Report on a "Study to establish a programme of detailed procedures for the assessment of risks to health and the environment from cadmium in fertilisers" [Study 2 - Establishment of procedures for carrying out a risk assessment - February 2000] carried out by Environment Resources Management, London, UK (hereinafter referred to as the "Study") - Opinion expressed at the 17th CSTEE plenary meeting, Brussels, 5 September 2000.

- Introduction. Terms of reference

Background

Legislative situation:

When Austria, Finland, and Sweden negotiated their EU membership, a derogation was agreed from Article 7 of Directive 76/116/EEC which establishes a single market for mineral fertilisers. The derogation concerned the cadmium content in fertilisers, which is limited in these Member States, and can therefore form an obstacle to trade as no limitations exist in other Member States. The Commission committed itself to review Article 7 of that Directive in order to establish a harmonised situation as far as cadmium in fertilisers is concerned. Consequently a study was carried out on the effects of cadmium in fertilisers. Based on the final report, issued at the end of 1997, the Commission decided that not enough data were available in order to come to a final conclusion. In 1999 Council and Parliament granted a further three year derogation combined with the reinforced commitment by the Commission to review the cadmium in fertilisers problem further and to come to conclusions by the end of 2001.

Commission's actions to prepare for risk assessment:

The Commission then launched two successive studies in order to (i) define the data requirements for filling-in the data gaps that were identified by the 1997 study, and (ii) establish a methodology for a risk assessment, in accordance with Community principles. The risk assessment is to be applied by individual MS for each own's territory.

Limitations:

Time and budgetary resources within which the studies as well as the future risk assessment had and will have to be performed, respectively, demanded a clear mission statement from the Commission, emphasising the need for pragmatism, as well as strong assistance from Member Sates. In addition it was decided that the report on the second study, i.e. "PROGRAMME OF PROCEDURES FOR THE ASSESSMENT OF RISK TO HEALTH AND THE ENVIRONMENT FROM CADMIUM IN FERTILISERS", should be submitted to the Commission's CSTEE for reviewing its scientific soundness.

The Question to the CSTEE

The Commission therefore has invited the CSTEE to answer the following question:

Can the ERM report "PROGRAMME OF PROCEDURES FOR THE ASSESSMENT OF RISK TO HEALTH AND THE ENVIRONMENT FROM CADMIUM IN FERTILISERS" be judged as an acceptable basis for a risk assessment to be carried out by competent authorities in Member States for their own territory in compliance with Community principles on risk assessment.

- Executive summary

The purpose of the Study is to establish the procedures to be applied in order to conduct an assessment of the current and possible future specific risks from cadmium in inorganic phosphate fertilisers, and also to consider in those procedures, whenever possible, vulnerable persons and vulnerable environments.

Those procedures or "risk assessment sequences" (p. 65) are proposed to follow a 3-step method, comprising: (i) an accumulation module; (ii) an exposure module; (iii) a risk characterisation module.

The aim of the Study, however, is not to carry out the risk assessment but rather to provide a framework for doing so (p. 6).

The procedures described in 1996 in the TGD (Technical Guidance Document in Support of Commission Directive 93/677/EEC on Risk Assessment for New notified Substances and Commission Regulation EC No. 1488/94 on Risk Assessment for Existing Substances) although forming the basis for designing the procedure for risk assessment, had been modified and adapted to "the circumstances and conditions under which this risk assessment will be carried out" (p.6).

The "accumulation module" is intended to derive the net accumulation of Cd in soil/soil solution over time and PEC using background soil Cd concentration, input of Cd to agricultural soils, leaching of Cd from soils and Cd offtake via crops.

The "exposure module", using exposure parameters (including dietary Cd intake by food type) and soil and plant factors, is intended to calculate the uptake of Cd from soil into crop plants and subsequent intake by humans and the relative contribution from fertilisers.

Finally, the "risk characterisation module" is intended to estimate the incidence and severity of adverse effects likely to occur due to actual or predicted exposure from the PEC, PNEC, NOAEL, LOAEL, exposure estimates and distribution of exposure.

To calculate Cd uptake by plants a set of algorithms is proposed assuming that the soil factors that increase uptake of Cd by plants are low pH, high salinity, high total Cd concentrations, low organic matter content, low cation exchange capacity, low clay, low iron and zinc and high temperature. The recommended algorithms (p.49, Table 5.4a) refer only to five types of plants, namely Winter wheat/cereals (grain), spring wheat/cereals (grain), potatoes, carrots/root vegetables and leafy vegetables, and assuming average Zn soil concentrations derives Cd concentrations in plant tissues as a function of pH and Cd concentration in soil, and using different constant values for each type of plant.

Then, using proposed equations (equ. 8, p. 30 and equ. 14, p. 50) an estimate of incremental dietary intake of Cd due to usage of fertiliser is computed. Equation 14 (p. 50) is a summation of products of the concentration of Cd in each food item, times its consumption by period of time, times the fraction of that food item that is grown on fertiliser-treated soil. Thus, even if the concentration of Cd in each considered food item can be accurately calculated by the proposed algorithms, it is mandatory that for each Member State (or even specific region in each country) there are accurate and updated statistical data to supply the remaining data.

Risk characterisation is proposed as an interpretation of ratios PEC/PNEC and NOAEL/human exposure.

General Comments

The aim of the Study (p.6) is not clear and reading it it did not become clear either to what extent and how each Member State could deviate from what is proposed. It is not clear whether the proposed algorithms and subsequent calculations might be practically applicable and of sufficient new value to circumvent the lack of accurate supplied values on types of foods really consumed by each specific population under consideration, and how much of those particular foods really consumed might have been grown in fertiliser-treated soils. It is questionable that the relevant data exist to calculate the Cd exposure module proposed, for example how can data for Equ. 14 (p. 50) be accurately obtained, specially the IbACKGROUND. It is not clear either how can the calculations proposed tackle the question of individual variability in Cd intake and absorption. There seems to exist no practical means available to validate the algorithms and calculations proposed, or to check if they are better than others, and even if they give a realistic picture of the situation. Practical examples of applications to real situations using data supplied by each Member State should have been given, and compared to existing approaches. These considerations are not to deny a potential value and interest in the proposed approach ipso facto. It contains a lot of valuable information but it is a mixture of guidelines and background information of varying quality. The aim of the report could be clearer if some of the background information would be deleted.

The assessment of the variation in dietary intake of Cd requires a more thorough evaluation. To make a risk assessment of the contribution of Cd from fertilisers one has to focus not only on special risk groups but also to study the 95 percentile of the dietary intake. This is neglected in the Study.

It is to be clarified if the content of Tables is to be considered as examples.

Specific issues that need to be considered

Title:

The title should make clear that the Study deals with Cd in inorganic fertilisers.

Page I:

Under Scope and Objective the limitations imposed by the availability of data have not been taken into account although mentioned in Page 1, Section 1.2.

Page II:

In Figure 1 Cd should be cadmium s.

Page 3, third point in Section 2.1:

A part of Cd in shellfish and fish could be from leaching?

Page 7:

The Figure is from the TGD (1996) but in a way is misleading since the document also asks for realistic worst case scenarios, which are not the case for the proposed algorithms.

Chapter 3:

Before risk groups can be discussed, the health effects of Cd must be described. Thus, a section with a description of human health effects of Cd should be inserted. Also, section 6.3, page 61, starts directly with a short discussion of dose-response assessment with no mentioning of the effects of Cd. Critical effects should be presented i.e. kidney damage. In addition, the recently published association between Cd exposure in the general population (mainly women) in Belgium and decreased bone density and increased risk of fractures (Staessen et al., Lancet 353, 1140-44, 1999) should be mentioned. Likewise, the similar Swedish study should also be mentioned (Alfvén et al, 2000).

Section 3.2.1:

High Cd concentrations are also found in high fibre content cereals, seeds and chocolate. The statement in the second sentence that the consumption of products with high Cd levels is low is not true for shellfish/crustaceans in some countries. Certainly, it is not true for high fibre content products. Typical groups with high dietary intake of Cd are those with preference also for high fibre content diets and chocolate. The support for some high Cd content foods being "relatively low in available cadmium" is not at all clear. Studies show contradictory results and most do not take into account iron status. Animal studies indicate similar uptake of shellfish Cd, mushroom Cd and CdCl2 at reasonable dose levels.

Section 3.2.2:

Second sentence under Environmental Exposure. Scientific studies indicate that the maximum dietary Cd intake may well exceed 3 times the average (Vahter et al, 1996; Thomas et al, 1999; Shimbo et al, 2000).

Page 11:

The second paragraph of section 3.2.1 needs a reference to the statement about the reflection in a proportional increase in body burden of Cd to allow for a verification of its validity.

Page 12:

All data on mean intake are only of interest if they could be estimated with 95% percentiles. Reference bottom line should be Alfvén et al, 2000.

Page 13, third paragraph:

Reference second line should be Vahter et al, 1996.

Page 14, second paragraph:

There is not only concern but indications from both human studies (Buchet et al, 1990) and experimental studies (Bernard et al, 1991).

Page 14, third paragraph:

The reference to the study referred is Berglund et al, 1994.

Page 15, third paragraph:

On the other hand, in most European countries cereals and vegetables are the most important sources of Cd intake. The concentrations of Cd in plants are highly dependent on Cd in soil, which has a very long half-life. Thus, major changes in the dietary intake of Cd in the population cannot be expected to change very much, unless the type of diet is changed. On a group basis this is not likely to happen rapidly. Cd in fertilisers and shellfish needs a reference. The influence of Cd in fertilisers has to be mentioned for shellfish.

Page 15, fourth paragraph:

The meaning of this paragraph is not clear.

Page 16, Table 3.1:

Increased Cd absorption at low iron stores is probably 2-3 fold. The meaning of the values given is not clear. In the text it is mentioned that the figures in the different risk groups have been made the same, for simplicity. This cannot be accepted. Exposure to Cd from drinking water was not considered, not even in this Table.

Page 19:

Concerning Cd from manure, it seems that almost half of the input may come from manure, but some of this may originally come from fertilisers. The algorithms do not take this into consideration.

Page 21:

Co should be Cs(o).

Page 27, Section 4.2.7:

The information in the first part is out of date due to marked changes in the draft report mentioned.

Page 32, Table 4.4c:

Cd concentrations in Belgium should be verified.

Page 42:

The concept of "vulnerable environment" is not defined in Section 3.1.2 but in Section 3.3.2.

Page 45, last line:

The order is not correct and a reference is lacking, or simply delete the line.

Page 47:

The number 14 in front of an equation at this point is a mistake, since the correct Equ. 14 is in p. 50.

Page 48:

The relative concentration is not in agreement with available information.

Page 51, Table 5.6a:

This information in not really interesting without 95% percentiles. Under heading 5.6 the word "background" should be deleted since it usually means exposure without influence from anthropogenic sources. The sources of the country average dietary intake values should be given. What methods are used? It should be pointed out that there are several studies and estimates from some countries showing a significant variation in average intake of Cd. There is an obvious need for a in-depth analysis of intake levels of Cd in various countries.

Page 51, last paragraph, sixth line:

It is not only "not entirely satisfactory", but rather "not at all satisfactory".

Page 52, Table 5.6b:

Since the values presented in Table 5.6a are highly uncertain it is misleading to prepare more calculations, in particular to present values with three decimals.

Page 52, last paragraph:

For females with low iron stores the Cd uptake (absorbed amount) can be assumed to be 2-3 times that of the general population.

Page 53, top paragraph:

It should be noted that the absorption of Cd from cigarette smoke in the lungs is much higher than in the gastrointestinal tract.

Section 6.3.2, page 61:

For several reasons the derivation of the NOAEL is not acceptable. References should be given for PTWI, both when it was set and when it was reviewed by the joint FAO/WHO expert Committee on Food Additives. Also the basis for the values should be given. It should be noted that the PTWI assumes 50 years of exposure and does not include any safety factor (see WHO Technical Report Series 837, 1993). The purpose of the PTWI is to protect the whole population. Therefore, it should not be compared with the average dietary intake in the countries as done in several instances in this chapter of the Study. It should be stressed that the CSTEE had already referred to a value of 50mg/kg Cd for effects in the kidney (CSTEE, Opinion on the report by WS Atkins International Ltd, expressed at the 6TH Plenary Meeting, Brussels, 26/27 November 1998). Only 50% of the population are below the average daily intake. Obviously, the 95th or the 99th percentile should be compared with the PTWI. Although this is to some extent discussed further in the Study it is not made clear. Any comparison of average intake with PTWI is unfortunate as it may give a false indication that there is a sufficient safety margin if the average is below the PTWI. A comparison must be done considering that half the population is above the average and one fraction is considerably higher as the distribution of daily intakes is skewed with a tailing towards high intakes.

Second paragraph, fourth line:

Should be "ranged" instead of "range". The data refer to three studies and are not representative for the total populations studied and definitely not for all Europe.

Paragraph three, fourth line, should read:

"...(Järup et al, 1998) that an average daily intake of 30microg Cd would lead to renal tubular damage in about 1% of the adult general population. In high-risk groups, e.g. women with low iron stores, up to 5% may develop such effects". It is no use referring to risk estimations at an average intake of 70microg, as this is not the situation for any European country. It should be mentioned that there are increasing data supporting this estimation, e.g. Alfvén et al, 2000. Also, there are several studies on kidney concentrations of Cd indicating that a fraction of the population has kidney Cd concentrations exceeding 50mg/kg.

Page 62:

First paragraph should be deleted. Considering the increasing support for an effect level of 50mg/kg in the kidneys, or even lower, the current PTWI cannot be proposed for this risk assessment.

Page 67, References:

Many references are useless since they are either private technical or commercial reports or drafts or even incompletely referenced publications.

Conclusions

It is not clear from the Study that these are guideline procedures and that alternative procedures may be used.

If the Study is to be judged for exposure assessment it may be useful, but not for risk assessment, as the title implies.

The document contains much useful information, which, however, is mixed with a lot of background information of varying or unknown quality. Also, it is not clear in the Study that much of the background data are not valid for most countries.

The aim of the report should be made clearer and background information not fitting the aim should be deleted.

The main information missing concerns the upper range of Cd intake. When evaluating the health risks due to Cd intake the upper range or 95th percentile should be considered, not only average data. The use of table 3.1 data (some should be corrected) should be made clearer.

Taking into account the above general points and the specific ones which are too numerous, each one being important in its own right, the Study should not be accepted in its present form. A thorough revision should be undertaken by the Authors.

References:

Alfvén T, Elinder CG, Carlsson D, Grubb A, Hellström I, Persson B, Petterson C, Spang G, Schütz A, Järup L. Low level cadmium exposure and osteoporosis. J Bone Mineral Res. 2000, In press

Bernard A, Schadeck C, Cardenas A, Buchet JP, Lauwerys R. Potentiation of diabetic glomerulopathy in uninephrectomized rats subchronically exposed to cadmium. Toxicol Lett. 58, 51-57, 1991

Shimbo S, Zhang Z-W, Moon C-S, Watanabe T, Nakasuka H, Matsuda-Inoguvhi N, Higashikawa K, Ikeda M. Correlation between urine and blood concentrations and dietary intake of cadmium and lead among women in the general population of Japan. Int Arch Occup Environ Health. 73, 163-170, 2000-08-31

Thomas KW, Pellizzari ED, Berry MR. Population-based dietary intakes and tap water concentrations for selected elements in the EPA Region V National Human Exposure Assessment Survey (NHEXAS). J. Exp Anl Environ Epidemiol. 9, 402-4113, 1999

Vahter M, Berglund M, Nermell B, Akesson A. Bioavailability of cadmium from shellfish and mixed diet in women. Toxicol Appl Pharmacol,.136, 332-241, 1996