
| Sex; no. (%) | ||
| Male | 2736 (57) | 1213 (50) |
| Female | 2029 (43) | 1214 (50) |
| Race; no. (%) | ||
| White | 4296 (90) | 2230 (92) |
| Black | 53 (1) | 19 (1) |
| Other | 254 (5) | 103 (4) |
| Unknown | 162 (3) | 75 (3) |
| Medical history a; no. (%) | ||
| Angina | 1046 (22) | 594 (25) |
| Diabetes mellitus | 1429 (30) | 901 (37) |
| Hypertension | 3052 (64) | 1671 (69) |
| Stroke | 515 (11) | 411 (17) |
| Family history of CHD b | 1437 (30) | 441 (18) |
| Current Smoker b; no. (%) | 860 (18) | 276 (11) |
| Type of MI; no. (%) | ||
| Q-wave | 1429 (30) | 585 (24) |
| Non Q-wave | 3336 (70) | 1842 (76) |
| Initial | 3356 (70) | 1522 (63) |
| Prior | 1409 (30) | 905 (37) |
| Clinical complications; no. (%) | ||
| Atrial fibrillation | 822 (17) | 538 (22) |
| Cardiogenic shock a | 299 (6) | 234 (10) |
| Heart failure | 1701 (36) | 1220 (50) |
| Elevation at residence, meters c; mean (SD) | 172.0 (51.8) | 172.2 (49.7) |
| Lives within 400 m of large (> 100,00 m2) water body c; no. (%) | 664 (14) | 336 (14) |
| Census block group median household income in 1999 c ($); mean (SD) | 48,465 (19,258) | 46,299 (18,654) |
| Census block group % of persons below the federally defined poverty line c; mean (SD) | 10.4 (11.6) | 11.2 (12.3) |
| % of census tract conservation / recreation c; mean (SD) | 11.0 (10.7) | 10.5 (10.8) |
| Normalized Difference Vegetation Index c; mean (SD) | 191.2 (29.1) | 189.4 (29.0) |
| Census tract year of building construction c; median | 1958 | 1957 |
| Census tract housing density (units/km2) c; mean (SD) | 638 (768) | 698 (798) |
| Census tract % of buildings containing > 4 units c; mean (SD) | 17.9 (15.4) | 19.3 (16.0) |
Missing data for one patient
Missing data for 896 patients in MI occurrence analysis and 554 patients in mortality analysis
Address could not be geocoded for 103 patients included in the MI occurrence analysis and 52 patients in the mortality analysis
Year 2000 census data were obtained from the US Bureau of the Census Summary File (III) at both the tract and block group level (for 101 census tracts and 375 block groups). Census tracts have an average population of about 4000 and census block groups have a population of about 1500. Both are designed to have relatively homogeneous socioeconomic characteristics. Census-block group data on economic poverty have been shown to be a relatively sensitive measure of socioeconomic inequalities in health outcomes.15 We used the proportion of the population with year 1999 income below the federally defined poverty level and median household income within a census-block group as area-based measures of socioeconomic position. To get a sense of area-level building characteristics, we obtained census-tract-level data on the median year of building construction, housing density, and percent of housing units that have more than four units. Data on open space, bodies of water, and elevation at patient's residence were downloaded from the MassGIS website.16 The Normalized Difference Vegetation Index was obtained as an additional measure of greenness. We chose a summer measurement from the year 2000 to match the census data in our analysis and assumed no change in the mean vegetation index in the same season by year.17 We used 250-m resolution 16-day composite vegetation-index data downloaded from the Global Land Cover Facility (http://glcf.umiacs.umd.edu/) and averaged it over 3 X 3 neighboring cells (750 m square) to capture conditions around subjects' homes.
Statistical Analysis
We used case-crossover analyses to compare temperature on the day of hospital presentation for acute MI and control days for each patient. To minimize potential biases due to season and long-term time trends,18 we selected control periods according to the time-stratified approach proposed by Lumley and Levy.19 Exposure during the 24 hours immediately preceding an event (and other corresponding moving averages) was compared with exposure within the same month of the same year by using every third day of that month as control days to reduce serial correlation (as compared with using adjacent days) in the exposure variable. We used conditional logistic regression models in which the individual patient was the conditioning factor. We also controlled our models for day of the week with indicator variables, regional daily average PM2.5 and ozone on the day of presentation, and absolute humidity. Because we used up to 6-day-moving-average windows of exposure, and some of the control periods would overlap with our exposure periods, we also performed a sensitivity analysis, in which we used as control days the same day in the other weeks within the same month of the same year.
We first examined acute MI as the outcome of interest, and next examined subsequent all-cause mortality after acute MI as an outcome. Importantly, the all-cause mortality analysis is on a subset of this population (all of whom previously had an acute MI), who died within the study period. We ran separate models for warmer (April – October) and colder (November – March) months. We examined apparent temperature as a continuous variable and as a binary indicator for extreme weather. To determine if an exposure period was “extreme,” the mean temperature was assessed against the 5th or 95th percentile during that season. For example, for a 2-day mean exposure, the mean of both days was calculated. If this mean value was greater than the 95th percentile of temperature during the warm season, it was considered an extreme period of heat.
We next examined effect modification between individual- and area-level characteristics and both continuous apparent temperature and temperature extremes. When residential address could not be geocoded at the street level, patients (n=103) were excluded from the effect-modification analysis of area-level characteristics. We calculated stratum-specific estimates for all modifying characteristics.
In sensitivity analysis, we examined daily mean temperature and ran models that included both continuous apparent temperature and binary temperature extremes in the same model. All statistical analyses were performed using SAS version 9.2 software (SAS Institute, Inc., Cary, North Carolina).
Results
Data were available for 4,774 persons presenting with acute MI during the study period. We excluded 2 patients living outside the Worcester metropolitan area and one patient without information on address. Another six patients were excluded because their admission dates were not available, leaving a total of 4,765 patients for our analysis of temperature and occurrence of acute MI. Of these, 2,427 died either in the hospital or over the follow-up period and were included in our analysis of mortality. Information on meteorologic variables for Worcester County during the study period is shown in Table 2. The mean apparent temperature in the study area was -1.3° C during the cold months and 14.7° C during the warm months.
Table 2. 24-hour averages of meteorologic data for Worcester, MA, 1 November 1994 – 31 December 2003.
| Overall Study Period | Cold Months a | Warm Months b | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Percentile | Percentile | Percentile | |||||||||||||
|
|
|
|
|||||||||||||
| Mean | Median | IQR | 5th | 95th | Mean | Median | IQR | 5t | 95th | Mean | Median | IQR | 5th | 95th | |
| Apparent temperature (°C) | 7.9 | 6.6 | 18.1 | -6.9 | 25.5 | - 1.3 | - 1.8 | 6.6 | - 8.6 | 8.2 | 14.7 | 15.3 | 12.3 | 2.0 | 26.9 |
| Temperature (°C) | 8.8 | 9.2 | 16.4 | -7.2 | 23.1 | 0.02 | 0.0 | 8.1 | -10.0 | 10.3 | 15.3 | 16.1 | 9.4 | 4.4 | 24.2 |
| Absolute humidity (kg/m3) | 0.0072 | 0.0058 | 0.01 | 0.0016 | 0.0158 | 0.0037 | 0.0033 | 0.0022 | 0.0013 | 0.0076 | 0.0098 | 0.0097 | 0.0065 | 0.0035 | 0.0169 |
| O3 (ppm) c | 0.02 | 0.02 | 0.02 | 0.01 | 0.05 | 0.02 | 0.02 | 0.01 | 0.004 | 0.03 | 0.03 | 0.03 | 0.02 | 0.01 | 0.05 |
| PM2.5 (μg/m3) c | 11.8 | 10.2 | 7.3 | 4.6 | 24.7 | 11.6 | 10.5 | 6.6 | 4.9 | 21.9 | 11.9 | 9.8 | 7.7 | 4.4 | 26.4 |
January, February, March, November and December
April, May, June, July, August, September, October
Data from Boston, MA
IQR indicates Interquartile range
Onset of acute MI
A decrease in an interquartile range (IQR) of apparent temperature on the same day of presentation was associated with an increased risk of acute MI during cold months (Table 3). Extreme cold during the 2-day moving average (average of mean apparent temperature for lag 0 and lag 1) was associated with an increased risk of acute MI during cold months (Table 4). We did not find any association between temperature and acute MI during the warm months.
Table 3. Occurrence of acute MI and All-Cause Mortality with an Interquartile Range (IQR) Change in Apparent Temperature.
| Cold Months a | Warm Months b | |||
|---|---|---|---|---|
| HR c | (95 % CI) | HR d | (95 % CI) | |
| Occurrence of acute MI | ||||
| Same day | 1.15 | (1.01–1.31) | 1.04 | (0.86–1.26) |
| 2 day average (Lag0 – Lag1) | 1.10 | (0.97–1.25) | 1.11 | (0.92–1.35) |
| 3 day average (Lag0 – Lag2) | 1.08 | (0.96–1.22) | 1.13 | (0.94–1.35) |
| 4 day average (Lag0 – Lag3) | 1.08 | (0.95–1.23) | 1.07 | (0.90–1.27) |
| 5 day average (Lag0 – Lag4) | 1.10 | (0.97–1.25) | 1.00 | (0.84–1.19) |
| 6 day average (Lag0 – Lag5) | 1.11 | (0.98–1.27) | 0.96 | (0.80–1.15) |
| All-Cause Mortality | ||||
| Same day | 0.93 | (0.80–1.10) | 0.92 | (0.71–1.20) |
| 2 day average (Lag0 – Lag1) | 1.00 | (0.85–1.17) | 0.88 | (0.67–1.16) |
| 3 day average (Lag0 – Lag2) | 1.03 | (0.88–1.20) | 0.88 | (0.68–1.13) |
| 4 day average (Lag0 – Lag3) | 1.07 | (0.92–1.25) | 0.85 | (0.66–1.08) |
| 5 day average (Lag0 – Lag4) | 1.12 | (0.95–1.32) | 0.84 | (0.66–1.07) |
| 6 day average (Lag0 – Lag5) | 1.16 | (0.98–1.37) | 0.88 | (0.68–1.12) |
January, February, March, November and December
April, May, June, July, August, September, October
Hazard Ratio for an IQR decrease in apparent temperature. The IQR for cold months was 6.6 °C. Models adjusted for day of week, PM2.5, and absolute humidity
Hazard Ratio for an IQR increase in apparent temperature. The IQR for warm months was 12.3 °C. Models adjusted for day of week, PM2.5, ozone, and absolute humidity
Table 4. Occurrence of acute MI and All-Cause Mortality with Extremea Temperature.
| Cold Months b | Warm Months c | |||
|---|---|---|---|---|
| HRd | (95 % CI) | HRe | (95 % CI) | |
| Occurrence of acute MI | ||||
| Same day | 1.23 | (0.98–1.53) | 1.07 | (0.86–1.33) |
| 2 day average (Lag0 – Lag1) | 1.36 | (1.07–1.74) | 0.93 | (0.73–1.19) |
| 3 day average (Lag0 – Lag2) | 1.10 | (0.83–1.46) | 1.11 | (0.86–1.43) |
| 4 day average (Lag0 – Lag3) | 1.03 | (0.77–1.36) | 1.03 | (0.78–1.36) |
| 5 day average (Lag0 – Lag4) | 1.01 | (0.74–1.37) | 1.20 | (0.89–1.62) |
| 6 day average (Lag0 – Lag5) | 1.05 | (0.77–1.44) | 1.10 | (0.78–1.54) |
| All-Cause Mortality | ||||
| Same day | 1.01 | (0.74–1.39) | 1.29 | (0.96–1.72) |
| 2 day average (Lag0 – Lag1) | 0.97 | (0.68–1.37) | 1.44 | (1.06–1.96) |
| 3 day average (Lag0 – Lag2) | 1.11 | (0.78–1.58) | 1.34 | (0.97–1.86) |
| 4 day average (Lag0 – Lag3) | 1.20 | (0.84–1.72) | 1.41 | (1.00–1.98) |
| 5 day average (Lag0 – Lag4) | 1.19 | (0.82–1.72) | 1.22 | (0.81–1.82) |
| 6 day average (Lag0 – Lag5) | 1.38 | (0.95–1.99) | 1.31 | (0.85–2.01) |
Extreme temperature days were those when the apparent temperature for that averaging periods was in the upper 5th percentile of all temperatures during the warm season or in the lower 5th percentile of all temperatures during the cold season.
January, February, March, November and December
April, May, June, July, August, September, October
Hazard Ratio comparing days when temperatures were in the lowest 5th percentile to other days. Models adjusted for day of week, PM2.5, and absolute humidity.
Hazard Ratio comparing days when temperatures were in the highest 5th percentile to other days. Models adjusted for day of week, PM2.5, ozone and absolute humidity.
All-Cause Mortality
We found no associations with continuous apparent temperature in either cold or warm months for subsequent mortality in people who had previously had an acute MI (Table 3). However, we found that extreme hot temperatures in the 2- and 4-day moving averages preceding death were associated with mortality (Table 4) in this population of patients who had previously had an acute MI. We also found a marginally increased risk for mortality during extreme cold days, particularly during longer averaging periods (6-day moving average). Our results were not materially different when we used day of the week to select control days.
Effect Modification
We evaluated effect modification for all of the individual- and area-level characteristics described in Table 1 (eTables 1 – 4). Persons who had a prior acute MI were more susceptible to the effects of decreases in apparent temperature during cold months (hazard ratio [HR] = 1.46 [95% confidence interval (CI)= 1.14–1.87]) compared with those who had an initial acute MI (1.04 [0.89–1.22]); test for interaction, P = 0.003. We also found that subjects who had a large (> 100,000 m2) lake or reservoir within a 400 m radius of their home were less susceptible to the effects of a decrease in temperature (0.90 [0.63–1.28]) than those who did not (1.20 [1.04–1.39]); P = 0.02. While we did not find any association between increases in temperature or extreme heat and acute MI in our main analysis, we did find that certain populations were more susceptible to the occurrence of acute MI from extreme heat. Persons living in census blocks with at least 14 % of families below the poverty line (the median proportion in our cohort) had a stronger association between extreme heat and acute MI (1.39 [0.90–2.14]) than those living in census blocks with less than 14 % of families below the poverty line (0.77 [0.57–1.04]); test for interaction, P = 0.01. Additionally, those living in more urban areas (census tracts with a housing density greater than or equal to 1000 units/km2) were more likely to suffer from an acute MI on extremely hot days (HR = 1.48 [95% CI= 0.88–2.49]) than those living in less dense census tracts (0.81 [0.61–1.08]); test for interaction, P = 0.02.
When examining vulnerability to all-cause mortality in these acute MI survivors, we found that younger patients (less than 65 years old) were more likely to die (1.32 [0.65–2.68]) on days with higher temperatures, compared with older patients (0.86 [0.64–1.14]); test for interaction, P = 0.02. We also found that patients who had a Q-Wave acute MI were more likely to die on hotter days (1.61 [0.92–2.82]) than those who had had a non-Q-wave acute MI (0.79 [0.58–1.06]); test for interaction, P = 0.02. Consistent with our results in the acute MI occurrence analysis, we found that those living in census blocks with above-median percent of families below the poverty line had a stronger association between heat and mortality (1.22 [0.74–2.01]) than those living in census blocks with fewer families below the poverty line (0.79 [0.57–1.08]); P = 0.03. Finally, we found that people without previously diagnosed heart failure were more likely to die after periods of extreme heat (2.15 [1.41–3.26]) compared with those who had a history of heart failure (0.92 [0.58–1.47]); P = 0.01.
When we examined models that simultaneously included continuous change in apparent temperature and a binary variable for extreme temperature (Table S.5), we observed similar results, with some reduction in power for the continuous exposures. When using daily mean temperature as an exposure metric (Table S.6), we found no associations with acute MI occurrence and mortality.
Discussion
We found that extreme cold and decreases in apparent temperature increased the risk of acute MI. We did not find any effects of heat on the risk of acute MI in the population as a whole, although certain susceptible groups had an increased risk of acute MI with heat. We also found that extreme heat was associated with an increased risk of dying in people with a prior acute MI.
The published literature on temperature and the development of MI is not as comprehensive as that on temperature and mortality. Findings from time-series studies have been inconsistent, with some reporting effects for cold, some for heat, and some for both.7 These inconsistencies may be due to a number of factors including methodology and study-population differences. Few studies adjusted for air pollution, which can vary with season and temperature and has been associated with the development of acute coronary disease. Further, a number of studies used MI-related mortality, which does not capture non-fatal events of MI and also may have poor specificity when the diagnosis is assumed. A recent systematic review and meta-analysis on ambient temperature and cardiorespiratory morbidity 20 found no effect of increasing temperature on cardiovascular morbidity (relative risk=0.999 [95% posterior interval: 0.982–1.016] per 1°C increase in temperature). This is consistent with our finding of no effect of heat on acute MI in the overall study population. This same meta-analysis concluded that there were too few studies on effects of cold temperatures to draw conclusions, although a number of studies that examined validated MI as an outcome 21-23 found an association between cold weather and MI. A study of hospital admissions in England and Wales 24 found that a 1° C decrease in daily mean temperature was associated with a 2% increase in risk of MI over the current and following 28 days. Another population-based registry in Augsburg, Germany 25 found that a 10°C decrease in 5-day average temperature was associated with a relative risk of MI of 1.10 (95% confidence interval= 1.04–1.15), similar in magnitude to our results. These studies adjusted for both air pollution and level of influenza and also validated most cases of MI.
Both high and low temperatures have been associated with all-cause mortality in a wide range of populations.26 While some studies have found evidence for only a heat effect,2 a comprehensive analysis of the effects of heat and cold on mortality found heat-related mortality for shorter lags and cold-related mortality for longer lags.27 Studies in the U.S. have found evidence for an approximately 2% increase in mortality associated with a 10°F increase in apparent temperature.28 Additionally, the effects of temperature have been found to vary by latitude, with northern cities in the US, such as Boston, experiencing more of a rise in mortality risk at higher temperatures than southern US cities.3 Our analysis of short-term effects, as well as the location of our study, may explain why we observed the strongest associations with mortality for heat and saw trends toward associations for cold in longer lags. A number of mechanisms to explain how temperature affects morbidity and mortality have been proposed, including increases in blood pressure 29 and increases in fibrinogen 30 on cold days, as well as increases in serum LDL levels with increasing ambient temperatures.31 These different mechanisms may partially explain the differences in results in morbidity and mortality.
Few studies have investigated susceptibility to the effects of temperature on acute MI occurrence. In one study of patients who had had a coronary event, there was no difference in the risk of MI due to temperature among those with and without a previous MI,32 whereas a French registry-based study 33 found that recurrent cases of MI were more susceptible to decreases in temperature, similar to our results. Danet et al.33 suggested that this subgroup might be particularly vulnerable to weather because of their chronic cardiac disease, and because analytic control of classic risk factors with secondary prevention might highlight the effort of less common risk factors such as weather.
We found that the effects of heat on both acute MI occurrence and mortality were greater in persons living in census blocks with a higher proportion of poverty. Little is known about how socioeconomic characteristics modify the effect of temperature on the occurrence of acute MI, but both individual and neighborhood characteristics related to lower socioeconomic position have been found to enhance the relationship between temperature and mortality.3,34-36 This may partially be explained by air conditioning or other adaptive measures. Consistent with our results on acute MI occurrence, previous multi-city studies have found that living in areas of higher population density enhances the temperature effect on mortality.37,38
We also found that persons living within 400 m of a large body of water were protected from the effects of decreasing temperature on MI incidence. Because it takes more energy to change the temperature of water than air, large water bodies moderate the local temperature. This finding suggests that the surrounding environment could be an important modifier in the relationship between temperature and morbidity. While we did not observe modification by other built-environment attributes (e.g. percentage recreation or conservation land), further studies should explore the potential public health impact of such modifable environmental features.
To our knowledge, no previous studies of temperature and mortality have been able to examine vulnerability based on prior type of acute MI. In general, patients with a Q-wave MI have been found to have worse in-hospital survival than patients with a non-Q-wave MI, although this difference decreases over time.39 Unexpectedly, we found that persons without previously diagnosed heart failure were more likely to die after periods of extreme heat than those who had a history of heart failure. This may be explained by the weakened condition of such patients, which may limit their mobility and confine them to indoors. A similar phenomenon was noted in a study of sudden cardiac death and seasonality among patients with and without prior CHD.40 However, our explanation is speculative and in contrast to an increase in heat-related in-hospital mortality among those hospitalized for heart failure.41 We also did not expect to find that younger persons would be more susceptible to the effects of heat, as this is in contrast with previous studies that have found increased age enhances vulnerability to heat.9,28,38,42 Our mortality analysis is limited to a population who has previously suffered from an acute MI, and therefore may represent a particularly vulnerable population.
Our study has several limitations. Ambient temperature was measured at only one monitoring station, which may have led to exposure error due to variability at individual residences and the amount of time spent indoors. However, the temporal correlation in ambient temperature is very high over wide spatial ranges in New England.31 A recent study that used both airport monitoring data and a spatial model to predict temperature at residence found little difference in the results.43 Given that temporal fluctuations in temperature are similar across this wide area, we do not expect error in our exposure metric to result in a large bias in this study of acute effects. Nonetheless, the inability to account for time spent in an environment with air conditioning or heating, which modify the relationship between temperature and mortality,27,44 is a limitation of our study. We also note that micro-scale climatic differences may explain some of the results of our effect-modification analysis.
We used a case-crossover approach, a design that controls for time-invariant confounding.45 We also controlled, through modeling, for such important time-varying confounders as absolute humidity and exposure to ozone and PM2.5. Because daily measurements of ozone and PM2.5 were not available for the Worcester metropolitan area, we used measurements from a monitoring site in Boston, MA, approximately 70 km away. Though concentrations of some pollutants have been shown to be homogeneous over a wide geographic region 46 and the ability to control for these pollutants to some degree is a strength of our study, this may still be a source of residual confounding. Seasonal infection may be a source of unmeasured confounding in the relationship between temperature and development of acute MI during winter. Also, our mortality analysis was performed on a cohort of patients who had already suffered from an acute MI and may be particularly vulnerable. Therefore, our mortality results are not necessarily generalizable the general population.
Few studies have examined the short-term effect of hot and cold temperature on the occurrence of acute MI, and no other study, to our knowledge, has been able to examine such a broad set of individual- and area-level modifying characteristics. We found associations between cold and acute MI incidence, as well as between heat and subsequent all-cause mortality. We also found that certain individual and area-level factors modified these associations. These findings point to the importance of accounting for sociodemographic vulnerability as cities and towns prepare to adapt to weather fluctuations that result from climate change.
Supplementary Material
Acknowledgments
Funding source: the National Institutes of Health (RO1 HL35434, T32ES07069, and T32ES016645), and USEPA grant RD 83479801.
We thank the participating hospitals for their cooperation in this project.
Footnotes
Conflicts of Interest: The authors declare no conflicts of interest. Stephanie von Klot is currently employed by Boehringer Ingelheim GmbH, which did not contribute any direct or indirect financing to this study.
SDC Supplemental digital content is available through direct URL citations in the HTML and PDF versions of this article (www.epidem.com). This content is not peer-reviewed or copy-edited; it is the sole responsibility of the author.
References
- 1.Meehl GA, Tebaldi C. More intense, more frequent, and longer lasting heat waves in the 21st century. Science. 2004;305:994–997. doi: 10.1126/science.1098704. [DOI] [PubMed] [Google Scholar]
- 2.Zanobetti A, Schwartz J. Temperature and mortality in nine US cities. Epidemiology. 2008;19:563–570. doi: 10.1097/EDE.0b013e31816d652d. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Curriero FC, Heiner KS, Samet JM, Zeger SL, Strug L, Patz JA. Temperature and mortality in 11 cities of the eastern United States. Am J Epidemiol. 2002;155:80–7. doi: 10.1093/aje/155.1.80. [DOI] [PubMed] [Google Scholar]
- 4.Braga AL, Zanobetti A, Schwartz J. The time course of weather-related deaths. Epidemiology. 2001;12:662–667. doi: 10.1097/00001648-200111000-00014. [DOI] [PubMed] [Google Scholar]
- 5.Armstrong B. Models for the relationship between ambient temperature and daily mortality. Epidemiology. 2006;17:624–631. doi: 10.1097/01.ede.0000239732.50999.8f. [DOI] [PubMed] [Google Scholar]
- 6.Medina-Ramon M, Zanobetti A, Cavanagh DP, Schwartz J. Extreme temperatures and mortality: assessing effect modification by personal characteristics and specific cause of death in a multi-city case-only analysis. Environ Health Perspect. 2006;114:1331–1336. doi: 10.1289/ehp.9074. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Bhaskaran K, Hajat S, Haines A, Herrett E, Wilkinson P, Smeeth L. Effects of ambient temperature on the incidence of myocardial infarction. Heart. 2009;95:1760–1769. doi: 10.1136/hrt.2009.175000. [DOI] [PubMed] [Google Scholar]
- 8.Portier CJ, Thigpen Tart K, Carter SR, Dilworth CH, Grambsch AE, Gohlke J, et al. A Human Health Perspective On Climate Change: A Report Outlining the Research Needs on the Human Health Effects of Climate Change. 2010 [Google Scholar]
- 9.Schwartz J. Who is sensitive to extremes of temperature?: A case-only analysis. Epidemiology. 2005;16:67–72. doi: 10.1097/01.ede.0000147114.25957.71. [DOI] [PubMed] [Google Scholar]
- 10.Goldberg RJ, Gore JM, Alpert JS, Dalen JE. Incidence and case fatality rates of acute myocardial infarction (1975-1984): the Worcester Heart Attack Study. Am Heart J. 1988;115:761–767. doi: 10.1016/0002-8703(88)90876-9. [DOI] [PubMed] [Google Scholar]
- 11.Goldberg RJ, Yarzebski J, Lessard D, Gore JM. A two-decades (1975 to 1995) long experience in the incidence, in-hospital and long-term case-fatality rates of acute myocardial infarction: a community-wide perspective. J Am Coll Cardiol. 1999;33:1533–9. doi: 10.1016/s0735-1097(99)00040-6. [DOI] [PubMed] [Google Scholar]
- 12.Floyd KC, Yarzebski J, Spencer FA, et al. A 30-year perspective (1975-2005) into the changing landscape of patients hospitalized with initial acute myocardial infarction: Worcester Heart Attack Study. Circ Cardiovasc Qual Outcomes. 2009;2:88–95. doi: 10.1161/CIRCOUTCOMES.108.811828. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Kovats RS, Hajat S. Heat stress and public health: a critical review. Annu Rev Public Health. 2008;29:41–55. doi: 10.1146/annurev.publhealth.29.020907.090843. [DOI] [PubMed] [Google Scholar]
- 14.Michelozzi P, Accetta G, De Sario M, et al. High temperature and hospitalizations for cardiovascular and respiratory causes in 12 European cities. Am J Respir Crit Care Med. 2009;179:383–389. doi: 10.1164/rccm.200802-217OC. [DOI] [PubMed] [Google Scholar]
- 15.Krieger N, Chen JT, Waterman PD, Soobader MJ, Subramanian SV, Carson R. Geocoding and monitoring of US socioeconomic inequalities in mortality and cancer incidence: does the choice of area-based measure and geographic level matter?: the Public Health Disparities Geocoding Project. Am J Epidemiol. 2002;156:471–482. doi: 10.1093/aje/kwf068. [DOI] [PubMed] [Google Scholar]
- 16.Commonwealth of Masachusetts Information Technology Division. Office of Geographic Information (MassGIS) Available at: http://www.mass.gov/mgis/
- 17.Bell JF, Wilson JS, Liu GC. Neighborhood greenness and 2-year changes in body mass index of children and youth. Am J Prev Med. 2008;35:547–553. doi: 10.1016/j.amepre.2008.07.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Bateson TF, Schwartz J. Control for seasonal variation and time trend in case-crossover studies of acute effects of environmental exposures. Epidemiology. 1999;10:539–544. [PubMed] [Google Scholar]
- 19.Lumley T, Levy D. Bias in the case - crossover design: implications for studies of air pollution. Environmetrics. 2000;11:689–704. [Google Scholar]
- 20.Turner LR, Barnett AG, Connell D, Tong S. Ambient temperature and cardiorespiratory morbidity: a systematic review and meta-analysis. Epidemiology. 2012;23:594–606. doi: 10.1097/EDE.0b013e3182572795. [DOI] [PubMed] [Google Scholar]
- 21.Danet S, Richard F, Montaye M, et al. Unhealthy effects of atmospheric temperature and pressure on the occurrence of myocardial infarction and coronary deaths. A 10-year survey: the Lille-World Health Organization MONICA project (Monitoring trends and determinants in cardiovascular disease) Circulation. 1999;100:E1–7. doi: 10.1161/01.cir.100.1.e1. [DOI] [PubMed] [Google Scholar]
- 22.Enquselassie F, Dobson AJ, Alexander HM, Steele PL. Seasons, temperature and coronary disease. Int J Epidemiol. 1993;22:632–636. doi: 10.1093/ije/22.4.632. [DOI] [PubMed] [Google Scholar]
- 23.Wang H, Matsumura M, Kakehashi M, Eboshida A. Effects of atmospheric temperature and pressure on the occurrence of acute myocardial infarction in Hiroshima City, Japan. Hiroshima J Med Sci. 2006;55:45–51. [PubMed] [Google Scholar]
- 24.Bhaskaran K, Hajat S, Haines A, Herrett E, Wilkinson P, Smeeth L. Short term effects of temperature on risk of myocardial infarction in England and Wales: time series regression analysis of the Myocardial Ischaemia National Audit Project (MINAP) registry. BMJ. 2010;341:c3823. doi: 10.1136/bmj.c3823. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Wolf K, Schneider A, Breitner S, et al. Air temperature and the occurrence of myocardial infarction in Augsburg, Germany. Circulation. 2009;120:735–742. doi: 10.1161/CIRCULATIONAHA.108.815860. [DOI] [PubMed] [Google Scholar]
- 26.McMichael AJ, Wilkinson P, Kovats RS, et al. International study of temperature, heat and urban mortality: the 'ISOTHURM' project. Int J Epidemiol. 2008;37:1121–31. doi: 10.1093/ije/dyn086. [DOI] [PubMed] [Google Scholar]
- 27.Anderson BG, Bell ML. Weather-related mortality: how heat, cold, and heat waves affect mortality in the United States. Epidemiology. 2009;20:205–13. doi: 10.1097/EDE.0b013e318190ee08. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Basu R. High ambient temperature and mortality: a review of epidemiologic studies from 2001 to 2008. Environ Health. 2009;8:40. doi: 10.1186/1476-069X-8-40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Modesti PA, Morabito M, Bertolozzi I, et al. Weather-related changes in 24-hour blood pressure profile: effects of age and implications for hypertension management. Hypertension. 2006;47:155–161. doi: 10.1161/01.HYP.0000199192.17126.d4. [DOI] [PubMed] [Google Scholar]
- 30.Wilmshurst P. Temperature and cardiovascular mortality. BMJ. 1994;309:1029–1030. doi: 10.1136/bmj.309.6961.1029. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Halonen JI, Zanobetti A, Sparrow D, Vokonas PS, Schwartz J. Outdoor temperature is associated with serum HDL and LDL. Environ Res. 2010 doi: 10.1016/j.envres.2010.12.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Barnett AG, Dobson AJ, McElduff P, et al. Cold periods and coronary events: an analysis of populations worldwide. J Epidemiol Community Health. 2005;59:551–557. doi: 10.1136/jech.2004.028514. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Danet S, Richard F, Montaye M, et al. Unhealthy effects of atmospheric temperature and pressure on the occurrence of myocardial infarction and coronary deaths. A 10-year survey: the Lille-World Health Organization MONICA project (Monitoring trends and determinants in cardiovascular disease) Circulation. 1999;100:E1–7. doi: 10.1161/01.cir.100.1.e1. [DOI] [PubMed] [Google Scholar]
- 34.Stafoggia M, Forastiere F, Agostini D, et al. Vulnerability to heat-related mortality: a multicity, population-based, case-crossover analysis. Epidemiology. 2006;17:315–323. doi: 10.1097/01.ede.0000208477.36665.34. [DOI] [PubMed] [Google Scholar]
- 35.O'Neill MS, Zanobetti A, Schwartz J. Modifiers of the temperature and mortality association in seven US cities. Am J Epidemiol. 2003;157:1074–1082. doi: 10.1093/aje/kwg096. [DOI] [PubMed] [Google Scholar]
- 36.Uejio CK, Wilhelmi OV, Golden JS, Mills DM, Gulino SP, Samenow JP. Intra-urban societal vulnerability to extreme heat: the role of heat exposure and the built environment, socioeconomics, and neighborhood stability. Health Place. 2011;17:498–507. doi: 10.1016/j.healthplace.2010.12.005. [DOI] [PubMed] [Google Scholar]
- 37.Medina-Ramon M, Schwartz J. Temperature, Temperature Extremes, and Mortality: A Study of Acclimatization and Effect Modification in 50 United States Cities. Occup Environ Med. 2007 doi: 10.1136/oem.2007.033175. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Hajat S, Kosatky T. Heat-related mortality: a review and exploration of heterogeneity. J Epidemiol Community Health. 2010;64:753–760. doi: 10.1136/jech.2009.087999. [DOI] [PubMed] [Google Scholar]
- 39.Behar S, Haim M, Hod H, et al. Long-term prognosis of patients after a Q wave compared with a non-Q wave first acute myocardial infarction. Data from the SPRINT Registry. Eur Heart J. 1996;17:1532–1537. doi: 10.1093/oxfordjournals.eurheartj.a014717. [DOI] [PubMed] [Google Scholar]
- 40.Gerber Y, Jacobsen SJ, Killian JM, Weston SA, Roger VL. Seasonality and daily weather conditions in relation to myocardial infarction and sudden cardiac death in Olmsted County, Minnesota, 1979 to 2002. J Am Coll Cardiol. 2006;48:287–92. doi: 10.1016/j.jacc.2006.02.065. [DOI] [PubMed] [Google Scholar]
- 41.Stafoggia M, Forastiere F, Agostini D, et al. Factors affecting in-hospital heat-related mortality: a multi-city case-crossover analysis. J Epidemiol Community Health. 2008;62:209–215. doi: 10.1136/jech.2007.060715. [DOI] [PubMed] [Google Scholar]
- 42.Baccini M, Biggeri A, Accetta G, et al. Heat effects on mortality in 15 European cities. Epidemiology. 2008;19:711–719. doi: 10.1097/EDE.0b013e318176bfcd. [DOI] [PubMed] [Google Scholar]
- 43.von Klot S, Paciorek C, Melly S, Coull B, Dutton J, Schwartz J. Association of Temperature at Residence Vs Central Site Temperature with Mortality in Eastern Massachusetts-A Case Crossover Analysis. Epidemiology. 2009;20:S75–S75. [Google Scholar]
- 44.O'Neill MS, Zanobetti A, Schwartz J. Disparities by race in heat-related mortality in four US cities: the role of air conditioning prevalence. J Urban Health. 2005;82:191–197. doi: 10.1093/jurban/jti043. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Maclure M. The case-crossover design: a method for studying transient effects on the risk of acute events. Am J Epidemiol. 1991;133:144–153. doi: 10.1093/oxfordjournals.aje.a115853. [DOI] [PubMed] [Google Scholar]
- 46.Suh HH, Nishioka Y, Allen GA, Koutrakis P, Burton RM. The metropolitan acid aerosol characterization study: results from the summer 1994 Washington, D.C. field study. Environ Health Perspect. 1997;105:826–834. doi: 10.1289/ehp.97105826. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.