Environmental and Health Assesment of Alternatives to Phthalates and to
flexible PVC

7. Combined Assessment of Use, Exposure and Effects

7.1 Chemical Hazard Evaluation
7.1.1 Data availability
7.1.2 Physical-chemical data
7.1.3 Humans
7.1.4 Environment
7.2 Risk evaluation
7.2.1 Working environment
7.2.2 Consumer exposure
7.2.3 Human exposure in environment/secondary poisoning
7.2.4 Aquatic ecosystems
7.2.5 Sediment
7.2.6 Groundwater, soil and microorganisms
7.3 Overview

7.1 Chemical Hazard Evaluation

7.1.1 Data availability

Data pattern
The data availability is very variable among the suggested alternatives for phthalate plasticisers and materials. A majority of information is collected based on the CAS number of the suggested compound. For DEHA, ATBC, TEHPA and TETM information is available covering a range of results from tests on toxicological and ecotoxicological properties. However, only DEHA can be considered adequately covered, although some areas need further investigation.

DEHPA, OTSA, TXIB, ESBO, DGB and DOS are covered in less detail, either because of lack of information or because of inferiour quality of the tests. For the substance polyadipate no CAS number is available and information has been searched in bibliographic databases. For this substance no information has been located. A similar lack of data is seen for LDPE. However, the MDI base for PU is well described.

The type of data that are missing varies between compounds. Typically missing data on the environment side are biodegradation data and measured bioaccumulation data. On the health side a less clear pattern is observed, although adequate studies on longterm effects, e.g. reproductive toxicity studies are often lacking.

Data sources
The sources of the data are given primarily in the data sheets in the report appendix and for core information also in the main report. The information includes peer reviewed original papers, databases, previous reviews and reports, books, and proprietary information from suppliers.

It has been attempted to prioritise studies performed after standard test methods and guidelines for inclusion. In a number of cases the database IUCLID (European Commission Joint Research Center, 1996 and 2000), which contains information submitted by the industry, is almost the sole data source (e.g. TXIB). Again standardised tests have been selected whenever possible.

Attention is drawn to the fact that the majority of data are evaluated on the basis of databases on physical-chemical, toxicity and ecotoxicity studies. Although the studies as a rule are reviewed before inclusion in the databases the quality cannot be guaranteed a priori, nor is it possible to scrutinise the testing conditions of the original studies. Especially, for older studies the relation to modern guideline based experiments can be difficult to assess and consequently compliance with e.g. classification criteria may not be obvious.

7.1.2 Physical-chemical data

The available data show that none of the substances display hazardous physical-chemical properties, such as flammability etc. The typical substance has low water solubility and a moderate to high lipophilicity (LogPow 4 and higher). Vapour pressures are generally low (a tentative grouping is shown in Table 7.1).

Table 7.1 Relative volatility of substances suggested as alternative to phthalates in PVC.
Tentative estimates for substances for which data is not available are given in parenthesis. DEHP is included for comparison.

Hydrolysis
Many of the alternative plasticisers are, similarly to DEHP, esters of car-boxylic acid compounds. Information on hydrolysis, which potentially may be an important environmental fate property for this type of substances, is rarely available and only very limited information has been found.

In general, hydrolysis of the carboxylic acid esters is rather slow except when the sidechain contains halogens or unsubstituted carbons. The process is also slower with the length of the alkyl chain. The dicarboxylic acid esters proposed as alternatives belong to groups of substances with relatively long alkyl chains. In Schwarzenbach et al. (1993) the estimated half time for hydrolysis of the relevant bond types range from 38 days to 140 years. In the same reference dimethyl phthalates are estimated to have hydrolysis half lives of 12 years at 10 °C and pH 7. A similar slow hydrolysis of the dialkyl acid ester bonds may be the case for DEHA, TETM, TXIB, DGD and DOS. For DEHA the BUA-review (BUA 1997) concludes on the prolonged reaction in a study performed at elevated pH and temperature, that hydrolysis under environmental conditions will proceed extremely slow.

Also for the tri-phosphate (BUA 1996) no significant hydrolysis is to be expected at typical environmental pH and temperature, which may also apply to DEHPA. An evaluation of the possible hydrolysis is not made for the remaining substances.

Migration
The substances display a range of migration potentials. The lipophilic substances such as DEHA, TETM, TXIB and DOS migrate to organic solvents and oil, whereas those with relatively high aqueous solubility migrate to water and weak acids (see Table 7.2).

Table 7.2 Comparison of migration potential for assessed substances into fatty food simulant (bold) or water/acid. Tentative estimates for substances for which data is not available are given in parenthesis. DEHP is included for comparison

7.1.3 Humans

Four of the possible phthalate substitutes fulfil the criteria for classification with regard to acute toxicity or local effects. Based on the available literature DEHPA should be classified as Corrosive (C) and Harmful (Xn) with the risk phrases R34 (Causes burns) and R21 (Harmful in contact with skin). This classification was suggested by Bayer AG (Bayer, 1993) and is supported by the toxicological findings in the literature. TEHPA should be classified as Irritant (Xi) with the risk phrase R36/38 (Irritating to eyes and skin) also according to Bayer (1993). TETM fulfils the classification criteria with respect to acute toxicity as Harmful (Xn) with the risk phrase R20 (Harmful by inhalation) and DOS as Harmful (Xn) with the risk phrase R22 (Harmful if swallowed) based on LC50 and LD50 values. There are apparently no substances with severe organ effects, but the data set is very limited. It has not been possible to evaluate all effects according to their possible classification. The data are presented in Table 7.3.

The citrate, mellitate, epoxidised soybean oil, sebacate, and di-phosphate have been tested and found without CMR effects. One study showed foeto-toxicity (reduced ossification) for DEHA in mice, but results were not statistically significant. The toluene sulfonamide may be the only of the substances having effects of the CMR type. However, the suspicion for OTSA is based on tests done in connection with assessments of saccharine and its impurities, among others OTSA. Here it was found that the impurities are responsible for the reproductive effects of impure saccharine. No results are available on the pure substance.

Only weak mutagenic activity was described and there is limited evidence that OTSA is carcinogenic when administered orally to rats. Based on the available data it cannot be assessed whether OTSA is responsible for these effects, although it is suggested in the studies.

The sensitisation effects have been tested for many of the substances and the adipate, citrate, di-phosphate, trimellitate, epoxidised soybean oil, and sebacate have been found not to have this effect. Only the PU precursor MDI is a recognised sensitiser.

It must be stressed that for the majority of the compounds an insufficient data set is available for a complete human health risk assessment.

7.1.4 Environment

The combination of high persistence and high bioaccumulation potential does warrant attention to uses that leads to emission to the environment. Such substances are possibly the mellitate, the citrate, the dibenzoate and the sebacate.

The compounds for which ecotoxicity data are available (only data for the aquatic environment available) show relativly high acute ecotoxicity, that in all cases would lead to an environmental hazard classification. For the trimellitate and the sebacate, the low aqueous solubility in combination with persistence and bioaccumulation potential would lead to a classification as 'May cause long term effects in the aquatic environment' (R53).

The polymer materials and the polyadipate are estimated as unlikely to give rise to effects in the aquatic environment. No data was identified for the terrestrial environment.

7.2 Risk evaluation

It is beyond the scope of the present report to evaluate the risks associated with the use of the chemicals or materials in specific production, formulation or processing activities, since such evaluation must be coupled to a detailed knowledge of the particular technical and occupational environment. However, core properties such as volatility and migration are included. The data on risk is presented in Table 7.4.

7.2.1 Working environment

The exposure in the working has not been estimated at values above toxic values in the various scenarios, except for the adipate, where the selected scenario results in concentrations in workplace air 104 times the concentration resulting in more pronounced reactions in workers with an allergy or asthma case history. In general, the loss of plasticiser will depend on the volatility of the compound. OECD has made an allocation of plasticisers into low, medium and high classes of volatility (OECD 1998). Based in this the 11 plasticisers have been grouped relative to each other at standard 20-25 C.

7.2.2 Consumer exposure

Migration from PVC products has been measured for several of the alternative. In Table 7.2 it is attempted to show the migratory properties for the substances in a fatty food simulant (typically olive oil) and in an aqueous solvent (water or weak acids).

In the special teething ring scenario the citrate does reach 37% of a preliminary ADI of 1 mg/kg bw/day. The preliminary ADI is not officially recognised and a closer investigation of the citrate exposure conditions and human toxicity may be warranted.

7.2.3 Human exposure in environment/secondary poisoning

Several of the assessed substances have lipophilic properties based estimated LogPow values, and they may consequently have a high tendency for accumulation in biota. This is particularly clear in the estimation of concentrations of the adipate and sebacate in root crops, and the ADI is exceeded for sebacate in the regional worst case scenario. Virtually all the daily dose of these substances to humans from the environment arises in the root crops. The EUSES model is not well calibrated at high LogPow values and may overestimate the accumulation. However, some plants do accumulate anthropogenic substances and EUSES does not model this very precisely (Trapp, Schwartz, 2000). No data on terrestrial toxicity were identified to determine whether this accumulation may take place for these substances.

7.2.4 Aquatic ecosystems

The combination of high persistence and high bioaccumulation potential does warrant attention to uses that leads to emission to the environment. Such substances are possibly the mellitate, the citrate, the dibenzoate and the sebacate. Toxicity in the environment (only data for aquatic organisms available) is also of concern. The adipate, the tri-phosphate and the epoxidised soybean oil display acute aquatic toxicities below 10 mg/l.

7.2.5 Sediment

The DEHA exceeds the risk quotient of one for the sediment compartment due to its sorptive properties, but only in the scenario with complete substitution to this substance. ATBC (limited data set), DEHPA, TEHPA, and TETM (limited data set) had risk quotients less than one. Several other substances could not be quantitatively assessed for risk in the sediment (or the aquatic) environment: TXIB, ESBO, OTSA, DGD, DOS. This applies to the materials as well.

7.2.6 Groundwater, soil and microorganisms

Only the toluene sulfonamide has a water solubility suggesting transport to groundwater. However, not only dissolved species are found in groundwater. Substances bound to dissolved organic matter are also found in the groundwater.

It must be stressed that a number of the assessed substances are lipophilic and may have a high affinity for sludge particles similar to that of DEHP. No data on terrestrial toxicity has been identified and very limited information on effects on microorganisms in the sewage treatment plant is found.

7.3 Overview

Assessment of chemicals is challenging when few and not necessarily the same parameters are available for all substances. A profound and comprehensive or quantitative ranking is by far a possibility with the data set presented for the substances and materials included in the present project. However, to allow for comparison among the substances and materials a compressed overview of the data and the (occasionally tentative) assessments is provided. It must be emphasised that the data sets rarely allow hazard and risk assessment strictly according to the various applicable guidelines, and that the assessment to some extent relies on data obtained in databases published by European and American authorities.

In the following two tables the properties of the alternatives to phthalates and to flexible PVC are considered. The choice of properties shown in Table 7.3 has been based on the hazard indicators for humans as mentioned in CSTEE (2000), i.e. carcinogenicity, reproductive and developmental effects, mutagenicity, sensitisation and severe organ toxicity supplemented here with assessment of acute and/or local effects. For the substances and materials evaluated none of three with sufficient data exhibited 'Severe organ toxicity' and this column has therefore been omitted (data was available for DEHA, ATBC and TETM). It should, however, be mentioned that one study from 1964 showed signs of CNS toxicity in rats and mice after intraperitoneal injection of 400 mg ATBC/kg bw. No supporting evidence for this effect has been found.

In addition to evaluating hazards, the risk is also assessed (Table 7.4). For humans this is achieved by comparing the estimated dose of the substance in consumer and environmental exposure with existing or estimated ADI. For the environment the environmental risk quotient is calculated from PNEC and estimated environmental concentrations.

Table 7.3 The inherent properties for the investigated subtances are summarised using key parameters: acute and local effects, carcinogenicity(C), genetic toxicity (M), reproductive toxicity (R), sensitisation, persistance, bioaccumulation and aquatic toxicity. If data are not available for all parameters or only from non standard test results a tentative assessment is given (shown in parentheses). For the materials an evaluation is given based on general polymer properties. The symbols: ● identified potential hazard, ○ no identified potential hazard, and – no data available.

^a Foetotoxicity (reduced ossification) has been identified as the most sensitive effect in a developmental toxicity study. 
^b QSAR estimates by Danish EPA leads to the classification N; R50/53 (May cause long term effects in the aquatic environment).
^c A test on reproductive effects performed on a product containing OTSA as impurity attributes effect to OTSA. No substance specific data available.
^d C,M,R indicated that the effect is investigated but no effects are seen.

Table 7.4 The evaluated risks to humans or the environment are summarised for the investigated substances (the polymer materials are not included). The estimated exposure of humans is compared to the Acceptable Daily Intake (ADI). Predicted environmental concentrations in the aquatic environment (PEC) are compared to predicted no-effect concentrations (PNEC). "Worst case" scenarios are used. The reader is referred to the main text and the data sheets for further explanations to the table. Parentheses show an assigned ADI. The symbols: ● ratio >1 (identified potential risk), ○ ratio <1 (no identified potential risk), and –no data available

^a Dose reaches 37% of preliminary ADI in teething ring scenario.
^b Tentative estimate based on only one ecotoxicity study.
^c Preliminary ADI from Nikiforov (1999)
^d Data set comprise only two acute values and one chronic NOEC value.