Occupational health surveillance: Pulmonary function testing in emergency responders.

Emergency responders may be exposed to a variety of fumes, gases, and particulates during the course of their job that can affect pulmonary function (PF) and require the use of respiratory protection. This investigation used occupational health monitoring examination data to characterize PF in a population currently employed as emergency responders. PF tests for workers who required health examinations to ensure fitness for continued respirator use were compared to the National Health and Nutrition Examination Survey (NHANES) III Raw Spirometry database to determine if decreased PF was associated with employment as an emergency responder. The results of this research indicated that the emergency responders experienced a modest, but statistically significant, increase in forced expiratory volume in 1 second (FEV1) and forced vital capacity (FVC) mean values over the NHANES III population in both total and stratified analyses, including stratification by age, gender, height, and smoking history. Results are likely due to a combination of effectively controlled exposures in the workplace, and the healthy worker effect among long-term workers. PF testing required by the Occupational and Safety Health Administration (OSHA) has substantial utility for conducting occupational surveillance at the population level. In this investigation, we were able to quickly evaluate if abnormal PF existed in an industrial sector known to have exposures that, when uncontrolled, can lead to PF impairment.

INTRODUCTION

Airborne occupational exposures to irritants, dusts, and gases can cause pulmonary function (PF) impairment during high-level acute conditions, as well as prolonged periods of low-dose exposure. For occupations that entail potential uncontrolled exposure to substances known to be associated with PF impairment, respirators may be the principal method for exposure control. Therefore, governmental standards have been established to ensure the protection of workers when the elimination of the airborne hazard cannot be removed and engineering controls are not possible.

In situations where uncontrolled pulmonary exposures may exist, it is universally recognized that a respiratory protection program should be established. A medical evaluation of respirator wearers is a required element of a respiratory protection program. To comply with this element of the program, employees may be required to undergo PF testing to evaluate whether they are able to safely wear a respirator. It is important to note that the Occupational and Safety Health Administration (OSHA) Respiratory Protection Standard only requires a physician to establish the necessary health and physical conditions for a worker to be able to perform their assigned job functions while wearing a respirator. The standard does not require a particular evaluation procedure, such as a PF test except in certain provisions of specific OSHA standards.

While PF testing is not required for all employees who wear respirators, it is a best practice that is utilized in many industries, particularly those with a high likelihood of uncontrolled, unknown or high-level exposures. Spirometry is the most frequently performed PF test and is the cornerstone of most occupational respiratory evaluation programs.[1] Spirometry data collected as a result of either mandatory testing, or respiratory evaluation programs, provide a unique opportunity to perform occupational health surveillance among workers in targeted industrial sectors known to have potentially harmful exposures in the workplace. Unfortunately, the vast majority of this data is used to simply validate individual capacity for respirator use, and is ignored for population level analysis.

PF testing is particularly well suited for occupational surveillance, given the availability of the NHANES III Raw Spirometry data set. NHANES III allows for population level analysis of worker spirometry data to be compared to a standard population adjusted for age, height, tobacco smoking, and other factors that impact PF not related to the occupational environment. The NHANES III Raw Spirometry data contain over 15,000 individual spirograms matched to standard NHANES demographic and survey data. Once compiled, PF data from exposed workers can be quickly analyzed for comparison to NHANES III data to determine if a population level abnormality exists within a specific industrial sector.[2] This investigation will determine if PF is impaired in a population currently employed as emergency responders and evaluate the feasibility of using PF data collected for purposes of compliance and/or best practices for workers who use respiratory protection because they are potentially exposed to pulmonary toxicants in the workplace.

MATERIALS AND METHODS

Study population

A record review was conducted on 127 PF tests from emergency responders in the state of Florida based upon the inclusion criteria of any worker over the age of 18 years whose respirator use required PF testing. Records included data for principal confounding factors regarding PF outcomes including smoking history, age, gender, and height. A standard population for comparison consisted of the NHANES III Raw Spirometry cohort, which consists of PF tests for 16,606 individuals sampled in the USA. The Raw Spirometry file was merged by respondent identification number with the NHANES III Household Adult Data file to obtain demographic and behavioral confounder data. The NHANES III control population was further restricted by age to reflect the age range of the study population, and unacceptable tests were removed from analysis by technician quality code resulting in a final control population of 9,792 subjects. Record reviews were approved under the University of South Florida Institutional Review Board # 00001348.

Pulmonary function

All study population PF testing was conducted using the Koko spirometry system. The best attempt of a minimum of three spirometry trials was used for analysis in both the study population and the control population. The PF test outcomes used for analysis included forced expiratory volume in 1 second (FEV1) and forced vital capacity (FVC). All results are expressed as liters (L). All spirograms were reviewed by a licensed physician, and spirograms not meeting American Thoracic Society acceptability and reproducibility criteria were removed.

Statistical analysis

To determine if emergency responders experienced abnormal PF compared to the standard population, univariate analysis was conducted. Mean values were produced for FEV1 and FVC, and the significance of the differences was evaluated using the Student's t-test. These analyses were further stratified by median age, median height, gender, and smoking history.

To determine which factors were most predictive of PF, multivariate linear regression analysis was performed for the outcomes of FEV1 and FVC. Multivariate analysis was conducted by constructing linear regression models including all data elements known to impact PF. The parameter estimates identify the magnitude of effect each predictor has on either increasing or decreasing PF in the total population. Multivariate analysis evaluated the following variables as predictors of PF outcomes: Age, gender, height, pack-years of smoking, and status as an emergency responder.

There is currently an active debate regarding the use of the FEV1/FVC ratio as a definitive criterion for the diagnosis of obstructive disorders, but it is generally acknowledged that a lowered FEV1/FVC ratio is indicative of obstruction when taken into context with other PF testing data for the individual and patient demographics.[3] In the current investigation, we evaluated the study population for deficits at the higher end of the normal FEV1/FVC range, 0.80. A categorical approach was used to evaluate potential pre-clinical pulmonary obstruction using multivariate logistic regression to evaluate associations with producing an abnormal FEV1/FVC ratio, defined as less than 0.80. Categories for independent variables were defined as above and below median height and median age, females vs. males, non-smokers vs. those with a smoking history.

Statistical significance was determined by P ≤ 0.05 for all analytical tests. All statistical analyses were performed using Statistical Analysis System (SAS) version 9.1.2.

RESULTS

Univariate analysis

The population demographics for both the study population (emergency responders) and the NHANES III segment used for analysis are reported in Table 1. The study population was largely male and approximately 16% had a history of tobacco smoking. The study population was slightly younger and taller overall, compared to the NHANES III median age and height.

Table 1

Summary of study population and NHANES III control population

Table 2 provides the results of means testing for FEV1 and FVC comparing the total study population to the NHANES III segment. The study population demonstrated modestly higher mean FEV1 and FVC values when comparing the populations as a whole. The differences in age and disproportionate gender represented in the study population necessitated the use of stratified analysis to determine the effect of these population differences on evaluating the effect of emergency responder status on PF.

Table 2

Pulmonary function means for the total population

Stratification by age, gender, height, and smoking status yielded statistically significant larger mean values for FEV1 and FVC measurements for the study population. The results of the analysis are reported in Table 3.

Table 3

Pulmonary function mean values stratified by salient cofactors

Multivariate linear regression analysis

The results of the linear regression analysis for FEV1 are reported in Table 4. The analysis identified age, height, gender, and smoking history as statistically significant predictors of FEV1. The adjusted outcome for status as an emergency responder was also a significant predictor of FEV1 in this analysis. With a parameter estimate of 0.44757, status as an emergency responder conferred a modest increase to FEV1 compared to the control population.

Table 4

Predictors of FEV1 from linear regression analysis

The results of the linear regression analysis for FVC are reported in Table 5. The analysis identified age, height, gender, but not smoking history as statistically significant predictors of FVC. The adjusted outcome for status as an emergency responder was also a significant predictor of FVC in this analysis. With a parameter estimate of 0.53683, status as an emergency responder conferred a modest increase to FVC compared to the control population.

Table 5

Predictors of FVC from linear regression analysis

Multivariate logistic regression analysis

Logistic regression analysis was used to determine the effect of PF predictors on generating an FEV1/FVC ratio less than 0.80 [Table 6]. From this analysis, four statistically significant factors impacted the FEV1/FVC ratio: Age, height, gender, and smoking history. Status as an emergency responder was not associated with the production of an FEV1/FVC ratio less than 0.80. Those in the population over the median age (38) and height (69 inches) were more likely to produce an FEV1/FVC ratio less than 0.80. Similarly, those with no smoking history and female gender were less likely to produce an FEV1/FVC ratio less than 0.80.

Table 6

Logistic regression analysis of FEV1/FVC to examine the effect of predictors on producing an abnormal ratio (<0.80 FEV1/FVC).

DISCUSSION

Part of the strategic plan for the National Institute of Occupational Safety and Health (NIOSH) includes the development and expansion of mechanisms for occupational health surveillance on both the state and federal levels.[4] There is a need to develop and utilize surveillance methodologies that are capable of efficiently evaluating occupational populations for health status, identifying changes in health status over time, and comparing the health status of occupational populations to baseline populations. The use of existing health data to quickly evaluate the health status of a population provides efficiency in both cost and time by limiting the need to perform prospective data collection on a population of interest.

Various industries currently use respiratory protection for workers to control inhalation exposures. In order to ensure that workers who are employed in such fields are healthy enough to use respiratory protection, workers may be required by OSHA, or best practice, to undergo periodic PF testing. The records of this PF testing may also provide a useful cross section of pulmonary health at the population level for a specific industrial sector, though the data is not currently exploited in this fashion. In addition, the NHANES III spirometry data set provides a robust control population that can be limited to closely reflect the occupational population's salient demographics, and adjusted for confounding factors that impact PF measurements, such as tobacco smoking history.

Emergency responders are potentially exposed to a variety of fumes, gases, and particulates during the course of their job. They may choose to disregard the risk of potential chemical exposure in order to attend to injured victims and/or contain a release. Just such a situation can be found in the first responders who were on the scene immediately after the collapse of the World Trade Center. They encountered high levels of airborne pollutants and have since reported various respiratory symptoms and conditions. In one study of this phenomena, Banauch et al.[5] reported that World Trade Center-exposed workers experienced a substantial reduction in adjusted average FEV1 during the year after 09/11/2001, and that this exposure-related FEV1 decrement equaled 12 years of aging-related FEV1 decline.

Emergency responders include such diverse groups as fire fighters, police, and emergency medical personnel. All of these occupations are physically demanding and challenging and have the potential for exposure to toxic agents in a dynamic and uncontrolled environment. For example, toxic chemicals emitted as a result of incomplete combustion and pyrolysis include hydrochloric acids, carbon monoxide, vinyl chloride, hydrogen sulfide, etc.[6] These chemicals have the potential to cause PF impairment if exposures are not properly controlled. Burgess et al.[7] state:

Occupational smoke exposure may result in acute adverse health effects, particularly during periods when respiratory protection is not worn. These changes include transient reductions in spirometric measurements and increased airway reactivity. Chronic respiratory effects may also occur, although the increased use of respiratory protection appears to have had a beneficial effect.

Musk et al.[8] found that following smoke exposure, the average decrease in FEV1 was 0.05 L among firefighters and that this decline in FEV1 was related to the severity of smoke exposure as estimated by the firefighter, as well as, the measured particulate concentration of the smoke to which he was exposed. In addition, the effect of PF impairment following exposure to house fires has shown a decrease in FEV1 with a small subgroup of firefighters who develop more substantial and potentially clinically important decreases in PF after smoke exposure.[9]

Chattopadhyay et al.[10] conducted a study on the effects to respiratory function in firefighters after working at a chemical warehouse fire. They found restrictive, obstructive as well as combined restrictive and obstructive types of pulmonary dysfunction that they related to this exposure. Research conducted by Sparrow et al.[11] also found that firefighters had a greater loss of PF (FVC and FEV1) than non-firefighters, which helps to substantiate earlier reports of a chronic effect of firefighting on PF, and suggest an association of this occupation with increased respiratory symptoms and disease.

Burgess et al.[7] evaluated adverse respiratory effects following overhaul operations (i.e., opening walls, ceilings and voids to evaluate the extension of fire) in firefighters. A total of 51 firefighters were monitored for exposure to products of combustion in order to determine any changes in spirometric measurements and lung permeability. Approximately half of the firefighters in the study wore no respiratory protection during overhaul, while the other half wore cartridge respirators. The results of the study revealed acute changes in spriometric measurements and lung permeability following firefighter overhaul, even in those participants wearing full-face cartridge respirators. Changes in FEV1 were associated with levels of specific products of combustion, and demonstrated a dose-response relationship. Based on the results of the study, the authors recommend that a self-contained breathing apparatus, which has a significantly higher protection factor, continue to be used during overhaul operations.

The results of this study indicated that emergency responders experienced a modest, but statistically significant, increase in FEV1 and FVC mean values over the NHANES III population in both total and stratified analyses, including stratification by age, gender, height, and smoking history. No analysis that examined mean values demonstrated better PF in the control group versus the target worker population.

In the linear regression analysis performed to examine the effect of salient cofactors on PF, FEV1 analysis demonstrated that age, gender, height, and smoking pack-years all significantly affected PF of the population in the expected direction. That is to say, increased age, female gender, and increased pack-years of smoking decreased FEV1, while increased height increased FEV1. As well, status as an emergency responder conferred a modest, significant increase in FEV1.

Similar results were reported for the analysis of FVC, with the exception of smoking history, which did not demonstrate statistical significance in this analysis. The results of the linear regression analysis for FEV1 and FVC outcomes indicate that the predominate factors that affect PF values are those traditionally known to impact lung volume and clearance, e.g., age, gender, height, and smoking history. A modest positive effect on FEV1 and FVC was observed for emergency responders in both the stratified analysis, as well as the linear regression analysis. This may indicate the presence of the “healthy worker effect” in the occupational population related to more time spent in active labor, compared to the NHANES III population which may contain unemployed persons or those engaged in more sedentary labor.

Logistic regression was performed on outcome of the FEV1/FVC ratio to evaluate the potential for obstructive disorders among the target occupational population compared to the NHANES III population. A cut-off point of <0.80 FEV1/FVC was used to classify persons with abnormal FEV1/FVC values that could potentially be an indicator of pre-clinical pulmonary obstruction. In this analysis, status as an emergency responderwas not significantly associated with an FEV1/FVC value of <0.80. However, the analysis clearly demonstrated that the older and taller half of the population was more likely to produce a lower FEV1/FVC ratio. Also, non-smokers and females were less likely to produce a lower FEV1/FVC ratio compared to the smokers and males in the population.

OSHA mandated PF testing, as well as testing included as a best practice or standard operating procedure, represent a potentially powerful surveillance tool to evaluate at-risk populations who have known inhalation exposures that require respiratory personal protective equipment and regular spirometry evaluations. Evaluation of this type of occupational health data can lead to the advancement of occupational exposure controls as well as regulatory and policy changes that can lead to safer workplace environments. In addition, this type of research can be expanded and/or tailored to include additional indicators of occupational disease, which allows for greater utility and the potential to identify various adverse health effects associated with occupational exposure to toxicants.

The methodology presented in this study provided an opportunity to determine the feasibility of using PF data collected for purposes of compliance and/or best practices for workers who use respiratory protection. The various methods used for data collection and data analysis in this study indicated the feasibility of using occupational health data to quickly and efficiently conduct a population level analysis and draw conclusions. Limitations of the interpretation of surveillance data can include under-reporting, reporting bias, and inconsistent case definitions. While procedures were established to address such limitations, further methods could be included to enhance the results. For example, additional analyses for consideration could include time comparative studies that track health outcomes throughout a work history. Inclusion of this additional analysis may help address biases associated with a one-time study. This study not only demonstrates the feasibility of using occupational health data for population level analysis, but also illustrates the flexibility provided by the methodology presented.

Two main limitations to conducting this line of research include data collection and data warehousing. The principal limitation to conducting this line of research is access to occupational health data. While all states require disease reporting, state and federal laws do not mandate that all occupational health data is reported and pooled for analysis. Convincing industry to share this type of data is necessary to perform widescale analyses and draw strong conclusions. Ethical and legal issues related to occupational health surveillance include right of access, public trust, confidentiality, informed consent, etc. Such concerns by industry need to be addressed, and safeguards need to be established to ensure that dissemination of such data is only provided to those who need it for surveillance purposes. Cooperation and education between governmental and private sectors to create acceptable data collection procedures that ensure ethical and legal concerns must be addressed, and are key to the promotion of occupational health surveillance.

The second principal limitation to conducting this line of research is the current lack of infrastructure to aggregate both required and voluntary PF testing data. OSHA required PF testing is conducted under standard PF testing guidelines and the resulting data is maintained with employers for compliance purposes. If this data were also transmitted to a local, state, or federal database to be used in population level analysis, the availability and efficacy of this method of surveillance would be greatly enhanced. Data sharing has many limitations, which include coding, formatting, definitions, etc. Standardization of occupational health data for the purpose of data sharing and data processing need to be addressed, and guidance created to eliminate the many different inconsistencies that exist in the equipment used and the data produced. Technological advancement and the increased awareness of the importance of data sharing should drive the needed advancements and collaboration to ensure that occupational health surveillance can be used to monitor health, which is in everyone's best interest.

CONCLUSION

Through the use of PF testing data, and the available NHANES III spirometry data set, this study was able to efficiently evaluate the pulmonary health of a substantive cross section of a specific industry: Emergency responders. The data collected from PFtesting and the NHANES III spirometry data allows for the control of confounding factors that impact measures of PF so that statistical comparisons can identify deficits in PF and indicate whether or not those deficits are associated with an occupational sector. The current study did not identify any PF deficits in the target occupational population and it demonstrated that in all cases emergency responders had equivalent or modestly superior PF compared to a baseline population.

There is a need to develop and utilize surveillance methodologies that are capable of efficiently evaluating occupational populations for health status changes over time and comparing the health status of occupational populations to baseline populations. As shown by the methodology described in this study, the use of spirometry data to quickly evaluate the pulmonary health status of selected occupational populations is both a feasible and efficient way to conduct occupational health surveillance.