We conducted a pre- and posttest evaluation of aluminum potroom workers at a smelting facility in Texas, over a 4-day period in July. The study population was based on a nonrandomized convenience sample. Inclusion criteria included aluminum potroom workers identified by the employer and union as working in the hottest jobs, and who worked during the first or second shift (8 am to 4 pm or 4 pm to 12 am). An attempt was made to select an equal number of employees working 8-hour and 16-hour shifts, and those who worked in the same crew (to facilitate observation of employee activities). Participation was voluntary, and written informed consent was obtained from all participants. This evaluation was conducted under a blanket institutional review board approval for the NIOSH Health Hazard Evaluation program. In general, Health Hazard Evaluations are considered workplace evaluations and not research.

Participants completed a questionnaire about their medical history, work history, and symptoms experienced and amount of fluids consumed during the shift on which they were monitored. Symptoms of interest included muscle cramps, headache, lightheadedness or dizziness, nausea or vomiting, racing heartbeat or palpitations, unsteady walk, and confusion or disorientation. These symptoms were selected on the basis of the extant literature and common clinical presentations of heat-related illnesses. Participants were considered unacclimatized if they had been off work or worked completely outside the potrooms for 4 or more consecutive days in the past 2 weeks.²

Urine specific gravity and blood electrolytes were measured in the field immediately after collection before and after each participant’s shift. The i-STAT 1 handheld analyzer and i-STAT 8+ cartridges (Abbott Point of Care, Inc, East Windsor, NJ) were used to measure serum electrolytes. Two to three drops of whole blood was collected via finger prick and placed in the well of the cartridge before inserting the cartridge in the analyzer. The Atago Uricon-Ne Refractometer (Atago USA, Inc, Bellevue, WA)was used to measure urine specific gravity. The refractometer measures specific gravity from 1.000 to 1.050 and is accurate to 0.001. Urine was collected in sterile urine collection cups.

Shortly before their shift, participants swallowed a VitalSense (Mini-Mitter Company, Inc, Bend, OR) temperature sensor pill, and a data logger was attached to the participant waist belt. The VitalSense monitoring system continuously measured T_c at 1-minute intervals to ±0.2°F. At the end of shift, data loggers were removed, and T_c data were downloaded. Core body temperature data were included in analysis once T_c readings stabilized (ie, readings no longer affected by the intake of fluids and food). Continuous HR was measured at 1-minute intervals throughout a work shift, using an ActiHeart HR monitor (Mini-Mitter Company, Inc). The HR monitors were attached to participants before their shift and removed after their shift.

Wet bulb globe temperatures (WBGTs) were monitored at several locations throughout the potrooms, simulating the locations where employees normally worked, using QUESTemp°36 and QUESTemp°34 instruments (Quest Technologies, Inc, Oconomowoc, WI). These instruments measure temperatures from 23°F to 212°F and are accurate to ±0.9°F. These instruments were allowed to equilibrate for a minimum of 15 minutes before data were collected, unless otherwise noted.

Outdoor weather conditions, including temperature, relative humidity, and wind speed, at the plant were measured using a HOBO® weather station (Onset Computer Corporation, Bourne, MA). The weather station was set up near the entrance to the plant and allowed to continually monitor conditions throughout the study.

In this evaluation, signs of excessive heat strain include (1) a T_c greater than 100.4°F for unselected, unacclimatized personnel and greater than 101.3°F for medically fit, heat-acclimatized personnel; (2) sustained (greater than 5minutes)HR exceeding 180 BPM minus the individual’s age in years (180 BPM − age) for those with normal cardiac performance; or (3) daily average HR exceeding 115 BPM. This HR criteria was used in this evaluation instead of recovery HR because the nature of the job did not permit assessment of recovery HR.

Statistical analysis was carried out with SAS version 9.1.3 software (SAS Institute, Cary, NC). A P ≤ 0.05 was considered statistically significant. Paired t tests were used to compare pre- and postshift blood and urine laboratory means. When the data were normally distributed, the t test was used to compare body mass index and changes in pre- and postshift laboratory values between employees who did and did not have signs of heat strain. In cases where the data were not normally distributed, the nonparametric Wilcoxon test was used to compare differences in pre- and postshift laboratory values between employees who did and did not have signs of heat strain. The t test was also used to compare average T_c, peak T_c, and average HR between employees who were and were not acclimatized. The Fisher exact test was used to compare differences in the percentages of employees having each sign of heat strain in employees who were and were not acclimatized. t tests or Wilcoxon tests were used to compare average T_c, peak T_c, and average HR between employees with and without work-related symptoms.

At the time, this facility produced up to 1.67 million pounds of aluminum per day. Products included sheet ingots, primary ingots, and aluminum powder. The smelter had six smelting lines. Aluminum production used two pot styles: (1) older T-51 pots containing 20 carbon anodes; and (2) newer, larger P-100 pots containing 26 carbon anodes. Each T-51 potroom housed 72 pots, while the P-100 potrooms housed 80 pots per room. Each crew oversaw 12 T-51 pots or 13 P-100 pots per shift.

Approximately 600 employees worked in the potrooms over three 8-hour shifts. Volunteers were solicited from the prior shift to work an additional 8-hour shift if there was a shortage of employees. If too few employees volunteered, mandatory overtime was enforced. No employee was allowed to work more than 64 hours per week.

Workers in the potrooms performed various tasks. Tapper carbon changers removed molten aluminum and replaced carbon anodes from the pots. Overhead crane operators helped replace carbon anodes, transported and added alumina via portable bins, and moved all pot tapping equipment. Pot tenders periodically broke the insulating crust that formed at the surface of the bath mixture, added aluminum ore, and aligned the anodes. They used wooden probes to break up air gaps in the molten material. Potroom helpers added ore and other materials to the pots and performed miscellaneous tasks. Heat-stress carbon setters were added to potroom crews during the warmer summer months. They performed the same duties as the tapper carbon changers, pot tenders, and potroom helpers.

Sixty-one potroom workers agreed to participate in the study, and 60 were included in the analysis. One employee completed the questionnaire but was excluded from analysis because that individual worked only 30 minutes. Personal characteristics of the participants are summarized in Table 1. Six of the 60 participants worked an overtime shift. Three participants performed equally hot job duties during their overtime shift as during their regular shift. Two worked as a tapper carbon changer (one of the hottest jobs) for two 8-hour shifts in a row, and one worked as a head tapper carbon changer during the first shift and then as a tapper carbon changer/crane operator for 2 overtime hours.

The most common symptoms during the work shift were racing heartbeat or palpitations (30%), headache (20%), muscle cramps (13%), and lightheadedness or dizziness (13%). Less common symptoms included confusion or disorientation (8%), unsteady walk (5%), and nausea or vomiting (2%). The only participant who reported nausea or vomiting had the highest peak T_c value of 102.83°F.

Study participants’ pre- and postshift changes in serum and urine measurements are shown in Table 2. Serum sodium was corrected for participants with glucose concentrations greater than 200 mg/dL.³ This applied to one participant who had a preshift glucose concentration of 263 mg/dL. Three participants showed marked increases in pre- to postshift blood creatinine measurements. Their pre- and postshift blood urea nitrogen to creatinine ratios were 14.0/1.1 to 20.0/2.3, 11.0/1.3 to 13.0/2.4, and 12.0/1.0 to 15.0/2.6. These participants were advised to drink plenty of fluids and have their blood urea nitrogen and creatinine levels rechecked the following day. One of these participants returned the following day, and the creatinine had dropped from 2.3 to 1.4 mg/dL.

The average amount of fluids consumed during the shift was 180 ± 102 ounces. Electrolyte packets, provided by the employer to replace electrolytes lost from excessive sweating, contained 40 mg potassium chloride, 18 mg calcium phosphate, and 9 mg magnesium carbonate. Among those reporting consumption of the electrolyte packets during their shifts (12 of 60), the average number of packets used was 3.1. Participants who reported drinking alcoholic beverages during the 24 hours before starting work (11 of 60) consumed an average of 3.5 alcoholic beverages.

The mean ambient temperature for the 5 days of our evaluation was 78°F. The mean temperature for July for that geographic area from 1971 to 2000 was 85°F.⁴

Thirty-one WBGT measurements were collected in areas where employees performed routine job tasks. These included a T-51 and a P-100 pot, a pot with side covers removed, a tapping crucible next to a T-51 and a P-100 pot, an aisle with no spent anodes and one with spent anodes, and the operator’s enclosure on the overhead crane. Protective clothing worn by employees in the potrooms included a flame-resistant long sleeve shirt and pants, long sleeve cotton t-shirt, steel toe boots, and a hard hat.

The WBGT measurements ranged from 83°F to 120°F, with dry bulb air temperatures reaching 134°F and radiant temperatures reaching 188°F. The highest WBGT measurement in the potroom was taken between a T-51 pot and a tapping crucible during tapping. The second-highest WBGT measurement was taken on the center of the catwalk between pots with one side open. This measurement may have underestimated the actual temperature, because the WBGT monitor was not fully equilibrated because the high radiant heat load was starting to melt the monitor’s tripod.

Radiant heat (globe bulb temperature) measured at all except one location inside the potrooms equaled or exceeded 96°F. At these temperatures, employees absorb rather than radiate heat unless proper shielding is provided. The highest globe bulb temperature was measured on the catwalk in the center of two pots with the side shields removed on one pot. The second-highest globe bulb temperature was measured between a T-51 pot and a crucible during tapping.

A clothing adjustment factor was estimated at 5°F for the clothing worn by employees in the potrooms. Metabolic rates for employees in the potrooms were estimated to be in the moderate to light category, ranging from 115 to 360 W. Clothing adjustment factor and metabolic heat production rates were estimated using the NIOSH method.⁵ When compared with the WBGT measurements taken in the potrooms, portions of all tasks, except for the crane operator, exceeded the ACGIH threshold limit value and NIOSH ceiling limit for working in a hot environment.^2,5

The number of participants showing signs of heat strain is shown in Table 3. Most participants (54%) had at least one sign of heat strain, and 17 participants (29%) exceeded the T_c criterion for their acclimatization status.

Comparisons of the differences in pre- and postshift laboratory values for participants who did and did not have signs of heat strain are shown in Tables 4 (T_c) and 5 (sustained peak HR and average HR). Participants who exceeded the T_c for their acclimatization status also had, on average, a significantly higher increase in blood creatinine.

Unacclimatized participants (n = 8) were significantly more likely to exceed the recommended maximum T_c, as shown in Table 6. Among unacclimatized participants, 88% (7 of 8) exceeded their ACGIH heat strain criterion for T_c compared with 20% (10 of 50) of acclimatized participants (P < 0.01). Among participants who exceeded ACGIH T_c criteria, the average percentage of the period of time monitored above criteria was 22% for unacclimatized employees and 9%for acclimatized employees. In addition, participants who exceeded the T_c for their acclimatization status and/or exceeded the sustained peak HR criterion had a significantly lower average body mass index than those who did not exceed those ACGIH criteria (27.6 vs 31.8 [P = 0.01] and 28.4 vs 32.4 [P = 0.01], respectively). Unacclimatized potroom participants also had a significantly higher average HR than the acclimatized participants (118 vs 107 BPM; P < 0.01).