aluminum infant formula-1

8/6/2019 Aluminum Infant Formula-1

http://slidepdf.com/reader/full/aluminum-infant-formula-1 1/13

PLEASE SCROLL DOWN FOR ARTICLE

This article was downloaded by: [Consorci de Biblioteques Universitaries de Catalunya]

On: 26 April 2011

Access details: Access Details: [subscription number 919083124]

Publisher Taylor & Francis

Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-

41 Mortimer Street, London W1T 3JH, UK

Food Additives & Contaminants: Part APublication details, including instructions for authors and subscription information:http://www.informaworld.com/smpp/title~content=t713599661

Aluminium content of Spanish infant formulaI. Navarro-Blascoa; J. I. Alvarez-Galindoa

a University of Navarra Faculty of Sciences Department of Chemistry Pamplona (Navarra) E-31080

Spain,

Online publication date: 10 November 2010

To cite this Article Navarro-Blasco, I. and Alvarez-Galindo, J. I.(2003) 'Aluminium content of Spanish infant formula',Food Additives & Contaminants: Part A, 20: 5, 470 — 481

To link to this Article: DOI: 10.1080/0265203031000098704

URL: http://dx.doi.org/10.1080/0265203031000098704

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf

This article may be used for research, teaching and private study purposes. Any substantial orsystematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply ordistribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contentswill be complete or accurate or up to date. The accuracy of any instructions, formulae and drug dosesshould be independently verified with primary sources. The publisher shall not be liable for any loss,actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directlyor indirectly in connection with or arising out of the use of this material.

http://www.informaworld.com/smpp/title~content=t713599661

http://dx.doi.org/10.1080/0265203031000098704

http://www.informaworld.com/terms-and-conditions-of-access.pdf

http://www.informaworld.com/terms-and-conditions-of-access.pdf

http://dx.doi.org/10.1080/0265203031000098704

http://www.informaworld.com/smpp/title~content=t713599661



Aluminium content of Spanish infant formula

I. Navarro-Blasco* and J. I. Alvarez-GalindoUniversity of Navarra, Faculty of Sciences, Department of Chemistry, Pamplona (Navarra), E-31080, Spain

(Received 31 July 2002; resubmitted 19 November 2002;revised 3 February 2003; accepted 12 February 2003)

Levels of aluminium in 82 different infant formulae from nine different manufacturers in Spain weredetermined by acid-microwave digestion and graphite

furnace atomic absorption spectrophotometry. Theinfluence of aluminium content in tap water inreconstituted powder formulae was examined and an

estimate was made of the theoretical toxic aluminiumintake in comparison with the provisional tolerableweekly intake (PTWI). Possible interactions betweenaluminium and certain essential trace elements added toinfant formulations have been studied according to thetype or main protein-based infant formula. In general,the infant formulae contained a higher aluminiumcontent than that found in human milk, especially inthe case of soya, preterm or hydrolysed casein-based

formulae. Standard formulae gave lower aluminiumintakes amounting to about 4% PTWI. Specialized and

preterm formulae resulted in a moderate intake (11–12and 8–10% PTWI, respectively) and soya formulae

contributed the highest intake (15% PTWI).Aluminium exposure from drinking water used for

powder formula reconstitution was not considered a potential risk. In accordance with the present state of knowledge about aluminium toxicity, it seems prudentto call for continued efforts to standardize routinequality control and reduce aluminium levels in infant

formula as well as to keep the aluminium concentrationunder 300 g l À1 for all infant formulae, most specifi-cally those formulae for premature and low birthneonates.

Keywords: aluminium, infant formula, drinkingwater, daily intake, provisional tolerable weeklyintake (PTWI)

Introduction

Exposure of adults to aluminium occurs throughdrinking water, food additives and antacids orbuffered analgesics. However, aluminium and itscompounds appear to be poorly absorbed and areeliminated effectively in the urine. It has been reportedthat neonates are more susceptible to exposurebecause of their greater intestinal absorption becauseof an immature gastrointestinal tract (Sedman et al.1985). In infants, aluminium toxicity has beenrelatively well documented in the case of neonateswith impaired renal function, and ill, premature or lowbirth weight neonates (Sedman et al. 1984a, b, Morenoet al. 1994, Puntis et al. 1986, Bishop et al. 1997). Highaluminium levels in infant formulae have beenimplicated in aluminium intoxication in two infantswith neonatal uraemia (Freundlich et al. 1985, 1990).

Infants, and especially preterm neonates, display anarrow tolerance to aluminium because of immaturityin the tissues and organs involved in its metabolism.Human milk provides a balance of elements suitablefor infants and thus a knowledge of the aluminiumcontent of human milk serves as the basis forformulating appropriate substitutes. Regardless of

the wide variability, the aluminium content in humanmilk is between three and 160 times lower than thatfound in infant formulae (Koo et al. 1988, Coni et al.1990, Simmer et al. 1990, Baxter et al. 1991, Ballabrigaet al. 1994, Fernandez-Lorenzo et al. 1999, Krachleret al. 2000). It is thus possible to suggest a referencealuminium range of 3–79mg lÀ1 for human milk.Ballabriga et al. (1994) and Ferna ´ ndez-Lorenzo et al.(1999) found an aluminium content in Spanish humanmilk of 6.5Æ 5.3 mg lÀ1 (n¼ 16, range 0.9–19.8mg lÀ1)and 23.9Æ 9.6 mg lÀ1 (n¼ 45, 7–42 mg lÀ1), respectively.

Given the evident toxicological impact of aluminiumon neonates, it is desirable that infant formulae should

be proportionally similar or lower in aluminiumconcentrations to human milk. The high aluminiumcontent found in certain infant formulae derived fromcomplex manufacturing processes has led to a call foran evaluation of aluminium levels, mainly in both soyaand preterm formulae (AAP 1996).

* To whom correspondence should be addressed.e-mail: [email protected]

Food Additives and Contaminants, Vol. 20, No. 5 (May 2003), pp. 470–481

Food Additives and Contaminants ISSN 0265–203X print/ISSN 1464–5122 online ß 2003 Taylor & Francis Ltdhttp://www.tandf.co.uk/journals

DOI: 10.1080/0265203031000098704



The aims in the present study were to analyse theconcentration of aluminium in the majority of infantformulae sold commercially in Spain to determine theinfluence of aluminium content in the tap water onfinal concentration in reconstituted powder formulae,to estimate the theoretical toxic aluminium intake incomparison with the provisional tolerable weekly

intake (PTWI) established by the joint FAO/WHOExpert Committee on Food Additives (WHO 1989)and, finally, to discuss the possible interactions of certain essential trace elements added to formulaewith aluminium according to the type or mainprotein-based infant formula.

Material and methods

Sample collection

Most infant formulae were purchased directly frommanufacturers. The remainder were obtained from adistribution company in Pamplona, Spain. A total of 82 different infant formulae from nine differentmanufacturers were studied. Formulae includedboth powder (n¼ 61) and ready-to-use preparations(14), such as those based on cow’s milk (68) or soy-based formulae (7). Cow’s milk-based formulaeincluded: preterm formula (n¼ 7), starter formula(adapted, n¼ 16—formula for infants from the firstday to 4–5 months of age following the EuropeanSociety of Paediatric Gastroenterology and Nutrition

recommendations—and non-adapted, n¼ 4—formulaprepared for infants from birth to 12 months of ageunder American Academy of Pediatrics regulations),follow-up formula (19) and specialized formula— hypoallergenic (12), designed for lactose intolerant(7), or inborn errors of metabolism (10) formulae.Infant formulae were stored in accordance with thedirections on the label. Containers were kept in thedark at room temperature in a humidity controlledroom.

Samples of drinking water were collected in duplicatefrom the Community of Navarra, Spain, in bothurban and rural areas. Thirty-nine tap water sampling

points were selected according to a population censusprovided by the local office of the National StatisticalInstitute and piping information obtained fromseveral public water-treatment plants in the commu-nity. Eighteen samples were taken from popula-tions in rural areas and 21 samples in Pamplona

(provincial capital). A strict protocol was thusestablished to carry out tap water sampling.

Sample handling

Special care was devoted to this phase to minimize therisk of adventitious contamination when handling.All plastic material (low-density polyethylene) orimplements used that came into contact with thesamples (infant formulae or tap water) were cleanedpreviously in 5% nitric acid solution (Merck,Darmstadt, Germany) for 6 days and later rinsedthree times with ultrapure water before use. Infantformula containers were opened in the clean labora-tory under a flow laminar bench, using vinyl talc-free gloves (RotiprotectÕ, Carl Roth, Karlsruhe,Germany) and plastic material (PlastibrandÕ, Brand,Wertheim, Germany), when performing the sampling.

Chemicals

Suprapure nitric acid was purchased from Merck andpurified by sub-boiling distillation before use.Ultrapure deionized water type Milli Q was used forpreparation and/or dilution of the treated sample andstandard solutions. Magnesium nitrate was used as amatrix modifier for aluminium determination(1.4 g Mg(NO3)2 6H2O (Merck) and 2 ml Triton X-100 (Sigma, St Louis, MO, USA) were diluted in 1litre with ultrapure water).

Sample treatment

Infant formula samples (0.3000 g powder, 1.500 mlliquid) were placed in high-pressure Teflon digestionbombs and digested with 4 ml sub-boiling nitric acid(Merck) in a closed acid-decomposition microwavesystem (Milestone MLS 1200, Millestone s.r.l.,Sorisole, Italy). The solutions obtained were thendiluted up to 10 ml with ultrapure deionized waterand kept in frozen storage at –20C until analysis.Samples were digested in triplicate. Sampling andtreatment operations were described in greater detailby Navarro and Alvarez (2002). In the case of

drinking water samples, to avoid flocculation orlosses by adsorption to the plastic walls in samplingcontainers always after measuring pH, 1 ml of sub-boiling nitric acid per litre was added until approxi-mately pH 2, and the solution was stored at 4C untilanalysis.

Aluminium content of Spanish infant formula 471



Instrumental analysis

Aluminium concentration was determined by graphitefurnace atomic absorption spectrometry (GFAAS,GBC GF 2000, Dandenong, Victoria, Australia). Theoperating parameters and optimizing temperatureprogramme of the instrument are given in table 1.

Digested samples were diluted in matrix modifiersolution. Injections (20 ml) were made in triplicate onL’vov platforms positioned inside pyrolytically coatedtubes. Samples were quantified by reference to acalibration curve obtained for aqueous standards.Working standard solutions (0–80 mg lÀ1) were madeup each day by dilution from stock 1000 mglÀ1

standard solution (Merck) in enough sub-boilingnitric acid to a final acid concentration similar toprepared samples. All solutions were kept in thecovered carousel throughout the analysis to preventany contamination.

Iron and zinc concentrations in digested acid solutionswere analysed by flame atomic absorption spectro-photometry (FAAS, GBC GF 2000) and inductivelycoupled plasma atomic emission spectrometry withMeinhard nebulizer (ICP-AES, Jobin Ybon JY 38SPlus Sequential) was used for manganese. Iron, zincand manganese calibration curves were accomplishedusing direct calibration against aqueous standards.Details of measurements, operating parameters and

trace element quantification were described byNavarro et al. (1996, 2000).

Quality assurance

Two blank reagents, an aqueous internal standard anda replicate of infant formula control, as well as 10infant formula samples were included in each analyti-cal batch to provide an on-going quality controlinformation. The relative standard deviation deter-mined for repeatability (precision for triplicates with arun) with the blank reagent (13.9%), internal standard(1.6%) and infant formula control (1.6%) werecomparable with those calculated for the reproduci-bility (day-to-day precision), 15.5, 2.6 and 2.7%,respectively.

Analytical blanks corrected for any possible contami-nation during analysis. When expressed in terms of

infant formula, the aluminium content was 2.3Æ0.3 mg lÀ1. The detection limit was defined as theaverage of three times the standard deviation of the reagent blanks and corresponded to 3.3mg lÀ1

(wet weight) for the infant formulae. In the case of drinking water, blank reagents consisted of ultrapurewater which was subjected to the same procedures of treatment, storing and mixing with matrix modifier.Aluminium level in analytical blanks was 0.2Æ0.5 mg lÀ1, resulting in a detection limit of 1.7 mg lÀ1.

Reference infant formulae were run throughout thecourse of the study (174.8Æ 4.2 mg lÀ1). The control

sample was previously analysed by the standardaddition calibration to minimize matrix effects andaluminium at 176Æ 7 mg lÀ1 was obtained. Recoveryassays of spiked aluminium at different amounts (200,400 and 800 ng aluminium) before digestion in thisin-house control formula were satisfactory (96–104%).An internal aqueous quality control (20mg lÀ1) wasrun concurrently with analytical samples. ThemeanÆ SD of aluminium determined was 20.0Æ1.1 mg lÀ1 (n¼ 34, range 18.2–22.3 mg lÀ1). The accept-able range established in the quality programme was17.6–22.4mg lÀ1.

Given the lack of adequate standard reference

material for aluminium based on milk powder withcertified values, IAEA (International Atomic EnergyAgency) 155 whey powder was analysed to provide anestimate of accuracy and guard against instrumentbias. The obtained result (43.2Æ 2.6mgkgÀ1) showedan acceptable agreement with the information value

Table 1. Instrumental parameters and optimizing furnace

programme for aluminium determination.Instrumental parameters

Wavelength (nm) 309.3Slit width (nm) 1.0Lamp current (mA) 10Sample volume (ml) 20Measurement mode peak areaSource hollow cathode lampBackground correction deuterium lamp

Temperature programme

StepTemperature

(C)Ramp

(s)Hold

(s)Argon flow(ml minÀ1)

Readon

Drying 1 110 5 15 300 – Drying 2 250 10 15 300 – Charring 1500 20 5 300 – Atomization 2500 1 4 0 yesCleaning 2600 1 5 300 – Cooling 20 5 3 300 –

472 I. Navarro-Blasco and J. I. Alvarez-Galindo



(non-certified) provided from the reference material(53mgkgÀ1, range 38–68 mg kgÀ1).

The IAEA milk powder A11 was assayed as a certifiedreference material to validate the analytical methodfor metals other than aluminium. The values deter-mined (12 independent analytical runs) were: iron,3.62Æ 0.09 mg gÀ1; zinc, 37.10Æ 0.38 mg gÀ1; and man-

ganese, 0.250Æ 0.021mg gÀ1

; in comparison withcertified values (confidence intervals) the valueswere (plus range): iron, 3.65mg gÀ1 (2.89–4.41); zinc,38.9 mg gÀ1 (36.6–41.2); and manganese 0.257mg gÀ1

(0.248–0.266).

Statistical analysis

Statistical analysis was performed with SPSS v.9.0 forWindows. Different groups of samples were com-pared through non-parametric Kruskal–Wallis andMann–Whitney U -tests or a Wilcoxon test for pairedgroups, with statistical significance set at p<0.05.Observed tendencies in some trace element contentswith regard to aluminium were analysed with theSpearman coefficient.

Results and discussion

Aluminium contents in infant formula

Table 2 shows the aluminium concentrations foreach of the different types of infant formulae. Thealuminium concentration found in powdered infantformula was calculated according to the manufac-turer’s dilution instructions and expressed in mg lÀ1.

The wide variability in aluminium content found insome of the formulations included in this study is of special relevance. Because of this, both means andmedians are included, although we have only used themedian as the most representative parameter. Therange of aluminium found in the infant formulae wasconsistent with those in the literature.

The variability of aluminium content is a result of theway it can be incorporated in the infant formulae.The total aluminium content can originate from:(1) the raw material—cow’s milk or isolated soy pro-tein; (2) contamination during processing from sur-faces of metallic equipment or utensils; (3) additivesor mineral supplements or (4) migration from thecontainers during storage. Each factor could explainthe high variability in the aluminium concentrationsfound in the different infant formulae.

Lower aluminium values in standard formulae (starterand follow-up formulae) in comparison with special-ized, preterm and soya formulae are readily apparent.This is confirmed by the Kruskal–Wallis test( p¼ 0.003). As previous studies have reported (Kooet al. 1988, Dabeka and McKenzie 1990, Simmer et al.1990, Baxter et al. 1991, Biego et al. 1998, Bloodworthet al. 1991, Ballabriga et al. 1994, Ikem et al. 2002), thelower aluminium values were detected in milk-basedformulae, such as starter formulae (adapted, 196 Æ152mg lÀ1; non-adapted, 231Æ 128mg lÀ1) and follow-up formulae (272Æ 189mg lÀ1). These finding levelsare in agreement with those found by others (Coniet al. 1993, Ballabriga et al. 1994, Hawkins et al. 1994,Ferna ´ ndez-Lorenzo et al. 1999, Ikem et al. 2002).Lower levels found in raw cow’s milk show that this byitself is not necessarily the greatest source of alumin-ium. However, the large ranges of aluminium contentobserved in standard formulae are indicative of thepotential for contamination during manufacture.

It is well known that aluminium is associated withproteins. Table 3 summarizes the aluminium levelsfound in different infant formulae, focusing on themain protein.

There is clear evidence of the involvement of intrinsicaluminium from a protein source on the final levelfound in standard formulae. The highest aluminiumcontent is provided by those formulae based on whole

milk (adapted starter formulae, 338Æ 114mg lÀ1; non-adapted starter formulae, 368mg lÀ1; follow-up for-mulae, 296Æ 185mg lÀ1), followed by skim-milk-basedformulae (adapted starter formulae, 190Æ 83mg lÀ1;non-adapted starter formulae, 136Æ 115mg lÀ1;follow-up formulae, 223Æ 220 mg lÀ1) and lastly,

Table 2. Aluminium content in different types of studied infant formulae ( g l À1).

Infant formula n Mean MedianÆSD Range

Preterm formula 7 449 421Æ 137 317–726Starter formula

Non-adapted 4 237 231Æ 128 118–368Adapted 16 252 196Æ 152 68–573

Follow-up formula 19 292 272Æ 189 66–788Specialized formula

Without lactose formula 7 574 399Æ 451 102–1439Hypoallergenic formula 12 687 294Æ 945 105–2720Inborn errors diet 10 453 443Æ 112 307–655

Soya formula 7 930 573Æ 1132 313–3479




formulae that include whey proteins (adaptedstarter formulae, 181Æ 177mg lÀ1) or casein (adaptedstarter formulae, 126Æ 38mg lÀ1) in its composition(table 3). The present tendency to reduce the proteincontent and replace different proteins with wheyprotein to mimic the protein profile found in humanmilk may advantageously influence the aluminium

content of newer types of infant formula.The complex manufacturing process of specializedformulae seems to play an important role in the degreeof aluminium contamination. The highest aluminiumvalue was found in formulae designed for inbornerrors of metabolism (443Æ 112mg lÀ1), an intermedi-ate level was found for formulae without lactose

(399Æ 451mg lÀ1) and, finally, the lowest content wasfound in hypoallergenic formulae (294Æ 945mg lÀ1).The probable source of aluminium contaminationmight be the added formula ingredients (calcium andphosphate salts, vitamins, and other minerals) and thecomplex processing operations in which infant for-mulae come into contact with aluminium-containing

surfaces and equipment for the long periods that theirpreparation requires.

The high levels of aluminium in formulae designed fordiverse pathologies is an additional source of alumin-ium for infants fed on these. This may represent ahealth hazard for this risk group of infants. Among theformulae without lactose, mainly casein-based for-mulae, one formula had an aluminium content of 1439mg lÀ1. Similarly, there were two hypoallergenicformulae with aluminium values of 2649 and2720mg lÀ1, both based on hydrolysed protein (caseinhydrolysate)andfromthesamemanufacturer(figure1).

Table 3 shows that the aluminium content of casein-hydrolysed hypoallergenic formulae (654Æ1233mg lÀ1) is significantly higher ( p¼ 0.015, Mann– Whitney U -test) than whey-hydrolysed hypo-allergenic formulae (190Æ 110mg lÀ1). These contentsare explained by the necessity for aggressive treatmentin order to modify highly the raw material, whichmeans that formulae can be exposed to a significantamount of aluminium during manufacture fromthe chemicals used, machinery and dust particles.Specialized formulae also contained higher levels of substances such as calcium, iron and citrate com-plexes that may enhance aluminium absorption fromthe gastrointestinal tract (Sahin et al. 1995). Thus, theincreased bioavailability of aluminium and theelevated plasma concentration noted in the infantsfed on casein-hydrolysed formulae (Hawkins et al.1994) are not unexpected findings. These may turnout to be recognized risk factors that must be takeninto consideration.

The aluminium content of premature formulae (421Æ137mg lÀ1) is significantly higher than that of formulaefor infants at term (non-adapted starter, p¼ 0.042;adapted starter, p¼ 0.010). These results are consis-tent with those reported elsewhere (Coni et al. 1993,Ballabriga et al. 1994, Hawkins et al. 1994, Plessi et al.

1997) and explained to be a result of the typeof processing or use of aluminium-containing foodadditives.

Preterm infant formulae are regarded as a potentialsource of aluminium exposure. The risk of aluminiumaccumulating in these formulae may be raised

Table 3. Aluminium concentrations in different types of infant formulae with regard to the main protein contained ( g l À1).

Infant formula

Aluminium

n MedianÆSD Range

Preterm formulaWhey-based 4 402Æ43 339–421Skim-milk-based 1 726 – Whole-milk-based 1 436 – Casein hydrolysed 1 317 –

Starter formulaNon adapted

Skim-milk-based 3 136Æ115 118–325Whole-milk-based 1 368 –

AdaptedWhey-based 6 181Æ177 158–573Casein-based 3 126Æ38 68–139Skim-milk-based 2 190Æ83 132–249Whole-milk-based 5 338Æ114 261–541

Follow-up formulaWhey-based 1 334 – Skim-milk-based 13 223Æ220 66–788Whole-milk-based 5 296Æ185 121–459

Specialized formulaWithout lactose formula

Casein-based 6 460Æ439 309–1439Skim-milk-based 1 102 –

Hypoallergenic formulaWhey hydrolysed 7 190Æ110 105–412Casein hydrolysed 5 654Æ1233 298–2720

Inborn errors dietCasein-based 1 307 – Skim-milk-based 1 518 – Whey hydrolysed 1 655 – Casein hydrolysed 4 449Æ98 335–529Free aminoacids 2 401Æ51 364–438No protein 1 370 –

Soya formulaSoy-based 7 573Æ1132 313–3479




especially in preterm infants with renal immaturitylinked to a low aluminium clearance due to decreasedurine excretion (Sedman et al. 1984b). Moreover, thehypothesis that aluminium overload may be due to anincreased intestinal absorption during the first weeksof neonate life has been reported. Both increasedabsorption and low excretion could explain the highplasma values observed in healthy premature and lowbirth weight newborns (Sedman et al. 1984b, Puntiset al. 1986, Stockhausen et al. 1990).

Aluminium absorption in infancy has been investi-gated in a previous study on antacids (Chedid et al.1991), and the risk of toxicity as a result of exposureto drugs containing aluminium or aluminium con-tamination of infant formulae has been confirmed. Inaddition, adverse effects on bone mineralizationassociated with high plasma aluminium levels ininfants have been reported (Bougle et al. 1997).

Although direct inhibiting effects of aluminium onbone mineralization have not fully been explained,they seem to be a consequence of interactions withboth calcium and phosphorus enhanced by the highaluminium content in premature formulae. In spite of this, taking into account the nutritional benefits of

premature formulae, it is necessary to keep up to datewith these formulations and request manufacturers tototal restructure older production processes in orderto achieve a further reduction in the aluminium levelsin this kind of infant formula.

Concentrations of aluminium in the soya formulaeranged from 313 to 3479mg lÀ1 (median 573Æ1132mg lÀ1). These formulae provide the highestaluminium levels and as table 2 shows, the aluminiumcontent is significantly higher than in standardformulae (adapted formulae, p¼ 0.001; non-adaptedformulae, p¼ 0.024; follow-up formulae, p¼ 0.004).High levels have also been reported elsewhere, callinginto question the nutritional safety of soya formulaefor neonate feeding.

Aluminium is naturally present in soybeans, andaluminium impurities that are contained in otherbasic components of the soya formulae or caused by

contamination during manufacture represent themost probable reasons for such high aluminiumlevels in soya formulae. The manufacture of soyaformulae requires crushing, refining and washingprocedures to isolate the soy protein from the soyabean. Treatment with calcium hydroxide is carried

Figure 1. Aluminium distributions in infant formulae provided by different manufacturers ( g l À1).




out, which often contains high amount of aluminiumas an impurity. It therefore seems likely that asignificant contribution to the final aluminium con-centration in soya formulae may be attributed toadventitious contamination during the isolationrather than the intrinsic aluminium content of soyabeans. In this survey, we found one formula with an

aluminium content of 3479 mg l

À1

that was very highwith respect to other infant formulae. This level iscomparable with other reported values where, forexample, several formulae had aluminium content of anything up to 5076 mg lÀ1 (Woollard et al. 1990).

In order to know the influence of the aggregationstate on the amount of aluminium found in infantformula, we evaluated statistically the aluminiumcontents in both powder and ready-to-use liquidformulae. The industrial production process of milk-based powder formulae involves several operations(warming, centrifugation, drying, homogenizing,bleeding, sterilization, packing, etc.), which bring it

into contact it with metallic surfaces of industrialmachinery. In the case of ready-to-use formulae,processing is quite different. The critical process forliquid formulae is the treatment in cationic inter-change columns, besides the addition of vitamins ormineral salts. For 14 pairs of infant formulae (powderand liquid forms), the aluminium content didnot differ significantly ( p¼ 0.060, Wilcoxon’s test)between powder and liquid data groups, although thealuminium content in ready-to-use formulae had atendency to be lower than in powder formulae, as inrecent studies (Ikem et al. 2002) but in contrast toearlier findings (Gruskin 1991).

Aluminium levels for both powder and liquidformulae for the different types of infant formulaeare summarized in table 4. Although the differencefound in aluminium might seem surprising, it shouldbe noted that liquid formulae started to be sold in

Spain only during the past decade, and are manu-factured with novel processing operations and betterattention to reduced contamination, in contrast toolder powder infant formulae.

Manufacturer and determined values

Figure 1 shows the concentrations of aluminiumdetermined in the different infant formula belongingto the nine manufacturers which can be examined forindications of aluminium content being related toprocessing operations. The meticulous care shown bymanufacturers 7 and 2 (150Æ 90mg lÀ1, n¼ 5 years,142Æ 929mg lÀ1, n¼ 14, respectively) is curious, whilemanufacturers 3 and 6 (436Æ 67mg lÀ1, n¼ 7 years,416Æ 126mg lÀ1, n¼ 8) show quite the opposite with alarge range of aluminium concentration (338–541 and261–573mg lÀ1, respectively). It is also necessary tomention manufacturer 8, which includes numerous

(n¼ 16) and complex formulations in its stock with adiscrete aluminium contribution from prepared infantformulae.

When the different infant formulae were consideredas a whole, median aluminium content was 340 mg lÀ1

(where percentiles 25 and 75 were 177 and 478 mg lÀ1,respectively). However, seen separately, median alu-minium values of different brands ranged from 142 to436mg lÀ1. Considering both points, recommendingan upper limit to be set in infant formulae around300–400mg lÀ1 is a reasonable target that manufac-turers could take on without excessive economic cost.

In view of these results and moves to reduce the levelof aluminium in infant formula in other Europeancountries (Baxter et al. 1991), it seems suitable to callfor an effort (1) to control as far as possible thecritical points of aluminium contamination, (2) tomodify the industrial handling process in order to

Table 4. Aluminium levels ( g l À1) from different types of infant formulae studied attending to aggregation state (powder orliquid formulae).

Formula

Powder Ready-to-use

n Mean Median Range n Mean Median Range

Preterm formula 6 467 428Æ 20 317–726 1 340 – – Starter formula

Non-adapted 3 271 325Æ 18 118–368 1 140 – – Adapted 12 273 215Æ 26 130–573 4 187 172Æ 14 68–340

Follow-up formula 13 313 340Æ 46 66–788 6 245 223Æ 14 150–460Specialized formula

Hypoallergenic 10 784 338Æ 1014 120–2720 2 201 201Æ 19 105–300




obtain formulations in which the concentration of aluminium is less than 300mg lÀ1 in premature andstandard formulae, and 400 mg lÀ1 in the otherformulae, a criterion which is a little more restrictivethan the recommendation suggested by Simmer et al.(1990), and (3) to analyse frequently or routinely thealuminium contents, especially in those formulae with

potential impact such as premature or soya formulae.

Estimated dietary aluminium intake

Assuming that the newborns in each age periodobserve a similar feeding regime, the aluminiumintake for infants fed on different types of infantformulae has been estimated according to recom-mended doses and the feeding tables specified by themanufacturers. Table 5 shows the daily amount of aluminium supplied by different infant formulae.

Dietary aluminium intake estimated by Dabeka andMcKenzie (1990) through human milk was 2 and3 mg dayÀ1 for Canadian infants 0–1 and 1–3 monthsold, respectively. Assuming a human milk referencevalue from published levels (3–79mg lÀ1), a daily milkintake of 200 mlkgÀ1 implies a daily aluminiumintake of 2.4–63.2 mgdayÀ1 (0.6–15.8 mg kgÀ1 dayÀ1),which is slightly higher than previous estimatedintakes.

Figure 2 compares the weekly aluminium intake(percentage of PTWI) estimated for each type of infant formulae studied.

Starter and follow-up formulae contributed thelowest aluminium intake (about 4% PTWI), special-ized formulae, such as hypoallergenic formulae gavean intermediate intake (11–12% PTWI) and soyaformulae contributed the highest intake (15% PTWI).Preterm intakes were calculated taking into account

the paediatric specifications or manufacturers guide-lines and infant weight. Figure 3 contains the esti-mated weekly aluminium intake by infants fed onpreterm formulae.

These formulae provide an intake similar to specialformulae (around 8–10% PTWI), which is far greaterthan that supplied on average by Spanish human milk

(0.1%) or the upper limit human milk reference value(1.3%). Aluminium levels ranging from 6000 to8000mg lÀ1 would be required for the hypotheticalcase of an infant formula to exceed the PTWI. In fact,for the highest aluminium level determined in theinfant formulae analysed (3479 mg lÀ1, soya formula),the daily aluminium intakes would be approximately45–55% PTWI.

Complementary contribution to daily aluminiumintake from drinking water

The aluminium concentration in natural waters variessignificantly depending on numerous physicochemi-cal, mineralogical and geochemical factors. Watertreatment in purifying plants includes a coagulationprocess using aluminium sulphate to remove organicmatter. The beneficial effect of the use of this metal intreated water is recognized. Nevertheless, in spite of good operating conditions, a residual amount isretained and supplied in drinking water. EuropeanCommunity legislation (Directive 80/778/EEC) andWHO guidelines have established an acceptablealuminium concentration of 200mg lÀ1 in drinkingwater (Commission of the European Communities1980, WHO 1998). This authorized aluminium level issimilar to that found in standard infant formula. As aconsequence, it is of interest to evaluate aluminiumexposure from tap water and its relative contributionto dietary intake of infants fed on reconstituted

Table 5. Daily intakes of aluminium for infants fed on infant formulae and drinking water used in the reconstitution of powder formulae ( g dayÀ1).

Starter formulaFollow-up

formula

Specialized formulaSoya

formulaDrinking

waterAge Non-adapted Adapted Without lactose Hypoallergenic Inborn errors diet

0–2 weeks 148 155 – 387 510 381 538 12.23–4 weeks 188 203 – 516 608 508 731 16.22 months 238 244 – 597 754 574 798 20.33 months 238 264 – 654 778 680 915 20.34–5 months 283 295 – 761 986 785 1055 23.76 months 257 302 296 793 854 846 1133 14.27 months – – 249 – – – – 10.8




powder formulae. In accordance with the legislationcurrently in force, taking into consideration the highdaily intake of newborns with respect to body weight,tap water used in prepared formulae could supplysimilar aluminium amounts to those provided by theinfant formulae itself.

The aluminium content of tap water was determinedas 22.6Æ 4.5mg lÀ1 (range 4.5–172.4 mg lÀ1). There wasno significant difference between aluminium levelsfound in rural (24.2Æ 9.3 mg lÀ1) and urban (22.3Æ2.2 mg lÀ1) sources. The mean aluminium value wasused to establish the influence of aluminium content in

Figure 2. PTWI (%) for aluminium estimated from infant formulae.

Figure 3. Weekly dietary aluminium intake for infants fed on premature infant formulae and human milk (mg weekÀ1).




drinking water on the final concentration of recon-stituted infant formulae. The amount provided by tapwater used in reconstitution of premature formulae issimilar to the lower limit found in human milk,approximately 0.05–0.06% of the PTWI. At the sametime, the aluminium supply to another infant for-mulae, is also low, about 10% of standard formulae

and 2–3% of special and soya formulae (table 5).Our results indicate that aluminium exposure fromtap water used for formula preparation is not clearlya potential risk. However, it should be appreciatedthat the aluminium concentration in drinking watercan vary significantly depending on the geographicand geochemical medium where the water supply orspring are sited.

Interactions between aluminium and other traceelements

The levels of trace elements in infant formulae aregenerally higher than in human milk. In recent years,changes in the levels of these elements in infantformulae have been introduced in the light of improved knowledge of infant requirements. Traceelements normally added to the raw material (cow’smilk or soy) are supplied in an inorganic form,sometimes at a high level to compensate for thelower bioavailability of infant formulae (Bra ¨ tter1996). There is little information on the impact of these changes or of other added elements on the finalconcentration of potential toxic trace elements,

including aluminium.In order to study trends in the levels of certainessential trace elements in relation to the aluminiumlevel in infant formulae, a comparative statisticalanalysis using the Spearman coefficient between thealuminium and different trace element contents wascarried out. Aluminium behaves similarly to iron inmany biological systems (Goyer 1997). As can be seenin figure 4, an inverse correlation was established forfollow-up formulae between both concentrations( p¼ 0.044). Cow’s milk is a very poor source of iron. A total of 14% of iron occurs in milk fat, about24% is bound to casein, while 29% is bound to whey

protein and 32% is associated with a low molecularweight fraction (Fransson and Lonnerdal 1983).Follow-up formulae are mainly based on skimmedand whole cow’s milk. This interaction could beattributed to the lower amount of iron added towhole-milk-based formulae, which is linked to a

higher aluminium content as discussed above. Onthe other hand, for those follow-up formulae based onskimmed milk, the iron supplementation is importantand the manufacturing process is simpler.

The distribution of iron added to infant formuladepends on the chemical form of iron supplementused. Supplementing with Fe(II) results in a formula

with a high iron content in the fat fraction(Hegenauer et al. 1979), whereas if Fe(III) is used,the iron is bound to casein micelles (Demott andDincer 1976). Probably, the existing correlation inwhole-milk-based formulae ( p¼ 0.010) could corre-spond with the different chemical form used in ironsupplementing and the technological process used bymanufacturers. The same consideration might beestablished for formulae designed for inborn errorsof metabolism ( p¼ 0.013), although the complexity informulations included in these groups does not allowa clear explanation.

Another negative correlation ( p¼

0.004) was foundbetween zinc and aluminium concentrations in stan-dard infant formulae (figure 5). This corresponds to

Figure 5. Zinc versus aluminium in standard infant formulae (starter and follow-up formulae).

Figure 4. Iron versus aluminium in follow-up infant formulae.




changes in the protein profile by manufacturers, sinceas in cow’s milk most of the zinc is in the skimmedmilk fraction and 95% is associated with caseinmicelles (Blakeborough et al. 1983, Sing et al. 1989).Lower aluminium and higher zinc concentrations werefound in casein-based formulae, followed by inter-mediate values of both elements in whey- and

skimmed-milk-based formulae and, finally, whole-milk-based formulae provided the highest aluminiumand lowest zinc added concentrations. This same trendwas observed for starter and follow-up formulae bythemselves with a statistical significance correlation

p¼ 0.004 and 0.030, respectively.

Finally, a positive correlation was established forhypoallergenic formulae between manganese andaluminium contents ( p¼ 0.006). Manganese is dis-tributed in cow’s milk with 67% binding to casein,18% to low molecular weight proteins and 1% to thelipid fraction (Lonnerdal et al. 1985). This fits withboth higher aluminium and manganese contents found

in hydrolysed-casein-based formulae and lower valuesof these elements in hydrolysed-whey-based formulae.

Conclusions

It has been shown that of the infant formulae studied,the aluminium level was higher than that found inhuman milk, especially for complex formulae such aspremature, specialized and soya. International pae-diatric organizations (AAP 1996) have requested thatmanufacturers reduce the aluminium level in infant

formulae, particularly with respect to soya andpreterm formulae. The results show that specializedformulae based on hydrolysed protein like caseinmust also be included in this risk formulae groupowing to their high aluminium content.

High aluminium infant formulae should be evaluatedby manufacturers, which should routinely monitoraluminium concentrations, especially in those formu-lations prepared for premature or low-birth neonates.Based on the present knowledge about aluminiumtoxicity, it is appropriate to set a maximum guidelinevalue for infant formula at 300 mg lÀ1, which is achiev-

able by manufacturers. Most infants who consumedinfant formulae containing more than this limit had araised plasma aluminium concentration and so couldbe at risk of aluminium toxicity (Hawkins et al. 1994).

Speciation studies are now required to establish andcharacterize the chemical forms in which aluminium

is present in human milk and infant formulae (Bra ¨ tteret al. 1998), with the goal of determining the realtoxicity of aluminium compounds and evaluating thetrue risk of feeding neonates on infant formulae.

Acknowledgements

The authors thank Gobierno de Navarra for financial

support.

References

AAP (AMERICAN ACADEMY OF PEDIATRICS), 1996, Aluminum toxicityin infants and children. Pediatrics, 97, 413–416.

BALLABRIGA, A., MORENO, A., and DOMI ´NGUEZ, C., 1994,Aluminium. Annales Nestle , 52, 118–131.

BAXTER, M. J., BURRELL, J. A., CREWS, H., and MASSEY, R. C., 1991,Aluminium levels in milk and infant formulae. Food Additivesand Contaminants, 8, 653–660.

BIEGO, G. H., JOYEUX, M., HARTEMANN, P., and DEBRY, G., 1998,Determination of mineral contents in different kinds of milkand estimation of dietary intake in infants. Food Additives and Contaminants, 15, 775–781.

BISHOP, H. J., MOLEY, R., CHIR, B., DAY, J. P., and LUCAS, A., 1997,Aluminium neurotoxicity in preterm infants receiving intrave-nous-feeding solutions. New England Journal of Medicine, 336,

1557–1561.BLAKEBOROUGH, P., SALTER, D. N., and GURR, M. I., 1983, Zinc

binding in cow’s and human milk. Biochemical Journal , 209,505–512.

BLOODWORTH, B. C., HOCK, C. T., and BOON, T. O., 1991, Aluminiumcontent in milk powders by inductively-coupled argon plasma-optical emission spectrometry. Food Additives and Contaminants, 8, 749–754.

BOUGLE, D., BUREAU, F., MORELLO, R., GUILLOIS, B., and SABATIER,J. P., 1997, Aluminium in the premature infant. Trace Elementsand Electrolytes, 14, 24–26.

BRA ¨ TTER, P., 1996, Essential trace elements in the nutrition of infants.Therapeutic Uses of Trace Elements, edited by J. Ne ` ve, P.Chappuis and M. Lamand (New York: Plenum), pp. 59–62.

BRA ¨ TTER, P., NAVARRO, I., NEGRETTI DE BRA ¨ TTER, V., and RAAB, A.,1998, Speciation as an analytical aid in trace element research ininfant nutrition. Analyst, 123, 821–826.

CHEDID, F., FUDGE, A., TEUBNER, J., JAMES, S. L., and SIMMER, K.,1991, Aluminium absorption in infancy. Journal of Paediatricsand Child Health, 27, 164–166.

COMMISSION OF THE EUROPEAN COMMUNITIES, 1980, CouncilDirective of 15 July 1980 relating to the quality of waterintended for water consumption (80/778/EEC). Official Journal

of European Communities Legislation, L229, 11–29.CONI, E., BELLOMONTE, G., and CAROLI, S., 1993, Aluminium content

of infant formulas. Journal of Trace Elements and Electrolytes inHealth and Disease, 7, 83–86.

CONI, E., FALCONIERI, P., FERRANTE, E., SEMERARO, P., BECCALONI,E., STACCHINI, A., and CAROLI, S., 1990, Reference values foressential and toxic elements in human milk. Annali dell’IstitutoSuperiore di Sanita `, 26, 119–130.




DABEKA, R. W., and MCKENZIE, A. D., 1990, Aluminium levels inCanadian infant formulae and estimation of aluminium intakesfrom formulae by infants 0–3 months old. Food Additives and Contaminants, 7, 275–282.

DEMOTT, B. J., and DINCER, B., 1976, Binding added iron to variousmilk proteins. Journal of Dairy Science, 59, 1557–1559.

FERNA ´ NDEZ-LORENZO, J. R., COCHO, J . A ., REY-GOLDAR, M. L.,COUCE, M., and FRAGA, J. M., 1999, Aluminum contents of human milk, cow’s milk and infant formulas. Journal of Pediatric Gastroenterology and Nutrition, 28, 270–275.

FRANSSON, G. B., and LONNERDAL, B., 1983, Distributions of traceelements and minerals in human and cow’s milk. PediatricResearch, 17, 912–915.

FREUNDLICH, M., ZILLERUELO, G., ABITBOL, C., STRAUSS, J.,FAUGERE, M. C., and MALLUCHE, H. H., 1985, Infant formulaas a cause of aluminum toxicity in neonatal uraemia. Lancet, ii,

527–529.FREUNDLICH, M ., ZILLERUELO, G ., STRAUSS, J . , ABITOL, C., and

MALLUCHE, H. H., 1990, More on aluminum toxic effects inchildren with uraemia. Journal of Pediatrics, 117, 1007–1009.

GOYER, R. A., 1997, Toxic and essential metal interactions. Annual Review of Nutrition, 17, 37–50.

GRUSKIN, A. B., 1991, Aluminium toxicity in infants and children.Trace Elements in Nutrition of Children II , edited by R. J.Chandra (New York: Raven), pp. 15–25.

HAWKINS, N. M., COFFEY, S., LAWSON, M. S., and DELVES, T., 1994,

Potential aluminium toxicity in infant fed special infant formula.Journal of Pediatric Gastroenterology and Nutrition, 19, 377–381.HEGENAUER, J., SALTMAN, P., LUDWIG, D., RIPLEY, L., and BAJO, P.,

1979, Effects of supplemental iron and copperon lipid oxidationin milk. I. Comparison of metal complexes in emulsified andhomogenized milk. Journal of Agricultural and Food Chemistry,27, 860–867.

IKEM, A., NWANKWOALA, A., ODUEYUNGBO, S., NYAVOR, K., andEGIEBOR, N., 2002, Levels of 26 elements in infant formulafrom USA, UK, and Nigeria by microwave digestion and ICP-OES. Food Chemistry, 77, 439–447.

KOO, W. W., KAPLAN, L. A., and KRUG-WISPE, S. K., 1988,Aluminium contamination of infant formulas. Journal of Parenteral and Enteral Nutrition, 12, 170–173.

KRACHLER, M., PROHASKA, T., KOELLENSPERGER, G., ROSSIPAL, E.,and STINGEDER, G., 2000, Concentrations of selected traceelements in human milk and in infant formulas determined by

magnetic sector field inductively coupled plasma-mass spectro-metry. Biological Trace Element Research, 76, 97–112.

LONNERDAL, B., KEEN, C. L., and HURLEY, L. S., 1985, Manganesebinding proteins in human and cow’s milk. American Journal of Clinical Nutrition, 41, 550–559.

MORENO, A., DOMI ´ NGUEZ, C., and BALLABRIGA, A., 1994, Aluminiumin neonate related to parenteral nutrition. Acta Paediatrica, 83,

25–29.

NAVARRO, I., and ALVAREZ, J. I., 2002, Selenium content of Spanishinfant formula and human milk: Influence of protein matrix,interactions with other trace elements and estimation of dietaryintake by infants. Journal of Trace Elements in Medicine and Biology (in press).

NAVARRO, I., ALVAREZ, J. I., and MARTI ´N, A., 2000, Estimated dailyintakes and concentrations of essential trace elements in infantformulas. Trace Elements in Man and Animals (TEMA 10),edited by A. M. Roussel, R. A. Anderson and A. E. Favier(New York: Kluwer/Plenum).

NAVARRO, I ., MARTI ´N, A., and VILLA, I., 1996, Dietary intake of essential trace elements in infant nutrition. Prenatal and Neonatal Medicine, 1 (suppl. 1), 365.

PLESSI, M., BERTELLI, D., and MONZANI, A., 1997, Determination of aluminium and zinc in infant formulas and infant foods.Journal of Food Composition and Analysis, 10, 36–42.

PUNTIS, J. W., HALL, K., and BOOTH, I. W., 1986, Plasma aluminiumand prolonged parenteral nutrition in infancy. Lancet, ii, 1332– 1333.

SAHIN, G., AYDIN, A., ISIMER, A., O ¨ ZALP, I., and DURU, S., 1995,Aluminium content of infant formulas used in Turkey.Biological Trace Element Research, 50, 87–96.

SEDMAN, A. B., KLEIN, G. C., MERRIT, R. J., MILLER, K. O., WEBER,W., GILL, W . L ., ANAUD, H., and ALFREY, A. C., 1985,Evidence of aluminum loading in infants receiving intravenoustherapy. New England Journal of Medicine, 312, 1337–1343.

SEDMAN, A. B., MILLER, N. L., WARADY, B. A., LUM, G. M., andALFREY, A. C., 1984a, Aluminum loading in children withchronic renal failure. Kidney International , 26, 201–204.

SEDMAN, A. B., WILKENING, G. N., WARADY, B. A., LUM, G. M., andALFREY, A. C., 1984b, Encephalopathy in childhood secondaryto aluminum toxicity. Journal of Pediatrics, 105, 836–838.

SIMMER, K . , FUDGE, A . , TEUBNER, J., and JAMES, S. L., 1990,Aluminium concentrations in infant formulae. Journal of Paediatrics and Child Health, 26, 9–11.

SING, H., FLYNN, A., and FOX, P. F., 1989, Zinc binding in bovinemilk. Journal of Dairy Research, 56, 249–263.

STOCKHAUSEN, H. B., SCCHROD, L., BRA ¨ TTER, P., and RO ¨ SICK, U.,1990, Aluminium loading in premature infants during intensivecare as related to clinical aspects. Journal of Trace Elements and Electrolytes in Health and Disease, 4, 209–213.

WHO (WORLD HEALTH ORGANIZATION), 1989, Evaluation of CertainFood Additives and Contaminants. Thirty-Third Report of the

Joint FAO/WHO Expert Committee on Food Additives. WHOTechnical Representative Series, 776, 1–64.

WHO (WORLD HEALTH ORGANIZATION), 1998, Aluminium. Guide-lines for Drinking-water Quality, 2nd edn (Geneva: WHO),pp. 3–13.

WOOLLARD, D. C., PYBUS, J., and WOOLLARD, G. A., 1990, Aluminiumconcentrations in infant formulae. Food Chemistry, 37, 81–94.


aluminum infant formula-1

Documents