The statistical merits of various methods of calculating transfer coefficients between environmental media--development of the ideal formula for data-sets with a log-normal distribution

The statistical treatment of data-sets from environmental pollutant studies in which different measurements are combined to produce averages or comparative factors (e.g., transfer coefficients (TCs), input-output balance values) are considered here, with particular reference to the analysis of data from input-output balance studies of pollutants such as PCBs in animals and humans. Many methods of statistical analysis ignore the fact that all measurements are subject to error, and generally assume that the normal distribution applies to all data-sets, which is commonly inappropriate for environmental (and particularly biological system) data. Considerably different estimations can be obtained by applying different, commonly used, statistical methods, as shown in a simulation study presented here and when applied to data from an input-output balance study of PCBs in humans. Alternative average and combined factor estimators for the treatment of data from these types of studies that give considerable advantages in terms of bias and the ease of assessment of accuracy are proposed.     

Method of median semi-variance for the analysis of left-censored data: Comparison with other techniques using environmental data

In environmental monitoring, variables with analytically non-detected values are commonly encountered. For the statistical evaluation of these data, most of the methods that produce a less biased performance require specific computer programs. In this paper, a statistical method based on the median semi-variance (SemiV) is proposed to estimate the position and spread statistics in a dataset with single left-censoring. The performances of the SemiV method and 12 other statistical methods are evaluated using real and complete datasets. The performances of all the methods are influenced by the percentage of censored data. In general, the simple substitution and deletion methods showed biased performance, with exceptions for L/2, Inter and L/√2 methods that can be used with caution under specific conditions. In general, the SemiV method and other parametric methods showed similar performances and were less biased than other methods. The SemiV method is a simple and accurate procedure that can be used in the analysis of datasets with less than 50% of left-censored data.     

Assigning concentration values for dioxin and furan congeners in human serum when measurements are below limits of detection: an observational approach

Concentrations of polychlorinated dibenzo-p-dioxins (PCDDs) and dibenzofurans (PCDFs) in dietary sources for humans have been declining over the previous two decades. These declines have been accompanied by decreases in concentrations of these compounds in humans, as evidenced by measurements in blood and milk. Because of the decreasing concentrations of PCDD/PCDFs in the environment and in humans, measuring PCDD/PCDF congeners in humans has become increasingly difficult, despite advances in analytic methods. An observational approach was recently described to address the quandary of non-detectable results in determining toxic equivalents. This approach, called the congener ratio approach, is specifically for cases where concentrations of 2,3,7,8-TCDD (TCDD) are below the limit of detection (LOD), and where concentrations of 1,2,3,7,8-pentachlorodibenzo-p-dioxin (PeCDD) are equal to or above the LOD. Development of this approach relied on evaluating data on measured concentrations of TCDD and PeCDD in human serum from the general population. The congener ratio approach for TCDD and PeCDD was based on the concentration of TCDD being approximately 40% that of PeCDD in serum from the general population. Additional analyses presented here reveal that when concentrations of both congeners are above the LOD, the data appear to generally support the congener ratio approach for TCDD and PeCDD, with the caveat that gender may affect the ratio. However, the TCDD/PeCDD relation is less clear when TCDD is less than the LOD; in this situation, the relation overpredicts levels of TCDD approximately 80% of the time for the 2001-2002 NHANES database. Using the congener ratio approach for other PCDD/PCDF congeners requires assessing the correlation and the frequency of detection for both TCDD and PeCDD.     

A context-intensive approach to imputation of missing values in data sets from networks of environmental monitors

Although networks of environmental monitors are constantly improving through advances in technology and management, instances of missing data still occur. Many methods of imputing values for missing data are available, but they are often difficult to use or produce unsatisfactory results. I-Bot (short for “Imputation Robot”) is a context-intensive approach to the imputation of missing data in data sets from networks of environmental monitors. I-Bot is easy to use and routinely produces imputed values that are highly reliable. I-Bot is described and demonstrated using more than 10 years of California data for daily maximum 8-hr ozone, 24-hr PM_2.5 (particulate matter with an aerodynamic diameter <2.5 μm), mid-day average surface temperature, and mid-day average wind speed. I-Bot performance is evaluated by imputing values for observed data as if they were missing, and then comparing the imputed values with the observed values. In many cases, I-Bot is able to impute values for long periods with missing data, such as a week, a month, a year, or even longer. Qualitative visual methods and standard quantitative metrics demonstrate the effectiveness of the I-Bot methodology.Implications: Many resources are expended every year to analyze and interpret data sets from networks of environmental monitors. A large fraction of those resources is used to cope with difficulties due to the presence of missing data. The I-Bot method of imputing values for such missing data may help convert incomplete data sets into virtually complete data sets that facilitate the analysis and reliable interpretation of vital environmental data.     

Revised relative potency values for PCDDs, PCDFs, and non-ortho-substituted PCBs for the optimized H4IIE-luc in vitro bioassay

While the World Health Organization 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) equivalency factors are useful estimates of relative potencies of mixtures when conducting risk assessments, they are not useful when comparing the results of bioassays such as the H4IIE-luc to concentrations of TCDD equivalents calculated from instrumental analyses. Since there are thousands of dioxin-like compounds (DLCs), one use of screening assays is to determine if all of the aryl hydrocarbon receptor (AhR) active DLCs in a mixture have been accounted for in instrumental analyses. For this purpose, bioassay-specific relative potency (ReP) values are needed. RePs of 21 polychlorinated dibenzo-p-dioxins, polychlorinated dibenzofurans, and dioxin-like polychlorinated biphenyls that exhibit effects mediated through the AhR were determined by use of the H4IIE-luc assay. Different values of RePs are derived, depending on the statistical, curve-fitting methods used to derive them from the dose–response relationships. Here, we discuss the various methods for deriving RePs from in vitro data and their assumptions and effects on values of RePs. Full dose–response curves of 2,3,7,8-TCDD and other representative DLCs were used to estimate effective concentrations at multiple points (e.g., EC20-50-80), which were then used to estimate ReP of each DLC to 2,3,7,8-TCDD.     

Fabricating data: how substituting values for nondetects can ruin results, and what can be done about it

The most commonly used method in environmental chemistry to deal with values below detection limits is to substitute a fraction of the detection limit for each nondetect. Two decades of research has shown that this fabrication of values produces poor estimates of statistics, and commonly obscures patterns and trends in the data. Papers using substitution may conclude that significant differences, correlations, and regression relationships do not exist, when in fact they do. The reverse may also be true. Fortunately, good alternative methods for dealing with nondetects already exist, and are summarized here with references to original sources. Substituting values for nondetects should be used rarely, and should generally be considered unacceptable in scientific research. There are better ways.     

Environmental fate and bioavailability of agent orange and its associated dioxin during the vietnam war

BACKGROUND: In 1996, the Committee on the Assessment of Wartime Exposure to Herbicides in Vietnam of the National Academy of Sciences' Institute of Medicine (IOM) issued a report on an exposure model for use in epidemiological studies of Vietnam veterans. This exposure model would consider troop locations based on military records; aerial spray mission data; estimated ground spraying activity; estimated exposure opportunity factors; military indications for herbicide use; and considerations of the composition and environmental fate of herbicides, including changes in the TCDD content of the herbicides over time, the persistence of TCDD and herbicides in the environment, and the degree of likely penetration of the herbicides into the ground. When the final report of the IOM Committee was released in October 2003, several components of the exposure model envisioned by the Committee were not addressed. These components included the environmental fate of the herbicides, including changes in the TCDD content over time, the persistence of TCDD and herbicides in the environment, and the degree of likely penetration of herbicides into the ground. This paper is intended to help investigators understand better the fate and transport of herbicides and TCDD from spray missions, particularly in performing epidemiological studies. METHODS: This paper reviews the published scientific literature related to the environmental fate of Agent Orange and the contaminant, 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD), and discusses how this affected the potential exposure to TCDD of ground troops in Vietnam. Specifically, the mechanisms of dissipation and degradation as they relate to environmental distribution and bioavailability are addressed. RESULTS: The evaluation of the spray systems used to disseminate herbicides in Vietnam showed that they were capable of highly precise applications both in terms of concentrations sprayed and area treated. Research on tropical forest canopies with leaf area indices (a measure of foliage density) from 2 to 5 indicated that the amount of herbicide and associated TCDD reaching the forest floor would have been between 1 and 6% of the total aerial spray. Studies of the properties of plant surface waxes of the cuticle layer suggested that Agent Orange, including the TCDD, would have dried (i.e., be absorbed into the wax layer of the plant cuticle) upon spraying within minutes and could not be physically dislodged. Studies of Agent Orange and the associated TCDD on both leaf and soil surface have demonstrated that photolysis by sunlight would have rapidly decreased the concentration of TCDD, and this process continued in shade. Studies of 'dislodgeable foliar residues' (DFR, the fraction of a substance that is available for cutaneous uptake from the plant leaves) showed that only 8% of the DFR was present 1 hr after application. This dropped to 1% of the total 24 hrs after application. Studies with human volunteers confirmed that after 2 hrs of saturated contact with bare skin, only 0.15-0.46% of 2,4,5-T, one of the phenoxy acetic acid compounds that was an active ingredient of Agent Orange, entered the body and was eliminated in the urine. CONCLUSIONS: The prospect of exposure to TCDD from Agent Orange in ground troops in Vietnam seems unlikely in light of the environmental dissipation of TCDD, little bioavailability, and the properties of the herbicides and circumstances of application that occurred. Photochemical degradation of TCDD and limited bioavailability of any residual TCDD present in soil or on vegetation suggest that dioxin concentrations in ground troops who served in Vietnam would have been small and indistinguishable from background levels even if they had been in recently treated areas. Laboratory and field data reported in the literature provide compelling evidence on the fate and dislodgeability of herbicide and TCDD in the environment. This evidence of the environmental fate and poor bioavailability of TCDD from Agent Orange is consistent with the observation of little or no exposure in the veterans who served in Vietnam. Appreciable accumulation of TCDD in veterans would have required repeated long-term direct skin contact of the type experienced by United States (US) Air Force RANCH HAND and US Army Chemical Corps personnel who handled or otherwise had direct contact with liquid herbicide, not from incidental exposure under field conditions where Agent Orange had been sprayed.     

Hemopoietic cell kinetics after intraperitoneal single injection of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) in mice

2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD) is a widely spread environmental pollutant. Homopoietic system is one of the targets of TCDD in laboratory animals including monkeys. The present study is the hemopoietic cell kinetics in mice, from the severe depression in cellularity of bone marrow and CFU-GM, to their recovery after the intraperitoneal injection of high dosage of 2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD). The bone-marrow cellularity and CFU-GM were severely decreased to 37.8% and 48% of the control, respectively until day 1 after exposure to TCDD. They were, however, soon recovered, even overshot the control value. Subsequently, they tended to show decrease and oscillation again to and under the control value. In conclusion, our cell kinetic study has proven the oscillation in bone-marrow cellularity and CFU-GM during the recovery period, of which the observation seems to be useful to extend our understanding in the hematotoxicity of TCDD.     

The reliability of 2,3,7,8-tetrachlorodibenzo-p-dioxin and toxicity equivalence measurements in Missouri residents

This study evaluates the reliability of serum measurements of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) and toxicity equivalence (TEQ) in individuals with no prior exposure to TCDD or dioxin-like compounds above background levels. Reliability was measured over three time-periods in the same individuals. TCDD and TEQ were measured in 127 and 128 participants. The values declined during the testing periods. Reliability coefficients for TCDD was 0.69 and for TEQ 0.91 but increased to 0.79 and 0.92, respectively, after removal of three outlying cases. Study results show good reliability over 22 months suggesting a consistency in the uptake and excretion of TCDD and other dioxin-like compounds.     

Statistical Methods and Pitfalls in Environmental Data Analysis

This paper reviews four commonly used statistical methods for environmental data analysis and discusses potential pitfalls associated with application of these methods through real case study data. The four statistical methods are percentile and confidence interval, correlation coefficient, regression analysis, and analysis of variance (ANOVA). The potential pitfall for estimation of percentile and confidence interval includes the automatic assumption of a normal distribution to environmental data, which so often show a log-normal distribution. The potential pitfall for correlation coefficient includes the use of a wide range of data points in which the maximum in value may trivialize other smaller data points and consequently skew the correlation coefficient. The potential pitfall for regression analysis includes the propagation of uncertainties of input variables to the regression model prediction, which may be even more uncertain. The potential pitfall for ANOVA includes the acceptance of a hypothesis as a weak argument to imply a strong conclusion. As demonstrated in this paper, we may draw very different conclusions based on statistical analysis if the pitfalls are not identified. Reminder and enlightenment obtained from the pitfalls are given at the end of this article.     

Statistical Methods and Pitfalls in Environmental Data Analysis

This paper reviews four commonly used statistical methods for environmental data analysis and discusses potential pitfalls associated with application of these methods through real case study data. The four statistical methods are percentile and confidence interval, correlation coefficient, regression analysis, and analysis of variance (ANOVA). The potential pitfall for estimation of percentile and confidence interval includes the automatic assumption of a normal distribution to environmental data, which so often show a log-normal distribution. The potential pitfall for correlation coefficient includes the use of a wide range of data points in which the maximum in value may trivialize other smaller data points and consequently skew the correlation coefficient. The potential pitfall for regression analysis includes the propagation of uncertainties of input variables to the regression model prediction, which may be even more uncertain. The potential pitfall for ANOVA includes the acceptance of a hypothesis as a weak argument to imply a strong conclusion. As demonstrated in this paper, we may draw very different conclusions based on statistical analysis if the pitfalls are not identified. Reminder and enlightenment obtained from the pitfalls are given at the end of this article.     

On the translation of uncertainty from toxicokinetic to toxicodynamic models--the TCDD example

When estimating human health risks from exposure to TCDD using toxicokinetic and toxicodynamic models, it is important to understand how model choice and assumptions necessary for modeling add to the uncertainty of risk estimates. Several toxicokinetic models have been proposed for the risk assessment of dioxins, in particular the elimination kinetics in humans has been a matter of constant debate. For a long time, a simple linear elimination kinetics has been common choice. Thus, it was used for the statistical analysis of the largest occupationally exposed cohort, the German Boehringer cohort. We challenge this assumption by considering, amongst others, a nonlinear modified Michaelis-Menten-type elimination kinetics, the so-called Carrier kinetics. Using the area under the lipid TCDD concentration time curve as dose metrics, we model the time to cancer-related death using the Cox proportional hazards model as toxicodynamic model. This risk assessment set-up was simulated in order to quantify uncertainty of both the dose (TCDD body burden) and the risk estimates, depending on the use of the kinetic model, variations of carcinogenic effect of TCDD and variations of latency period (lag time). If past exposure is estimated assuming a linear elimination kinetics although a Carrier kinetics actually holds, then high exposures in reality will be underestimated through statistical analysis and low exposures will be overestimated, respectively. This bias will carry over on the estimated individual concentration-time curves and the therefrom derived TCDD dose metric values. Using biased dose values when estimating a dose-response relationship will finally lead to biased risk estimates. The extent of bias and the decrease of precision are quantified in selected scenarios through this simulation approach. Our findings are in concordance with recent results in the field of dioxin risk assessment. They also reinforce the general demand for the scheduled uncertainty assessments in risk analyses.     

Seveso Women's Health Study: a study of the effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin on reproductive health

Although reproductive effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) exposure have been reported in numerous investigations of animals, studies of this association in humans are limited. In 1976, an explosion in Seveso, Italy exposed the surrounding population to among the highest levels of TCDD recorded in humans. The relatively pure exposure to TCDD and the ability to quantify individual level TCDD exposure from sera collected in 1976 for the Seveso cohort affords a unique opportunity to evaluate the potential dose-response relationship between TCDD exposure and a spectrum of reproductive endpoints. The Seveso Women's Health Study (SWHS) is the first comprehensive study of the reproductive health of a human population exposed to TCDD. The primary objectives of the study are to investigate the relationship of TCDD and the following endpoints: (1) endometriosis; (2) menstrual cycle characteristics; (3) age at menarche; (4) birth outcomes of pregnancies conceived after 1976; (5) time to conception and clinical infertility; and (6) age at menopause. Included in the SWHS cohort are women who were 0-40 yr old in 1976, who have adequate stored sera collected between 1976 and 1980, and who resided in Zones A or B at the time of the accident. All women were interviewed extensively about their reproductive and pregnancy history and had a blood draw. For an eligible subset of women, a pelvic exam and transvaginal ultrasound were conducted and a menstrual diary was completed. More than 95% of the women were located 20 yr after the accident and roughly 80% of the cohort agreed to participate. Data collection was completed in July 1998, serum TCDD analysis of samples for analysis of endometriosis as a nested case-control study was completed in October 1998, and statistical analysis of these data should be completed in early 1999. Serum samples are now being analyzed in order to relate TCDD levels with the remaining reproductive outcomes.     

T cell-derived IL-5 production is a sensitive target of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD)

The immune system is one of the organs most vulnerable to the toxicity of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). Among the various immunotoxic effects of TCDD, the thymus involution and suppression of IgM antibody production are well known sensitive reactions of the thymocytes and B cells affected by TCDD. Recently, we reported that TCDD greatly inhibits the production of type-2 helper T (Th2) cell-derived cytokines, especially IL-5, by the splenocytes in mice immunized with ovalbumin (OVA). In the present study, we investigated the dose-dependency of these TCDD immunotoxic effects in OVA-immunized mice to identify the most sensitive target. Mice of two age groups, 6 weeks old and 3 weeks old, were dosed with 0.3, 1.0, or 3.0 microg TCDD/kg and immunized with OVA using alum as an adjuvant. Seven days later, the thymus weight, thymocyte population, antigen-specific IgM in the plasma, and IL-5 production by the splenocytes were examined. Among them, IL-5 production was significantly suppressed by all three doses of TCDD and reduced to about 30% by even a small dose of 0.3 microg TCDD/kg in both age groups. The thymus weight was significantly reduced by 1.0 microg or 3.0 microg TCDD/kg, but IgM production was not affected by up to 3.0 microg/kg of TCDD in both age groups. Taken together, the Th2 cell-derived IL-5 production was the most sensitive endpoint detecting TCDD toxicity among those examined. Our results also suggest that effector T cells are targets more vulnerable to TCDD toxicity than thymocytes or antibody-producing B cells in the OVA-immunized mice.     

On the statistical analysis of ambient ozone data when measurements are missing

Two methods are proposed to deal with missing observations, which frequently pose problems in the statistical analysis of ambient ozone data. The first is based on exceedances of the data over thresholds and provides a flexible and general class of models for statistical analysis of air pollution data. The second uses the measured values of related variables to impute missing observations. They are applied to data for three sites in E Texas.     

Evaluation of the thermal/optical reflectance method for quantification of elemental carbon in sediments

The IMPROVE thermal/optical reflectance (TOR) method, commonly used for EC quantification in atmospheric aerosols, is applied to soils and sediments and compared with a thermochemical method commonly applied to these non-atmospheric samples. TOR determines elemental carbon (EC) by an optical method, but it also yields thermally defined EC fractions in a 2% O2/98% He oxidizing atmosphere at 550 degrees C (EC1), 700 degrees C (EC2), and 800 degrees C (EC3). Replicate TOR TC, OC, and EC values exhibited precisions of approximately +/-10% as determined from multiple analyses of the same samples. EC abundances relative to total mass concentrations were within the ranges reported by other methods for diesel exhaust soot, n-hexane soot, wood and rice chars, and coals, as well as for environmental matrices. A direct comparison with the chemothermal (CTO) method of Gustafson et al. for ten soil and sediment samples demonstrated that almost all of the OC and EC1 are eliminated, as is part of the EC2. The CTO soot carbon is bounded by the EC3 and EC2+EC3 fractions of the IMPROVE TOR analysis. It might be possible to adjust these fractions to obtain better agreement between atmospheric aerosol and soil/sediment analysis methods. Given its linking the EC measurement in the atmosphere to sediments, the TOR method will not only provide useful information on the explanation and comparison between different environmental matrices, but also can be used to derive information on global cycling of EC.     

Relationships between Aerosol Extinction Coefficients Derived from Airport Visual Range Observations and Alternative Measures of Airborne Particle Mass

Epidemiological assessments of population exposures to airborne particles are often hampered by the scarcity of available fine particle mass measurements. In an attempt to overcome this serious problem, we analyze In this paper methods for predicting fine particle (M ₁) and Inhalable particle (IP) mass concentrations using relative humidity corrected light extinction coefficient (b _ext) estimated from airport visual range (V _r) observations. The analyses presented are based on theoretical determinations as well as statistical investigations utilizing EPA's NASN and Inhalable Particle Monitoring Network (IPMN) data bases and routine airport visual range observations In twelve large U.S. cities. Our results Indicate that, after controlling for certain limitations of airport visual range data, most of the regression models developed in this paper can be applied satisfactorily to predict M _t and IP. Furthermore, our findings Indicate that a more representative formula than the commonly used meteorological range formula to predict atmospheric b _ext values in urban areas may be b _ext = (1.8 ± 0.04)/V _r. Because of known local or regional influences, however, we do suggest calibration of any predictive model which utilizes airport visibility data against site-specific aerometric data on particle mass concentrations or scattering coefficient measurements.     

Substituting missing data in compositional analysis

Multivariate analysis of environmental data sets requires the absence of missing values or their substitution by small values. However, if the data is transformed logarithmically prior to the analysis, this solution cannot be applied because the logarithm of a small value might become an outlier. Several methods for substituting the missing values can be found in the literature although none of them guarantees that no distortion of the structure of the data set is produced. We propose a method for the assessment of these distortions which can be used for deciding whether to retain or not the samples or variables containing missing values and for the investigation of the performance of different substitution techniques. The method analyzes the structure of the distances among samples using Mantel tests. We present an application of the method to PCDD/F data measured in samples of terrestrial moss as part of a biomonitoring study.     

In vitro assessment of AhR-mediated activities of TCDD in mixture with humic substances

Humic substances (HS) are ubiquitous natural products of decomposition of dead organic matter. HS is present in most freshwaters at concentrations ranging from 0.5 to 50 mg L⁻¹. Organic carbon can represent 20% dry weight of sediments. Recently, the interaction of dissolved HS with the aryl hydrocarbon receptor (AhR) has been demonstrated. The AhR is a cytosolic receptor to which persistent organic pollutants (POPs) can bind and many of their toxic effects are mediated through interactions with this receptor. We describe in vitro effects (using H4IIE-luc cells) of binary mixtures of various HS with 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD), since in contaminated environments these compounds occur simultaneously. Six out of 12 HS samples activated AhR even at environmentally relevant concentrations (17 mg L⁻¹), but did not reach the full AhR-activation even at excessive concentration. In simultaneous exposure of H4IIE-luc to HS (17 mg L⁻¹) and TCDD (1.2 pM) without any preincubation prior to exposure, either significant additive or facilitative effects were observed. No negative interactions, due to possible sorption of TCDD to HS was observed. Nevertheless, if the HS–TCDD binary mixture was preincubated for 6 days prior to the exposure on H4IIE-luc cells, the additive and facilitative effects were less due to possible sorption of TCDD onto HS. Similar results were obtained from analogous experiments with greater concentrations of both TCDD and HS.