Morbidity and mortality experience in St. Louis dw-ing the heat wave ofJuly, 1966, present problems pertinent to general health care and tohealth care in enclosed spaces such as shelters.

AN ANALYSIS OF THE HEAT DEATHS IN ST. LOUIS

DURING JULY, 1966

Austin Henschel, Ph.D.; Linda L. Burton, M.D.; Leonard Margolies; and J. Earl Smith, M.D.,F.A.PJL.A.

Introduction Methods and Materials

XPOSURE to a combination of high en-Lvironmental temperature and physical

work that exceed the levels duringthe preceding few days has been con-

sidered the precipitating factor in heatstroke in many instances.9-12,13,15-18 Ithas also been observed that deaths fromall causes increased during heat waves.1,2,4,5,10,11,14

From these two series of reports it isconcluded that in certain situations en-

vironmental heat can be a primarycause of death, but in others it may beonly a contributing factor. The extentto which heat factors are responsible forthe increased death rate associated withnaturally occurring summer heat waves

has not been fully documented.The present study was undertaken as

part of Work Order OCS-PS-64-126 be-tween the Office of Civil Defense andthe Division of Health Mobilization,U.S. Public Health Service. The objec-tive of the study was to evaluate the ef-fects of excessive environmental condi-tions upon deviations from normal ex-

perience of mortality and morbidity inwhich heat was an identifiable primaryor contributing factor. Data are primar-ily from a study of the role of heat inthe increased death rate during heatwaves of July, 1966, in St. Louis, Mo.

The main sources of data were thedeath certificates and hospital recordsthat were made available to us for re-view, and the July, 1966, Local Cli-matological Data compilation for Lam-bert Field in St. Louis which was ob-tained from the National Weather Rec-ords Center in Asheville, N. C. The perti-nent items of information were ab-stracted from the records and enteredon punched cards for computer analysis.In addition, on a map of the greater St.Louis area, the home address of eachheat death was identified to show theresidence-area distribution of heat-deathcases.

Results and DiscussionThe weather data for the periods be-

fore, during, and after the heat wavein July, 1966, are presented in Figure1 and Table 1. The data were abstractedfrom the official Weather Bureau recordsfor Lambert Field, the St. Louis Munic-ipal Airport located a few miles north-west of the central core area of St.Louis. The maximum and minimumtemperatures were probably higher inthe high-density, central-core area thanin the outlying suburbs where LambertField is located. Within the core area

a horizontal and vertical variation in

2232 VOL. 59. NO. 12. A.J.P.H.

HEAT DEATHS IN ST. LOUIS

temperatures would also be expected.Consequently, actual temperatures inmany places may have been 2 to 5 de-grees higher than those recorded by theWeather Bureau station at LambertField.The temperatures each day were

above 900 F from June 22 until July20 except for July 7. During this periodthe maximum temperatures exceededthe normal for the date by 2 to 17 de-grees. Maximum temperatures were con-sistently in the low to middle 90s, 2 to7 degrees above normal, until July 9.For the next six days the maximumtemperatures were above 1000 F (1010-1060 F) and were about 10 degrees

higher than during the preceding twoweeks of hot weather and 12 to 17 de-grees higher than normal. After July14 temperatures were lower again andmore nearly approximated the condi-tions that preceded the six-day heatwave.The minimum temperatures for the

nights were also 10 to 15 degrees higherduring the July 9-14 heat wave thaneither before or after. The higher mini-mum temperatures would slow the rateof night-time cooling of the buildingsand would, in effect, increase the totaldaily heat load. In addition, the highnight temperatures would interfere withproper sleep, rest, and recuperation.

Figure 1-Daily temperature in °F at Lambert Field, St. Louis, Mo., during July, 1966

105

95- .%

5,-

I II'

- / %% *j I SI U . I% - I S a

,,,..-5, . I

I S SI I .S I

5' A

i I I I '1 I I I I ILEGEND

. .MAXIMUMMINIMUMAVERAGE

- DEW POINT

1.

.. 1... : 1. .:I

.

850

z

4

0-

Lii

I. ;

I IgI I~~II. , "'1I '''

I I AI I%

% I %~~~~~~~~II I %I.75 1-l

65 F-

i

i

l li

I I I I

55

l l l lI I I I III I

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 27 28 29 30 31

DAY OF MONTH

DECEMBER, 1969 2233

Table 1-Daily temperatures in °F at Lambert Field, St. Louis, duringJune-August, 1966

Temperature, °FDate

June 24252627282930

July 12345678

91011121314

1516171819202122232425262728293031

Maximum

939594949192949494939599978596

101104106105106105

90909410292848589918986949494888489

Minimum

677373727269707371727375736964

787779848280

7472697874635857697469757775686261

Average

808484838281828483838487857780

909193959493

8281829083747273808278858685787375

+ Normal

37765

45755697

-12

121315171615

423114

-5-7-5240

7870

-5-3

Averagedew point

596064676564666968676868676265

696868686766

7065646973564748587069686872695855

We know that dew point temperaturesexpress the amount of moisture in theair and are directly relatable to the va-

por pressure of the moisture in the air.We also know that the amount of mois-ture that can be evaporated from theskin to keep the body from overheatingis determined by the vapor pressure ofthe air (and can be readily calculatedfrom dew point). However, relative hu-

midity is not a good indicator of theease with which sweat can be evapo-rated. Even though humidity was mod-erately high for the six-week periodshown in Table 1, the moisture condi-tions of the air that existed in St. Louisbefore, during, and after the heat wave

should not have been a limiting factorin sweat evaporation.

Calculations employing various heat-

2234 VOL. 59. NO. 12. A.J.P.H.

HEAT DEATHS IN ST. LOUIS

stress indexes show that the total heatload was high during the heat wave.The corrected effective temperature(ET) was between 830 and 850 F whenbased on the maximum daily tempera-tures and between 80° and 810 F whenbased on daily mean temperatures. Ac-cording to the relative strain index(RSI),8 the conditions during the heatwave would be equal to an RSI of about0.4 to 0.5. At that level of heat stress,nonacclimatized, middle-aged or olderpersons would be expected to exhibitdistress and progressive inability tomaintain normal thermal balance. Theless heat-tolerant individuals would be-come heat casualties and, if relief werenot forthcoming, some deaths would beexpected.

Increase in Death Rates

A total of 1,428 death certificates forJuly were analyzed. Included were cer-tificates for all deaths that occurredfrom July 1 through July 25. For theperiod of July 26-29, some of the deathcertificates were filed later and were notincluded on the original microfilm rec-ord used in making the study.The distribution of deaths from all

causes by sex and race is shown inTable 2. For total deaths, there wereequal numbers of males and females andthe proportion of Negro to Caucasian ineach sex group was about the same.

Daily deaths by age groups for themonth of July are given in Table 3. Asexpected, the total deaths increased withage; about half of the deaths occurredin the over-69-year age group. Duringthe days of the heat wave, July 9-14,there was some increase in the numberof deaths in each age group except the1-19-year group. The increase was pro-gressively more pronounced with in-creasing age. The average daily deathrate was about the same for the dayspreceding as for those following theheat wave. There was not a lower-than-normal death rate after the heat wavewhich could compensate for the highrate during the heat. It did not appear,therefore, that only terminal cases wereaffected by the heat, and that all thatthe heat wave did was to hasten thedeaths of those who were destined to diewithin a few days anyway.

Heat was listed as the primary causeof death on 246 of the 1,428 death cer-tificates analyzed. On an additional 40,heat was listed as a contributing factor.The daily totals of deaths from all causesand from heat for July, 1966, are shownin Figure 2. The expected number ofdeaths in St. Louis based on the preced-ing five-year average is about 35 a dayduring July. The observed average dailynumber of deaths for the first nine daysin July, 1966, was 41. By most stand-ards, the weather preceding the heatwave would be considered hot. The

Table 2-St. Louis deaths for July, 1966

All causes Due to heatMale Female Total Male Female Total

Caucasian 525 514 1,039 69 87 156

Negro 186 201 387 32 56 88

Indian 1 1 1 1

Unknowvn 1 1 1 1

Totals 711 717 1,428 101 145 246

DECEMBER, 1969 2235

Table 3-Total daily deaths by age group, St. Louis, July, 1966

July Under 1 yr 1-19 20-39 40-49 50-59 60-69 Over 69

Preheat wave12345678

1 0 3 6 7 8 184 0 1 1 4 14 114 1 0 1 2 10 292 3 1 3 2 2 251 1 0 2 8 14 261 1 0 3 3 10 210 0 3 3 5 8 191 1 1 0 6 10 11

Heat wave91011121314

Post-heat wave151617181920212223242526272829

Totals

1 0 2 4 4 6 271 2 2 1 6 10 272 0 0 3 10 19 434 1 2 5 12 18 536 3 4 7 17 31 840 0 2 9 14 23 78

3 1 3 5 14 223 0 1 1 4 111 2 4 3 10 92 0 1 0 4 71 1 0 1 4 100 0 2 1 3 51 0 3 2 5 61 0 3 2 6 70 0 4 3 0 72 3 1 3 4 30 1 2 3 9 50 0 2 4 3 62 0 1 1 2 82 0 1 0 4 40 0 0 1 1 3

46 21 49 78 173 296

49292425191612221326228

12115

765

above average death rate observed dur-ing these first days of July may reflectthe higher than normal temperatures.A striking increase in daily deaths oc-

curred on July 11, 12, 13, 14, and 15.The beginning of the increase in deathrate followed the start of the heat wave

by a day or two and returned to normalon the second day after the heat wave

broke. The delay of one to two days be-tween the start of a heat wave and theappearance of the effects reflects the timerequired for heat to build up in thebuildings in which people live and work

and the time required for the physio-logical reserve capacities of individualsto be overwhelmed.3'7'8The first of the 246 deaths certified

as due to heat was reported on July 10,the second day of the heat wave. OnJuly 11 there were 11 certified primaryheat deaths, 35 on July 12, 73 on July13, 61 on July 14, 31 on July 15, and5 to 7 on each of the next four days(Figure 2). The remaining eight re-

ported heat deaths occurred after July19 when air temperatures were againdown to levels of the period prior to

2236 VOL. 59. NO. 12, A.J.P.H.

HEAT DEATHS IN ST. LOUIS

the heat wave. For the heat wave period,the 246 certified primary heat deathsaccounted for about 65 per cent of thetotal excess deaths for the six days fromJuly 10 to 15.The incidence of certified primary

heat deaths is presented by age and dayin Table 4. Primary heat deaths oc-curred in all age groups except the 1-19-year category. The number of heatdeaths increased with age as did thetotal number of deaths. The proportionof the total deaths that was attributableto heat did not show a consistent rela-tionship to age. For the period July10-19, 26 to 37 per cent of the total

deaths were primary heat deaths in in-dividuals 20 years of age or older.As seen in Table 2, more females than

males were primary heat deaths eventhough there was no sex difference intotal deaths for July. The proportion ofheat deaths to total deaths was alsogreater for Negroes than for Caucasians.The same sex difference in the propor-tion of Negro heat deaths was observed.For total deaths, 26 per cent of the malesand 28 per cent of the females wereNegro. Among the certified heat deaths,32 per cent of the males and 39 per centof the females were Negro. The higherproportion of Negro to Caucasian heat

Figure 2-Daily deaths in St. Louis-July, 1966180

LEGEND_- TOTALo...oHEAT150 F

120

cnI

4wa

90

60 _

30 -I-

i

A

I*. -

.

.1II t<<I no--o*

0 -

1 3 5 7 9 11 13--v- T It I I I I I I I I 1

15 17 19 21 23 25 27 29 31DAYS

DECEMBER, 1969 2237

Table 4-Heat deaths by ageSt. Louis, July, 1966

July Under 1 yr 1-19 20-39 40-49 50-59 60-69 Over 69

Preheat wave8 0 0 0 0 0 0 0

Heat wave9 0 0 0 0 0 0 010 0 0 1 0 1 0 011 0 0 0 0 3 3 512 1 0 0 2 6 8 1813 1 0 1 3 9 23 3614 0 0 2 5 9 12 33

Post-heat wave15 1 0 1 0 5 7 1716 0 0 0 0 1 0 617 1 0 0 0 0 2 318 0 0 0 0 0 0 719 0 0 0 0 0 1 420 0 0 0 0 0 0 021 0 0 0 0 0 0 122 0 0 0 0 1 0 123 0 0 0 0 0 0 124 0 0 0 0 0 0 325 0 0 0 0 0 0 126 0 0 0 0 0 0 027 0 0 0 0 0 0 028 0 0 0 0 0 0 0

Totals 4 0 5 10 35 56 136

deaths could reflect the area of the cityin which they lived and the general liv-ing conditions.

Location of Heat Casualties

The location within the greater St.Louis area where the 246 heat-death in.dividuals lived at the time they becameheat casualties is shown in Figure 3. Itis obvious that the heat deaths were notrandomly distributed throughout thearea but were clustered within the city-core area. Many factors might accountfor fewer heat deaths occurring in thesuburbs. The population density is lowerin a suburban residential area. Themore open areas with grass, trees, andso on, in the suburbs would tend toreduce the air temperature and ac-celerate radiative heat loss, particularly

at night. It is also likely that more air-conditioning is in use in the suburbanhomes.The population estimate by census

tracts for July 1, 1966, for the City ofSt. Louis was 724,000: 454,000 whiteand 270,000 nonwhite. These estimatesdo not include the suburban areas wherethe population is predominantly white.The ratio of white to nonwhite was cal-culated for each of the 26 census tractsin the City of St. Louis. In 15 of thecensus tracts, 0 to 25 per cent of thepopulation was nonwhite (average2.4%). These census tracts made upmost of the southern and northern areasof the city from the river to the citylimits on the west. The two census tractswith a nonwhite population of 26 to 50per cent (average 36%) bordered onthe east of Forest Park. Six census tracts

2238 VOL. 59. NO. 12. A.J.P.H.

HEAT DEATHS IN ST. LOUIS

had a nonwhite majority of 51 to 75per cent (average 54%). In the remain-ing three census tracts, more than 75per cent of the inhabitants were non-white (92%).Of the 246 certified heat deaths, 209

occurred within the City of St. Louis.The incidence of certified heat deathsfor July 9-15, 1966, within the citylimits was 25 per 100,000 in the pre-dominantly white census tract areas, and32 per 100,000 in the nine census tractsin which more than 50 per cent of the

population was nonwhite. The rate was29 per 100,000 in the census tracts with26 to 50 per cent (average 36%) non-white. These observations tend to sub-stantiate the conclusion that the non-white population of the City of St. Louiscontributed more than its proportionateshare to the certified heat deaths. Thisdoes not constitute proof, however, thatthe nonwhite individual may be lesstolerant to heat. Overcrowding, poorerhousing conditions, lower economicstatus, and poorer general health might

Figure 3-Distribution of heat deaths, St. Louis, July, 1966

41.

DECEMBER, 1969 2239

Table 5-Daily deaths by selected disease category, St. Louis, July, 1966

Dayof mo CV Cer Pul Hep Car Hem Endo Renal Acc Heat Post op GI Inf

Preheatwave1 15 7 10 0 2 4 1 12 18 3 5 0 1 0 0 03 24 7 5 0 4 0 0 14 20 4 5 0 1 1 0 15 23 10 5 3 5 1 2 06 22 8 4 0 1 0 0 17 16 2 10 1 4 0 0 28 7 7 4 0 7 1 0 2

Post-heatwave9 18 7 4 2 6 3 0 210 19 8 4 0 9 1 1 211 28 8 12 3 7 0 2 012 27 14 9 1 2 0 0 113 35 11 12 1 7 0 1 314 42 9 5 1 6 0 0 1

Post-heatwave15 45 7 7 1 2 0 0 016 18 6 8 1 3 1 1 117 18 6 8 1 3 1 0 118 13 4 4 0 5 1 2 219 15 1 7 0 3 2 0 020 16 2 3 0 1 1 0 121 9 2 2 0 6 1 0 322 14 4 3 2 5 0 1 423 10 4 5 0 3 0 0 124 12 6 4 0 7 1 0 025 14 5 8 1 5 0 1 126 8 1 5 0 7 1 0 027 5 3 4 0 5 1 2 128 6 3 5 1 4 0 0 029 5 1 1 0 1 0 0 1

Totals 522 160 168 19 122 21 14 33

2 0 0 01 0 0 21 0 0 02 0 0 10 0 0 02 0 0 01 0 2 00 0 0 0

14421101

1 0 02 2 02 11 30 35 00 73 40 61 1

1

14

1

0

20

13430

210

37

31 07 16 17 05 0

0 0

1 02 21 03 01 00 0

0 0

0 0

0 0

246 14

0 10 10 11 30 40 0

1 20 10 0

0 0

0 1o o

2 11 1o 0

0 10 0

0 0

2 10 2o 0

10 34

also contribute to the observed differ- Symptoms of Heat Strokeences.

Daily deaths listed by major disease Hospital records were reviewed forcategories are given in Table 5. The fre- 200 of the 246 certified heat deaths dur-quency of death from cardiovascular ing the heat wave. Of these 2009 80 were

disease, from cerebral accidents, and to dead on arrival at the hospital. Almosta less degree from pulmonary disorders, all of those admitted to the hospitalwas increased during the heat wave as during the weeks of the heat wave hadcompared with the periods before and body (oral or rectal) temperatures fromafter. 1030 to 1060 F on admission. These

2240 VOL. 59. NO. 12. A.J.P.H.

HEAT DEATHS IN ST. LOUIS

would appear to be heat-stroke cases.Many persons who died had been in ahospital for months; suddenly, with highenvironmental temperatures, they de-veloped body temperatures of 1030 For above and died within a few days.

Laboratory data could be obtained ononly about 50 per cent of the deceasedpatients whose charts were reviewed. Inmost cases normal plasma electrolyte(Na, K, Cl) values were found. Hyper-kalemia has been reported as an im-portant electrolyte alteration in heatstroke. Most of the liver function testsand enzymes (SGPT, SGOT, LDH)showed abnormally high values, indi-cating definite intracellular damage.Fasting blood sugars and blood ureanitrogens were also consistently abovenormal. Many gave abnormal reactionsto blood tests, such as increased pro-thrombin times and bleeding problemsmanifested by guaiac positive stools,melena, and hemoptysis. Similar bio-chemical changes have been reportedin fatal heat-stroke cases by others.6'17

Only about one-third of the 200 certi-fied heat-death cases whose hospitalcharts were reviewed had any previoushospital admissions or available medicalhistories. Most of the others were ad-mitted either unconscious or were deadon arrival. Because of this, no informa-tion could be obtained concerning theirprevious diseases, medications taken, orgeneral state of health.For the persons with available hos-

pital records, the main diseases were:No. ofpersons

Arteriosclerotic heart disease 19Congestive heart failure 10Diabetes 7

Four patients were described as ex-tremely obese; one had multiple scle-rosis; one, epilepsy. Eight were taking adigitalis preparation and seven were ondiuretics. However, many of these per-sons had more than one disease, andover one-third of those with medicalhistories were over 65 years of age andhad been chronically ill for years.One interesting feature was that one

hospital was divided into three mainunits: one was air-conditioned, and theother two were without air-conditioning.(Each patient was confined to one ofthe three units.) All 35 heat deaths inthe hospital occurred in the two unitswithout air-conditioning.

ConclusionsThe morbidity and mortality experi-

ences in St. Louis during the July, 1966,heat wave illustrate problems that arepertinent to general health care and tocivil defense shelter management. In 12to 36 hours, many individuals becomeheat casualties if the temperature withinthe shelter rises to about 950 F or above.The older and middle-aged groups areaffected more than younger people. Par-ticularly susceptible are those with acuteor chronic cardiovascular and pulmonarydisorders. Excessive heat can be a pri-mary cause of death from typical heatstroke or a contributing cause in al-ready ill patients. The high environmen-tal temperatures in shelters will magnifythe problems of health support and ofshelter management. Of particular con-cern will be the hospitalization andmedical care for the load of heat casual-ties that will develop.

REFERENCESHypertensionSyphilisCerebral vascular accidents

(embolism and thrombosis)AlcoholismEmphysemaCancer (uterus, lip, breast)

65 1. Anon. Excess Mortality Related to Heat

Wave. USPHS Weekly Rep. 15, No. 29,5 1966, pp. 245-246.

2. Anon. Heat Deaths in Two Cities Last3

July are Put at 1,751 by Epidemiologist.3 Medical News 1,22:5 (May 29), 1967.3 3. Bass, D. E.; Klieman, C. R.; Henschel,

DECEMBER. 1969 2241

A.; and Hagnauer, A. Mechanisms ofAcclimatization to Heat in Man. Medicine34:323-380, 1955.

4. Brown, E. G. Death from Excessive Heatin Kansas. Pub. Health Rep. 50, 1934,p. 546.

5. Grover, M. Mortality During Periods ofExcessive Temperature. Ibid. 53, 1938,pp. 1112-1143.

6. Knochel, J. P.; Beisel, W. R.; Herndon,E. G.; Gerard, E. S.; and Barry, K. G.The Renal, Cardiovascular, Hematologicand Serum Electrolyte Abnormalities ofHeat Stroke. Am. J. Med. 30:299-309,1961.

7. Lee, D. H. K., and Henschel, A. Effectsof Physiological and Clinical Factors inResponse to Heat. Ann. New York Acad.Sc. 134:743-749, 1966.

8. . Evaluation of Thermal En-vironment in Shelters. USPHS Div. Occup.Health Tech. Rep. 8, 1963.

9. Leithead, C. S., and Lind, A. R. HeatStress and Heat Disorders. Philadelphia:F. A. Davis, 1964.

10. Macpherson, R. K., and Offner, F. Heatand Survival of the Aged and ChronicallyIll. M. J. Australia 1:292-295, 1965.

11. Meyers, J., and Calonje, M. A. Heat Syn-drome Data from Selected Hospital Rec-ord Survey. USPHS Rep., 1965.

12. Minard, D. Prevention of Heat Casualtiesin Marine Recruits. Mil. Med. 126:261-272, 1961.

13. Schickele, E. Environment and Fatal HeatStroke. Mil. Surgeon 100:235-256, 1947.

14. Schuman, S. H.; Anderson, C. P.; andOliver, J. T. Epidemiology of SuccessiveHeat Waves in Michigan 1962 and 1963.J.A.M.A. 189:733-738, 1964.

15. Shattuck, G. C., and Hilferty, M. M. SunStroke and Allied Conditions in U. S.Am. J. Trop. Med. 12:223-245, 1932.

16. . Causes of Deaths from Heatin Massachusetts. New England J. Med.209:319-329, 1933.

17. Sohar, E. Personal communication, 1967.18. Shibolet, E.; Gilat, T.; and Sohar, E.

Physical Effort as a Main Cause of HeatStroke. UNESCO Symposium, Lucknow,India, 1962.

Dr. Henschel, Dr. Burton. and Mr. Margolies are with the Department ofHealth, Education, and Welfare, PHS, Buireau of Disease Prevention and En-vironmental Control, National Center for Urban and Industrial Health, Occupa-tional Health Program (1014 Broadway), Cincinnati, Ohio 45202. Dr. Smith isCity Health Commissioner, St. Louis, Mo.

This paper was presented before a Joint Session of the Statistics and Radi-ological Health Sections of the American Public Health Association at theNinety-Sixth Annuial Meeting in Detroit, November 14, 1968.

2242 VOL. 59. NO. 12, A.J.P.H.