Exposures to High Levels of Carbon Monoxide from Wood-fired Temazcal (Steam Bath) Use in Highland Guatemala

Abstract

The temazcal is a wood-fired steam bath used in the rural highlands of Guatemala for bathing and healing. We measured carbon monoxide (CO) among 288 participants in 72 temazcales. Participants were drawn from communities who participated in the RESPIRE (Randomized Exposure Study of Pollution Indoors and Respiratory Effects) chimney stove intervention trial. Temazcal CO exposures were extremely high, averaging 431 parts per million (time-weighted average). Compared to kitchen wood-smoke exposures, the temazcal contributes significantly to weekly exposures, despite the fact that the population spends less time in the temazcal than in the kitchen. This report 1) describes temazcal use patterns; 2) reports participants’ signs and symptoms during temazcal use; 3) models the distribution of temazcal CO concentrations; 4) assesses reliability of exhaled breath CO as a biomarker of CO exposure; and 5) provides a proportional analysis of CO concentrations from temazcal use, as compared to kitchen concentrations.

INTRODUCTION

In February 2010, in the highland region of Comitancillo, Guatemala, two women and two young children died while using a traditional wood-fired steam bath referred to as the temazcal (spelling variations include temescal, temascal). A Guatemalan newspaper reported that the family, ranging in age from 10 months to 48 years, was overcome by the steam vapor while practicing a “Mayan ritual.”¹ A second article reported the actual cause of death—poisoning from carbon monoxide (CO), one of the products of incomplete combustion from burning wood used to heat rocks in the steam bath.² The temazcal is used extensively in the western highlands of Guatemala, where an estimated 1 to 2 million rural indigenous people of Mayan descent use the steam bath primarily for bathing purposes. The temazcal is also believed to have healing properties; the traditional hot-cold theory of disease is an important concept in these indigenous communities, especially during the postpartum period.³ To restore humoral balance lost during childbirth, mothers and their newborns spend a large portion of the four-week postpartum period near an open fire and typically enter the temazcal once or twice a day.⁴ Women and newborns who use the temazcal in the postpartum period are at greater risk for increased exposure to CO, but all users face an acute risk for serious injury or death that needs to be addressed.

Our current interest in quantifying CO levels among temazcal users, and specifically among women of reproductive age, arose during the RESPIRE chimney stove intervention trial conducted in the rural highlands of Guatemala between 2002 and 2004. The study’s primary aim was to estimate the impact of reduced household air pollution on the incidence of acute lower respiratory infection among young children under 18 months of age.⁵ During RESPIRE, we found elevated exhaled breath CO measurements among adult women who had a cleaner-burning chimney stove but reported entering the temazcal within the 48 hours before the breath CO measurement. Despite 90% reductions in kitchen air pollution exposures from the chimney stove, 86% of the 534 study households reported using the temazcal once or twice a week on the baseline questionnaire. Thus, we were concerned about exposures to high levels of carbon monoxide (CO) in the temazcal.⁶

Construction Characteristics and Use Patterns

Temazcal is a Spanish word derived from the Nahautl language, consisting of two parts—tema, to bathe, and calli, house. A Quiché dictionary, compiled shortly after the Spanish Conquest of Guatemala, uses the term zumpulche to describe a sweat bath used by women after childbirth and by sick persons to cast out the cold in their bodies.⁷ In the Mam language, the temazcal is referred to as the chuj. The temazcal has been an integral part of indigenous cultural practices throughout Mexico and Central America since at least 900 BCE.⁸ Archeological excavations of a 12th-century temazcal, located in the highland Quiché region of Guatemala, uncovered a hearth separated from the main bathing chamber and a ventilation hole in the wall to remove smoke from the chamber.⁹ This construction differs dramatically from the rustic temazcales used today, where the hearth and the bathers share the same small chamber.

In the rural Guatemalan highlands, the temazcal is a small (averaging 5.5 m³ volume), stand-alone structure typically constructed of adobe bricks with a packed dirt floor partly covered by a wooden platform (Figure 1). The roof is made of mud with wood support beams. An unvented firebox faces the door. A wood fire is built in the firebox to heat rocks suspended above the fire on a metal grate. Women are usually in charge of building the fire in the temazcal, which requires crawling into the structure and kneeling in front of the fire, blowing on it until it starts to blaze. At that point, the fire-builder leaves, returning periodically to add more wood to the fire. When the fire burns down, and few, if any, embers remain, family members take turns entering the temazcal to bathe. Water is poured on the hot rocks to produce steam for sauna-like conditions. There are no windows or eave spaces to allow smoke to escape, and in order to retain heat inside the chamber, the door is covered with a blanket.

Figure 1—

The temazcal

Human Health Effects from Carbon Monoxide Exposures

The toxicity of CO depends primarily on CO concentration, exposure duration, and breathing rate. High and/or sustained exposures to CO can cause large reductions in oxygen carrying capacity in the blood-stream, because CO displaces oxygen molecules attached to hemoglobin, replacing oxyhemoglobin with carboxyhemoglobin [COHb]. Acute, high CO exposures leading to hemoglobin saturation with COHb levels greater than 20% can cause cardiovascular and neurologic injury.^10–12 Carboxyhemoglobin levels between 3% and 5% have been associated with angina complaints among patients with coronary artery disease. However, studies have not found long-term neurobehavioral sequelae with COHb levels below 5%.^10,12 Active and passive cigarette smoking, as well as exposure to unvented household cook stoves, represent chronic CO exposures that may cause some of the physiological changes described previously.¹³ The average COHb of nonsmokers is < 1%, whereas active tobacco smokers average 4%.¹⁴ CO exposures among women cooking with solid fuels in poorly ventilated homes are sometimes high enough to result in COHb between 2.5% and 16%, with upper limits comparable to upper limits seen in heavy smokers.^15,16

Signs and Symptoms of Carbon Monoxide Toxicity

Table 1 outlines potential signs and symptoms of CO toxicity with the corresponding estimated range of COHb levels. Most of these findings are drawn from clinical case reports of acute CO poisoning where exposure concentration and duration, prior to the determination of COHb, is typically unknown.^17,18 Given the variability in exposure and human response to these exposures, caution should be exercised when estimating COHb based only on reported signs and symptoms.

TABLE 1.

Signs and Symptoms Suggestive of Carbon Monoxide Poisoning and Corresponding Carboxyhemoglobin Saturation^22,26

Signs and Symptoms	COHb%
MILD	0–30
Mild/moderate headache
Decreased exercise tolerance with pulmonary disease
Chest pain with cardiac disease
Shortness of breath
Nausea/vomiting
Dizziness
Fatigue
Blurred vision
MODERATE	30–40
Severe headache
Chest pain
Nausea/vomiting
Disorientation
Memory loss/confusion
Syncope/loss of consciousness
SEVERE	> 40
Cardiac dysrhythmias
Cardiac arrest
Hypotension
Coma
Seizures
Apnea
Death

Air Quality Guidelines and Standards

The World Health Organization Air Quality Guideline (AQG) values for different time periods were chosen such that a COHb level of 2.5% is not exceeded, in order to protect the most sensitive groups within the general population (rounded ppm):¹⁴

• 100 mg/m³ (87 ppm) for 15 minutes

• 60 mg/m³ (52 ppm) for 30 minutes

• 30 mg/m³ (26 ppm) for 1 hour

• 10 mg/m³ (9 ppm) for 8 hours

Worker exposure limits are set at higher thresholds than population limits and illustrate the maximum recommended exposures for human populations. The National Institute for Occupational Safety and Health (NIOSH) set a recommended CO exposure limit of 35 ppm (eight-hour time-weighted average) and 200 ppm as a maximum permissible ceiling, so COHb in workers does not surpass 5%.¹⁹

Exposure Assessment Pilot Study

In 2003, during RESPIRE, we conducted a pilot study to quantify CO concentrations in the temazcal.⁴ Before and after entering the temazcal, 23 participants provided a sample of exhaled breath CO. Exhaled breath CO is a biomarker of exposure that reflects CO concentrations in the pulmonary alveolar spaces. The average pre-temazcal measure was 4.1 parts per million (ppm) (median 3.5, range 1–11 ppm); the post-temazcal measure was 67.5 ppm (median 52 ppm; range 10–222 ppm). Although exhaled breath CO is not directly comparable to ambient concentrations, the observed mean change in breath CO (~63 ppm) were exceedingly high. Investigators remained in the vicinity of the temazcal while the participants bathed; thus we suspect that participants may have altered and specifically shortened their time in the temazcal. Because the pilot study was small, and temazcal use patterns were probably not typical, in 2005, we undertook a larger self-monitoring study to measure the acute, high CO exposures during bathing events; these results are reported here.

Purpose

Preliminary data suggested that the magnitude of CO exposures in the temazcal resulted in a large contribution to weekly exposures as compared to CO exposures from kitchen wood smoke, despite the fact that the population spends much less time in the temazcal than in the kitchen. We found that the person who built the temazcal fire, typically a woman of reproductive age, had the highest exposures because she built the fire and spent more time inside bathing her children. The purpose of the current investigation is to describe temazcal use patterns; report participants’ signs and symptoms during temazcal use; model the distribution of temazcal CO concentrations; assess the reliability of exhaled breath CO as an accurate biomarker of exposure to CO; and provide a proportional analysis of CO concentrations from temazcal use, as compared to kitchen CO concentrations.

PARTICIPANTS AND METHODS

Baseline Questionnaire and Recruitment

Prior to monitoring, a focus group was conducted in Spanish with 11 local, Mam-speaking fieldworkers who worked for the RESPIRE study. The fieldworkers were asked a series of open-ended questions to ascertain not only why one might increase or decrease use of the temazcal, but also the feasibility of a self-monitoring study with community participants. A questionnaire was developed based on the focus group results. After a phase of pilot-testing and subsequent modification, the fieldworkers verbally administered a baseline questionnaire to a convenience sample of 97 households in six RESPIRE communities. Questions included frequency and duration of temazcal use among all household members, and signs and symptoms experienced during temazcal use that might indicate CO poisoning. At the end of the questionnaire, the family member who answered the survey (usually the mother of the RESPIRE study child) was asked if she and other members of her family (typically other children, grandparents, in-laws, etc.) would be willing to participate in a monitoring study during a temazcal bathing event later in the week. If she was, a return date was scheduled for the self-monitoring session.

Self-monitoring Training Session

Five fieldworkers were trained in equipment use. They pilot-tested the equipment and recording forms with their family during a single temazcal event. Based on this pilot test, modifications were made to the final questionnaire and sampling protocol. The fieldworker made two visits to 72 out of 97 homes that participated in the baseline questionnaire and were willing to take part in the monitoring study. At the first visit, the fieldworker designated a member of the family as the “recorder” and trained him or her to use the equipment. Selection of this family member (usually an older child or father) was based on who would be able to read and record numbers on the recording sheets. Many of the participants have limited literacy; therefore, each of the family members was assigned a different color corresponding to his or her own recording sheet. The fieldworker inspected the temazcal, placed a nail on the back wall so equipment could be positioned there during the bathing event, and left the equipment with the family. Up to four members were asked to participate in each household. Because we were interested in monitoring participants of all ages, only participants who were too young to use the equipment (< 4 years) were excluded.

Carbon Monoxide Exposure Measurements: Exhaled Breath CO

Device.

A measure of personal exposure was obtained before and after temazcal use, using a Micro Medical Micro CO handheld exhaled breath CO monitor (Micro Direct, Inc., 803 Webster St., Lewiston, ME 04240). The monitor uses an electrochemical fuel cell sensor that generates an electrical current proportional to the CO concentration of alveolar gas. The CO level in ppm units (sensitivity to within 1 ppm) is displayed digitally. These devices were calibrated by the manufacturer prior to the study.

Field Procedure.

During the initial self-monitoring visit, the fieldworker trained the family members to use the breath CO device. The fieldworker recorded the digital reading for the before-temazcal breath CO measurements on the recording sheets for all family members who planned to use the temazcal. Participants were instructed to take a deep breath and hold it for 17 seconds, as indicated by a timer on the device, and to exhale slowly and steadily into the single-use mouthpiece attached to the monitoring device until they felt that all of the air had been expelled from their lungs. This maneuver was repeated twice, with a one-minute interval between measurements. Each of the family members repeated this procedure after the temazcal event, noting the digital reading and time of measurement (using digital watches provided to them) on his or her color-coded recording sheet.

Carbon Monoxide Exposure Measurements: Electrochemical CO Data-logging Devices

Device.

Temazcal CO concentrations were measured using the HOBO/Onset data-logging H11–001 CO monitor (Onset Computer Corporation, PO Box 3450, Pocasset, MA 02559–3450). This device has an electrochemical sensor that uses both oxygen and CO to generate an electrical current proportional to CO concentrations. The CO monitors in the present study were set to record CO (ppm) at 10-second intervals and at the maximum 0–2,000 ppm range (resolution of 8 ppm, typical accuracy of ±12 ppm or ±7% of readings, whichever is greater). Although the CO sensor compensates for temperature changes at temperatures between 20°C and 40°C (typical of the temazcal conditions), errors in readings may by increased by an additional ±5 ppm or ±5% of readings, whichever is greater. The CO monitors were calibrated before and after the study with span gases of 5, 19.9, and 60 ppm CO, and responded linearly over this range (r²= 0.82–0.99), with no change in linear response after the temazcal monitoring session.

Field procedure.

The first person to enter the temazcal (usually the fire-builder, or the mother) was asked to take the CO data-logging monitor mounted on a clipboard into the structure and to place it on the nail on the back wall of the structure, opposite the fire and the door, at breathing height, approximately one foot from the ceiling. The last person to leave the temazcal removed the clipboard. At the second visit, usually the next day, the fieldworker assured that the recorder had completed the forms correctly and retrieved all equipment, checking to see if the monitor was still functioning.

Statistical Analysis

Carbon monoxide concentrations (ppm) were measured continuously for all temazcales in the study and CO exposures for each temazcal participant were estimated. Four exposure metrics were calculated based on the time the participant recorded entering and exiting the temazcal: CO cumulative exposure; average concentrations for each person’s 15-minute peak, based on a moving average; maximum 10-second peak; and time weighted average (TWA). To estimate the relative contribution of temazcal exposures as compared to kitchen exposures, we used the population-averaged distribution of personal CO exposures for women (n = 488) and children (n = 500) who were followed during the RESPIRE trial intervention period. (These results have already been reported elsewhere.⁶) Briefly, women and their children wore passive diffusion CO tubes every three months (Gastec Corp.) for a 48-hour period. CO tubes were removed when they entered the temazcal to bathe, thus the tubes did not include CO exposures from the temazcal. Personal CO exposures were calculated for both chimney stove (women = 247; children = 256) and open fire (women = 241; children = 244) households. A “typical” daily distribution of CO exposures for open fire and chimney stove households was constructed using these data, and the proportional contribution of the temazcal from the current study was added to this estimated exposure profile. Data were analyzed using Stata, version 11.0.

Ethical Considerations

The project was approved by the Committee for the Protection of Human Subjects at the University of California at Berkeley, and by the Universidad del Valle in Guatemala City. Informed consents were verbally administered to all participants and signed consents and assents were obtained.

RESULTS

Participant Baseline Questionnaire

All 97 households who were approached by the fieldworkers were willing to respond to the questionnaire. Among those that answered the questionnaire, 72 households (74.2%) participated in the self-monitoring of a single temazcal event.

Frequency of Temazcal Use

Among the 97 households that responded to the baseline questionnaire, 53.8% used the temazcal once a week, 38.2% twice a week, and 4% less than twice a month. The remaining 4% used the temazcal three times a week. This is comparable to RESPIRE baseline data, where 52.3% of 534 households reported using the temazcal once a week and 35.6% used it twice a week. Ninety-six percent of the fire builders were women of reproductive age. The women estimated that they spent an average of 70 minutes (range 20–120) inside the temazcal during each event. They estimated that their children under the age of 12 years averaged only 20 minutes (range 5–90) inside during each event.

Use During Postpartum Period

All respondents were asked to report if they used the temazcal during a previous pregnancy and whether they visited the temazcal more frequently during the postpartum period (referred to as the dieta and usually lasting four weeks) (Table 2). Women reported that they did not increase temazcal use during pregnancy, but did increase it in the immediate postpartum, entering the temazcal up to two times a day. By the fourth week, 20% of the women reported they no longer used the temazcal for postpartum healing and resumed customary bathing practices. The newborns were taken into the temazcal slightly more than three times per week over the entire four-week period.

TABLE 2.

Reported Frequency and Duration of Temazcal Use by Mothers and their Children from Baseline Questionnaire (n = 97) and Self-monitoring Session (n = 72)

	Baseline Questionnaire n = 97	Self-monitored Session n = 72
Fire builder
Female, n (%)	96 (99)	72 (100)
Age, mean (SD)	30.6 (8.2)	29 (7.8)
Minutes inside the temazcal
Fire builder, mean (SD)	70 (28)*	80 (35)**
Child < 12 years, mean (SD)	20 (18)*	58 (37)**
Temazcal Use During Postpartum Period	Mother	Newborn
Week 1, n (%):	97 (100)	95 (98)
Events, mean (SD)	8.9 (3.5)	3.4 (2.6)
Minutes each use, mean (SD)	34.8 (13.8)	9.8 (5.9)
Week 2, n (%):	96 (99)	97 (100)
Events, mean (SD)	6.5 (1.4)	3.4 (2.1)
Minutes each use, mean (SD)	35.3 (13.2)	11.3 (7.2)
Week 3, n (%):	95 (98)	95 (98)
Events, mean (SD)	5.8 (1.9)	3.4 (2.2)
Minutes each use, mean (SD)	37.3 (15.8)	11.8 (7.4)
Week 4, n (%):	78 (80)	79 (81)
Events, mean (SD)	3.6 (2)	3 (2.3)
Minutes each use, mean (SD)	45.8 (16.5)	14.4 (9.6)

^*Respondents’ estimated time

^**Time from recording sheets during temazcal session

Cultural Beliefs and Practices

In addition to bathing and hygiene use, interviewees were asked to identify all health or illness states that would lead them, or their family members, to increase or avoid using the temazcal (Figure 2).

Figure 2—

Participants’ Reasons for Using and Avoiding the Temazcal from Baseline Questionnaire (n = 97)

Signs and Symptoms of Potential Carbon Monoxide Toxicity

Over two-thirds (70.1%) of the 97 interviewees reported having experienced signs and symptoms potentially related to CO poisoning in the past year (Table 3). Among these women, 26% reported two or more of the eight signs or symptoms of possible CO toxicity (“weak” and “tired” symptoms were excluded because the temazcal is used at night before going to bed). Thirty-eight percent of the women reported experiencing a headache after the temazcal, but headache may not be specific enough to indicate CO toxicity. Clustering other signs or symptoms may improve detection of actual CO toxicity. Twenty of the 37 women (54%) who experienced headache also experienced dizziness (n = 11), difficulty walking (n = 9), loss of consciousness (n = 7), nausea (n = 6) and/or vomiting (n = 4), confusion (n = 3), or fell asleep inside (n = 1). Women who reported “loss of consciousness/ fell asleep,” “difficulty walking,” or becoming “confused” (symptoms associated with COHB exceeding 30%) during a temazcal event in the past year were more likely to report, on average, 4.2 concomitant symptoms.

TABLE 3.

Self-reported Signs or Symptoms of Potential CO Poisoning in Women and their Children in the Past Year, from Baseline Questionnaire (n = 97)

	Mother		Her Child
	Experienced in Last Year n (%)	Number of Times in Past Year Mean (SD)	Experienced in Last Year n (%)	Number of Times in Past Year Mean (SD)
None reported	29 (30)		70 (72)
Individual symptoms reported
Tired	38 (39)	3.7 (3.7)	13 (13)	2.6 (2.4)
Headache	37 (38)	4.5 (3)	8 (8)	4.3 (3.9)
Dizziness/lightheaded	15 (15)	2.5 (1.9)	5 (5)	3.2 (3.8)
Difficulty walking	11 (11)	2 (1)	5 (5)	1.6 (0.9)
Weak	11 (11)	2.4 (1.7)	6 (6)	2.1 (0.7)
Loss of consciousness	10 (10)	2.2 (1.2)	9 (9)	1.7 (1)
Vomiting	10 (10)	2.4 (1.1)	9 (9)	1.2 (0.4)
Nausea	7 (7)	1.9 (0.7)	4 (4)	1.5 (0.6)
Confusion	4 (4)	1.8 (1.5)	3 (3)	1.7 (0.6)
Falling asleep	3 (3)	1.7 (1.2)	7 (7)	1.4 (0.5)
Multiple symptoms reported*
2 symptoms	16 (17)	5.1 (3.1)	2 (2)	2(1)
3–4 symptoms	4 (4)	10.7 (6.2)	3 (3)	4.3 (1.3)
5–7 symptoms	5 (5)	9.4 (4.2)	4 (4)	10 (3.6)

^*“Tired” and “weak” are not included in multiple symptoms—non-specific for CO toxicity because temazcal is used at night before bedtime

Self-monitoring Sessions

Among the 72 households who agreed to participate in the self-monitoring event, information was collected on 288 participants. Among these participants, seven were excluded because they did not record entrance and exit times. Sixty-seven percent, or 187, were female and 33%, or 94, were male. The mean age was 25 years (range 4–69 years). Although children under the age of 4 years used the temazcal during monitoring sessions, we did not collect exposure data on them because they could not perform the exhaled breath CO test.

Temazcal Carbon Monoxide Concentrations and Personal Exposures

Mean time in the temazcal was 80 minutes (SD, 35) for the fire builder and 58 minutes (SD, 37) for 4–12-year-old children. Summary statistics for 281 individuals obtained from the CO data logger are presented in Table 4. Ten of the 281 participants (representing two temazcal events) had maximum peaks that exceeded the recording capacity of the CO monitor at 1,998 ppm. The mean 15-minute peak for the 275 participants who remained in the temazcal for at least 15 minutes was 823 ppm, with only one person having a 15-minute exposure concentration below the WHO guidelines of 87 ppm.

TABLE 4.

Summary Statistics of Personal Carbon Monoxide Using CO Data Logger in 72 Temazcal Monitoring Events

Exposure Measure	n	Mean	Min	25%	Median	75%	Max
Cumulative (ppm-hr)	281	525	4	233	388	665	2,563
15-min peak (ppm)*	275**	823	41	510	781	1,054	2,127***
Peak (ppm)****	281	950	27	562	923	1,202	2,127***
TWA (ppm)	281	431	18	256	385	588	1,436
Time in temazcal (min)	281	71	5	46	62	87	180

^*15 min-peak refers to the highest average 15-minute ambient concentration inside the temazcal during each participant’s bath

^**6 participants did not stay in for the full 15 minutes, missing values

^***10 participants exceeded HOBO device limit of detection = 1,988 ppm; calibrated value presented in table

^****Peak 10-second value reported by HOBO

Exhaled Breath CO: Biomarker of Personal Exposure

We compared the reliability of predicted COHb based on self-monitored exhaled breath CO measures to predicted COHb based on time-weighted average temazcal CO concentrations using the following process. First, we estimated COHb from breath CO using the linear regression equation given by: COHb = 0.63 + 0.16 (CO breath, ppm)²⁰ (based on a laboratory study of exhaled breath in smokers and non-smokers regressed against blood carboxyhemoglobin ranging from 0.3 to 12% (r = 0.95)). Second, we predicted COHb from the total, time-weighted temazcal CO concentrations using Peterson and Stewart’s (1970) simplified regression equation²¹ (based on a one-minute to 24-hour exposure to CO level averaging between 1 and 500 ppm for resting, nonsmoking population) and given by:

$$\%\mathit{\text{COHb}}=\frac{C{O}^{0.358}{t}^{0.63}}{197}$$

Where:

t = time in minutes and

CO = concentration of CO in inhaled ambient air in parts per million.

Finally, the observed breath CO measures, the predicted COHb from exhaled breath CO, and the predicted COHb from ambient CO concentrations for 281 temazcal users were compared (see Table 5).

TABLE 5.

Measured Carbon Monoxide (CO) (ppm), Predicted CarboxyHemoglobin (COHb) from Exhaled Breath among 281^* Temazcal Users upon Exiting the Temazcal, and Predicted Carboxyhemoglobin (COHb) Based on Measured Concentrations during a Single Temazcal Event

Exposure Measure	Mean	Median	Minimum	25%	75%	Maximum
Measured exhaled breath CO (ppm)	61	49	3	22	83	277
Predicted from breath CO: COHb20	10	8	1	4	14	45
Predicted from temazcal concentrations: COHb21	13	11	0.3	2	16	44

^*Missing data = 7

COHb estimated from breath CO tended to be somewhat lower than COHb predicted from time-weighted average temazcal CO concentrations. COHb predicted from time-weighted average temazcal concentrations (using the equation given previously) is normally distributed (data not shown), but COHb predicted from breath CO is right-skewed (Figure 3). This right-skew of the data would likely result from misuse of the breath CO monitor such that a complete reading was not taken (e.g., incomplete exhalation, not holding breath for full amount of time, or incomplete contact between device and participant’s mouth). Because investigators were not present at the time of the breath CO recording, we are unable to estimate the extent of user error.

Figure 3—

Predicted COHb Based on Breath CO Measurements after Exiting Temazcal with Superimposed Normal Distribution Line

It appears that temazcal users did not accurately measure or record post-temazcal exhaled breath CO, as these data should have reflected the normal distribution of predicted COHb based on temazcal CO concentrations. Log-transformation did not improve the correlation between the two measures, with moderately low Spearman (rho) rank correlation coefficients ranging between 0.40 and 0.44. Diagnostics checking using regression analysis of estimated COHb from breath CO on predicted COHb from temazcal CO concentrations demonstrated that while linearity could be corrected through log-transformation, homogeneity of the variance and the assumption of normality were violated (data not shown), thus the reliability of the breath CO data could not be assured. Therefore, COHb predicted from temazcal time-weighted CO concentrations illustrates a more complete exposure profile.

Maximum predicted COHb from exhaled breath upon exiting the temazcal and from temazcal concentrations from a single monitoring event, are in the severe CO toxicity range of > 40% and are consistent with signs and symptoms reported on the baseline questionnaire over the past year. Levels greater than 60% lead to apnea, coma, or death.²²

We did not monitor temazcal use during the post-partum period, nor did we monitor newborns or small infants. However, using information from 97 baseline questionnaires of self-reported temazcal use during the postpartum period, and the concentrations monitored during the study (using the log-transformed, CO concentrations for women of reproductive age [15 to 48 years of age] and children [4 to 8 years of age]), CO exposures during the postpartum period from temazcal use alone would increase maternal exposures by 8 ppm per week (reportedly averaging 3.8 hours inside the temazcal per week) and newborn exposures by 1.3 ppm per week (averaging 39 minutes per week inside the temazcal). Table 6 presents the estimated proportional contribution of the temazcal among women and children in open fire and improved chimney stove households. Population level averages of 48-hour personal exposure measurements among women and children in open fire or chimney stove households have been reported previously.⁶ When comparing chimney stove to open fire households, as kitchen exposures decrease, the proportion of weekly CO exposure from the temazcal increases.

TABLE 6.

Estimated Proportional Contribution of the Temazcal to Weekly Carbon Monoxide (CO) Exposures among Women of Reproductive Age and their Young Children

	Chimney Stove6		Open Fire6		Temazcal Use During Postpartum Period
	CO, ppm	Time, Hours	CO, ppm	Time, Hours	CO, ppm	Time, Hours	Weekly Contribution from Temazcal, % of Total
Women of reproductive age	2.2	164	4.8	164	354	4	Chimney stove: 78% Open fire: 62%
Young children	1.5	167	2.8	167	337	1	Chimney stove: 46% Open fire: 32%

DISCUSSION

This appears to be the first study to document the extremely high exposures to CO among a rural indigenous population of temazcal users in the highland communities of Guatemala. With a mean average 15-minute peak of 823 ppm, 98% of the temazcal participants exceeded the WHO standard of 87 ppm for a 15-minute period. For a single peak exposure, the mean for the 281 participants was 950 ppm; 97% of the 281 participants exceed the NIOSH standards of 200 ppm for a peak exposure. Even if this were the only CO exposure for these participants, their mean exposure over an eight-hour period would be 48.5 ppm (assuming a median time of 60 minutes in the temazcal), five times greater than the WHO AQG of 9 ppm for an eight-hour period.

In recent years, an increasing number of studies have found an association between adverse pregnancy outcomes and maternal exposure to outdoor air pollution. These levels of pollution are substantially lower than levels found in households using solid fuel for cooking and heating. A systematic review of 12 studies examined the association between outdoor air pollution and fetal health.²³ The authors concluded that CO (an asphyxiant gas) is the only pollutant with biologically plausible toxic effects on the fetus. At least two large ambient pollution studies have demonstrated that pregnant women exposed to CO of < 6 ppm during the first²⁴ and the third²⁵ trimesters have an increased risk of low birth weight babies. The current study explored the potential high exposures to CO that the gestating mother and her fetus might experience if temazcal use were to be continued throughout pregnancy.

Overall, COHb estimated from exhaled breath CO measures used in this study tended to be somewhat lower than COHb predicted from the Peterson and Stewart equation (based on temazcal CO concentrations). There are several factors that might influence the lack of agreement between some measurement pairs and may actually underestimate exposure levels for predicted COHb. First, the Peterson and Stewart equation,²¹ which was used to estimate COHb from temazcal CO concentrations, is based on an exposure limit of 500 ppm. Ninety-one participants, or 32.3%, had time-weighted averages that exceeded 500 ppm. Second, the position of the CO monitor in the temazcal may not have accurately reflected the location of the users in the temazcal (mixing of air in the temazcal, personal breathing zone, etc.) and thus may not reflect their true exposure. Finally, participants’ technique when using the exhaled breath CO monitor and the lack of supervision by fieldworkers during this self-monitoring exercise undoubtedly played a role in underestimation of exposure. Because we were unable to verify who used the breath CO monitor correctly, the data-logging CO monitor was used as the “gold standard” for measuring exposure during temazcal use.

Within the temazcales, extremely high CO concentrations exist in the presence of high heat and humidity. The CO monitor specifications indicate that for temperatures ranging between 20°C to 40°C, the actual readings may be elevated by as much as 5%, or 5 ppm, whichever is greater. Given that a previous monitoring study measured CO under super-saturation conditions with a mean temperature of 71°C (averaged over the fire-lighting stage and while participants were occupying the temazcal), we cannot be entirely confident of the accuracy of the CO readings.⁴ However, even if we were to reduce our estimates by 5–10%, the proportional contributions to CO exposures from temazcal use would still be extremely high.

The majority of the women in the study had never heard of the term “carbon monoxide.” However, from the baseline questionnaire, respondents reported that people with headache (38%), nausea (7%), and weakness/tiredness (11%) should avoid temazcal use. As these may be potential signs and symptoms of CO toxicity, it appears that some of the women are aware that these signs and symptoms only worsen with temazcal use, indicating an awareness of potential health risk. One of the limitations of this study is that we were not able to determine the relationship between reported signs and symptoms and elevated CO concentrations. We did not ask participants to report adverse signs or symptoms immediately following the monitored temazcal event. On the baseline questionnaire, mothers reported less CO toxicity symptoms in their children, which could be related to children’s shorter exposure time in the temazcal or mother’s lack of awareness about their children’s symptoms (headache and nausea cannot be objectively measured).

During the focus groups, the fieldworkers reported that “some people like it to be really hot, others don’t; it all depends.” Future studies should monitor CO exposures, and concomitant signs and symptoms of CO toxicity, during several temazcal events in the same household to see if certain families do experience elevated concentrations, or an increased number of CO signs and symptoms, more consistently than others. To reduce exposures in the “high-end users” (75th percentile time-weighted average CO concentration of 588 ppm), future studies should characterize temazcal use in “low-end users” (25th-percentile time-weighted average CO concentration of 256 ppm). From this, recommendations could be made to shift the distribution of exposures, such that current “high-end users” shift to levels characteristic of current “low-end users.”

RECOMMENDATIONS

Given the extreme, high exposures that we monitored in over 70 temazcales, and the reports of recent deaths in communities in the Guatemalan highlands, there is a need to develop and test modifications to the current temazcal structure that vent the combustion by-products to the exterior. In addition, an alternative technology for heating the rocks used in the temazcal might also be well received in these local communities. Motivation for this stems not only from the high acute exposures described previously, but also from the potentially serious chronic health effects resulting from continuous exposures to CO over many years, starting in early infancy. Motivation for change has come from several of the participants in the present study, who, after finding out about the high CO levels and potential health repercussions, have requested that this be further explored. Temazcal use will not be discontinued in the population; it is the primary bathing practice and is deeply grounded in local tradition. However, by working with community members, such as the traditional birth attendants who customarily recommend increased use by the mother and her newborn during the post-partum period, and local entrepreneurs, such as the chimney cook stove builders, a practical, sustainable, and socially desirable “improved” temazcal that preserves traditional practices should be developed and implemented.