Reaction mass of diethyl hydrogen 2-hydroxypropane-1,2,3-tricarboxylate and ethyl dihydrogen 2-hydroxypropane-1,2,3-tricarboxylate and triethyl citrate

Endpoint summary

Description of key information

LC 50 (oral, rat): 3500 - 8643 mg/kg

Key value for chemical safety assessment

Acute toxicity: via oral route

Link to relevant study records

Reference

Reference 1

Endpoint:

acute toxicity: oral

Type of information:

(Q)SAR

Adequacy of study:

weight of evidence

Study period:

January 16, 2018

Reliability:

2 (reliable with restrictions)

Rationale for reliability incl. deficiencies:

results derived from a valid (Q)SAR model and falling into its applicability domain, with adequate and reliable documentation / justification

Justification for type of information:

The ACD/Labs Acute Toxicity LD50 module provides fast and accurate predictions of LD50 value for rats according to various routes of administration.
For the validation studies, predictions that were considered inaccurate or unreliable were not included in the statistical results. To filter out unreliable predictions, the Reliability Index value provided by the software with each prediction was considered. The Reliability Index (RI) is a value between 0 and 1and represents an indicator of how well the local chemical space around a particular compound is represented within the training set of the model.
The RI index represents an effective way for users of the software to focus only on those predictions that are of high confidence.
Generally speaking, during the model validation process, predictions that provided an RI value less than 0.3 were considered unreliable and excluded from the results.

Qualifier:

equivalent or similar to guideline

Guideline:

other: ECHA Guidance on information requirements and chemical safety assessment - Chapter R.06: QSARs and grouping of chemicals

Principles of method if other than guideline:

SCOPE
ACD/Labs Percepta provides predictions of LD50 values for rats and mice according to various routes of administration. The program requires as input the molecular structure (SMILES) of the substance.

(Q)SAR PREDICTION METHODOLOGIES
Prior to modelling, the original experimental data were converted to logarithmic form (pLD50) in order to maintain linear relationship with used descriptors. The final prediction results returned to the user are converted back to LD50 value (mg/kg).
The predictive model for pLD50 were derived using GALAS (Global, Adjusted Locally According to Similarity) modelling methodology. 404 fragmental descriptors were used for the development of the GALAS model, which is a combination of two systems:

• Global (baseline) statistical model based on PLS with multiple bootstrapping, using a predefined set of fragmental descriptors.
• Local correction to baseline prediction based on the analysis of model performance for similar compounds from the training set (the so called Self-training Library).

The fragmental descriptor set was identified based on general knowledge and considerations regarding all possible chemical structures and include all the fragments, even those that are not detected in the training set molecules at all. The major part of the utilized fragment set was intended for the description of the general chemical constitution of any compound and comprised conventional fragmental descriptors, such as atoms, functional groups, molecular ‘shape fragments’, etc. This initial set was expanded with a group of more complex fragments, called toxicophores, i.e. substructures identified to be responsible for the toxic action of the molecules possessing them.

The local part of the model provides the basis for estimating reliability of prediction by the means of calculated Reliability Index (RI) values.
Reliability Index (RI) is given as a product of two indices: RI = SI (Similarity Index) * DMCI (Data-Model Consistency Index). The following two criteria are applied for reliability estimation:

• Similarity of the analyzed molecule to compounds in the Self-training Library (a reliable prediction cannot be made if no similar compounds have been found in the Library).
• Consistency of model predictions with experimental data for similar compounds (highly variable LD50 values for similar molecules lead to lower RI values).

The Applicability Domain limits are the following:
a) RI < 0.3: unreliable prediction
b) 0.3 < RI < 0.5: borderline reliability of prediction
c) 0.5 < RI < 0.75: moderate reliable prediction
d) RI > 0.75: high reliable prediction

GLP compliance:

not specified

Test type:

other: in-silico

Species:

rat

Route of administration:

oral: unspecified

Statistics:

Internal validation
Data Set R2 RMSE Coverage of the entire
test set (N= 4893)
Test Set (RI > 0.3) N= 1976 0.56 0.59 91.2%
Test Set (RI > 0.5) N= 1335 0.64 0.54 61.6%
Test Set (RI > 0.75) N= 290 0.75 0.43 13.4%

External validation
RI range R2 RMSE MAE(*) Coverage of the entire
validation set (N= 2718)
RI > 0.3 (N= 2501) 0.63 0.60 0.44 92.0%
RI > 0.5 (N= 1804) 0.70 0.55 0.40 66.4%
RI > 0.75 (N= 430) 0.81 0.44 0.30 15.8%

(*) Mean Absolute Error (MAE) values, being a primary measure of average prediction error in the original study involving the external validation dataset, have been also determined in this case and are reported in the above table for the result comparison purposes.

Sex:

not specified

Dose descriptor:

LD50

Effect level:

5 800 mg/kg bw

Based on:

not specified

Remarks on result:

other: Reliability Index (RI)= 0.73 (moderate)

Interpretation of results:

GHS criteria not met

Conclusions:

The oral median lethal dose (LD50) in the rat of 1,2-diethyl citrate was estimated, with the software tool ACD/Labs Percepta, at 5800 mg/Kg bw.

Executive summary:

The oral median lethal dose (LD50) in the rat of 1,2-diethyl citrate was estimated, with the software tool ACD/Labs Percepta, at 5800 mg/Kg bw.

Reference 2

Endpoint:

acute toxicity: oral

Type of information:

(Q)SAR

Adequacy of study:

weight of evidence

Study period:

January 16, 2018

Reliability:

2 (reliable with restrictions)

Rationale for reliability incl. deficiencies:

results derived from a valid (Q)SAR model and falling into its applicability domain, with adequate and reliable documentation / justification

Justification for type of information:

The ACD/Labs Acute Toxicity LD50 module provides fast and accurate predictions of LD50 value for rats according to various routes of administration.
For the validation studies, predictions that were considered inaccurate or unreliable were not included in the statistical results. To filter out unreliable predictions, the Reliability Index value provided by the software with each prediction was considered. The Reliability Index (RI) is a value between 0 and 1and represents an indicator of how well the local chemical space around a particular compound is represented within the training set of the model.
The RI index represents an effective way for users of the software to focus only on those predictions that are of high confidence.
Generally speaking, during the model validation process, predictions that provided an RI value less than 0.3 were considered unreliable and excluded from the results.

Qualifier:

equivalent or similar to guideline

Guideline:

other: ECHA Guidance on information requirements and chemical safety assessment - Chapter R.06: QSARs and grouping of chemicals

Principles of method if other than guideline:

SCOPE
ACD/Labs Percepta provides predictions of LD50 values for rats and mice according to various routes of administration. The program requires as input the molecular structure (SMILES) of the substance.

(Q)SAR PREDICTION METHODOLOGIES
Prior to modelling, the original experimental data were converted to logarithmic form (pLD50) in order to maintain linear relationship with used descriptors. The final prediction results returned to the user are converted back to LD50 value (mg/kg).
The predictive model for pLD50 were derived using GALAS (Global, Adjusted Locally According to Similarity) modelling methodology. 404 fragmental descriptors were used for the development of the GALAS model, which is a combination of two systems:

• Global (baseline) statistical model based on PLS with multiple bootstrapping, using a predefined set of fragmental descriptors.
• Local correction to baseline prediction based on the analysis of model performance for similar compounds from the training set (the so called Self-training Library).

The fragmental descriptor set was identified based on general knowledge and considerations regarding all possible chemical structures and include all the fragments, even those that are not detected in the training set molecules at all. The major part of the utilized fragment set was intended for the description of the general chemical constitution of any compound and comprised conventional fragmental descriptors, such as atoms, functional groups, molecular ‘shape fragments’, etc. This initial set was expanded with a group of more complex fragments, called toxicophores, i.e. substructures identified to be responsible for the toxic action of the molecules possessing them.

The local part of the model provides the basis for estimating reliability of prediction by the means of calculated Reliability Index (RI) values.
Reliability Index (RI) is given as a product of two indices: RI = SI (Similarity Index) * DMCI (Data-Model Consistency Index). The following two criteria are applied for reliability estimation:

• Similarity of the analyzed molecule to compounds in the Self-training Library (a reliable prediction cannot be made if no similar compounds have been found in the Library).
• Consistency of model predictions with experimental data for similar compounds (highly variable LD50 values for similar molecules lead to lower RI values).

The Applicability Domain limits are the following:
a) RI < 0.3: unreliable prediction
b) 0.3 < RI < 0.5: borderline reliability of prediction
c) 0.5 < RI < 0.75: moderate reliable prediction
d) RI > 0.75: high reliable prediction

GLP compliance:

not specified

Test type:

other: in-silico

Species:

rat

Route of administration:

oral: unspecified

Statistics:

Internal validation
Data Set R2 RMSE Coverage of the entire
test set (N= 4893)
Test Set (RI > 0.3) N= 1976 0.56 0.59 91.2%
Test Set (RI > 0.5) N= 1335 0.64 0.54 61.6%
Test Set (RI > 0.75) N= 290 0.75 0.43 13.4%

External validation
RI range R2 RMSE MAE(*) Coverage of the entire
validation set (N= 2718)
RI > 0.3 (N= 2501) 0.63 0.60 0.44 92.0%
RI > 0.5 (N= 1804) 0.70 0.55 0.40 66.4%
RI > 0.75 (N= 430) 0.81 0.44 0.30 15.8%

(*) Mean Absolute Error (MAE) values, being a primary measure of average prediction error in the original study involving the external validation dataset, have been also determined in this case and are reported in the above table for the result comparison purposes.

Sex:

not specified

Dose descriptor:

LD50

Effect level:

3 500 mg/kg bw

Based on:

not specified

Remarks on result:

other: Reliability Index (RI)= 0.54 (moderate)

Interpretation of results:

GHS criteria not met

Conclusions:

The oral median lethal dose (LD50) in the rat of 1-ethyl citrate was estimated, with the software tool ACD/Labs Percepta, at 3500 mg/Kg bw.

Executive summary:

The oral median lethal dose (LD50) in the rat of 1-ethyl citrate was estimated, with the software tool ACD/Labs Percepta, at 3500 mg/Kg bw.

Reference 3

Endpoint:

acute toxicity: oral

Type of information:

experimental study

Adequacy of study:

weight of evidence

Reliability:

2 (reliable with restrictions)

Rationale for reliability incl. deficiencies:

data from handbook or collection of data

Qualifier:

no guideline available

Principles of method if other than guideline:

No more details available, database match.

GLP compliance:

not specified

Species:

rat

Route of administration:

oral: unspecified

Sex:

not specified

Dose descriptor:

LD50

Effect level:

5 900 mg/kg bw

Interpretation of results:

GHS criteria not met

Conclusions:

The oral median lethal dose (LD50) in the mouse of triethyl citrate was determined to be 5900 mg/Kg bw.

Executive summary:

The oral median lethal dose (LD50) in the mouse of triethyl citrate was determined to be 5900 mg/Kg bw.

Reference 4

Endpoint:

acute toxicity: oral

Type of information:

experimental study

Adequacy of study:

weight of evidence

Reliability:

2 (reliable with restrictions)

Rationale for reliability incl. deficiencies:

data from handbook or collection of data

Qualifier:

equivalent or similar to guideline

Guideline:

OECD Guideline 401 (Acute Oral Toxicity)

GLP compliance:

not specified

Test type:

standard acute method

Limit test:

no

Species:

rat

Route of administration:

oral: gavage

Vehicle:

not specified

Doses:

5 - 15 mL/kg bw

No. of animals per sex per dose:

5

Sex:

not specified

Dose descriptor:

LD50

Effect level:

7 mL/kg bw

Based on:

test mat.

Sex:

not specified

Dose descriptor:

LD50

Effect level:

8 643 mg/kg bw

Based on:

test mat.

Remarks on result:

other: volume conversion based on a density of 1.2347

Signs of toxicity occurred within 1 hour of dosing. These included weakness, depression, ataxia, hyperexcitability, unrest, urinary incontinence, irregular and labored respiration. In some animals, convulsions preceeding death. Mortalities occurred from 2 hours to 3 days of dosing, while survivors recovered

within 15 hours to 4 days of dosing.

Interpretation of results:

GHS criteria not met

Executive summary:

The acute oral median lethal dose (LD50) in the rat of triethyl citrate has been determined to be 7 mL/kg body weight (8643 mg/kg body weight).

Endpoint conclusion

Endpoint conclusion:

no adverse effect observed

Dose descriptor:

LD50

Value:

3 500 mg/kg bw

Additional information

The acute oral toxicity LD 50 of diethyl citrate technical on rat was estimated to be in range 3500 - 8643 mg/kg. This is supported by a murine study via the intraperitoneal route which estimated the LD50 to be greater than 9900 mg/kg.

For the assessment was conservatively selected the lowest value.

Justification for classification or non-classification

No classification because of GHS criteria not met.