Developing wipe sampling strategy guidance for assessing environmental contamination of antineoplastic drugs.

Surveillance for environmental contamination of antineoplastic drugs has been recommended by authoritative bodies such as the United States Pharmacopeia and the National Association of Pharmacy Regulatory Authorities. Clear guidance is needed on how to develop sampling strategies that align with surveillance objectives efficiently and effectively. We conducted a series of simulations using previously collected surveillance data from nine cancer treatment centers to evaluate different sampling strategies. We evaluated the impact of sampling 2, 5, 10, or 20 surfaces, at monthly, quarterly, semi-annual, and annual frequencies, while employing either a random or sentinel surface selection strategy to assess contamination by a single antineoplastic drug (AD) or by a panel of three ADs. We applied two different benchmarks: a binary benchmark of above or below the limit of detection and AD-specific hygienic guidance values, based on 90th percentile values as quantitative benchmarks. The use of sentinel surfaces to evaluate a three-drug panel relative to 90th percentile hygienic guidance values (HGVs) resulted in the most efficient and effective surveillance strategy.

Spatial and Temporal Variability in Antineoplastic Drug Surface Contamination in Cancer Care Centers in Alberta and Minnesota.

The health risks of exposure to antineoplastic drugs (ADs) are well established, and healthcare professionals can be exposed while caring for cancer patients receiving AD therapy. Studies conducted worldwide over the past two decades indicate continuing widespread surface contamination by ADs. No occupational exposure limits have been established for ADs, but concerns over exposures have led to the development of guidelines, such as United States Pharmacopeia (USP) General Chapter <800> Hazardous Drugs-Handling in Healthcare. While recommending regular surveillance for surface contamination by ADs these guidelines do not provide guidance on sampling strategies. Better characterization of spatial and temporal variability of multidrug contamination would help to inform such strategies. We conducted surface-wipe monitoring of nine cancer care centers in Alberta, Canada and Minnesota, USA, with each center sampled eight times over a 12-month period. Twenty surfaces from within pharmacy and drug administration areas were sampled, and 11 drugs were analyzed from each wipe sample. Exposure data were highly left-censored which restricted data analysis; we examined prevalence of samples above limit of detection (LOD), and used the 90th percentile of the exposure distribution as a measure of level of contamination. We collected 1984 wipe samples over a total of 75 sampling days resulting in 21 824 observations. Forty-five percent of wipe samples detected at least one drug above the LOD, but only three of the drugs had more than 10% of observations above the LOD: gemcitabine (GEM) (24%), cyclophosphamide (CP) (16%), and paclitaxel (13%). Of 741 wipe samples with at least one drug above LOD, 60% had a single drug above LOD, 19% had two drugs, and 21% had three drugs or more; the maximum number of drugs found above LOD on one wipe was 8. Surfaces in the compounding area of the pharmacy and in the patient area showed the highest prevalence of samples above the LOD, including the compounding work surface, drug fridge handle, clean room cart, passthrough tray, and hazardous drug room temperature storage, the IV pump keypad, patient washroom toilet handle, patient washroom door handle, nurses' storage shelf/tray, and patient side table. Over the course of the study, both 90th percentiles and prevalence above LOD varied without clear temporal patterns, although some centers appeared to show decreasing levels with time. Within centers, the degree of variability was high, with some centers showing changes of two to three orders of magnitude in the 90th percentile of drug concentrations month to month. A clear difference was observed between the six centers located in Alberta and the three in Minnesota, with Minnesota centers having substantially higher percentages of samples above the LOD for CP and GEM. Other factors that were associated with significant variability in exposures were drug compounding volume, size of center, number of patients seen, and age of the center. We hope that demonstrating variability associated with drug, surface, clinic-factors, and time will aid in a better understanding of the nature of AD contamination, and inform improved sampling strategies.

The application of novel field measurement and field evaluation protocols for assessing health care workers' exposure risk to antineoplastic drugs.

Contamination of multiple antineoplastic drugs (ADs) on work surfaces presents an exposure concern for health care workers. Surface wipe sampling is a recognized method to evaluate the degree of contamination present. Our research team has previously reported on wipe sampling and analytical methods to simultaneously detect 10 commonly used ADs from a single wipe. Our objectives here were: to field test a protocol consisting of the wipe sampling method and an accompanying wipe sample collection tool kit and confirm this protocol can be effectively used by health care workers to assess drug contamination levels in their facilities; and, to confirm the potential for simultaneous exposure to multiple antineoplastic drugs. Three facilities within one health authority in British Columbia, Canada participated in this field study. In collaboration with the site health and safety advisors, up to 25 surfaces within each facility were considered for sampling. Collected wipe samples were analyzed using HPLC-MS/MS to quantify the 10 analyte, resulting in 750 potential analyses. Following the sampling, each of the three facilities' safety advisors provided feedback regarding the usability of the protocols. Among the 72 wipe samples actually collected (or 720 analyses conducted), detectable levels and simultaneous contamination of work surfaces of five of the 10 analytes were found at all three participating sites: 5-fluorouracil, cyclophosphamide, vincristine, paclitaxel, and methotrexate; (range < LoD to 33.0 ng/cm2) with 5-fluorouracil having the highest concentration in every instance. Drug contamination was found on a variety of different work surfaces in pharmacies and patient care areas among all three sites. Users of the sampling protocols were generally satisfied with the wipe sample collection toolkit with some minor suggestions for improvement. Our findings support the hypothesis that health care workers may be simultaneously at risk of exposure to several ADs. Our toolkit was found to be user-friendly and manageable by those who were not experienced in collecting wipe samples to monitor contamination of ADs on the work surfaces in their facilities.

Bioaerosols and Transmission, a Diverse and Growing Community of Practice.

The transmission of infectious microbes via bioaerosols is of significant concern for both human and animal health. However, gaps in our understanding of respiratory pathogen transmission and methodological heterogeneity persist. New developments have enabled progress in this domain, and one of the major turning points has been the recognition that cross-disciplinary collaborations across spheres of human and animal health, microbiology, biophysics, engineering, aerobiology, infection control, public health, occupational health, and industrial hygiene are essential. Collaborative initiatives support advances in topics such as bioaerosol behavior, dispersion models, risk assessment, risk/exposure effects, and mitigation strategies in clinical, experimental, agricultural, and other field settings. There is a need to enhance the knowledge translation for researchers, stakeholders, and private partners to support a growing network of individuals and agencies to achieve common goals to mitigate inter- and intra-species pathogen transmission via bioaerosols.

Wipe Sampling Method and Evaluation of Environmental Variables for Assessing Surface Contamination of 10 Antineoplastic Drugs by Liquid Chromatography/Tandem Mass Spectrometry.

This paper describes a novel wipe sampling and high-performance liquid chromatography/tandem mass spectrometry (HPLC-MS/MS) method capable of simultaneously detecting 10 antineoplastic drugs (5-fluorouracil, oxaliplatin, methotrexate, vindesine, ifosfamide, cyclophosphamide, vincristine, vinblastine, docetaxel, and paclitaxel). The good overall recoveries and sensitivity values of this method along with the comparatively short run time (8 min) allows for its use in routine monitoring in health care facilities. The long-term behavior of the studied drugs on contaminated surfaces and the effect of surface roughness on drug recoveries were studied to gain insights about how these environmental variables influence the detection, cleaning, and occupational exposure of these drugs. Surfaces with higher roughness parameter (Ra) values (rougher) had the lowest recoveries while those with lower Ra (smoother) presented the highest recoveries. Long-term assessments evidence distinctive drug behaviors with oxaliplatin, vindesine, vincristine, and vinblastine being the less persistent drugs (~20% was recovered after 24 h) and docetaxel and paclitaxel the most persistent drugs with recoveries of 40% and 80% after 1 month. This information indicates the importance of collecting ancillary information about drug usage (throughput, timing, cleaning procedures, etc.) to interpret the results in the context of potential exposure. Finally, the method was successfully applied to evaluate trace surface contamination down to the single picogram per square centimeter in multiple work areas within three local health care centers on Vancouver Island, Canada.

Estimating National-Level Exposure to Antineoplastic Agents in the Workplace: CAREX Canada Findings and Future Research Needs.

OBJECTIVES: Occupational exposure to antineoplastic agents occurs in various environments and is associated with increased cancer risk and adverse reproductive outcomes. National-level information describing the location and extent of occupational exposure to antineoplastic agents is unavailable in Canada and most other countries. CAREX Canada aimed to estimate the prevalence and relative levels of occupational exposures to antineoplastic agents across work setting, occupation, and sex. METHODS: 'Exposure' was defined as any potential for worker contact with antineoplastic agents. Baseline numbers of licensed workers were obtained from their respective professional bodies. For unlicensed workers, Census data or data extrapolated from human resources reports (e.g., staffing ratios) were used. Prevalence was estimated by combining population estimates with exposure proportions from peer-reviewed and grey literature. Exposure levels (classified as low, moderate, and high) by occupation and work setting were estimated qualitatively by combining estimates of contact frequency and exposure control practices. RESULTS: Approximately 75000 Canadians (0.42% of the total workforce) are estimated as occupationally exposed to antineoplastic agents; over 75% are female. The largest occupational group exposed to antineoplastic agents is community pharmacy workers, with 30200 exposed. By work setting, 39000 workers (52% of all exposed) are located in non-hospital settings; the remaining 48% are exposed in hospitals. The majority (75%) of workers are in the moderate exposure category. CONCLUSIONS: These estimates of the prevalence and location of occupational exposures to antineoplastic agents could be used to identify high-risk groups, estimate disease burden, and target new research and prevention activities. The limited secondary data available for developing these estimates highlights the need for increased quantitative measurement and documentation of antineoplastic agent contamination and exposure, particularly in work environments where use is emerging.

Comparison of qualitative and quantitative fit-testing results for three commonly used respirators in the healthcare sector.

N95 filtering facepiece respirators are used by healthcare workers when there is a risk of exposure to airborne hazards during aerosol-generating procedures. Respirator fit-testing is required prior to use to ensure that the selected respirator provides an adequate face seal. Two common fit-test methods can be employed: qualitative fit-test (QLFT) or quantitative fit-test (QNFT). Respiratory protection standards deem both fit-tests to be acceptable. However, previous studies have indicated that fit-test results may differ between QLFT and QNFT and that the outcomes may also be influenced by the type of respirator model. The aim of this study was to determine if there is a difference in fit-test outcomes with our suite of respirators, 3M - 1860S, 1860, AND 1870, and whether the model impacts the fit-test results. Subjects were recruited from residential care facilities. Each participant was assigned a respirator and underwent sequential QLFT and QNFT fit-tests and the results (either pass or fail) were recorded. To ascertain the degree of agreement between the two fit-tests, a Kappa (Κ) statistic was conducted as per the American National Standards Institute (ANSI) respiratory protection standard. The pass-fail rates were stratified by respirator model and a Kappa statistic was calculated for each to determine effect of model on fit-test outcomes. We had 619 participants and the aggregate Κ statistic for all respirators was 0.63 which is below the suggested ANSI threshold of 0.70. There was no statistically significant difference in results when stratified by respirator model. QNFT and QLFT produced different fit-test outcomes for the three respirator models examined. The disagreement in outcomes between the two fit-test methods with our suite of N95 filtering facepiece respirators was approximately 12%. Our findings may benefit other healthcare organizations that use these three respirators.

Hospital support services and the impacts of outsourcing on occupational health and safety.

BACKGROUND: Outsourcing labor is linked to negative impacts on occupational health and safety (OHS). In British Columbia, Canada, provincial health care service providers outsource support services such as cleaners and food service workers (CFSWs) to external contractors. OBJECTIVES: This study investigates the impact of outsourcing on the occupational health safety of hospital CFSWs through a mixed methods approach. METHODS: Worker's compensation data for hospital CFSWs were analyzed by negative binomial and multiple linear regressions supplemented by iterative thematic analysis of telephone interviews of the same job groups. RESULTS: Non-significant decreases in injury rates and days lost per injury were observed in outsourced CFSWs post outsourcing. Significant decreases (P < 0.05) were observed in average costs per injury for cleaners post outsourcing. Outsourced workers interviewed implied instances of underreporting workplace injuries. CONCLUSIONS: This mixed methods study describes the impact of outsourcing on OHS of healthcare workers in British Columbia. Results will be helpful for policy-makers and workplace regulators to assess program effectiveness for outsourced workers.

Efficacy of an automated ultraviolet C device in a shared hospital bathroom.

Toilet flushing can contribute to disease transmission by generating aerosolized bacteria and viruses that can land on nearby surfaces or follow air currents. Aerobic and anaerobic bacterial bioaerosol loads, and bacterial counts on 2 surfaces in a bathroom with a permanently installed, automated ultraviolet C (UVC) irradiation device, were significantly lower than in a comparable bathroom without the UVC device. Permanently installed UVC lights may be a useful supplementary decontamination tool in shared patient bathrooms.

Influenza virus emitted by naturally-infected hosts in a healthcare setting.

BACKGROUND: The emergence of novel respiratory viruses such as avian influenza A(H7N9) virus and the Middle East Respiratory Syndrome Coronavirus (MERS-CoV) highlights the importance of understanding determinants of transmission to healthcare workers (HCWs) and the public. OBJECTIVES: We aim to determine the viral content of the air emitted by symptomatic inpatients or long-term care residents with laboratory-confirmed influenza virus infection (emitters), and in the breathing zones of healthcare workers who attend to them. DESIGN: A prospective pilot study of patients with laboratory-confirmed influenza virus infection was undertaken. Air within 1m of the patient was sampled using a high volume air sampler. In addition, a lower volume air sampler was placed <1 m from the patient, with another >1 m from the patient. Viral RNA was recovered from the samplers and submitted for quantitative real time PCR. In addition, personal button samplers were provided to HCWs. RESULTS: The air emitted by 15 participants with laboratory-confirmed influenza virus infection was sampled. Of the patients infected with influenza A, viral RNA was recovered from the air emitted by 9/12 patients using the low-volume sampler; no viral RNA was detected from air emitted by patients with influenza B (n=3). Influenza virus RNA was recovered from one HCW's sampler. CONCLUSIONS: Patients with respiratory virus infection emit virus into the air which disperses to >1 m and may reach the breathing zone of a HCW. This pilot study highlights the feasibility and importance of conducting a larger-scale study to identify determinants of exposure and transmission from patient to HCW.

A surface wipe sampling and LC-MS/MS method for the simultaneous detection of six antineoplastic drugs commonly handled by healthcare workers.

An effective wipe sampling and LC-MS/MS method was developed to simultaneously analyze six commonly administered antineoplastic drugs in stainless steel surface. The analyzed drugs were methotrexate, paclitaxel, cyclophosphamide, 5-fluorouracil, vincristine, and oxaliplatin, a frequently prepared antineoplastic drug that has not been included among any of the published simultaneous detection methods. The established method was used to evaluate the recoveries of antineoplastic drugs on brand new and worn stainless steel surfaces by wiping the plates with a Whatman filter paper wetted with 0.5 mL of water/methanol (20:80) with 0.1% formic acid followed by LC-MS/MS before desorbing the filter with a water/methanol (50:50) solution. A significant decrease in the recovery of all evaluated drugs was found when worn plates were used. Additionally, the inter-personnel variability on drug recoveries during wiping procedures was evaluated. Significantly higher recoveries were achieved by the personnel with more training and experience versus personnel without prior experience. Finally, a laboratory stability test was developed to assess the degradation of the antineoplastic drugs during replicated shipping conditions. With the exception of vincristine sulfate which exhibited a significant (p < 0.05) degradation after 48 h, all evaluated drugs were stable during the first 24-48 h. However, after 144 h, an increase in the degradation of all evaluated drugs was observed, with oxaliplatin and 5-fluorouracil exhibiting the most degradation.

Corrigendum to "Identification of Knowledge Gaps Regarding Healthcare Workers' Exposure to Antineoplastic Drugs: Review of Literature, North America versus Europe" [Saf Health Work 5 (4) (2014) 169-174].

[This corrects the article DOI: 10.1016/j.shaw.2014.06.001.].

Identification of Knowledge Gaps Regarding Healthcare Workers' Exposure to Antineoplastic Drugs: Review of Literature, North America versus Europe.

We have been examining the issue of healthcare workers' exposure to antineoplastic drugs for nearly a decade and have observed that there appears to be more publications on the subject matter originating from Europe than from North America. The concern is that findings from Europe may not be generalizable to North America because of differences in handling practices, regulatory requirements, and training. Our objective was to perform a literature review to confirm our observation and, in turn, identify gaps in knowledge that warrants addressing in North America. Using select keywords, we searched for publications in PubMed and Web of Science. All papers were initially classified according to the originating continent and then categorized into one or more subject categories (analytical methods, biological monitoring, occupational exposure, surface contamination, and probability of risk/exposure). Our review identified 16 papers originating from North America and 55 papers from Europe with surface contamination being the subject matter most often studied overall. Based on our results, we are of the opinion that North American researchers need to further conduct dermal and/or urinary drug contamination studies as well as assess the exposure risk faced by healthcare workers who handle antineoplastic drugs. Trends in exposure levels should also be explored.

Examining factors that influence the effectiveness of cleaning antineoplastic drugs from drug preparation surfaces: a pilot study.

Occupational exposure to antineoplastic drugs has been documented to result in various adverse health effects. Despite the implementation of control measures to minimize exposure, detectable levels of drug residual are still found on hospital work surfaces. Cleaning these surfaces is considered as one means to minimize the exposure potential. However, there are no consistent guiding principles related to cleaning of contaminated surfaces resulting in hospitals to adopt varying practices. As such, this pilot study sought to evaluate current cleaning protocols and identify those factors that were most effective in reducing contamination on drug preparation surfaces. Three cleaning variables were examined: (1) type of cleaning agent (CaviCide®, Phenokil II™, bleach and chlorhexidine), (2) application method of cleaning agent (directly onto surface or indirectly onto a wipe) and (3) use of isopropyl alcohol after cleaning agent application. Known concentrations of antineoplastic drugs (either methotrexate or cyclophosphamide) were placed on a stainless steel swatch and then, systematically, each of the three cleaning variables was tested. Surface wipes were collected and quantified using high-performance liquid chromatography-tandem mass spectrometry to determine the percent residual of drug remaining (with 100% being complete elimination of the drug). No one single cleaning agent proved to be effective in completely eliminating all drug contamination. The method of application had minimal effect on the amount of drug residual. In general, application of isopropyl alcohol after the use of cleaning agent further reduced the level of drug contamination although measureable levels of drug were still found in some cases.

Increasing sharp safety device use in healthcare: a semi-structured interview study.

The use of sharp safety devices in healthcare is considered the most important means of preventing occupational percutaneous injuries and has been mandated for use in most hospitals in industrialized countries including in Canada. However, clinical personnel's perceptions on the use of safety devices needs further characterization to improve compliance. This study's objective was to identify healthcare provider perspectives on different aspects of sharp safety device use and on how use could be increased. Using a constant comparison approach, data from semi-structured interviews with 39 nurses, physicians and phlebotomists providing direct patient care, and six nurses acting as the interface between clinical personnel and purchasing departments, were analyzed. Study participants were from three of the six health authorities in British Columbia. The four major categories that emerged from the data were: selection processes; identification and replacement; training; and multi-level barriers and facilitators. Findings highlighted the importance of including personnel regularly using safety devices at each stage of their selection including when they are being considered for replacement with superior devices, as well as the need for appropriate initial and refresher training, and how practices at the hospital, ward and individual level facilitate safety device use.

Evidence of a paradoxical relationship between endotoxin and lung cancer after accounting for left truncation in a study of Chinese female textile workers.

INTRODUCTION: Occupational exposure to endotoxin, found in Gram-negative bacteria in organic material, has been associated predominantly with a reduced risk of lung cancer among workers. An inverse exposure-response gradient among women textile workers in Shanghai, China, has been reported previously. In this case-cohort study, we investigated the influence of left truncation, which can itself induce a downward trend, on the observed association. METHODS: Subjects were enrolled between 1989 and 1991 and followed until 1998. The data were left-truncated as all subjects were hired before baseline. An analysis was performed with 3038 subcohort members and 602 cases of incident lung cancer. To evaluate left truncation, we compared lung cancer rates in those hired longer ago with those hired more recently among unexposed subjects. Cox proportional hazards modelling was used to estimate incident rate ratios (IRRs) and 95% CIs. RESULTS: Among those who were never exposed to workplace endotoxin, we compared lung cancer rates in those hired >35 years before enrolment with workers hired ≤35 years before enrolment and observed a reduced risk in the former group, IRR=0.74, 95% CI (0.51 to 1.07). After accounting for this downward bias from left truncation, the reduced risk associated with endotoxin remained among those hired ≤50 years before enrolment. In contrast, there was suggestion of an increased risk of lung cancer among those hired >50 years ago. CONCLUSIONS: After examination of left truncation bias, an inverse dose-response between endotoxin and lung cancer remained for all subjects except those hired longest ago.

Are health care workers protected? An observational study of selection and removal of personal protective equipment in Canadian acute care hospitals.

BACKGROUND: The proper use of personal protective equipment (PPE) by health care workers (HCWs) is vital in preventing the spread of infection and has implications for HCW safety. METHODS: An observational study was performed in 11 hospitals participating in the Canadian Nosocomial Infection Surveillance Program between January 7 and March 30, 2011. Using a standardized data collection tool, observers recorded HCWs selecting and removing PPE and performing hand hygiene on entry into the rooms of febrile respiratory illness patients. RESULTS: The majority of HCWs put on gloves (88%, n = 390), gown (83%, n = 368), and mask (88%, n = 386). Only 37% (n = 163) were observed to have put on eye protection. Working in a pediatric unit was significantly associated with not wearing eye protection (7%), gown (70%), gloves (77%), or mask (79%). Half of the observed HCWs (54%, n = 206) removed their PPE in the correct sequence. Twenty-six percent performed hand hygiene after removing their gloves, 46% after removing their gown, and 57% after removing their mask and/or eye protection. CONCLUSION: Overall adherence with appropriate PPE use in health care settings involving febrile respiratory illness patients was modest, particularly on pediatric units. Interventions to improve PPE use should be targeted toward the use of recommended precautions (eg, eye protection), HCWs working in pediatric units, the correct sequence of PPE removal, and performing hand hygiene.

Impact of the 2009 influenza A (H1N1) pandemic on Canadian health care workers: a survey on vaccination, illness, absenteeism, and personal protective equipment.

BACKGROUND: Data are limited on the impact of the 2009 H1N1 influenza A pandemic on health care worker (HCW) vaccination, illness, absenteeism, and personal protective equipment (PPE) use. METHODS: A survey was completed by HCWs from 14 hospitals participating in the Canadian Nosocomial Infection Surveillance Program who provided direct care to patients with pH1N1 influenza in high-risk units between September and December 2009. RESULTS: Surveys were returned from 986 HCWs (80% nurses, 14% respiratory therapists, and 6% physicians). HCWs working in an intensive care unit (78%) or a designated influenza ward (67%) were more compliant with wearing an N95 respirator for aerosol-generating medical procedures than those working in an emergency department (47%; P < .001). HCWs who worked in health care for >11 years were more compliant with wearing protective eyewear than those who worked for ≤11 years (69% vs 54%; P < .001). A total of 815 HCWs (83%) reported having received the pH1N1 influenza vaccine, and 372 (38%) reported having received the 2009-2010 seasonal influenza vaccine. Influenza-like illness was reported by 236 (24%) HCWs, 170 of whom (72%) reported missing work. CONCLUSIONS: Experience working in health care improves PPE use and HCWs in emergency departments should be targeted for interventions to improve PPE compliance. pH1N1 influenza vaccine coverage was high, but seasonal influenza vaccine coverage was low, and significant HCW illness and absenteeism were reported.

Pilot assessment of the antineoplastic drug contamination levels in British Columbian hospitals pre- and post-cleaning.

OBJECTIVE: We undertook a pilot study involving six British Columbian hospital pharmacies to determine if antineoplastic drug contamination of surfaces exists and whether residual drugs remain on these surfaces despite being cleaned. METHODS: At each site, the pharmacy technician responsible for preparing the antineoplastic drugs was observed to determine which surfaces were contacted and to ascertain the frequency of contact. Surfaces observed to be most frequently contacted were subsequently wiped after drug preparation pre- and post-clean. The wipe samples were then analyzed by liquid chromatography tandem mass spectrometry to determine the amount of contamination. Cyclophosphamide (CP) and methotrexate (MTX) were used as representative markers to reflect overall antineoplastic drug contamination levels. RESULTS: Fourteen of the 23 surfaces sampled pre-clean (61%) were contaminated with either MTX or CP. The pre-clean wipe samples had a geometric mean concentration of 0.0135 ng/cm(2) for MTX (range

Risk of lung cancer in relation to contiguous windows of endotoxin exposure among female textile workers in Shanghai.

OBJECTIVE: Exposure to endotoxin has been consistently associated with a reduced risk of lung cancer. However, there is a paucity of information regarding temporal aspects of this relationship. The objective of this study was to investigate the associations between contiguous windows of endotoxin exposure and risk of lung cancer. METHODS: Data were reanalyzed from a case-cohort study (602 cases, 3,038 subcohort) of female textile workers in Shanghai, China. Cumulative endotoxin exposure was partitioned into two windows: ≥20 and <20 years before risk. Exposure-response relations were examined using categorical and non-linear (semi-parametric) models, accounting for confounding by previous exposure windows. RESULTS: There was an inverse trend of decreasing risk of lung cancer associated with increasing levels of endotoxin exposure ≥20 years before risk (p trend = 0.02). Women in the highest two categories of cumulative exposures had hazard ratios of 0.78 (95% CI 0.60-1.03) and 0.77 (95% CI 0.58-1.02) for lung cancer, respectively, in comparison with unexposed textile workers. There was, however, a weaker association and not statistically significant between lung cancer and endotoxin exposure accumulated in the more recent window (<20 years before risk). CONCLUSION: Results provide further evidence that endotoxin exposure that occurred 20 years or more before risk confers the strongest protection against lung cancer, indicating a possible early anti-carcinogenic effect. Further studies are needed to better understand the underlying biological mechanisms for this effect.