• Home
• About ATSDR
• Press Room
• A-Z Index
• Glossary
• Employment
• Contact Us
• CDC

ATSDR en Español

Search:

Skip directly to: content | left navigation | search

---

Find a Publication

1. ATSDR > Public Health Assessments & Consultations

HEALTH CONSULTATION

Historical VOC Levels

TRANGUCH GASOLINE SITE
HAZELTON, LUZERNE COUNTY, PENNSYLVANIA

INTRODUCTION

The Agency for Toxic Substances and Disease Registry (ATSDR) Strike team was asked by the ATSDR Region 3 representative to provide a rapid, focused assessment of the following question posed by an internal ATSDR workgroup:

Were historical exposures via vapor intrusions into homes sufficiently high to pose a public health concern for leukemia in this community [1]?

This document only assesses the potential for the described exposures to have resulted in a public health concern for leukemia. It does not contain other considerations, such as noncancer health effects.

BACKGROUND AND STATEMENT OF ISSUES

The Pennsylvania Department of Environmental Protection (PADEP) first received complaints of gasoline odors in two homes in the site area in February 1990. The site area, running along and east of Routes 309/940, between 23rd and 17th Streets, includes the Laurel Gardens neighborhood. PADEP continued to receive scattered complaints about gasoline odors in various properties in 1992 and 1993. In April 1993, PADEP found gasoline components in water that had infiltrated a basement on W. 22nd Street. In response to persistent gasoline-like vapors at the residence on W 22nd Street, hydrogeologic investigations and further sampling were begun in the area. PADEP concluded that volatile organic carbons (VOCs) adversely affected 28 residences and one commercial building from late September through the beginning of December 1993. PADEP initiated an emergency action and began installing interim remedial systems (vapor extraction and vapor barrier systems) in the affected homes in October 1993. In 1996 and at the request of the PADEP, the U.S. Environmental Protection Agency (EPA) began sampling and remediating homes affected by the gasoline spill and defining the extent of groundwater contamination. Sewer main breaks in the area of the groundwater contamination are believed to be the pathway of exposure to residents living beyond the immediate area of the groundwater plume. EPA also investigated whether contaminants could have traveled to homes via underground coal mines in the site area, but concluded that these mines did not represent a significant pathway of exposure for community members. At the request of EPA, public health agencies have been involved with the site since May 2000.

The University of Pittsburgh, School of Public Health, was contracted by local officials to study the possibility of heath effects in the affected community. The university released its first report on the affected portion of the population in Hazle Township in August 2001. They released their second report on the affected portion of the population in the City of Hazleton in January 2003 and an executive summary of their analysis of the combined, total population in May 2003. The Pennsylvania Department of Health (PADOH) has also been reviewing the cancer statistics for this area. The primary public health issue of concern is the possibility that leukemia rates are elevated in this community. In May 2003, ATSDR committed to reviewing the university's final summary report and to make any necessary follow-up recommendations for this community. ATSDR formed an internal workgroup with representatives from the Division of Health Assessment and Consultation (DHAC), the Division of Health Studies (DHS), and the Division (proposed) of Regional Operations (DRO). DHS determined that additional assessment information was needed on past environmental data (preremediation) related to residential exposures, hence this request.

Public health officials have already addressed the potential for current or future exposures at this site. During the winter of 2001, ATSDR collaborated with PADOH to review the environmental sampling data available for 368 homes and businesses in the site area. Letters provided to each of the property owners reported that no public heath threat exists under current (and future) post-remediation conditions.

DISCUSSION

Environmental Data
Environmental agencies have collected extensive sampling data from this site, starting with the PADEP's initial sampling in 1993. ATSDR selected the PADEP's 1993-1994 sampling as the best example of environmental levels in affected homes before any remediation actions. PADEP's sampling is not as specific or as comprehensive as subsequent sampling conducted by EPA in some of the homes at this site. However, ATSDR determined that the early preremediation sampling provides the most conservative look at possible historical exposures in the affected homes. ATSDR reviewed the available EPA sampling information and determined that the results supported the assumptions described in this document. None of the results from the EPA sampling (i.e., maximum of 43 parts per billion (ppb) or 140 µg/m³) are as elevated as the approximated results we calculate in the following section for the pre-remediation PADEP sampling (i.e., maximum of 42 parts per million (ppm) or 136 mg/m³) [2,3]. This is consistent with the following: (1) the bulk of the spilled gasoline probably moved through the sewers (and most likely produced the highest concentrations of vapors in homes) early in the history of the spill; and (2) the PADEP results reflect conditions prior to the emergency remedial actions implemented by PADEP in some of the most impacted homes.

In this document, ATSDR specifically focused on the indoor air data collected by PADEP during a 72-day period in 1993 (included in Appendix A). PADEP collected indoor air data from 43 houses on 49 sampling days during this period using a Hnu photo ionization detector (PID). The highest total VOC concentration from any room in each house was recorded in PID units. The limit of detection for total VOCs by the Hnu PID was 1 part per million (ppm). Hnu PID readings for total VOCs ranged from 0-202 ppm.

The Hnu PID measures total VOCs and not individual species of VOCs such as benzene. Thus, estimating exposure to specific chemicals involves using educated assumptions based on PID readings, calibration of the PID, and the source of VOCs. It is reasonable to assume the detected VOCs emanated from the known underground gasoline plume and were largely composed of benzene, toluene, ethylbenzene, and xylenes (BTEX). This assumption is reasonable based on gasoline odors reported by PADEP personnel in homes with elevated PID readings and more recent sampling confirming the presence of BTEX. In addition, gasoline components were detected in water that infiltrated a basement in the site area [3]. If the PID-detected VOCs were mostly due to vapor intrusion from the underground gasoline plume and consisted largely of BTEX, and the PID was calibrated with isobutylene, the potential concentration of individual BTEX components can be estimated from the PID response to individual VOCs. Although containing much uncertainty, this screening method enables a worst-case estimate of VOCs of concern.

The frequency and duration of exposure is uncertain as the data were not continuously collected. PADEP took individual readings over approximately 10 weeks. PID readings reported on different days were often inconsistent in each house. The frequency of detection ranged from 0% to 33% of the total number of days sampled, with an average detection frequency of 6%. PADEP staff took instantaneous PID readings once per day at each residence sampled. The duration of each exposure is unknown.

Exposure

The limited ambient indoor air data and the nature of the collected data preclude dose/response characterization because of the following limitations:

• The duration of exposure is unknown.
• The exposure concentration is known only for the sample taken and may not reflect the actual exposure concentration over time. Thus, the reported concentration from each sampling event could overestimate or underestimate the daily exposure concentration.
• The sampling period is only 10 weeks. That may not reflect indoor air concentrations that may fluctuate over time from changes in temperature, barometric pressure, soil saturation, rainfall, and depth of water table.
• Data are expressed in PID units reporting response to total VOCs. The speciated composition is unknown.
• While assumed to be from volatiles of gasoline in groundwater, the source of PID detections is unknown and may include household VOC sources.

Benzene is a VOC in gasoline that has been associated with leukemia in occupational exposures [4]. Benzene has been most consistently associated with acute myelogenous (or nonlymphocytic) leukemia, but other hematopoietic associations have also been reported [5-9]. If the previously-mentioned assumptions regarding gasoline as the source for the PID indoor air measurements are valid, benzene may represent a potential 21% of PID units measured, after adjusting for volatility and BTEX response factors [10]. Applying this conversion yields estimated benzene concentrations of 0-42 ppm (136 mg/m³), with an estimated average of 0-2 ppm (6 mg/m³) benzene and a mean of 0.16 ppm (0.5 mg/m³). All but one residence (2 ppm) averaged < 1 ppm (3 mg/m³).

At these estimated indoor air concentrations, benzene could exceed the ATSDR cancer risk evaluation guideline (CREG) for air (0.0001 mg/m³) [11]. The CREG concentration represents a comparison value estimating a level of exposure where cancer health effects would not be expected, even over a lifetime of continuous exposure. If exposures are not continuous for a lifetime, values exceeding a comparison value are further evaluated for other parameters. Those include frequency and duration of exposure and toxicological and epidemiological considerations. A previous health consultation at this site reported using an action level for cancer risk of 0.008 mg/m³ (0.003 ppm) for maximum benzene concentrations detected using SUMMA canisters for collection over an 8-hour period [12]. For clarification of other values previously used, this value (0.008 mg/m³) was used as an action level for cancer health effects and reflects the further evaluation of values exceeding a CREG with site-specific exposure information. Both values are protective of public health for cancer health effects under the defined exposure scenarios. Another action level (0.032 mg/m³, 0.01 ppm), was previously used for noncancer health effects, but does not apply in this evaluation, which targets a cancer endpoint.

All comparison values for benzene would be exceeded by all measured PID units because the PID limit for detection was reported to be 1 ppm total VOCs. The limit of detection for benzene is estimated to have been 0.21 ppm (0.7 mg/m³), if previously addressed assumptions are valid.

Estimates of benzene exposure were converted to ppm-years to compare with reported occupational health effects on an equivalent quantitative basis. If a three-year exposure is assumed at the maximum (42 ppm, 136 mg/m³) and average (0.16 ppm, 0.5 mg/m³) levels estimated for benzene, maximum exposure estimates would approximate 2 ppm-years, with average exposure estimates of 0.03 ppm-years. Many of the epidemiological studies associating benzene and leukemia contain limitations that render them unsuitable for quantitative risk estimation. Suitable studies are available that show ample power, latency, few confounders, and reasonably good estimates of a wide range of exposure to benzene. Exposure to benzene has been estimated from these studies to increase the risk of leukemia at the level of 40 ppm-years of cumulative exposure [4, 5, 13]. This risk estimate could overestimate or underestimate the actual risk because of uncertainties in exposure concentration, duration, and frequency. In addition, this estimate was based on mortality in occupationally exposed workers. It may not be protective of more sensitive populations or less serious adverse health outcomes.

The lowest level of cumulative exposure associated with an increased incidence of acute nonlymphocytic leukemia among occupationally exposed workers has been in the range of 10-25 ppm [14-16]. Therefore, the levels in some residences could be of concern if residents were persistently exposed, similar to occupational exposures (8 hours/day, 5 days/week, 50 weeks/year). However, the limited data do not suggest that such exposures would have been likely. In addition, occupational exposures may not be applicable to exposures in more sensitive populations. Children may be more vulnerable to leukemogenesis because their hemopoietic cells are differentiating and maturing [4]. Recent evidence suggests in utero exposures may be of concern [17].

Limitations in the described environmental data do not allow meaningful estimates of long-term cumulative exposure. The actual benzene levels and the duration and frequency of exposure are unknown. However, available environmental data and site-specific information suggest the likely presence of levels of benzene that could constitute a hazard in some locations at consistent exposure to those levels. The exposure appears to have lasted about three years, but it may have been longer. The limited indoor air sampling provides a "snapshot" of VOC concentrations at one point during a day; it may have been greater or less than this concentration at other times. How long these concentrations persisted during a day or through several days is unknown. The limited data indicate periods of peaks and subsidence. Relative risk has usually been expressed as a function of cumulative exposure (ppm-years). Recent investigations have suggested, however, that increased risk tends to be associated with "peaked" exposures to benzene, but not with cumulative exposure or mean intensity, if treated as a continuous variable [18, 19].

CONCLUSIONS

ATSDR considers exposure to preremedial indoor air levels, as described by data in this consult, to be an indeterminate health hazard (see Appendix B for an explanation of an indeterminate health hazard). In previous, individual house-by-house evaluations, ATSDR and PADOH concluded that no apparent health hazard existed for current and future exposures. However, the uncertainty in exposure frequency, duration, and chemical speciation preclude a conclusion as to whether preremedial exposures via vapor intrusions into homes were sufficiently high to pose a public health concern for leukemia in this community.

The potential exists that some exposures may have been of public health concern. Historical data are insufficient to exclude such exposures from further investigation. Some levels may have been above comparison values where health effects would not be expected, but data are not available to indicate how long and how often those levels may have been present. Additionally, limited data do not preclude that levels may have been higher than observed. Finally, application of cancer effect levels in occupational exposures may not be protective of other populations.

RECOMMENDATIONS

Although the historical PADEP environmental data limitations preclude characterization of this site as a public health hazard in this document, the available information suggests further investigation has merit as prudent public health practice. Therefore, ATSDR concurs with the approach the environmental agencies have taken at this site to characterize the groundwater plume and its location and identify potential preferential pathways and location. ATSDR also supports efforts to evaluate biologically plausible health outcome data, especially reported leukemia cases.

PREPARERS OF REPORT

David A. Fowler, PhD
Consultations Section
Exposure Investigations and Consultations Branch
Division of Health Assessment and Consultations

Lora Siegmann Werner, MPH
Environmental Health Scientist
Office of Regional Operations

Technical Assistance

Gregory Zarus
Atmospheric Scientist
Exposure Investigations Section
Exposure Investigations and Consultations Branch
Division of Health Assessment and Consultations

Reviewed by

Gregory Zarus
Atmospheric Scientist
Strike Team Leader
Exposure Investigations and Consultation Branch
Division of Health Assessment and Consultations

Don Joe, PE
Chief, Petitions Response Section
Exposure Investigations and Consultations Branch
Division of Health Assessment and Consultation

Tina Forrester
Deputy Director
Office of Regional Operations

Dhelia Williamson
Epidemiologist
Health Investigations Branch
Division of Health Studies

Robin Lee, MPH
Epidemiologist
Health Investigations Branch
Division of Health Studies

APPENDIX A

Indoor air data collected using an Hnu PID in 43 houses in Hazleton, PA by PADEP over a 49 sampling days period in 1993. The highest total VOC concentration from any room in each house was recorded in PID units. The limit of detection for total VOCs by the Hnu PID was 1 part per million (ppm). Hnu PID readings for total VOCs ranged from 0-202 ppm.

Next Section     Table of Contents

Agency for Toxic Substances and Disease Registry, 1825 Century Blvd, Atlanta, GA 30345
Contact CDC: 800-232-4636 / TTY: 888-232-6348 

Department of Health and Human Services