Deaths in Australia from Work-Related Heat Stress, 2000-2015.

The infrequency of deaths from work-related heat stress may be due to self-pacing, whereby workers adjust their work rate in response to thermal discomfort. Thirteen cases attributed after coronial investigation to work-related heat stress were studied to evaluate the causal contribution of environmental and personal risk factors. Meteorological records and coronial records were examined to estimate environmental and metabolic heat loads and to identify any personal risk factors likely to have contributed to death. Seven deaths occurred in workers within one week of hiring, demonstrating not only the importance of acclimatisation but also the likelihood of compromised self-pacing in recently-hired workers. Personal risk factors identified included intercurrent illness, cardiovascular disease and obesity. Four deaths occurred following indoor work, where the probable critical risk parameter was low air velocity. Cerebral and pulmonary oedema were reported in some autopsy reports, and uncal herniation was found in one case. Modified work rates and close supervision are essential in recently-hired workers. The risk of death from raised intracranial pressure suggests the need for specific remediation of cerebral oedema in hyperthermic individuals.

1. Introduction

Death from occupational heat stress is not common. In the U.S., only 13 such deaths were reported for the period 2012–2013 [1]. In Australia, the National Coronial Information System has recorded only 17 cases between 2000 and 2015, even though most workers in the mining, agriculture and construction sectors work outdoors in the hotter regions of Australia [2]. The infrequency of deaths from this cause may be attributable to self-pacing, which is the natural behavioural response to heat whereby discomfort arising from a combination of metabolic and environmental heat load leads to slowing of the work rate [3,4].

This study was based on reports in which the coroner had found death to be attributable to work-related heat stress. The aim was to confirm (or not confirm) the coroners’ conclusion that death was attributable to work-related heat stress and, in those cases where heat stress was found not to be a sufficient cause, to identify other contributing factors.

2. Materials and Methods

Details of deaths from work-related heat stress were obtained from coronial records. Seven deaths were the subject of coronial inquests with reports on the public domain and were obtained from SafeWork Australia. Reports of the other deaths were obtained from State coronial offices. The study was approved by the Human Research Ethics Committee of the University of Adelaide.

For each death, weather observations were obtained for the day when exposure occurred (not always the date of death) from the nearest monitoring station of the Bureau of Meteorology (BOM). The observations include air (shade) temperature, dew point, relative humidity, air velocity, and total daily solar exposure. Information on work performed and personal information relating to the deceased worker were obtained from coronial records.

Records of each case were examined: (i) to confirm (or not confirm) the coronial finding that work-related heat stress was a necessary cause of the death, i.e. whether the death would not have occurred but for that level of heat stress; and (ii) to determine whether the level of heat stress was a sufficient cause of the death.

The main criterion to confirm heat stress as a cause of death was heat stroke, identified by a body core temperature greater than 40.0 °C. Where no body temperature was recorded, the decision on whether, on the balance of probability, heat stress was a contributory cause of death was based on computation of metabolic and environmental heat load, likely fluid deficit and maximum sweat evaporative capacity (Emax).

The decision on whether heat stress was a sufficient cause depends primarily on whether other workers performing the same work in the same environment developed heat-related illness: if not, heat stress is self-evidently not a sufficient cause. Where there were no others performing similar work, the decision was based on information from BOM data and from coronial documents on the presence of known risk factors for heat-related death or illness.

BOM data on air temperature and humidity were used to derive the Heat Index, using the calculator of the US National Weather Service https://www.wpc.ncep.noaa.gov/html/heatindex.shtml.

BOM data do not include mean radiant temperature in direct sunlight. An approximation was obtained by assuming that mean radiant temperature in the absence of cloud cover or shade was double the shade temperature, an assumption based on direct observations made by this author in previous field studies [5]. Radiant Heat Load was then estimated from equations derived experimentally by Hertig [6]. Corroboration of the level of radiant heat was obtained by BOM records, based on satellite data, of total daily solar load in kWh·m−2.

Metabolic heat load was estimated from the Table of Metabolic Rate according to activity level from the Heat Stress and Strain document of the American Conference of Governmental Industrial Hygienists (ACGIH).

Heat load and consequent fluid requirement from evaporative sweat loss were estimated from combined radiant and metabolic heat load. Convective heat load was not included because of the relatively small gradient between air and skin temperatures.

An equation by Hertig was also used to estimate maximum sweat evaporative capacity (Emax) [6]. This function is derived from air velocity, partial pressure of water vapour (PPA), and a factor for clothing. PPA was computed from dew point on a website of the U.S. National Weather Service https://www.weather.gov/epz/wxcalc_vaporpressure.

3. Results

3.1. Study Population

Altogether 13 reports were obtained for the period 2000–2015. Four more deaths occurred from work-related heat stress in one State, but the coroner’s office in that jurisdiction declined the request to supply the documentation. All cases but one (#12) were male. Ages ranged from 19 to 72 years, and six subjects were aged 25 years or less. Ten of the deaths occurred in January and one in February, i.e., at the height of the Australian summer, and two in November.

3.2. Deaths in Which Heat Stress Was a Causal Factor

Table 1 shows a summary of heat stress parameters and risk factors, and a summary of each case is given in Appendix B.

Table 1

Summary of heat stress parameters and risk factors.

Study Number	Age	Activity	Max Heat Index	Daily Solar Exposure	Mean Air Velocity	Workload	Body Core Temperature	Time Employed	Heat Sufficient Cause?	Risk Factors	Fellow Workers
1	35	Stranded in outback, walking	43	8.1	7	Moderate	NR	1 day	Y	Dehydration	None present
2	30	Construction	36	6.3	22	Heavy	NR	3 days	N	Intercurrent illness, anorexia?myocarditis	Unaffected
3	56	Door-to-door collection	36	7.0	38	Light -moderate	NR	Weeks or months	N	Piecework payment	None present
4	25	Military	37	6.9	NA	Moderate-heavy	41.7	Years	N	Hyponatremia	Unaffected
5	23	Supermarket trolley-boy	34	8.2	25	Moderate	40.5	4 days	N	Obesity, anti-psychotic medication	None present
6	38	Concrete formwork	41	8.7	17	Very heavy	37.1(oral)	Weeks or months	N	Possible steroid use	Unaffected
7	25	Stranded in outback, walking	41	7.6	37.1	Moderate	NR	1 year	Y	Dehydration	Affected
8	50	Farm work	33	7.5	9	Low	NR	6 years	N	Prolonged exposure; trauma	None present
9	19	Installing insulation	39	Indoors	Indoors	Very heavy	40.5	1 day	N	Possible steroid use	Unaffected
10	25	Ship’s engine room	*	Indoors	Indoors	Moderate-heavy	41.5	4 days	N		Unaffected
11	34	Furniture removal	38	Indoors	Indoors	Moderate	42.3	1 day	N	Antipsychotic medication	Unaffected
12	24	Fruit- picking	33	8.2	15	Moderate	NR	2 days	N	Obesity, high humidity, suspected snakebite	Unaffected
13	72	Carpet-laying	36	Indoors	Indoors	Moderate-heavy		Not known	N	Pulmonary hypertension	None present

Units: heat index and body temperature in °C, daily solar exposure in kWh·m−2, air velocity in km/h. * No BOM data available: reported air temperature 48 °C. NR: not recorded.

Heat stress was a probable contributing factor in 12 of the 13 deaths. In five cases, this conclusion was based on a diagnosis of heat stroke, i.e., body core temperature >40.0 °C. In six cases, the worker had already died when found, and no ante-mortem temperature measurement was recorded: in all six of these deaths, the heat stress parameters and the history justify the coroners’ conclusions that death resulted from heat stress. In one case (#6), several oral temperatures readings were obtained, the highest being 37.1 °C, ruling out a diagnosis of heat stroke; nevertheless, the high level of solar radiation and the heavy workload justify the conclusion that heat stress was a contributing factor to death. In one case (#11), while heat stress was a possible contributory cause of death, there was evidence of snakebite as a more likely explanation.

In the 12 cases where a causal relationship with heat stress was certain or probable, Heat Index (which reflects shade temperature and humidity) ranged from 33 °C to 43 °C. (BOM data were not available in Case #10: this was in the engine room of a ship, where the reported air temperature was 48 °C). In the deaths that occurred following outdoor work, there was a significant additional heat load from solar radiation: the daily aggregate solar exposure ranged from 6.9 to 8.7 kWh·m−2. (As a guide to intensity of solar radiation, the mean solar load in the city of Adelaide, South Australia is 7.6 kWh·m−2 in January and 2.1 kWh·m−2 in June: this difference is of course partly due to the large seasonal variation in the number of daylight hours.)

3.3. Deaths in Which Heat Stress Was a Sufficient Cause

In two cases, drivers had become bogged in remote areas and set out on foot: modelling clearly shows that both would have incurred prolonged exposure to extreme heat with severe dehydration. While in one of these cases a severely dehydrated fellow worker was found in time to be resuscitated, these circumstances are incompatible with survival in the absence of rescue.

3.4. Deaths in Which Heat Stress Was not a Sufficient Cause

In the remaining 10 cases (that is, excluding the one case not attributable to heat stress), it is probable that heat stress was not the sole explanation. In six of these cases, this judgement is strengthened by the observation that there were fellow workers who were unaffected. Known risk factors for heat-related illness were identified in all cases, including reduced capacity for self-pacing, lack of acclimatisation, obesity, intercurrent illness and cardiovascular disease. Nevertheless, a complete explanation of the cause of death is lacking for two cases (#3 and #6), and, in one case (#8), there was significant trauma, suggesting that the environmental heat load was not the major factor in the death.

Four of the deaths followed indoor work.

3.5. Risk Factors

Inability to self-pace is a variable that is not measurable a priori: nevertheless, there were circumstances in two cases where there was a factor which self-evidently gives an incentive to override this safety mechanism: in one case, the subject was undergoing military duties in simulated battle conditions, and in another a door-to-door collector for charity was paid partly according to his success.

In six deaths the worker was in the first week of the job, indicating lack of acclimatisation as a risk factor.

One subject was reported as being obese. Two subjects were receiving anti-psychotic medication. In one subject, there was strong evidence of an intercurrent illness, possible myocarditis, and, in another, previously unrecognised pulmonary hypertension was a likely causal factor. Dilutional hyponatremia was confirmed in one case by low serum sodium (and low potassium); the worker had reported drinking large volumes of fluid.

3.6. Autopsy Findings

Autopsy details were available for seven of the cases. Cerebral oedema was found in two cases, and in a further two the brain weight was above normal values (1520 g and 1465 g) [7]. Pulmonary oedema was noted in six of these seven cases.

4. Discussion

The finding that heat stress was not the sole cause of 10 of the deaths indicates the importance of identifying risk factors in any workforce subjected to hot conditions.

4.1. Self-Pacing

In at least two cases, self-pacing was compromised. The protective effect of self-pacing is most commonly overridden in sport, as evidenced by the often-reported occurrence of heat stroke in marathon runners [8]. In the occupational setting, attenuation of this protective mechanism can occur in military activity, where battle conditions or military discipline can inhibit the impulse to reduce physical workload. One such case occurred in this series, albeit in simulated battle conditions. Self-pacing may also be compromised under piecework conditions, as has been shown in studies of shearers and construction workers (although the consequences were not fatal) [3,4]. In this series, one death occurred in a collector for charity: his income was partly based on results.

4.2. Acclimatisation

A salient finding of this series was that in no fewer than six of the heat-related deaths the worker had been hired for less than a week. While this shows the importance of acclimatisation, it may also reflect compromised self-pacing: recently-hired workers may be especially motivated to push themselves beyond thermal comfort levels for fear of dismissal or of ridicule. This finding shows the need for special attention to recently-hired workers in hot weather. Expectations of work rates should be lowered during the acclimatisation period, and one-on-one supervision (the buddy system) is recommended.

4.3. Indoor Work

Four deaths occurred in the course of indoor work. Although indoor workers are spared the intense radiation load from solar exposure, they face the critical hazard of low air velocity: although the total heat load is less than with outdoor work, the capacity to dissipate body heat is reduced by the relatively still air, which reduces the evaporation of sweat. This is demonstrated in Appendix A by modelling, showing that the beneficial reduction in radiant heat load from working indoors can be offset by low air velocity, which reduces the capacity to evaporate sweat.

4.4. Intercurrent Illness

Case #2 appears to have had an intercurrent illness: some lymphocytes were present in the myocardium at autopsy, so that the worker’s condition may have been viral myocarditis. He had had almost no food or fluid intake over the four days leading to his death, and the likely dehydration, as well as the illness itself, were likely contributory factors. The lack of food intake over four days may also have reduced the worker’s heat tolerance. During work in the heat, there is normally an exaggerated rise in blood glucose, the source of which is breakdown of liver glycogen [9]. Liver glycogen levels run down quickly if carbohydrate intake is low or absent [10], thus it is quite likely that this worker’s tolerance to working in the heat was compromised by depletion of liver glycogen.

4.5. Obesity

Obesity, with accompanying lack of physical fitness, was a likely contributor to heat stroke in one case and is a well-recognised predisposing factor to heat-related disorders. In a series of fatal heat stroke cases, a 3.5-fold increase in risk was found in men 18 kg or more overweight compared with men who were 4.5 kg overweight [11]. The capacity to dissipate body heat by evaporation of sweat and convection is a function of skin surface area, whereas heat production is a function of body mass. Since obesity confers a lowered body surface area in relation to body mass, the capacity to dissipate heat is attenuated. Moreover, the extra weight carried by obese people requires more energy expenditure—hence more heat generation—to perform a given task compared with a lean individual [12,13].

4.6. Heart Disease

One subject was found at autopsy to have pulmonary hypertension with right ventricular hypertrophy. In another subject, discussed above, viral myocarditis was suspected. Cardiovascular disease has been observed to increase the risk of heat-related deaths. Cardiac output needs to increase in heat to increase perfusion of skin and increase sweat production. In heart disease, reduction in maximal oxygen uptake (VO2max) limits the capacity to increase cardiac output. During the European heat wave of 2003, daily mortality from cardiovascular disease in the German city of Essen was found to be 25% higher than the daily average over summer months [14].

4.7. Psychotropic Drugs

Two of the decedents were receiving psychotropic drugs, and in both instances the coroner gave much consideration to whether they had any causal role. Antipsychotic drugs have been associated with heat-related hospitalisation in a case-control study: the authors attributed the effect to the medication interfering with thermoregulation [15], although such dysregulation has also been attributed to mental disease itself [16]. On the other hand, convincing evidence of any effect of any individual psychotropic drug is lacking. One subject had been taking venlafaxine, clozapine and a low dose of amisulpride. The National Data of Adverse Event Notifications (DAEN) was consulted but this case was the only recorded occurrence of heat stroke or heat illness from clozapine: this subject had other risk factors. The other subject was taking quetiapine: although the coroner agreed with the forensic pathologist that this drug contributed to the worker’s death, interrogation of the DAEN identified only one notified case (non-fatal) of heat illness from quetiapine.

4.8. Hyponatremia

One subject, a soldier, was hyponatremic (#4): on admission to hospital, his serum sodium was 128 mm/L (reference range 135–145 mm/L). The serum potassium was also low (3 mm/L, reference range 3.5–5 mm/L), indicating that the low sodium level was dilutional. He had reported to the health centre in the middle of the day having drunk 8 L of fluid since 04:00. On admission to hospital that evening, his core temperature was 41.7 °C. Concurrent heat stroke and hyponatremia has previously been reported, also in a military situation [17].

Exertion-associated hyponatremia (EAH) is well recognised in marathon runners and in hikers [18] but is not common in the work environment. This is not surprising, since sweat has a lower sodium content than serum, so that uncompensated sweat loss leads to hypovolemic hypernatremia; only if the subject overcompensates by drinking excess water or hypotonic fluid can hyponatremia develop. It is commonly stated that the thirst mechanism is an inadequate driver for fluid replacement, and that drinking small amounts frequently is preferable to sole reliance on thirst [19]. On the contrary, an Exercise-associated Hyponatremia (EAH) Consensus Development Conference states “thirst should provide adequate stimulus for preventing excess dehydration and markedly reduce the risk of developing EAH in all sports” [20]. The occurrence in this series of a case of hyponatremia suggests the need for an awareness of risk from over-zealous fluid replacement.

4.9. Pulmonary and Cerebral Oedema

The findings of cerebral oedema and pulmonary oedema in heat stroke cases have been noted previously [21,22]. In a series of autopsies of military personnel who had died of heat stroke, the brain weight was usually higher than expected, often by several hundred grams, the average weight being 1493 g [23] compared with the average for normal adult males of 1336 g [7]. Since cerebral dysfunction is a constant feature of heat stroke, it is probable that cerebral oedema is the cause. In one case in this series, uncal and tonsillar herniation were found at autopsy, suggesting that the terminal event in some cases of heat stroke may be herniation into the foramen magnum from raised intracranial pressure. Consideration should therefore be given to specific remediation of cerebral oedema in hyperthermic individuals with signs of cerebral impairment.

5. Conclusions

The analysis supported the coroners’ conclusions that work-related heat stress was a causal factor in 12 of the 13 deaths. In only two cases heat stress was a sufficient cause of death, and risk factors other than heat stress were identified as potentially contributing to the other 10 deaths. The finding that 6 of the 13 deaths occurred in the worker’s first week on the job shows the need for special attention to recently-hired workers in hot weather. Expectations of work rates should be lowered during the acclimatisation period, and one-on-one supervision (the buddy system) is recommended. Other risk factors identified were compromised self-pacing, obesity, intercurrent illness and pre-existing heart disease. The autopsy finding of cerebral oedema is consistent with a large previous study. The possibility of uncal herniation as a precipitating cause of death suggests a possible benefit of measures to lower intracranial pressure in the emergency treatment of patients with heat stroke.