Biological test method for determining toxicity of sediment using luminescent bacteria: chapter 8

Section 6: Data Analysis and Interpretation

• 6.1 Data Analysis
• 6.2 Interpretation of Results

6.1 Data Analysis

The mean and standard deviation of the light readings for the control solutions used in the study (see Sections 4.6.4, 4.7, and 4.8) must be calculated. These values are used to calculate the coefficient of variation of the mean for the control solutions, which is used as the criterion described herein for judging if the test results are valid (Section 4.8).

A study performed according to this reference method should include one or more samples of test (contaminated or potentially contaminated) sediment together with one or more samples of reference sediment. Additionally, the inclusion of one or more samples of positive control sediment (see Sections 4.4 and 5) is recommended. Tests involving one or more samples of coarse-grained sediment (i.e., sediment with <20% fines) must also include one or more samples of negative control sediment (artificial or natural) or clean reference sediment with a percent fines content that does not differ by more than 30% from that of the coarse-grained test sediment(s) (see Section 6.2). In each instance, the statistical endpoint to be calculated for each of these test materials is the ICp (inhibiting concentration for a specified percent effect).

Unless specified otherwise by regulatory requirements or by design, the endpoint for this reference method is the concentration causing 50% inhibition of light, i.e., the IC50. The calculations to estimate the IC50 and its 95% confidence limit are included in the most recent (1999 or later) Omni^TM software packages (Version 1.18 or equivalent) formerly marketed by AZUR Environmental Ltd. and now available from Strategic Diagnostics Inc. (see Section 2.2 for contact information).^Footnote 13 Alternatively, guidance for estimating IC50 (together with its 95% confidence limit) is provided in EC (1992), and other statistical software packages are available which enable this calculation (EC, 1997b; 1997c; 2002b).

In the absence of a computer with appropriate software, the IC50 can be calculated using one of the following two equations and approaches.

For each test concentration, Gamma (Γ) is calculated (ASTM, 1995) as:

Γ = (I_c/I_t) - 1

where:

I_c = the average light reading of filtrates of the control solutions, and

I_t = the light reading of a filtrate of a particular concentration of the test material.

The Gamma values for each test filtrate falling within 0.02 < Gamma < 200 are plotted manually, and a line is fitted by eye. Then, the IC50 is read off as the concentration that corresponds to a Gamma of 1.0. The manually plotted points should be checked against the observed readings to guard against errors in entry and anomalous estimates of IC50. The manual plot and its estimated IC50 should also be checked against any computer-generated graph and the computer calculation of the IC50 (EC, 1992).

Alternatively, a linear regression of logC (concentration, on the ordinate) vs. logΓ (on the abscissa) is computed according to ASTM (1995):

logΓ = b(logC) + log(a)

In this equation, ‘b’ is the slope and ‘log(a)’ is the intercept of the regression line with the ordinate (y-axis) at logΓ = 0, corresponding to Γ = 1. Therefore, ‘a’ obtained as 10^log(a) is the IC50.

The IC50 and its associated values for the 95% confidence limits must be converted to and expressed as mg/L on a dry-weight basis. This is achieved using the dry-weight data (see Section 4.6.2) (ASTM, 1995). The IC50 (as well as the upper and lower value of the confidence limits) of the wet sediment is multiplied by the average of the ratios of the dry-to-wet subsample weights:

IC50 = IC50_w × [(S1_d/S1_w) + (S2_d/S2_w) + (S3_d/S3_w)]/3

where:

IC50_w is the calculated IC50 (or its 95% confidence limits) of the wet sediment sample,

S1_d through S3_d are the dry weights of the sediment subsamples from Section 4.6.2, and

S1_w through S3_w are the corresponding wet weights.

These calculations can be expedited by entering the weights and IC50 values into a spreadsheet and using the necessary formulae.

Investigators should consult Environment Canada (2002b) for detailed guidance regarding appropriate statistical endpoints and their calculation. The objectives of the data analysis are: to quantify contaminant effects on test organisms exposed to various samples of test (contaminated or potentially contaminated) sediment; to determine if these effects are statistically different from those occurring in a reference sediment; and to reach a decision as to sample toxicity (Section 6.2). Initially, ICp (normally, IC50) is calculated for each of the samples (including those representing the field-collected reference sediment).

Depending on the study design and objectives, an appropriate number (typically, ≥5/station, each depth of interest) of replicate samples of field-collected test and reference sediments should be collected and tested (see Section 4.1). Each series of toxicity tests must include a minimum of three replicate control solutions (see Sections 4.6 and 4.8), and one or more test sediments. Test sediment might be represented by replicate samples of dredged material from a particular depth or locale (sampling station) of interest, or replicate samples of field-collected sediment from a particular station within or adjacent to an ocean disposal site. Alternatively, test sediment might be represented by one or more subsamples (i.e., laboratory replicates) of a single (non-replicated) sample of sediment from a particular sampling station or site-specific depth (see Section 4.1). The same number of replicates per treatment (i.e., test sediment from a particular sampling station and site-specific depth) and per sample should be used in the test wherever possible, to maximize statistical power and robustness (EC, 2002b).

6.2 Interpretation of Results

Interpretation of results is not necessarily the sole responsibility of the laboratory personnel undertaking the test; this might be a shared task which includes an environmental consultant or other qualified persons responsible for reviewing and interpreting the findings.

Environment Canada (1999b) provides useful advice for interpreting and applying the results of toxicity tests with environmental samples, and should be referred to for guidance in these respects. Initially, the investigator should examine the results and determine if they are valid. In this regard, the criterion for a valid test (see Section 4.8) must be met. Additionally, it is recommended that the dose-response curve for each sample of test sediment be examined to confirm that light loss decreases as test concentration decreases, in an approximately monotonic manner. If not (e.g., if one or more data points appear to be “out of place” with respect to the others), consideration should be given to repeating the test for that sample.

The findings of the reference toxicity test which was initiated with the same lot of Bacterial Reagent as that used in the sediment toxicity test (see Section 5) should be considered during the interpretive phase of the investigation. These results, when compared with historic test results derived by the testing facility using the same reference toxicant and test procedure (i.e., by comparison against the laboratory’s warning chart for this reference toxicity test), will provide insight into the sensitivity of the test organisms as well as the laboratory’s testing precision and performance for a reference toxicity test with V. fischeri.

All data representing the known physicochemical characteristics of each sample of test material (including that for any samples of reference sediment or negative control sediment included in the study) should be reviewed and considered when interpreting the results. The analytical data determined for whole sediment and pore water (see Section 4.3) should be compared with the known influence of these variables on light production by V. fischeri.

Concentrations of porewater ammonia and/or hydrogen sulphide can be elevated in samples of dredged material or field-collected sediment. The elevated levels might be due to organic enrichment from natural and/or anthropogenic (man-made) sources. The known influence of ammonia (see, for example, Tay et al., 1998 and McLeay et al., 2001) and hydrogen sulphide (Brouwer and Murphy, 1995; Tay et al., 1998) on the inhibition of light production by V. fischeri should be considered together with measured concentrations of these variables in the pore water of test samples, when considering and interpreting results for field-collected samples of test and reference sediments.

Observations of turbid or highly coloured filtrates analyzed for light emission by V. fischeri (see Section 4.8) should be considered when reviewing and interpreting the test results.

A number of variables besides toxicity can interfere with readings of light production by V. fischeri surviving in the filtrate of each test concentration, and thus can confound the interpretation of the test results. Investigators performing this reference method, as well as those interpreting the findings, should be aware of these confounding factors and their implications in terms of judging if test materials are toxic or not. Variables which can interfere with the light production of V. fischeri in test filtrates include (ASTM, 1995; Ringwood et al., 1997; Tay et al., 1998):

• sorption of V. fischeri to sediment particles (especially fine-grained ones) retained on the filter; and the resulting loss of transfer of these luminescent bacteria to the filtrate;
• sorption of V. fischeri to the filter, and the resulting loss of transfer of these luminescent bacteria to the filtrate;
• optical interference of the filtrate, due to colour (light absorption) and/or turbidity (light scatter); and
• loss of metabolic activity of V. fischeri transferred to and surviving in the filtrate, due to toxic and/or handling and mechanical stress.

The grain size of test sediments can be a significant confounding factor, since an increasing percentage of clay in the test material has been demonstrated to cause a proportionate decrease in resulting IC50s determined for V. fischeri recovered in filtrates of uncontaminated sediment. Samples of uncontaminated sediment comprised primarily of sand-sized particles (e.g., 0-5% fines) characteristically yield an IC50 of 28 000 to >100 000 mg/L in a V. fischeri solid-phase assay (Cook and Wells, 1996; Ringwood et al., 1997; Tay et al., 1998). IC50s show a “precipitous drop” (Benton et al., 1995; Ringwood et al., 1997) when the percentage of fines in uncontaminated sediment increases from 5 to ~20%, whereupon the IC50 might range from, say, 5000 to 15 000 mg/L depending on the nature of the fines (e.g., percent clay and percent silt) (Ringwood et al., 1997; Tay et al., 1998; McLeay et al., 2001). Higher percentages of fines in uncontaminated sediment typically show a “leveling off” of further declines in IC50s associated with increasing sediment fines. V. fischeri solid-phase tests with 100% kaolin clay have reported IC50s ranging from 1373 to 2450 mg/L (Ringwood et al., 1997; Tay et al., 1998). In an interlaboratory study to validate this reference method, IC50s for a sample of 100% kaolin clay ranged from 1765 to 2450 mg/L (McLeay et al., 2001). Together, these findings support the following new interim guidelines for judging samples as toxic or not, according to the V. fischeri solid-phase assay. These new guidelines take into account the percentage of fines in the test sediment and the known sharp inflection of values when their fines content reaches or exceeds 20% (Ringwood et al., 1997), as well as the ability of a sample of test material comprised of 100% clay to reduce the IC50 to as low as 1765 mg/L using this reference method (McLeay et al., 2001).

Two interim guidelines for judging the toxicity of samples of test sediment using this reference method are discussed in the following paragraphs. The first one, which has been recommended and applied by Environment Canada in the past (EC, 1996b; Porebski and Osborne, 1998), is based on the premise that all samples are toxic according to this biological test method if their IC50 is <1000 mg/L, regardless of grain size characteristics. The second guideline is based on the premise that samples with <20% fines might be toxic at an IC50 ≥ 1000 mg/L, since confounding grain size effects are appreciably less in coarse-grained sediment.

The first interim guideline should be applied to all samples of test sediment with ≥20% fines, as well as to any sample with <20% fines which has an IC50 < 1000 mg/L. The second interim guideline should be applied to all samples of test sediment with <20% fines that have an IC50 ≥ 1000 mg/L. Applying the second interim guideline to samples of sediment with <20% fines and IC50s ≥ 1000 mg/L enables toxic coarse-grained sediments to be identified as such when their IC50 is appreciably higher than 1000 mg/L. It is recommended that the second interim guideline be applied to each sample of test sediment with <20% fines, except in the instance where the IC50 is <1000 mg/L in which case the sample should be judged as toxic and the second guideline does not apply.

For Guideline 2, and throughout this reference method, “fines” refers to (sediment) particles which are ≤ 0.063 mm in size. Measurements of percent fines include all particles defined as silt (i.e., particles ≤ 0.063 mm but ≥ 0.004 mm) or clay (i.e., particles <0.004 mm).

The first condition for Guideline 2 is tested to see if it is met using the following examples for calculations as a guide: If the sample of reference or negative control sediment used to judge the toxicity of the course-grained test sediment has an IC50 of 20 000 mg/L, the IC50 of the test sediment must be <10 000 mg/L. Similarly, if the sample of reference or negative control sediment used to judge the toxicity of the course-grained test sediment has an IC50 of 5050 mg/L, the IC50 of the test sediment must be <2025 mg/L.

The second condition for Guideline 2 must be tested to see if it is met using the pairwise comparison of the values for the two IC50s and their 95% confidence limits, which is described in Sprague and Fogels (1977) as a means of comparing two LC50s.^Footnote 16 When testing for a significant difference, p<0.05 is to be used as the distinguishing effect level.