Trends in Food Science & Technology 13 (2002) 293–311
Review
Intrinsic time
temperature
integrators for heat
treatment of milk
Wendie L. Claeys,
Ann M. Van Loey
and Marc E. Hendrickx*
Department of Food and Microbial Technology,
Faculty of Agricultural and Applied Biological
Sciences, Katholieke Universiteit Leuven, Kasteelpark
Arenberg, 22, B-3001 Heverlee, Belgium (tel: +32-16-
32-15-85; fax: +32-16-32-19-60; e-mail:
marc.hendrickx@agr.kuleuven.ac.be)
This paper reviews the available research information on
intrinsic time temperature integrators (TTIs) for thermally
processed milk. These are specific indicators present or
formed in milk that allow direct and quantitative measure-
ment of the impact of the process without knowledge of
the actual thermal history. General principles concerning
the characterization of intrinsic TTIs as well as the applic-
ability and limitations of some milk compounds considered
as potential intrinsic TTIs are discussed.
# 2002 Elsevier Science Ltd. All rights reserved.
Introduction
Milk contains approximately 87.3% water, 4.2% fat,
4.6% lactose, 3.25% protein, and 0.65% mineral sub-
stances. Because of its high nutritional value, milk is an
excellent medium for microbiological growth. Conse-
quently, fresh milk necessitates a heat treatment in order to
guarantee a safe and shelf-stable product. Unfortunately,
* Corresponding author. Tel.: +32-16-32-15-85; fax: +32-16-32-
19-60.
0924-2244/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S 0 92 4 - 2 24 4 ( 0 2 ) 0 0 16 4 - 4
next to preservation, heating generally causes a significant
loss of organoleptic and nutritional quality. Depending on
the intensity of the process, different changes take place
such as decrease of pH, precipitation of calcium phos-
phate, denaturation of whey proteins and their interac-
tion with casein, vitamin destruction, Maillard reaction,
etc. (Burton, 1984; Jeurnink & De Kruif, 1993; Lewis,
1994; Schaafsma, 1989). Furthermore, an undesirable
precipitation of denatured proteins and minerals can be
formed on the walls of heat exchangers (de Jong, Bou-
man, & Van Der Linden, 1992; Gotham, Fryer, &
Pritchard, 1992; Lewis, 1994). This deposit or ‘fouling’
decreases the heat transfer of the system so that higher
energy input is required leading to larger product losses
and higher wastewater production (Sandu & Singh,
1991). To overcome overprocessing of the product as
well as to control if heating was adequate from a
safety point of view, criteria need to be defined. Such
criteria not only must guarantee the correctness of a
heat treatment, but can also be applied for quality
management at all steps in the production process
(e.g. HACCP). Moreover, these criteria must result in
products on the market that comply with their label-
ling in terms of processing (thermization, pasteuriza-
tion, sterilization). In order to define such criteria,
the impact of a thermal process on milk has to be
known.
Due to inherent limitations of in situ methods (detec-
tion limit, sample size, recontamination, etc.) and of
physical mathematical methods (lack of accurate ther-
mophysical data, etc.), time temperature integrators
(TTIs) can offer some potential. TTIs are heat sensitive
components extrinsic or intrinsic to the food product
that allow direct and quantitative measurement of the
impact of the process on a safety or quality attribute
without knowledge of the actual thermal history
(Hendrickx et al., 1994; Taoukis & Labuza, 1989).
Generally, TTIs can be classified according to their
working principle (biological, chemical, physical),
response (single, multi), origin (extrinsic, intrinsic),
application (dispersed, permeable, isolated) and loca-
tion (volume average or single point) (Van Loey,
1996). In the context of the evaluation of processing
authenticity, only intrinsic TTIs, which are homo-
geneously distributed in the product allowing the
evaluation of the volume average process impact, are
of interest.
294 W.L. Claeys et al. / Trends in Food Science & Technology 13 (2002) 293–311
The aim of this paper is to discuss the concept of
intrinsic TTIs and their application in the specific con-
text of liquid milk.
General principles
A TTI can be defined as a small measuring device (a
food component or property) that shows a time tem-
perature dependent, easily, accurately and precisely
measurable, irreversible change that mimics the change
of a target attribute (a safety or quality characteristic)
undergoing the same treatment (Hendrickx et al., 1994;
Taoukis & Labuza, 1989). Besides practical and ther-
mophysical requirements (e.g. accurate, fast and user-
friendly read-out, no influence on heat transfer), this
definition implies kinetic requirements of the system.
The temperature sensitivity factor of TTI and target
attribute must correspond, and additionally, the reac-
tion rate of the TTI must guarantee a measurable
response within the time temperature combinations of
interest. Consequently, in order to characterize poten-
tial TTIs, a detailed study of their kinetics under defined
time temperature conditions needs to be performed.
Basic principles of kinetics
Kinetics describe the evolution of a reaction (inacti-
vation, denaturation, production) as a function of time.
Generally, the rate of a chemical reaction can be
described by:
dX
v¼ ¼ kX n ;
dt
with v the rate of the reaction, X the activity or con-
centration of the compound of concern at treatment
time t, k the reaction rate constant at the temperature
studied, and n the order of the reaction. The effect of
temperature can in many cases be expressed by the
Arrhenius relation (Arrhenius, 1889), in which the tem-
perature dependence of the rate constant k is quantified
by the activation energy Ea (J/mol) according to:
  
Ea 1 1 thermal phase of heating up and cooling down, kinetics
k ¼ kref exp  ; ð2Þ
R Tref T
where R is the universal gas constant (8.314 J/mol,K), T
the temperature concerned, and kref the reaction rate
constant at reference temperature Tref. When kinetics
are analysed under isothermal conditions (constant
temperature), where extrinsic/intrinsic factors and thus
k can be assumed constant, expression (1) can be simply
integrated giving the equation:
ðaÞ for a 1st order reaction ðn ¼ 1Þ:
 
X
ln ¼ kt;
X0
ðbÞ for an nth order reaction:
ð4Þ
X ð1nÞ ¼ X0ð1nÞ þ ð1  nÞkt:
In the area of food processing and preservation, 1st
order kinetics are often described by the Thermal Death
Time model (TDT) (Bigelow, 1921), a concept usually
applied to describe microbiological destruction kinetics.
In this model the decimal reduction time D is defined as
the time required to reduce the initial response value by
90% at a constant temperature. For 1st order kinetics D
and k are directly related. The temperature dependence
of the D-value is characterized by the z-value, which is
the temperature increase necessary to obtain a tenfold
decrease of D.
Another parameter often encountered in food science
literature is Q10, a value that describes the temperature
dependence of a reaction as the factor by which the
reaction rate changes when the temperature is altered by
10 C.
Identification of TTIs
Research on potential intrinsic TTIs can be divided
into three phases (Fig. 1). The first phase or inventory
phase consists of the identification (i) of the target
attribute, i.e. a search for relevant microbiological indi-
ces corresponding to currently applied heat treatments
and their destruction kinetics, and (ii) of possible TTIs,
i.e. a selection of relevant milk components with a view
ð1Þ
to the kinetic and technical base criteria a TTI has to
meet. Once components are selected as potential TTIs,
in a second phase a detailed study of their kinetics needs
to be performed. A first assessment of the kinetic model
and associated parameters describing formation, inacti-
vation or denaturation of the TTI under study is made
under idealized conditions by isothermal experiments or
experiments at a constant temperature. Since TTIs need
to quantify the integrated impact of time and tempera-
ture of a thermal process on the product and since such
processes are generally characterized by a non-iso-
need to be validated under variable temperature condi-
tions. Additionally, kinetics need to be evaluated for
variations associated with the raw material, since such
variations can influence the initial concentration or
activity of the TTI as well as its inactivation, denatura-
tion or formation kinetics.
In the final phase or implementation phase time-tem-
perature tolerance diagrams are constructed that com-
bine kinetic information concerning the destruction of
microbial and qualitative indices with the analysed
inactivation, denaturation or formation kinetics of the
ð3Þ selected TTIs. This phase also consists of a validation of
the TTIs on industrial samples of which the thermal
process history is known.
 W.L. Claeys et al. / Trends in Food Science & Technology 13 (2002) 293–311
Fig. 1. Overview of the strategy for the characterization of intrinsic TTIs.
Heat processes and microbiological indices
In order to select potential TTIs and to evaluate their
applicability, systematic data concerning thermal
destruction kinetics of target micro-organisms associated
with each heating process need to be available, con-
sidering safety (reduction of pathogens) as well as qual-
ity (reduction of spoilage organisms) of the product.
In the context of thermal processing of milk, generally
a distinction is made between thermization and heat
treatment (Lewis, 1994; Muir, 1996; Pearce, 1989;
Roberts, Pitt, Farkas, & Grau, 1998; Spreer, 1998).
Thermization is a mild heating (57–68 C/15–20 s) that
reduces viable numbers of vegetative psychrotrophic
micro-organisms, multiplication of which may generate
heat-stable proteinases and lipases. Psychrotrophic
growth found in raw milk includes species of Pseudo-
monas, Flavobacterium, Micrococcus, Bacillus, Enter-
obacter, Aeromonas and Alcaligenes. Thermization may
also provide a heat-shock to endospores of Bacillus
spp., which can germinate during succeeding storage
295
and are possibly inactivated when milk is subsequently
pasteurized (van den Berg, 1984). Although effective in
inhibiting spoilage, thermization offers no guarantee of
microbiological safety.
Heat treatment is a process resulting in the complete
inactivation of the indigenous milk enzyme alkaline
phosphatase. Dependent on the intensity of the process,
distinction is made between pasteurization and ster-
ilization. Pasteurization (minimum 71.7 C/15 s or
62.7 C/30 min, or any comparable time temperature
combination) aims to reduce populations of vegetative
bacterial pathogens to a level considered safe for public
health with a minimum of chemical, physical or orga-
noleptic changes. It specifically targets Mycobacterium
tuberculosis, a relatively heat-resistant non-sporogenic
bacterium that formerly was among the most serious
pathogenic bacteria present in milk. Amongst other
pathogens found in raw milk are Streptococcus mas-
titis, Salmonella, Staphylococcus aureus, Campylobacter
jejuni, Coxiella burnetii, Listeria monocytogenes,
296 W.L. Claeys et al. / Trends in Food Science & Technology 13 (2002) 293–311
Escherechia coli and Yersinia enterolitica (Spreer, 1998;
Walstra, Geurts, Noomen, Jellema, & van Boekel,
1999). Properly applied pasteurization not only kills all
common pathogens in milk, but also eliminates the
Gram-negative psychrotrophs, the most ordinary cause of
spoilage in raw and thermized milk (Muir, 1996). Pasteur-
ized milks can have well over 10 days of shelf-life under
refrigeration, and even more if milk is aseptically packed
(Roberts et al., 1998). Within the definition of pasteuriza-
tion additional distinction is made between ‘low tempera-
ture holding’ (LTH) and ‘high temperature short time’
(HTST) pasteurization, based on the inactivation of the
milk enzyme lactoperoxidase. HTST pasteurization is in the
range of 85–90 C for 15–20 s, although sometimes higher
temperatures, up to 100 C, are applied.
Pasteurization is inadequate to inactivate thermo-
resistant spores and to guarantee a commercially sterile
product. For example, endospores of Bacillus and Clos-
tridium spp. show varying levels of heat resistance, and
may survive minimal pasteurization, in particular if
present in high levels. The same applies for thermoduric
organisms, which include vegetative micro-organisms
showing an increased heat resistance such as Micro-
coccus spp., Enterococcus faecalis and faecium, and
some lactobacilli. And although pasteurization is cap-
able of killing S.aureus, enterotoxins formed in milk will
not be destroyed (Roberts et al., 1998). Pasteurization
does not destroy extracellular bacterial degradative
enzymes either; 60% or more of the proteinase and
lipase and about one third of the phospholipase activity
survive HTST pasteurization (Muir, 1996).
The main sterilization processes that can be dis-
tinguished are ultra-high-temperature (UHT) processing
(minimum 135 C/1 s) with direct or indirect heat trans-
fer, and the more classical in-bottle sterilization (110–
120 C/10–20 min). Their purpose is to destroy all
micro-organisms, including spores (e.g. the genera
Clostridium and Bacillus), and to ensure extended shelf-
life without need for refrigeration. Both processes pro-
duce a commercially sterile product and exhibit similar
kinetic data for discoloring, vitamin B1 and lysine loss,
hydroxymethylfurfural formation and whey protein
denaturation (Lewis, 1994). Nevertheless, UHT offers
some advantages in comparison to in-bottle steriliza-
tion, due to the fact that chemical reaction rates are less
sensitive to changes in temperature than thermal inacti-
vation of spores. Because of much shorter processing
times, UHT-milk is whiter and less caramelized, and has
undergone less protein denaturation and loss of heat-
sensitive vitamins compared to in-bottle sterilized milk.
One of the methods used to discriminate UHT and in-
bottle sterilization is the turbidity or Aschaffenburg test,
based on the different extents of protein denaturation
during both processes (FIL-IDF, 1972). However, the
test is reported to be incapable of distinguishing
between the two types of milk because some UHT pro-
cesses now used cause the same degree of whey protein
denaturation as conventional 2-stage sterilization (ster-
ilization before and after packaging) (Burton, 1984; de
Koning, 1984). Also, fat content is reported to markedly
influence the test (Fink & Kessler, 1988). Furthermore, it is
a semi-quantitative method, which gives no additional
information on the impact of the process, whereas a quan-
titative measurement of the impact is of great importance
in process design, evaluation, optimization and control.
There are also some drawbacks of UHT processing
such as the incomplete inactivation of the more heat-
resistant, proteolytic enzymes and the higher suscept-
ibility to gelation (Andrews, 1984; Lewis, 1994; Pearce,
1989). For example, most psychrotrophs such as Bacil-
lus spp. and Pseudomonas spp. are readily destroyed by
UHT treatment, but their lipolytic and proteolytic
enzymes may survive in sufficient quantities to cause
product deterioration during storage (Law, 1979). Next
to bacterial enzymes, also a number of native milk
enzymes detrimental to the milk quality or shelf-life
may survive UHT processing. A well-known example is
plasmin (EC 3.4.21.7), an alkaline milk proteinase
causing e.g. bitter flavor during storage of milk. During
keeping at room temperature, proteolysis may occur not
only because of the remaining plasmin after heating but
also because of plasmin converted from undenatured
plasminogen (de Koning, Badings, van der Pol, Kaper,
& Vos-Klompmaker, 1990; Saint-Denis, Humbert, &
Gaillard, 2001; Walstra & Jeness, 1984). The effects of
residual enzyme activity (bacterial and native) have
been reviewed by Muir (Muir, 1996). Also some spores
(e.g. Bacillus sporothermodurans) are able to survive
UHT (Petterson, Lemke, Hammer, Stackebrandt, &
Priest, 1996). The potential for survivors is critical rela-
tive to the initial spore load since time-temperature
parameters for UHT purposes are almost constant and
defined. Spore loads in raw milk are related to poor
farming practices, but may also fluctuate according to
the season (Phillips & Griffiths, 1986).
Destruction kinetics of some micro-organisms in milk
are presented in Table 1. Though, interpretation and use
of microbiological destruction data is hampered because
e.g. dependent on the source different data are reported.
Also heating conditions and medium must be con-
sidered, e.g. heat resistance of micro-organisms often
increases with the dry matter content of the medium.
Moreover, different species or even strains of one spe-
cies can have different heat resistances, and clumping of
cells can augment the heat resistance. An example is
Mycobacterium avium paratuberculosis, which has the
potential to survive commercial HTST pasteurization
processes (Grant, Hitchings, McCartney, Ferguson, &
Rowe, 2002; Stabel et al., 2001). Clumped cells of this
micro-organism are reported to be approximately twice
as heat resistant as single cells, with D-values of 97.7
and 47.5 s, respectively at 63 C (Grant, Rowe, Dundee,
 W.L. Claeys et al. / Trends in Food Science & Technology 13 (2002) 293–311 297

Table 1. Kinetic destruction data of some micro-organisms in heated milk

Micro-organism Tref ( C) Dref (min) z ( C) Ref.
Aeromonas 48 3.2–6.2 5.5–7.7 Roberts et al., 1996
Bacillus cereus 95 1.8 9.4 Roberts et al., 1996
Bacillus cereus spores 121 0.04 9.55 Walstra et al., 1999
Bacillus licheniformis 111 0.48 8 Walstra et al., 1999
Bacillus stearothermophilus spores 121 5.5 9.5 Walstra et al., 1999
Campylobacter jejuni 50 4.5 7 Walstra et al., 1999
Clostridium botulinum 115 0.21–0.3 7.9 Roberts et al., 1996
Clostridium botulinum spores 110 0.85 9.5 Walstra et al., 1999
Escherechia coli 62.8 0.13 4.6 Walstra et al., 1999
Listeria monocytogenes 65 0.1 6.6 Eckner, 1992; Walstra et al., 1999
Pseudomonas fluorescens 60 3.2 7.5 Walstra et al., 1999
Pseudomonas fragi 49 7–9 10–12 Walstra et al., 1999
Salmonella (6 spp.) 62.8 3 4.6 Walstra et al., 1999
Staphylococcus aureus 62.8 18.5 5.1 Walstra et al., 1999
60 0.9 9.5 Roberts et al., 1996
Yersinia enterolitica 58 1.6 4.26 Eckner, 1992

& Hitchings, 2001). Another point worth considering is
the possibility of microbial adaptation to more severe
processing conditions.
Study of kinetics
Characterization of kinetics describing inactivation,
denaturation or formation of potential TTIs starts with
analyzing the model and associated parameters under
idealized conditions by isothermal experiments. Next,
kinetics are validated for variability associated with
extrinsic (temperature) and intrinsic factors (raw material).
Constant temperature conditions
The quantification of a lethal impact by use of TTIs is
based on its kinetic parameters. It is therefore necessary
to measure these parameters with the greatest accuracy
and precision by well-designed experiments and proper
data-analysis. The kinetic model or the order of a reac-
tion can be defined by a trial-and-error procedure by
which linear behaviour is checked graphically on trends
or deviations, or by differential or integral methods
(Hill & Grieger-Block, 1980; van Boekel & Walstra,
1995). When the response value (X) changes only
slightly as a function of time, it is difficult to distinguish
between a 0, 1st and 2nd order model. According to
Arabshahi and Lund (1985), where possible reactions
should be followed through at least 4–5 half-lives (X
 X0). If the analytical method cannot measure such a
low response value, the longest treatment time should
be used (Lenz & Lund, 1980). According to Hill and
Grieger-Block (1980), reaction must be followed until at
least 50% conversion is reached. To measure linearity and
to verify the validity of the model, regression coefficients
(R2) can be calculated, residual plots checked for the
absence of trends or correlations, the F-value and its
probability level can be determined, t-statistics of the
slope can be carried out (Draper & Smith, 1981;
Motulsky & Ransas, 1987; Straume & Johnson, 1992).
Next to the model, distinction must be made between
different regression procedures. Commonly, model para-
meters are estimated by a two-step linear regression pro-
cedure on the experimental data-set, but alternatively, a
one-step non-linear regression procedure can be used.
Some authors proclaim the global approach to be more
efficient, primarily because in the individual two-step
analysis some unnecessary parameters are estimated
(Arabshahi & Lund, 1985; Cohen & Saguy, 1985; Har-
alampu, Saguy, & Karel, 1985), while others concluded
that overall, none of the possible methods is convin-
cingly superior to the other and that the performance of
a specific regression approach depends on the data-set
to which it is applied (De Cordt, 1994).
The precision of an estimate is mostly measured by
use of individual confidence intervals. In case of linear
models, these intervals are exactly defined and sym-
metric, and their determination is straightforward.
However, when model parameters are estimated simul-
taneously, resulting individual confidence intervals are
only approximate because covariances of the simulta-
neously estimated parameters are neglected (Motulsky
& Ransas, 1987; van Boekel, 1996). An alternative
technique taking the possible correlation between the
parameters into account, is the construction of 100
(1’) joint confidence regions (Bard, 1974; Draper &
Smith, 1981; Johnson, 1992; Metzler, 1981).
Although much more emphasis is often put on the
data analysis, the quality of the experimental data
greatly determines the quality of the parameters
obtained in terms of precision and accuracy. Moreover,
much of the data currently found in literature could
have been obtained with considerable less effort by
proper choice of experimental conditions. Box and
Lucas (1959) proposed an optimum design criterion for
non-linear models, also known as the D-optimal design.
Based on this principle, it has been shown that when the
model is known, maximal precision in determining p
298 W.L. Claeys et al. / Trends in Food Science & Technology 13 (2002) 293–311
parameter values is obtained if measurements are made
only in p experimental points. The number of replicates
per measurement should be at least 5 to make a reason-
able estimation of the variance possible. The location of
the p experimental points optimal for parameter esti-
mation is determined by the model and the nature of the
experimental error (Duggleby, 1981; Endrenyi, 1981;
Fedorov, 1972). Several authors applied the optimal
D-design criterion for establishing the location of sam-
pling points for different models, such as the Michaelis–
Menten equation (Duggleby, 1979), the TDT model under
both isothermal and non-isothermal conditions (Cunha,
Oliveira, Brandão, & Oliveira, 1997), the Weibull prob-
abilistic model (Cunha, Oliveira, & Oliveira, 1998). How-
ever, the optimal D-design does not take the nature or
variation of the experimental error into account.
Variable temperature conditions
Evaluation of kinetics for dynamic process conditions
can be performed by non-isothermal experiments.
Because in non-isothermal experiments temperature
varies with time, the assumption of a constant D- or
k-value does not longer apply and the integrated effect
of temperature on the reaction rate constant has to be
taken into account. When time–temperature profiles are
registered, kinetic parameters can be calculated by non-
linear regression using a numerical integration routine
(e.g. Simpson) (Carnahan, Luther, & Wilkes, 1969).
A concept to express the integrated time-temperature
impact of different heat processes is the ‘equivalent time
at reference temperature’ concept, referred to as the
processing value F (defined as the duration of a hypo-
thetical process at a constant reference temperature that
results in exactly the same impact on a specific food
attribute as the actual process to which the food was
subjected). In the canning industry, the reference tem-
perature is taken as 121.1 C and the resulting process
value is indicated as F0. For in-container sterilization of
milk, a reference temperature of 121.1 C can be used as
well, but not for UHT heating where temperature ran-
ges from 130 to 150 C (van Boekel & Walstra, 1995).
Variability of product related factors
In order to evaluate the general applicability of selec-
ted milk compounds as intrinsic TTIs for thermal pro-
cessed milk, kinetics need to be evaluated for variations
in milk composition. The widest variation usually
occurs in milk fat content; variation in protein is less and
in ash still less (Walstra & Jeness, 1984). The variation in
milk composition however is not precisely known,
mainly because so many factors affect it and because
many sources of variation are independent. Besides, not
only milk composition, but also the structure (e.g. the
size of fat globules) varies (Walstra et al., 1999).
Main factors affecting milk composition and proper-
ties are genetic (species, breed, individual), physiological
(stage of lactation, age of cow, illness), and environ-
mental factors (feed, climate). In the context of TTIs
environmental, and more specific seasonal variation is
of main interest.
Seasonal variations in milk composition are small to
negligible, with as a most pronounced change a decrease
in fat and protein content in the period May–June
(Johnson, 1978; Thorne, Hardin, Spain, & Hardin,
1995; Walstra et al., 1999). This decrease is often
attributed to a change in feed or temperature, but it has
been shown that cows fed the same ration month after
month and with only slight temperature variation still
exhibit this seasonal low (Johnson, 1978). Lactose is the
least variable of the gross components on a percentage
basis. Because of its osmotic relationship in milk, it
increases slightly as the other solids decrease, and vice
versa. As to the minor components, a marked increase is
observed in the level of colloidal calcium in raw milk
between December and February, whereas pH, inor-
ganic PO2-4 , Mg
2+
, citrate and K+ remained unchanged
(Pouliot & Boulet, 1995). (The months referred to are
appropriate for the Northern hemisphere.)
Also worth mentioning in the context of TTIs is arti-
ficially imposed variation in milk composition and
properties due to processing. An obvious example is
separation yielding skim milk and cream. Milk skimmed
by centrifugation has a very low fat content
(0.05–0.08%) compared to milk skimmed by means of
gravity creaming. By mixing skim milk and cream, milk
may be standardized to a desired fat content. Heat
treatment has a minor effect on milk fat itself. The fat
globule membranes with heat sensitive protein com-
pounds however, experience some modifications, hereby
changing agglomeration and creaming properties of the
fat globules (Spreer, 2001). Possible physical changes
in the fat globules during heating of milk are reviewed
by van Boekel and Walstra (1989). The role of fat in
protecting micro-organisms from thermal destruction
has been reported in literature (Davis & Wilbey, 1990).
Similarly, milk fat content can have an effect on the
concentration of the TTI in concern or its behaviour
during thermal processing. Several studies on the effect
of milk fat on a milk compound evaluated as an index
for heat treatment are reported, but contradictory
observations were made. The mechanism by which milk
fat affects heat-induced modifications, remains unclear;
it does not appear to be certain whether this is due
to viscous effects interfering with heat transfer and
reducing the holding time, differences in conductivity
or to more fundamental causes (Fernandez-Martin
& Montes, 1972; Morales & Jiménez-Pérez, 1999;
Pellegrino, 1994).
Available TTIs
Assessment of different classes of heat treatment of
milk can be performed by means of enzymes (alkaline
 W.L. Claeys et al. / Trends in Food Science & Technology 13 (2002) 293–311
phosphatase, g-glutamyl transferase, lactoperoxidase),
secondary products of heat treatment (hydro-
xymethylfurfural, lactulose, furosine), changes in whey
proteins (whey protein N index, heat number, relative
soluble whey protein N, b-lactoglobulin), loss of natural
milk constituents (vitamin destruction), and mis-
cellaneous parameters (color change, surface hydro-
phobicity and free sulphydryl in milk). However, in the
context of intrinsic TTIs the degradation, denaturation
or inactivation of heat-labile compounds (e.g. whey pro-
teins and enzymes) and the formation of ‘new’ substances
are most interesting to investigate. The first type of reac-
tions is a suitable tool for the evaluation of low-heat
treatments (thermal inactivation of vegetative bacteria as
well as protein denaturation and enzyme inactivation are
generally characterized by an Ea-value between 200 and
600 kJ/mol), whereas the second type of reactions is more
effective for the assessment of processes involving high
temperatures (the Ea-value for thermal destruction of
spores is situated between 250 and 330 kJ/mol and for
many chemical reactions between 80 and 125 kJ/mol).
An overview of some milk compounds that can be
considered as potential TTIs is given in Table 2.
Alkaline phosphatase
Alkaline phosphatase (ALP, EC 3.1.3.1) is a phos-
phomonoesterase that catalyzes the hydrolysis of phos-
phoric monoesters (Schlimme, Kiesner, Lorenzen, &
Martin, 1997). The enzyme is widely used in establishing
adequate pasteurization of milk; inactivation of ALP
takes place at temperatures slightly higher than neces-
sary to kill M. tuberculosis, S. senftenberg or L. mono-
cytogenes (Eckner, 1992; Girotti, Ferri, Ghini, Budini,
& Roda, 1994). Remaining ALP activity after pasteur-
ization points to improperly operating pasteurization
units or possible contamination by raw milk (Murthy,
Bradshaw, & Peeler, 1990; Painter & Bradley, 1997).
Mostly, the ALP activity is quantified spectro-
photometrically with phenylphosphate (AOAC 979.13;
FIL-IDF 63, 1971) or p-nitrophenyl phosphate
(FIL-IDF 82A, 1987; DIN 10 337, 1993) as substrate.
The disadvantage of these methods is their relative
complexity and long incubation time. Dependent on the
source, their ALP detection limit is equivalent to about
0.1–0.2 to 0.5% of the raw milk level, which seems to be
adequate in practice (Black, Kuzyk, & Duggan, 1992; de
Jong, Mahulette, & Ellen, 1993; Schlimme et al., 1997).
Nevertheless, a rapid, more sensitive procedure has been
developed, based on the fluorometric measurement of
the commercial substrate Fluorophos (Rocco, 1990).
This technique allows estimation of ALP at very low
levels in the equivalent of 0.006–0.01% raw milk (de
Jong et al., 1993; Schlimme et al., 1997). Conversion of
results is based on the fact that dilution of 0.1% (v/v)
raw milk gives an ALP activity of 500 mU/l by the
fluorometric method, of 1mg phenol/ml (per 15 min) by
299
the AOAC 979.13 procedure, and of 10 mg p-nitro-
phenolphosphate (per 2 h) by the FIL-IDF 82A (1987)
method (de Jong et al., 1993).
Two major isoenzymes of ALP have been identified,
a- and b-phosphatase, which are present mainly in the
milk plasma and in the membranes of fat globules
respectively (Girotti et al., 1994; Walstra & Jeness,
1984). The difference in ALP allocation is reflected by
the ALP activity. About 30–40% of ALP activity is
situated in the fat globule membrane. ALP activity is
consequently lower in raw skimmed than in raw whole
milk (Claeys, Van Loey, & Hendrickx, 2002a; Girotti et
al., 1994; Schlimme & Thiemann, 1992).
Next to milk fat content, the ALP activity in raw milk
is also reported to depend on the breed of cow, the stage
of lactation (increasing at the end of lactation), the
volume of milk produced and the age of cow (decreasing
with age) (Girotti et al., 1994; Schlimme & Thiemann,
1992). Moreover, seasonal variation in ALP activity has
been observed; activity increased during summer to
remain high in autumn, and to decrease again in the
middle of winter (Schlimme & Thiemann, 1992).
Although ALP is widely applied as an indicator of
efficient pasteurization, only few detailed quantitative
kinetic studies on thermal inactivation of ALP have
been published. Thermal inactivation of ALP follows
first order kinetics. Independent of the ALP assay
applied, reported z-values range between 5 and 6.7 C
(Claeys, Ludikhuyze, Van Loey, & Hendrickx, 2001a;
Eckner, 1992; Murthy et al., 1990) and between 8.5
and 9.5 C (Schlimme et al., 1997; Martin, Lorenzen,
Kiesner, & Schlimme, 1999).
The use of the ALP test in determining the accuracy
of milk pasteurization has been questioned, mainly
because of a partial reactivation of ALP possible by
exposure of the product to temperatures above 20 C in
the presence of Mg2+, causing a false positive result. A
false positive may also be caused by microbial phospha-
tase, which is less sensitive to heat than the indigenous
milk enzyme. Furthermore, the borderline between a
positive and a negative result of the phosphatase test on
heat treated milk depends somehow on the assay method
applied (Girotti et al., 1994; Schlimme et al., 1997).
Some authors suggest the ALP test to be less sensitive
when applied to skim milk since because of its lower fat
content skimmed milk tends to have less residual ALP
activity (Painter & Bradley, 1997; Walstra et al., 1999).
However, we observed comparable inactivation kinetics
in milk with different fat content. And although initial
ALP activity was lower in skimmed than in whole milk,
fat content seemed not to affect the ALP test result for
pasteurized milk substantially (Claeys et al., 2002a).
Inactivation kinetics were also independent of the sea-
son of the year, and use of ALP as TTI was not affected
by the seasonal variation of its activity in crude milk
(unpublished).
300 W.L. Claeys et al. / Trends in Food Science & Technology 13 (2002) 293–311

Table 2. Some milk components considered as potential intrinsic TTIs for thermally processed milk
Kinetics Analysis Remarks Ref.
Type I indicators

Acid phosphatase 1st order - Spectrophotometric Quick loss of activity at Balci and Wilbey, 1999;
(EC 3.1.3.2) D100 C=4.8–45s (p-nitrophenol phosphate) room temperature (60% Griffiths, 1986; Walstra
z=6.6–27.6 C - Fluorometric (Fluorophos) after 3 days at 17 C). et al., 1999
(Q10=10.5)

Adenosine deaminase 1st order (between HPLC (oxidative deamination Activity increases up to Martin, Kiesner, Lorenzen,
(EC 3.5.4.4) 82 and 90 C) of adenosine to inosine) 75 C, whereas at 85 C and Schlimme, 1989;
D84 C=93s the enzyme is inactive. Martin et al., 1999
z=9.2 C

Alkaline phosphatase 1st order - Spectrophotometric 30–40% of the activity is Claeys, Ludikhuyze,
(EC 3.1.3.1) D60 C=24.6 min (p-nitrophenyl situated in the cream. Van Loey, and Hendrickx,
z=5.3 C dinatriumphosphate) 2001a
- Fluorometric (Fluorophos)

N-Acetyl-b-glucosidase Spectrophotometric Levels in raw milk are Wilbey, 1996

(EC 3.2.1.30) (p-nitrophenyl-N-acetyl-b-d- variable, being elevated
glucosaminide) in cases of mastitis.

Catalase (EC 1.11.1.6) 1st order - Disk flotation - Catalase level depends Griffiths, 1986; Walstra
D77.5 C=15s - Polarographic on disease status of the et al., 1999
z=7.4 C (Q10=180) animal.
- Reactivation of catalase
is possible.

a-Fucosidase 1st order Reflectance colorimetry a-fucoside+H2O! McKellar and Piyasena,

(EC 3.2.1.51) Ea=378 kJ/mol (p-nitrophenyl-a-l- fructose-6-P 2000; Zehetner,
fucopyranoside Bareuther, Henle,
and Klostermeyer,
1996

g-Glutamyl-transferase 1st order Spectrophotometric 25% of the activity is Dos Anjos et al., 1998;
(EC 2.3.2.2) Ea=457 kJ/mol (g-glutamyl-p-nitroaniline) situated in the cream. Zehetner et al., 1995
z=4.8 C (Q10=123)

a-Lactalbumin 1st order HPLC Pompei and Rossi, 1994;

z=34.9 C Morales et al., 2000
Ea=89.1 kJ/mol

b-Lactoglobulin 1st order HPLC Shows a marked change Claeys et al., 2001a
D75 C=49.9 min; in temperature dependence
z1=7.9 C at 83 C.
D90 C=3.5 min ;
z2=24.2 C

Lactoperoxidase 1st order Spectrophotometric Claeys et al., 2001a

(EC 1.11.1.7) D71 C=38.6 min (ABTS and H2O2)
z=4.3 C

Lipoprotein lipase 1st order Walstra et al., 1999

(EC 3.1.1.34) D70 C=20s
Q10=13

a-Mannosidase 1st order Spectrophotometric Hydrolysis of the terminal, Zehetner et al., 1996
(EC 3.2.1.24) Ea=554 kJ/mol (p-nitrophenyl-a-d-mannoside) non-reducing ends of
mannosides

Phosphohexo-isomerase 1st order glucose-P $ fructose-P Zehetner et al., 1996

(EC 5.3.1.9) Ea=361kJ/mol

(continued on next page)

 W.L. Claeys et al. / Trends in Food Science & Technology 13 (2002) 293–311 301

Table 2 (continued)

Kinetics Analysis Remarks Ref.

Phosphodiesterase 1st order Hydrolysis of Zehetner et al., 1996
(EC 3.1.4.1) Ea=490 kJ/mol 5-oligonucleotides.

Superoxide dismutase 1st order Walstra et al., 1999

(EC 1.15.1.1) D80 C=345s
Q10=150

Xanthine oxidase 1st order Assay based on the oxidation - Activity in raw milk varies Griffiths, 1986; Zehetner
(EC 1.1.3.22) D80 C=17s of vanillin - Activated by heating et al., 1996
z=6.75 C between 60 and 70 C
(Q10=46) - Milk contains activators
and inhibitors of xanthine
oxidase

Type II indicators

Epilactose 1st order GC Formation 10% (lactulose Calvo and Olano, 1989;
Ea=131.3 kJ/mol formation) Olano and Calvo, 1989

Furosine Pseudo-0 order HPLC Formed through acid Claeys et al., 2001b
k110 C=12.1 mg / hydrolysis of lactulosyl-lysine
100 g prot, min
Ea=83.5 kJ/mol

Hydroxymethylfurfural Pseudo-0 order - Spectrophotometric

k110 C=0.8 mmol/l, (thiobarbituric acid)
min Claeys et al., 2001b
Ea=105.6 kJ/mol - HPLC

Lactulose Pseudo-0 order - GC Formed by isomerization Claeys et al., 2001b

k110 C=30.9 mg/l, - Spectrophotometric of lactose
min (d-glucose/d-fructose
Ea=105.6 kJ/mol enzymatic kit)

N6-Methyl adenosine 1st order HPLC Formed from 1-methyladenosine Schlimme et al., 1994
Ea=125 kJ/mol by Dimroth-rearrangement.
z=30.5 C
(Q10=2.13)

Next to fat, also sugar has been reported to alter heat
inactivation characteristics of ALP by increasing the heat
stability (Sanders & Sager, 1948). More specific, removal
of lactose during ultrafiltration increased ALP inactiva-
tion (Mistry, 1989). Thus, use of ALP to measure the effi-
ciency of pasteurization of dairy products with modified
composition has to be undertaken with some caution.
g-Glutamyl transferase
g-Glutamyl transferase (GGT, EC 2.3.2.2) catalyzes
the transfer of g-glutamyl residues from glutathione and
other g-glutamyl containing peptides to amino acids or
peptides (Baumrucker, 1980). GGT is more abundant in
milk than ALP rendering an enzymic assay based on
residual GGT activity more sensitive. GGT is also more
heat resistant than ALP, but less than lactoperoxidase
(Andrews, Anderson, & Goodenough, 1987; Dos Anjos,
Machado, Ferro, Otto, & Bogin, 1998; Martin et al.,
1989). This fact and the good correlation that exists
between reduction in GGT activity and destruction of
streptococci (Patel & Wilbey, 1994), demonstrates the
potential of GGT to describe the temperature region
between that covered by ALP and lactoperoxidase as
well as to define the limit between pasteurized and high-
temperature pasteurized milk (Herrmann, Lamprecht,
Frister, & Rudzik, 1996; Zehetner, Bareuther, Henle, &
Klostermeyer, 1995).
For the quantitative determination of GGT a com-
mercial test kit is available based on the procedure of
Baumrucker (1979). In the assay the enzyme acts on the
specific substrate, l-g-glutamyl-r-nitroanilide, transfer-
ring the glutamyl group to the receptor, glycylglycine.
The r-nitroaniline thus released is quantified spectro-
photometrically (McKellar, Emmons, & Farber, 1991).
About one quarter of the GGT activity is located
within the cream and three quarters within the milk
302 W.L. Claeys et al. / Trends in Food Science & Technology 13 (2002) 293–311
serum (Baumrucker, 1979; Zehetner et al., 1995).
Nevertheless, fat content does not appear to have any
significant effect on the reduction of GGT activity dur-
ing heating of milk (Patel & Wilbey, 1994). Little sea-
sonal variation was noted in either whole or skimmed
milk, with a higher level of GGT activity in the late part
of the year (McKellar et al., 1991; Patel & Wilbey,
1994).
Like ALP, GGT inactivation kinetics can be descri-
bed by a first order reaction, and are characterized by
an Ea-value of 457–473 kJ/mol (Andrews et al., 1987;
Zehetner et al., 1995).
Lactoperoxidase
Lactoperoxidase (Lpo, EC 1.11.1.7) is present in the
milk serum at a concentration of about 0.03 g/l,
although variations have been reported. The enzyme
catalyzes the oxidation of indigenous thiocyanate by
hydrogen peroxide to yield highly reactive, short-lived
antimicrobiological substances, mainly hypothiocyanate
(Fugslang, Johansen, Christgau, & Adler-Nissen, 1995;
Hernández, van Markwijk, & Vreeman, 1990; Muir,
1996).
Lpo has been suggested as a possible enzyme to
monitor thermal processes higher than 72 C and is used
to make a distinction between pasteurized and high-
temperature pasteurized milk (Griffiths, 1986; Guha &
Roy, 1973; Hernández et al., 1990). Though, there are
some limitations on the suitability of Lpo as a TTI for
liquid milk. Firstly, a cyclic variation of Lpo activity is
observed with a maximum in summer (August) and a
minimum in winter (February) probably explicable by
change in feed of cows during the year (Olszewski &
Reuter, 1992). Secondly, its thermostability is strongly
dependent on the nature of the medium (buffer < milk
or whey) and influenced by pH, casein, b-lg and ionic
strength (Hernández et al., 1990; Sato, Dosako, Naka-
jima, & Ido, 1992; Sciancalepore, De Stefano, & Piac-
quado, 1996). Thirdly, Lpo is sensitive for irreversible
photochemical inactivation and may regenerate during
storage (Hernández et al., 1990; Olszewski & Reuter,
1992). Finally, a validated analytical procedure to
quantify Lpo activity is not available yet. Existing
methods for estimating peroxidase activity are spectro-
photometrical assays based on different donors such as
p-phenylenediamine (Aurand, Roberts, & Cardwell,
1956), pyrogallol and guaiacol (Pruitt, Kamau, Miller,
Månsson-Rahemtulla, & Rahemtulla, 1990), and ABTS
(2,20 -azinobis(3-ethylbenzothiazoline-6-sulphonic acid)
(Hernández et al., 1990), of which the latter is most
commonly used.
Thermal inactivation of Lpo follows first order kinet-
ics with a z-value situated between 3.7 and 4.3 C
(Claeys, Ludikhuyze, Van Loey, et al., 2001a; Olszewski
& Reuter, 1992). Although the enzyme is mainly present
in the serum phase, initial Lpo activity as well as Lpo
inactivation kinetics are reported to be comparable in
milk with different fat contents (Claeys et al., 2002b).
b-Lactoglobulin
The whole whey protein fraction (e.g. whey protein
nitrogen index) as well as its individual components can
be used for the classification of heat treatments. The
main whey proteins involved are in order of decreasing
heat stability: a-lactalbumin > b-lactoglobulin >
serum albumin > immunoglobulins.
b-lactoglobulin (b-lg) is the most abundant whey
protein; on average, milk contains about 3.3–3.5 g b-lg
per litre (Schlimme et al., 1996). b-Lg is a globular pro-
tein containing two intramolecular disulfide bonds and
one thiol group. At room temperature and milk pH the
protein is predominantly present as a dimer. b-Lg is
important as it can have a marked influence on the
functional properties of milk products. Moreover, b-lg
plays an important role in fouling of heat exchangers
(de Jong et al., 1992; Jeurnik, 1991). The quantitative
determination of undenatured b-lg has been proposed
for distinguishing between different categories of heat-
treated milk. A minimum content of 2600 mg/l b-lg for
pasteurized milk and of 2000 mg/l b-lg for high-pasteur-
ized milk is within the limits proposed by the Interna-
tional Dairy Federation. As a clear-cut distinction of
UHT milk from in-bottle sterilized milk, a minimum
content of 50 mg/l for the latter has been proposed
(Schlimme, Buchheim, & Heeschen, 1994; Wilbey, 1996).
In literature, conflicting results exist regarding the
kinetics of b-lg denaturation, possibly reflecting the
complexity of the process. Several researchers have
reported first order kinetics (Claeys, Ludikhuyze, Van
Loey, et al., 2001a; deWit & Klarenbeek, 1981; Har-
walkar, 1986; Verheul, Roefs, & de Kruif, 1998), while
others reported second order kinetics (El-Shazly, Mah-
ran, & Hofi, 1978; Harwalkar, 1986; Lyster, 1970), 1.5
order kinetics (Dannenberg & Kessler, 1988; Schlimme
et al., 1996), or consecutive first order reaction kinetics
(Harwalkar, 1980). The discrepancies among these stu-
dies might also be explained by variation in methodol-
ogy and experimental conditions such as the medium of
the protein solution (skim milk, whey or isolated pro-
tein in buffer solution), the method and conditions of
heat treatment, and the method of assay to determine
the residual native protein concentration (Anema &
McKenna, 1996; El-Shazly et al., 1978; Prakabaran &
Damodaran, 1997).
Denatured b-lg is naturally not present in raw milk
and the amount of protein remains stable with respect
to total protein, so that it is not essential to analyze the
raw milk corresponding to each sample of processed
milk to measure the degree of denaturation (Morales,
Romero, & Jiménez-Pérez, 2000).
As to the effect of milk fat content on b-lg denatura-
tion, views diverge. Some observed a protective effect
 W.L. Claeys et al. / Trends in Food Science & Technology 13 (2002) 293–311
(Pellegrino, 1994), whereas others observed no effect of
milk fat on b-lg denaturation (López-Fandiño, Olano,
Corzo, & Ramos, 1993; Lyster, 1970). Still others
observed a synergistic effect of fat on b-lg denaturation
although the concentration of b-lg in crude whole and
skimmed milk was comparable (Claeys et al., 2002a; de
Koning et al., 1990). This positive correlation between
milk fat content and b-lg denaturation was suggested to
be related to the binding of b-lg on the fat globule sur-
face (Claeys et al., 2002a). When a b-lg molecule dena-
tures, it can follow three possible pathways: (i) it can
bind to another denatured serum protein molecule, (ii)
it can react with micellar k-casein, or (iii) it may bind to
a protein on, or adsorb to the surface of a fat globule.
From the rate of the reaction with fat globules, the lat-
ter pathway appears to be favored (Corredig & Dalgle-
ish, 1996; Dalgleish & Banks, 1991).
Similar to ALP inactivation, lactose content has been
reported to have a protective effect on b-lg denaturation
(de Wit & Klarenbeek, 1981; Pearce, 1989; Plock, Spie-
gel, & Kessler, 1998). Two mechanisms are held
responsible for this stabilization. Firstly, sugars and
similar substances tend to maintain or increase the
hydratation of protein molecules, enhancing the water
structure in the immediate surroundings of the protein
molecules and contributing to their stability. Secondly,
irreversible aggregation of the whey protein with casein
is stericly hindered by means of protein-lactose com-
plexes formed during heating.
Hydroxymethylfurfural
In the case of hydroxymethylfurfural (HMF), a dis-
tinction must be made between ‘total’ and ‘free’ HMF,
which are produced through different pathways. While
free HMF is formed only by decomposition of lactulo-
syllysine in the Maillard reaction, total HMF can also
be formed through the acid-catalyzed degradation of
lactose via 3-deoxyosulose, known as the Lobry De
Bruyn–Alberda van Ekenstein (LA) transformation
(Morales & Jiménez-Pérez, 1999; van Boekel, 1998). The
relative rates of the two routes depend on a number of
variables, like pH, water activity, temperature and time.
In milk, the route of HMF formation is mostly the
LA transformation (Berg, 1993; Morales, Romero, &
Jiménez-Pérez, 1997). Formation of free HMF is limited
compared to formation of total HMF, and probably not
a reliable measure for the occurrence of the Maillard
reaction (Morales & Jiménez-Pérez, 1999).
Traditionally, HMF is measured spectro-
photometrically with thiobarbituric acid (TBA) as sub-
strate (Keeney & Bassette, 1959). This method however,
suffers a lack of specificity because of the general reac-
tivity of TBA towards aldehydic groups. Also, the
formed HMF-TBA complex is fairly unstable and vari-
able in function of time. An accurate HPLC technique is
now available, but is reported to be time-consuming and
303
to show sometimes interference between HMF and
other compounds that can get co-eluted with the HMF
peak (Morales, Romero, & Jiménez-Pérez, 1995, 1997).
As for kinetics of formation of HMF, most studies in
model systems and milk reported zero-order kinetics
with an Ea-value between 104 and 135.1 kJ/mol (Claeys,
Ludikhuyze, & Hendrickx, 2001b; Morales et al., 1995;
Peri, Pagliarini, & Pierucci, 1988). It has to be remarked
that zero-order kinetics, implying that reaction rates are
independent of concentration, are rather artificial and a
consequence of either experimental conditions, sam-
pling procedure or the presence of competing or rate-
limiting intermediate reactions (O’Brien, 1997).
According to Berg (1993) and Claeys, Van Loey, and
Hendrickx (2002b) milk fat content has no effect on total
HMF formation in UHT-milk. Morales and Jiménez-
Pérez (1999) on the other hand, observed a protective
effect of milk fat on total HMF formation, after pas-
teurization as well as after direct UHT and sterilization.
Varying fat contents did not influence associated Ea-
values.
Next to fat content, it has been reported that HMF
values increase with increasing dry matter concentration
(Fink & Kessler, 1986) and initial lactose concentration
in the product (Kind & Reuter, 1990).
Although there is formation of HMF from its pre-
cursors during storage of milk, HMF content remains
constant between refrigeration and room temperature
(Fink & Kessler, 1986; Jiménez-Pérez, Corzo, Morales,
Delgado, & Olano, 1992). This has been attributed to
the fact that part of the HMF generated during storage
is lost through oxidations or other transformations. At
temperatures 530 C the equilibrium is moved towards
HMF formation (Jiménez-Pérez et al., 1992; Morales et
al., 1997).
HMF is a recognised indicator for distinguishing in-
bottle sterilized milk (HMF values of 30–140 mmol/l)
from UHT-milk (1–10 mmol/l) (Fink & Kessler, 1986;
Morales, Romero, & Jiménez-Pérez, 1996). Directly
heated UHT milk has lower HMF values than indirectly
heated milk, probably because the warming up and
cooling down periods in direct UHT are virtually
instantaneous so that the amount of heat this milk
receives is usually less than the amount of heat received
by indirectly heated UHT milk (Fink & Kessler, 1986).
Another explanation is that during direct UHT satu-
rated steam is injected into the preheated milk, diluting
the milk and thus reducing the concentrations of reac-
tants susceptible to heat damage (Kind & Reuter, 1990;
Nangpal & Reuter, 1990a). The suitability of HMF as
TTI for heat treated milk is however questioned because
of the variable amounts of HMF measured in raw milk
(3.6–7.3 mmol/l) (Burton, 1984; Fink & Kessler, 1986).
These variable HMF amounts detected in raw milk are
an artefact generated by the acidic digestion preceding
the spectrophotometric or chromatographic analysis,
304 W.L. Claeys et al. / Trends in Food Science & Technology 13 (2002) 293–311
since HMF should not be present in unheated milk
(Morales et al., 1996, 2000).
Lactulose
Lactulose may be present in heated milk as free lac-
tulose or covalently bound to milk protein as e-N-
deoxylactulosyl-l-lysine. Free lactulose is formed by the
LA transformation of lactose via a 1,2-enediol inter-
mediate. Increasing the lactose concentration increases
the formation of lactulose (Andrews, 1989; Berg, 1993).
Several methods for lactulose determination have
been described such as thin-layer chromatographic,
liquid chromatographic, gas chromatographic and
enzymatic methods (Amine, Moscone, Bernardo, Mar-
coni, & Palleschi, 2000; Geier & Klostermeyer, 1983;
Martinez-Castro & Olano, 1981; Morales et al., 2000;
Olano & Martinez-Castro, 1989; Parrish, Hicks, &
Doner, 1980). In recent years mainly the two latter
methods have been applied.
Published data describe lactose isomerization into
lactulose as an irreversible zero order (Claeys, Ludi-
khuyze, & Hendrickx, 2001b; De Rafael, Villamiel, &
Olano, 1997; Geier, Klostermeyer, 1983) or first order
reaction (Olano & Calvo, 1989; Schlimme et al., 1996)
with Ea-values between 105.6 and 118 kJ/mol, and
between 118.3 and 125 kJ/mol respectively. In fact, the
reaction is reversible, as expected for an isomerization
reaction, and lactulose degrades further to galactose
and other constituents. Furthermore, lactose also iso-
merizes into very small amounts of epilactose, and both
lactose and lactulose are involved in the Maillard reac-
tion (Berg & van Boekel, 1994; O’Brien, 1997; Olano &
Calvo, 1989). All such responses can be taken into
account by multiresponse modeling (van Boekel, 1998).
However, use of a simple reaction order for complex
formation pathways can be useful for describing chemi-
cal changes during processing or for modeling shelf-life,
when knowledge of pure chemistry or mechanism of the
reaction is of no importance.
Milk pH has been reported to affect lactose iso-
merization (Geier & Klostermeyer, 1983). In the pH
range between 6.7 and 6.8 the variation of lactulose
content is minimal, while below pH 6.7 the lactulose
content decreases and above pH 8 the lactulose content
increases substantially (Nangpal & Reuter, 1990a).
Although precipitation of tertiary calcium phosphate,
hydrolysis of organic casein phosphate and production
of organic acids all contribute to the decrease of pH in
heated milk, most important from a quantitative point
of view is the production of organic acids (formic acid)
(Berg, 1993; O’Brien, 1997). Nevertheless, pH values
within the range found in normal milk samples have
only a small influence, if any (Olano, Corzo, Paez, &
Martinez-Castro, 1987).
Soluble citrate and phosphate were found to catalyze
the formation of lactulose, presumably by acting as
bases (Andrews & Prasad, 1987). Probably, variation in
the phosphate content of milk (86–103 mg/100 ml) is
too small to affect the rate of lactulose formation in
commercially processed milks sufficiently as a result of
which the applicability of lactulose as a TTI could be
invalidated (Florence, Knight, Owen, Milner, & Harris,
1985). There are no published data on the variability in
the citrate content of milk. When the calcium content of
milk is increased by 50 ppm, formation of lactulose is
depressed, probably due to the increased complex for-
mation of citrate and phosphate with calcium, thus
reducing the levels of phosphate and citrate available to
catalyze lactulose formation (Andrews, 1989; Olano et
al., 1987).
As to the effect of proteins, comparing ultrafiltrate
(retentate) with milk of various protein concentrations,
it was observed that although a small amount increases
lactulose formation, further addition of protein reduces
the amount of lactulose formed. This effect was ascribed
to a decrease in pH of the ultrafiltrate when heated,
reduced substrate concentration (due to increased for-
mation of lactosyl-amino compounds) or removal of
free lactulose (due to condensation of lactulose with an
amino group) (Andrews & Prasad, 1987; Berg, 1993;
Calvo & Olano, 1989; Greig & Payne, 1985). Kinetics
for lactulose formation were almost certainly identical
in milk and ultrafiltrate indicating that milk protein
does not catalyze lactulose formation. Probably the salt
system is responsible for most of the lactulose formed
during thermal processing of milk (Andrews & Prasad,
1987; Montilla & Olano, 1997).
Contradictory data have been reported concerning
the influence of milk fat content on lactulose formation.
Some observed an enhancement of formation at higher
fat content (de Koning et al., 1990), while others found
a depressing effect of fat on formation (Pellegrino,
1994). Still others observed no alteration in lactulose
concentrations attributable to milk fat content
(Andrews, 1984; Andrews & Morant, 1987; Berg, 1993;
Claeys et al., 2002b; Geier & Klostmeyer, 1983).
Also on the effect of milk storage on lactulose forma-
tion, contradictory data have been reported. Nangpal
and Reuter (1990a) observed even at 20 C a system-
atically increase of lactulose concentration, which was
independently of the initial concentration. Andrews
(1989) on the other hand, observed no effect of milk
storage at room temperature on the lactulose con-
centration. There is probably a balance between forma-
tion and degradation of lactulose. This balance depends
on external and compositional factors, which can
explain the different observations during storage.
Determination of the lactulose concentration allows
the differentiation of UHT from in-bottle sterilized
milk, and of directly from indirectly processed UHT
milk (Jiménez-Pérez et al., 1992; Schlimme et al., 1996).
Tentative proposals being considered include lactulose
 W.L. Claeys et al. / Trends in Food Science & Technology 13 (2002) 293–311
levels of > 100 mg/l and > 600 mg/l for UHT and
sterilized milk respectively (Schlimme et al., 1994; Wil-
bey, 1996). However, an overlap between lactulose
levels in UHT and sterilized milk has been reported
rendering measurement of lactulose alone insufficient as
an index of heat treatment (Wilbey, 1996).
Furosine
The main first stable products of the Maillard reac-
tion are aminoketoses, known as the Amadori com-
pounds. In milk, these compounds are largely
represented by e-lactulosyl-lysine, an unstable com-
pound that after acid hydrolysis is converted into the
more stable furosine (Resmini & Pellegrino, 1994; van
Boekel, 1998).
Methods for quantifying furosine include ion-pair
reversed phase HPLC (Resmini, Pellegrino, & Battelli,
1990), ion-exchange chromatography (Hartkopf &
Erbersdobler, 1993) and capillary zone electrophoresis
(CZE) (Tirelli & Pellegrino, 1995). Comparison between
the HPLC and CZE technique showed that low levels of
furosine were underestimated by CZE (Tirelli & Pelle-
grino, 1995).
Studies dealing with furosine formation kinetics,
reported a zero order reaction with activation energies
between 81.6 kJ/mol and 104.1 kJ/mol (Claeys, Ludi-
khuyze, Van Loey et al., 2001b; De Rafael et al., 1997;
Montilla, Calvo, Santa-Maria, Corzo, & Olano, 1996;
Schlimme et al., 1996). The formation of furosine is
highly dependent on protein concentration (positively
correlated), and is therefore expressed as mg/100 g pro-
tein (Montilla & Olano, 1997; Rattray, Gallmann, &
Jelen, 1997). As the Amadori rearrangement is catalyzed
by acids, it can be expected that also the pH-value of the
raw milk would influence furosine formation. However,
no regular pattern was observed in the variation of fur-
osine concentration between pH 6.6 and 6.95 (Nangpal
& Reuter, 1990b).
Similar for whey protein denaturation and lactulose
formation, Pellegrino (1994) observed a higher furosine
formation in skimmed than in whole UHT milk, which
was explained by a difference in heat load between
whole and skimmed milk. Fat content would affect
viscosity of milk and thus hinder the heat transfer of the
process. However, excluding heat transfer phenomena
by working with a small sample size, we observed a
positive correlation between fat content and furosine
concentration (Claeys et al., 2002b). During heating,
denatured serum proteins initiate binding to micellar
k-casein via disulphide bonds. Native fat globule pro-
teins contain cysteine and can therefore also interact
with denatured serum protein when milk is heated. This
kind of interaction could explain the protective effect of
milk fat on lysine blocking and on furosine formation
when whole milk is heated (Morales et al., 1995). It has
to be remarked however, that according to our results
305
the effect of milk fat on furosine formation appeared to
be insignificant in the context of process impact quanti-
fication (Claeys et al., 2002b).
Formation of furosine has been observed during milk
storage at room temperature. This can lead to mis-
interpretation of the actual heat load when applying
furosine content as an index of heating. The amount of
furosine newly formed was however independent of the
furosine content immediately after heating and negli-
gible at 4 C (Nangpal & Reuter, 1990b; Pellegrino, De
Noni, & Resmini, 1995a; Rattray et al., 1997).
Furosine is a useful indicator of lysine damage in milk
products as well as a quality parameter to identify the
presence of reconstituted milk powder in raw or pas-
teurized milk (Corzo, Delgado, Troyano, & Olano,
1994). Furosine has also been proposed as a useful
index for heat-induced changes in milk products; a fur-
osine content of 8 mg/100 g protein has been suggested
as upper heating limit for pasteurization, of 20 mg/100 g
protein for high-pasteurization and of 250 mg/100 g
protein for UHT (Clawin-Rädecker & Schlimme, 1995;
Schlimme et al., 1996).
Multi-component TTI-approach
In most cases distinction of different heat-classes of
liquid milk cannot be achieved by one TTI alone due to
overlap of values and to the wide range of heat load
applied on liquid milk within the same heat-class. This
wide range can partly be explained by variability in
design of industrial plants, in working and heating con-
ditions, in milk recycling and in other factors. So, for
the differentiation of different classes of sterilized milk,
it was suggested to couple relatively heat-labile compo-
nents (protein, enzyme) with heat-stable components
(chemical compound formed), e.g. a-lactalbumin or b-lg
with lactulose. Due to slight variations caused by severe
heating, the residual low concentration of the whey
proteins cannot be used as a precise indicator. For milk
thus labeled as ‘UHT’ it was proposed that two criteria,
a b-lg content higher than 50 mg/l and a lactulose con-
tent equal or lower than 600 mg/l, must be fulfilled.
Otherwise, milk has to be designed as ‘sterilized’.
Another example is coupling of the chemical compound
furosine with the heat-labile proteins b-lg and/or Lpo
with the objective of identifying small amounts of
reconstituted milk powder in pasteurized milk (Pelle-
grino, Resmini, & Luf, 1995b).
Coupling of two chemical compounds, namely fur-
osine and lactulose, can also offer some potential since
the early Maillard reaction and lactose isomerization
have different kinetics and are affected differently by
milk composition and heating conditions (van Boekel,
1998). Because the formation of furosine is faster than
that of lactulose, an abnormal shift from the established
correlation line between lactulose and furosine indicates
incorrect thermal processing or poor storage conditions.
306 W.L. Claeys et al. / Trends in Food Science & Technology 13 (2002) 293–311
Since milk powder contains a considerable amount of
furosine compared to lactulose, coupling of both TTIs
allows a better distinction between genuine and adult-
erated milk as well. The detection of added milk powder
based only on furosine determination has an important
limitation seeing that there is a great range in the
amount of furosine formed with the varying time and
temperature limits in e.g. the UHT process (Corzo et al.,
1994; Montilla et al., 1996; Pellegrino et al., 1995a).
Time temperature tolerance diagrams
Based on the acquired kinetic information, the suit-
ability of the intrinsic TTIs studied with relation to the
microbiological safety of the product can be visualised
by means of time temperature tolerance (TTT-) dia-
grams. Construction of such diagrams is sometimes
hampered and interpretation must be done with some
caution. This is illustrated by an example in the domains
of pasteurization and sterilization.
Pasteurization
Milk designed for refrigerated storage (pasteurized
milk) should receive a heat treatment process designed
to reduce the probability of survival of L.monocytogenes
by at least a factor of 104 (FIL-IDF, 1994). Taking a
safety margin into account, usually a log reduction of 6
is considered. The effectiveness of pasteurization is
monitored based on the ALP test. However, the degree
of inactivation chosen as a ‘cut-off’-value for pasteur-
ized milk is not a given parameter, but depends on the
detection limit of the method applied. This was demon-
strated amongst others by Schlimme et al. (1997).
Assuming that a negative test reflects an ALP activity
equivalent or less than 0.5% of the initial activity, it
thus follows from the TTT-diagram in Fig. 2 that tem-
perature time combinations resulting in a negative ALP
test seem to be sufficient to inactivate micro-organisms
Fig. 2. Time temperature tolerance diagram for the domain of
pasteurization: implication of a negative alkaline phosphatase (ALP)
test on the microbiological safety of milk. Legend: *: 95.5% ALP
inactivation curve based on kinetic data from Table 2; 6 log
reduction curves of *: L. monocytogenes, x: Ps. fluorescens, &:
Salmonella spp., and ~ and ~: S. aureus 1 and 2 based on kinetic
parameters from Table 1.
with a heat resistance lower or equal to the one of
L.monocytogenes (for example E.coli, Y.enterolitica)
(Fig. 2). As for Ps.fluorescens and Salmonella spp., a 6
log reduction will be guaranteed by the ALP test only at
temperatures lower than 68–69 C and higher than 64 C
respectively. For S. aureus none of the temperature time
combinations resulting in a negative ALP test will be
adequate to achieve a sufficient reduction. However,
interpretation of the TTT-diagram is dependent on the lit-
erature source consulted for data. For example, when the
destruction curve of S.aureus was constructed by means of
kinetic data from Roberts, Baird-Parker, and Tompkin
(1996) instead of those from Walstra et al. (1999) (in Fig. 2
presented by the 6 log reduction curve of S. aureus 1 and 2
respectively), at temperatures below 72–73 C a negative
ALP test does guarantee the inactivation of S.aureus.
Sterilization
Low acid products with a pH of 4.5 or above intended
for storage under non-refrigerated conditions, should be
subjected to a scheduled heat treatment process
designed to reduce the probability of survival of Cl.
botulinum by at least a factor of 1012 (‘the minimum
botulinum process’) (FIL-IDF, 1994). This condition
together with the proposed minimal heating limits of
100 and 600 mg/l lactulose for UHT and sterilized milk
respectively, are presented by the TTT-diagram in Fig.
3. At temperatures higher than 120 C B.cereus spores
andCl.botulinum vegetative cells and spores are suffi-
ciently reduced by time-temperature conditions result-
ing in formation of 100 mg/l lactulose. The lower limit
of sterilization however, seems to be taken too broad;
spores of B.cereus and Cl.botulinum are killed at tem-
perature time conditions much less intensive than those
resulting in formation of 600 mg/l lactulose.
Because the temperature sensitivity factor of lactulose
formation is not similar to the one characterizing the
Fig. 3. Time temperature tolerance diagram for the domain of
sterilization: application of lactulose to guarantee the micro-
biological safety of milk. Legend: * and *: formation curve of
600 and 100 mg/l lactulose respectively, based on kinetic data
from Table 2; 12 log reduction curves of ^: B. cereus spores, ~:
Cl. botulinum, and ^: Cl. botulinum spores based on kinetic data
from Table 1.
 W.L. Claeys et al. / Trends in Food Science & Technology 13 (2002) 293–311
reduction of spores, curves intersect. In such a case,
more than one TTI is necessary to define clearly the
allowed boundary conditions for thermal processing.
Finally, it has to be remarked that criteria based on
the level of a TTI are easier to define for formation
products such as HMF, lactulose or furosine, than for
proteins. The application of proteins as heat markers
not only depends on the kinetics describing their dena-
turation or inactivation, but also on their initial con-
centration or activity. Moreover, the procentual
inactivation or denaturation of a protein is restricted
from 100 to 0% indicating a dependency on the detec-
tion limit of the analytical method.
Conclusions
The choice of a compound suitable for the assessment
of heating processes depends on its kinetics, the avail-
ability of an appropriate procedure of analysis, fluctua-
tion of its activity or concentration in crude milk, and its
stability after heat treatment (including shelf-life analy-
sis). The purpose of this paper was not to review all
published related studies, but only to indicate some of
the problems or choices with regard to the characteriza-
tion of intrinsic TTIs with liquid milk as a case-study.
Although many reports in literature discuss the
impact of heating on some food components, often
there is a lack of quantitative data that enable a correct
evaluation of the process impact. Mainly the qualitative
effect of temperature, the singular effect of certain heat
loads on components, the influence of preservation
conditions on component concentrations are discussed.
When quantitative data are available, comparing of
data is often obstructed because of different choice of
analytical procedures, variations in experimental set-up
(different temperature range, different medium: milk/
reconstituted milk powder/model system), variations in
the kinetic model applied, etc.
If it turns out that inactivation, denaturation or for-
mation of the TTI examined is controlled by kinetics
which differ from those for destruction of micro-organ-
isms or spores, the TTI in concern can still be useful as a
tool to evaluate if a plant is processing milk to specifi-
cation and if operating conditions have not drifted from
the design conditions. It also provides a means whereby
one process can be compared to another on the basis of
a simple test.
Generally, the development of TTIs answers to the
need of industry towards optimisation and a higher
economic efficiency, the consumers demand for micro-
biological safe and qualitative products, and the
requirements of legal authorities (criteria for authenti-
city, HACCP, etc.).
Finally, it has to be remarked that when reflecting the
notion of TTIs to a novel technology such as high pres-
sure processing, the complexity of the problem increases
from a two-dimensional (time and temperature) to a
307
three-dimensional domain (time, temperature and pres-
sure). Additional conditions must be fulfilled and the
question arises whether compounds suitable as TTIs for
the assessment of thermal processes of milk are still
useful for quantifying the impact of e.g. high pressure
processing of milk.
Acknowledgements
This research has been supported by the Office of the
Federal Office for Scientific, Technical and Cultural
Affaires as a part of project NP/01/034.
References
Amine, A., Moscone, D., Bernardo, R., Marconi, E., & Palleschi, G.
(2000). A new enzymatic spectrophotometric assay for the deter-
mination of lactulose in milk. Analytica Chimica Acta, 406, 217–224.
Andrews, G. (1984). Distinguishing pasteurized, UHT and sterilized
milks by their lactulose content. Journal of the Society of Dairy
Technology, 37, 92–95.
Andrews, A., Anderson, M., & Goodenough, P. (1987). A study of
the heat stabilities of a number of indigenous milk enzymes.
Journal of Dairy Research, 54, 237–246.
Andrews, G., & Morant, S. (1987). Lactulose content, colour and the
organoleptic assessment of UHT and sterilized milks. Journal of
Dairy Research, 54, 493–507.
Andrews, G., & Prasad, K. (1987). Effect of the protein, citrate and
phosphate content of milk on formation of lactulose during heat
treatment. Journal of Dairy Research, 54, 207–218.
Andrews, G. (1989). Lactulose in heated milk. Bulletin of the Inter-
national Dairy Federation, 238, 45–52.
Anema, S., & McKenna, A. (1996). Reaction kinetics of thermal
denaturation of whey proteins in heated reconstituted whole
milk. Journal of Agricultural and Food Chemistry, 44, 422–428.
Arabshahi, A., & Lund, D. (1985). Considerations in calculating
kinetic parameters from experimental data. Journal of Food
Process Engineering, 7, 239–251.
Arrhenius, S. (1889). [About the reaction rate of the inversion of
non-refined sugar at souring.] Zeitschrift für Physikalische
Chemie, 4, 226–248.
Aurand, L., Roberts, W., & Cardwell, J. (1956). A method for esti-
mation of peroxidase activity in milk. Journal of Dairy Science,
39, 568–573.
Balci, A., & Wilbey, R. (1999). Determination of acid phosphatase in
heat treated milks by the Fluorophos Test System. International
Journal of Dairy technology, 52, 56–58.
Bard, Y. (1974). Nonlinear parameter estimation. New York:
Academic Press.
Baumrucker, C. (1979). g-Glutamyl transpeptidase of bovine milk
membranes: distribution and characterization. Journal of Dairy
Science, 62, 253–258.
Baumrucker, C. (1980). Purification of g-glutamyl transpeptidase of
milk membranes. Journal of Dairy Science, 63, 49–54.
Berg, H. (1993). Reactions of lactose during heat treatment of milk:
a quantitative study. PhD thesis, Wageningen Agricultural
University, Wageningen, The Netherlands
Berg, H., & van Boekel, M. (1994). Degradation of lactose during
heating of milk. I. Reaction pathways. Netherlands Milk & Dairy
Journal, 48, 157–175.
Bigelow, W. (1921). The logarithmic nature of thermal death time
curves. Journal of Infectuous Diseases, 29, 528–536.
Black, R., Kuzyk, M., & Duggan, J. (1992). Evaluation of a fluoro-
metric assay for alkaline phosphatase in fluid dairy products.
The Australian Journal of Dairy Technology, 47, 64–67.
308 W.L. Claeys et al. / Trends in Food Science & Technology 13 (2002) 293–311
Box, G., & Lucas, H. (1959). Design of experiments for non-linear
situations. Biometrika, 46, 77–90.
Burton, H. (1984). Reviews of the progress of dairy science: The
bacteriological, chemical, biochemical and physical changes
that occur in milk at temperatures of 100–150 C. Journal of
Dairy Research, 51, 341–363.
Calvo, M., & Olano, A. (1989). Formation of galactose during heat
treatment of milk and model systems. Journal of Dairy Research,
56, 737–740.
Carnahan, B., Luther, H., & Wilkes, J. (1969). Numerical integration.
In: Applied numerical methods (pp. 69–140). New York : John
Wiley & Sons
Claeys, W., Ludikhuyze, L., Van Loey, A., & Hendrickx, M. (2001a).
Inactivation kinetics of alkaline phosphatase and lactoperox-
idase, and denaturation kinetics of b-lactoglobulin in raw milk
under isothermal and dynamic temperature conditions. Journal
of Dairy Research, 68, 95–107.
Claeys, W., Ludikhuyze, L., & Hendrickx, M. (2001b). Formation
kinetics of hydroxymethylfurfural, lactulose and furosine under
isothermal and non-isothermal conditions. Journal of Dairy
Research, 68, 287–301.
Claeys, W., Van Loey, A., & Hendrickx, M. (2002a). Kinetics of
alkaline phosphatase and lactoperoxidase inactivation, and of
b-lactoglobulin denaturation in milk with different fat content.
Journal of Dairy Research, 69, 541–553.
Claeys, W., Van Loey, A., & Hendrickx, M. (2002b). Kinetics of
hydroxymethylfurfural, lactulose and furosine in thermal pro-
cessing of milk with different fat content. Journal of Dairy
Research (in press).
Clawin-Rädecker, I., & Schlimme, E. (1995). [Determination of
furosine in pasteurised milk by use of ion-pair reversed-phase
liquid chromatography.]. Kieler Milchwirtschaftliche For-
schungsberichte, 47, 169–175.
Cohen, E., & Saguy, I. (1985). Statistical evaluation of Arrhenius
model and its applicability in prediction of food quality losses.
Journal of Food Processing & Preservation, 9, 273–290.
Corredig, M., & Dalgleish, D. (1996). Effect of different heat treat-
ments on the strong binding interactions between whey pro-
teins and milk fat globules in whole milk. Journal of Dairy
Research, 63, 441–449.
Corzo, N., Delgado, T., Troyano, E., & Olano, A. (1994). Ratio of
lactulose to furosine as indicator of quality of commercial milks.
Journal of Food Protection, 57, 737–739.
Cunha, L., Oliveira, F., Brandão, T., & Oliveira, J. (1997). Optimal
experimental design for estimating the kinetic parameters of the
Bigelow model. Journal of Food Engineering, 33, 111–128.
Cunha, L., Oliveira, F., & Oliveira, J. (1998). Optimal experimental
design fot estimating the kinetic parameters of processes
described by the Weibull probability distribution function. Jour-
nal of Food Engineering, 37, 175–191.
Dalgleish, D., & Banks, J. (1991). The formation of complexes
between serum proteins and fat globules during heating of
whole milk. Milchwissenschaft, 46, 75–78.
Dannenberg, F., & Kessler, H. (1988). Application of reaction
kinetics to the denaturation of whey proteins in heated milk.
Milchwissenschaft, 43, 3–7.
Davis, J., & Wilbey, R. (1990). Microbiology of cream and dairy
desserts. In K. Robinson (Ed.), Dairy microbiology. The micro-
biology of milk products (Vol. 2, pp. 41–108). London
De Cordt, S. (1994). Feasibility of development of protein-based
time-temperature-integrators for heat-process evaluation. PhD
thesis, K.U.Leuven, Laboratory of Food Technology, Leuven,
Belgium
de Jong, P., Bouman, S., & Van Der Linden, H. (1992). Fouling of
heat treatment equipment in relation to the denaturation of
b-lactoglobulin. Journal of the Society of Dairy Technology,
45, 3–8.
de Jong, E., Mahulette, G., & Ellen, G. (1993). Fluorometric assay of
alkaline phosphatase in dairy products. Voedingsmidde-
lentechnologie, 23, 11–16.
de Koning, P. (1984). The definition of UHT milk: product-based,
process-based, or both ? Voedingsmiddelentechnologie, 17,
25–27.
de Koning, P., Badings, H., van der Pol, J., Kaper, J., & Vos-Klomp-
maker, E. (1990). Effect of heat treatment and fat content on
UHT-milk. Voedingsmiddelentechnologie, 1, 11–14.
De Rafael, D., Villamiel, M., & Olano, A. (1997). Formation of lac-
tulose and furosine during heat treatment of milk at tempera-
tures of 100–120 C. Milchwissenschaft, 52, 76–78.
de Wit, J., & Klarenbeek, G. (1981). A differential scanning calori-
metric study of the thermal behaviour of bovine b-lactoglobulin
at temperatures up to 160 C. Journal of Dairy Research, 48,
293–302.
Dos Anjos, F., Machado, A., Ferro, C., Otto, F., & Bogin, E. (1998).
Gamma-glutamyltransferase as a marker for the pasteurization
of raw milk. Journal of Food Protection, 61, 1057–1059.
Draper, N., & Smith, H. (1981). Applied regression analysis. New
York: John Wiley & Sons, Inc.
Duggleby, R. (1979). Experimental designs for estimating the
kinetic parameters for enzyme-catalysed reactions. Journal of
Theoretical Biology, 81, 671–684.
Duggleby, R. (1981). Experimental designs for the distribution-free
analysis of enzyme kinetic data. In L. Endrenyi (Ed.), Kinetic data
analysis, design and analysis of enzyme and pharmacokinetic
experiments (pp. 169–179). NY-London: Plenum Press.
Eckner, K. (1992). Fluorometric analysis of alkaline phosphatase
inactivation correlated to Salmonella and Listeria inactivation.
Journal of Food Protection, 55, 960–963.
El-Shazly, A., Mahran, G., & Hofi, A. (1978). Kinetics of heat dena-
turation of b-lactoglobulin. Milchwissenschaft, 33, 166–170.
Endrenyi, L. (1981). Design of experiments for estimating enzyme
and pharmacokinetic parameters. In L. Endrenyi (Ed.), Kinetic
data analysis, design and analysis of enzyme and pharmacokinetic
experiments (pp. 137–167). NY-London: Plenum Press.
Fedorov, V. (1972). Theory of optimal experiments. NY - London:
Academic Press.
Fernandez-Martin, F., & Montes (1972). Influence of temperature
and composition on some physical properties of milk and milk
concentrates. III. Thermal conductivity. Milchwissenschaft, 27,
772–776.
FIL-IDF. (1971). Milk and milk powder, buttermilk and buttermilk
powder, whey and whey powder: determination of phosphatase
activity. Fédération Internationale de Laiterie-International Dairy
Federation, Standard 63
FIL-IDF. (1972). Monograph on UHT milk. Fédération Internationale
de Laiterie-International Dairy Federation, Brussels, Annual Bulle-
tin, 5, 158–164.
FIL-IDF. (1987). Milk and powdered milk, buttermilk and powdered
buttermilk, serum and powdered serum: detection of alkaline
phosphtase activity. Fédération Internationale de Laiterie-Interna-
tional Dairy Federation, Standard 82A
FIL-IDF. (1994). Recommendations for the hygienic manufacture of
milk and milk based products. Fédération Internationale de Lait-
erie-International Dairy Federation, Bulletin 292, Brussels, Belgium
Fink, R., & Kessler, H. (1986). HMF values in heat treated and stored
milk. Milchwissenschaft, 41, 638–641.
Fink, R., & Kessler, H. (1988). Comparison methods for distinguish-
ing UHT treatment and sterilization of milk. Milchwissenschaft,
43, 275–280.
Florence, E., Knight, D., Owen, J., Milner, D., & Harris, W. (1985).
Nutrient content of liquid milk as retailed in the UK. Journal of
the Society of Dairy Technology, 38, 121–127.
Fuglsang, C., Johansen, C., Christgau, S., & Adler-Nissen, J. (1995).
Antimicrobial enzymes: applications and future potential in
 W.L. Claeys et al. / Trends in Food Science & Technology 13 (2002) 293–311
the food industry. Trends in Food Science and Technology, 6,
390–396.
Geier, H., & Klostermeyer, H. (1983). Formation of lactulose during
heat treatment of milk. Milchwissenschaft, 38, 475–477.
Geier, H. (1984). Study of the analytical control of the impact of a heat
treatment on consumption milk. PhD thesis, Technische Uni-
versität, Munich, Germany
Girotti, S., Ferri, E., Ghini, S., Budini, R., & Roda, A. (1994). Chemi-
luminescent assay of alkaline phosphatase in milk. Netherlands
Milk & Dairy Journal, 48, 213–224.
Gotham, S., Fryer, P., & Pritchard, A. (1992). b-lactoglobulin dena-
turation and aggregation reactions and fouling deposit forma-
tion: a DSC study. International Journal of Food Science and
Technology, 27, 313–327.
Grant, I., Rowe, M., Dundee, L., & Hitchings, E. (2001). Myco-
bacterium avium spp. paratuberculosis: its incidence, heat
resistance and detection in milk and dairy products. Interna-
tional Journal of Dairy Technology, 54, 2–13.
Grant, I., Hitchings, E., Mc Cartney, A., Ferguson, F., & Rowe, M.
(2002). Effect of commercial-scale high-temeprature, short-time
pasteurization on the viability of Mycobacterium para-
tuberculosis in naturally infected cows’ milk. Applied and Envir-
onmental Microbiology, 68, 602–607.
Greig, B., & Payne, G. (1985). Epimerization of lactose to free lactose
in heated model milk solutions. Journal of Dairy Research, 52,
409–417.
Griffiths, W. (1986). Use of milk enzymes as indices of heat treat-
ment. Journal of Food Protection, 49, 696–705.
Guha, K., & Roy, B. (1973). Enzymic differentiation between curds of
heated and raw milk. II. Milk peroxidase. Journal of the Science of
Food and Agriculture, 24, 1–6.
Haralampu, S., Saguy, I., & Karel, M. (1985). Estimation of Arrhenius
model parameters using three least squares methods. Journal of
Food Processing and Preservation, 9, 129–143.
Hartkopf, J., & Erbersdobler, H. (1993). Stability of furosine during
ion-exchange chromatography in comparison with reversed-
phase HPLC. Journal of Chromatography, 635, 151–154.
Harwalkar, V. (1980). Kinetics of thermal denaturation of b-lacto-
globulin at pH 2.5. Journal of Dairy Science, 63, 1052–1057.
Harwalkar, V. (1986). Kinetic study of thermal denaturation of pro-
teins in whey. Milchwissenschaft, 41, 206–209.
Hendrickx, M., Maesmans, G., De Cordt, S., Noronha, J., Van Loey,
A., Willocx, F., & Tobback, P. (1994). Advances in process
modelling and assessment: the physical mathematical approach
and product history integrators. In R. Single, & F. Oliveira (Eds.),
Minimal processing of foods and process optimisation (pp. 315–
336). CRC Press
Hernández, M., van Markwijk, B., & Vreeman, H. (1990). Isolation
and properties of lactoperoxidase from bovine milk. Netherlands
Milk & Dairy Journal, 44, 213–231.
Herrmann, G., Lamprecht, S., Frister, H., & Rudzik, L. (1996). New
heat indicator for heat treated milk and cheese. Deutsche
Milchwirtschaft, 47, 121–122.
Hill, C., & Grieger-Block, R. (1980). Kinetic data: generation, inter-
pretation, and use. Food Technology, 34, 56–66.
Jeurnink, T. (1991). Effects of proteolysis in milk on fouling in heat
exchangers. The Netherlands Milk and Dairy Journal, 45, 23–32.
Jeurnink, T., & De Kruif, K. (1993). Changes in milk on heating:
viscosity measurements. Journal of Dairy Research, 60, 139–150.
Jiménez-Pérez, S., Corzo, N., Morales, F.-J., Delgado, T., & Olano, A.
(1992). Effect of storage temperature on lactulose and 5-hydro-
xymethyl-furfural formation in UHT milk. Journal of Food Pro-
tection, 55, 304–306.
Johnson, A. (1978). The composition of milk. In B. Webb,
A. Johnson, & J. Alford (Eds.), Fundamentals of dairy chemistry
(pp. 1–57). Westport, Connecticut, US: The Avi Publishing Com-
pany, Inc.
309
Johnson, M. (1992). Review: why, when and how biochemists
should use least squares. Analytical Biochemistry, 206, 215–225.
Keeney, M., & Bassette, R. (1959). Detection of intermediate com-
pounds in the early stages of browning reaction in milk pro-
ducts. Journal of Dairy Science, 42, 945–960.
Kind, E., & Reuter, H. (1990). [HMF formation during UHT-treatment
of milk.]. Kieler Milchwirtschaftliche Forschungsberichte, 42, 87–98.
Law, B. (1979). Reviews of the progress of dairy science: Enzymes of
psychrotrophic bacteria and their effects on milk and milk pro-
ducts. Journal of Dairy Research, 46, 573–578.
Lenz, M., & Lund, D. (1980). Experimental procedure for determin-
ing destruction kinetics of food components. Food Technology,
34, 51–55.
Lewis, M. (1994). Heat treatment of milk. In R. Robinson (Ed.),
Modern dairy technology. Vol. 1. Advances in milk processing
(pp. 1–60). GB: Chapman & Hall.
López-Fandiño, R., Olano, A., Corzo, N., & Ramos, M. (1993). Pro-
teolysis during storage of UHT milk: differences between whole
and skim milk. Journal of Dairy Research, 60, 339–347.
Lyster, R. (1970). The denaturation of a-lactalbumin and b-lacto-
globulin in heated milk. Journal of Dairy Research, 37, 233–243.
Martin, D., Kiesner, C., Lorenzen, P., & Schlimme, E. (1989). [Ade-
nosine deaminase (EC 3.5.4.4): A potential indicator of heat
treatment for the distinction of short-time and high temperature
pasteurized milk from the market.]. Kieler Milchwirtschaftliche
Forschungsberichte, 50, 225–233.
Martin, D., Lorenzen, P., Kiesner, C., & Schlimme, E. (1999). [Analysis
of reaction kinetics occurring at thermal inactivation of the milk
enzymes adenosine deaminase, alkaline phosphatase and
g-glutamyl transferase.]. Kieler Milchwirtschaftliche For-
schungsberichte, 51, 343–355.
Martinez-Castro, I., & Olano, A. (1981). Ready detection of small
amounts of lactulose in dairy products by thin-layer chromato-
graphy. Chromatographia, 14, 621–622.
Mc Kellar, R., Emmons, D., & Farber, J. (1991). Gamma-glutamyl
transpeptidase in milk and butter as an indicator of heat treat-
ment. International Dairy Journal, 1, 241–251.
Mc Kellar, R., & Piyasena, P. (2000). Predictive modelling of inacti-
vation of bovine milk a-l-fucosidase in high-temperature short-
time pasteurizer. International Dairy Journal, 10, 1–6.
Metzler, C. (1981). Statistical properties of kinetic estimates. In
L. Endrenyi (Ed.), Kinetic data analysis, design and analysis of
enzyme and pharmacokinetic experiments (pp. 25–37).
NY-London: Plenum Press.
Mistry, V. (1989). Thermal inactivation characteristics of alkaline
phosphatase in ultrafiltered milk. Journal of Dairy Science, 72,
1112–1117.
Montilla, A., Calvo, M., Santa-Maria, G., Corzo, N., & Olano, A.
(1996). Correlation between lactulose and furosine in UHT-
heated milk. Journal of Food Protection, 59, 1061–1064.
Montilla, A., & Olano, A. (1997). Effect of the protein content and
dilution on the lactulose/furosine ratio in heat-treated milk.
Milchwissenschaft, 52, 506–507.
Morales, F.-J., Romero, C., & Jiménez-Pérez, S. (1995). New
methodologies for kinetic study of 5-(hydroxymethyl)-furfural
formation and reactive lysine blockage in heat-treated milk
and model systems. Journal of Food Protection, 58,
310–315.
Morales, F.-J., Romero, C., & Jiménez-Pérez, S. (1996). Evaluation of
heat-induced changes in Spanish commercial milk: hydro-
xymethylfurfural and available lysine content. International
Journal of Food Science and Technology, 31, 411–418.
Morales, F.-J., Romero, C., & Jiménez-Pérez, S. (1997). Chromato-
graphic determination of bound hydroxymethylfurfural as an
index of milk protein glycosylation. Journal of Agricultural and
Food Chemistry, 45, 1570–1573.
Morales, F.-J., & Jiménez-Pérez, S. (1999). HMF formation during
310 W.L. Claeys et al. / Trends in Food Science & Technology 13 (2002) 293–311
heat-treatment of milk-type products as related to milkfat con-
tent. Journal of Food Science, 64, 855–859.
Morales, F.-J., Romero, C., & Jiménez-Pérez, S. (2000). Character-
ization of industrial processed milk by analysis of heat-induced
changes. International Journal of Food Science and Technology,
35, 193–200.
Motulsky, H., & Ransas, L. (1987). Fitting curves to data using non-
linear regression: a practical and nonmathematical review.
FASEB, 1, 365–374.
Muir, D. (1996). The shelf-life of dairy products: 1. Factors influen-
cing raw milk and fresh products. Journal of the Society of Dairy
Technology, 49, 24–32.
Murthy, G., Bradshaw, J., & Peeler, J. (1990). Thermal inactivation of
phosphatase by the AOAC-V method. Journal of Food Protec-
tion, 53, 969–971.
Nangpal, A., & Reuter, H. (1990a). Reference diagram for lactulose
content in UHT milk. Kieler Milchwirtschaftliche For-
schungsberichte, 42, 65–76.
Nangpal, A., & Reuter, H. (1990b). Reference diagram for furosine
content in UHT milk. Kieler Milchwirtschaftliche For-
schungsberichte, 42, 77–86.
O’Brien, J. (1997). Reaction chemistry of lactose: non-enzymatic
degradation pathways and their significance in dairy products.
In P. Fox (Ed.), Advanced dairy chemistry, Vol. 3: Lactose, water,
salts and vitamins (pp. 155–231). London: Chapman & Hall.
Olano, A., Corzo, N., Paez, M., & Martinez-Castro, I. (1987). Iso-
merization of lactose during heat treatment of liquid and freeze-
dried simulated milk ultrafiltrates. Effect of pH and calcium.
Milchwissenschaft, 42, 628–630.
Olano, A., & Calvo, M. (1989). Kinetics of lactulose, galactose and
epilactose formation during heat-treatment of milk. Food
Chemistry, 34, 239–248.
Olano, A., & Martinez-Castro, I. (1989). Modification & interactions
of lactose. Bulletin of the International Dairy Federation, 238, 35–
44.
Olszewski, E., & Reuter, H. (1992). The inactivation and reactivation
behaviour of lactoperoxidase in milk at temperatures between
50 C and 135 C. Zeitschrift für Lebensmittel-Untersuchung und-
Forschung, 194, 235–239.
Painter, C., & Bradley, R. (1997). Residual alkaline phosphatase
activity in milks subjected to various time-temperature treat-
ments. Journal of Food Protection, 60, 525–530.
Parrish, F., Hicks, K., & Doner, L. (1980). Analysis of lactulose pre-
parations by spectrophotometric and HPLC methods. Journal of
Dairy Science, 63, 1809–1814.
Patel, S., & Wilbey, R. (1994). Thermal inactivation of g-glutamyl
transpeptidase and Enterecoccus faecium in milk-based sys-
tems. Journal of Dairy Research, 61, 263–270.
Pearce, R. (1989). Thermal denaturation of whey protein. Bulletin of
the International Dairy Federation, 238, 17–23.
Pellegrino, L. (1994). Influence of fat content on some heat-induced
changes in milk and cream. Netherlands Milk & Dairy journal, 48,
71–80.
Pellegrino, L., De Noni, I., & Resmini, P. (1995a). Coupling of
lactulose and furosine indices for quality evaluation of sterilized
milk. International Dairy Journal, 5, 647–659.
Pellegrino L., Resmini, P., & Luf, W. (1995b). Assessment (indices) of
heat treatment of milk. In P. Fox (Ed.), Heat-induced changes in
milk (pp. 409–453). 2nd edn. IDF Special Issue 9501
Peri, C., Pagliarini, E., & Pierucci, S. (1988). A study on optimizing
heat treatment of milk. I. Pasteurization. Milchwissenschaft, 43,
636–639.
Petterson, B., Lemke, F., Hammer, P., Stackebrandt, E., & Priest, F.
(1996). Bacillus sporothermodurans: a new species producing
highly heat-resistant endospores. International Journal of Sys-
tematic Bacteriology, 46, 759–764.
Phillips, J., & Griffiths, M. (1986). Factors contributing to the
seasonal variation of Bacillus spp. in pasteurized dairy products.
Journal of Applied Bacteriology, 61, 275–285.
Plock, J., Spiegel, T., & Kessler, H. (1998). Influence of the lactose
concentration on the denaturation kinetics of whey proteins in
concentrated sweet whey. Milchwissenschaft, 53, 389–393.
Pompei, C., & Rossi, M. (1994). Use of a model solution for the
evaluation of heat damage in milk treated in an ultrahigh-tem-
perature heat exchanger. Journal of Agricultural and Food
Chemistry, 42, 360–365.
Pouliot, Y., & Boulet, M. (1995). Observations on the seasonal var-
iations in the salt balance of concentrated milk. International
Dairy Journal, 5, 75–85.
Prakabaran, S., & Damodaran, S. (1997). Thermal unfolding of
b-lactoglobulin: characterization of initial unfolding events
responsible for heat-induced aggregation. Journal of Agricultural
and Food Chemistry, 45, 4303–4308.
Pruitt, K., Kamau, D., Miller, K., Månsson-Rahemtulla, B., & Rahem-
tulla, F. (1990). Quantitative, standardized assays for determining
the concentrations of bovine lactoperoxidase, human salivary
peroxidase, and human myeloperoxidase. Analytical Biochem-
istry, 191, 278–286.
Rattray, W., Gallmann, P., & Jelen, P. (1997). Influence of protein
standardization and UHT heating on the furosine value and
freezing point of milk. Lait, 77, 297–305.
Resmini, P., Pellegrino, L., & Battelli. (1990). Accurate quantification
of furosine in milk and dairy products by a direct HPLC method.
Italian Journal of Food Science, 3, 173–183.
Resmini, P., & Pellegrino, L. (1994). HPLC of furosine for evaluating
Maillard reaction damage in skimmilk powders during proces-
sing and storage. Bulletin of the International Dairy Federation,
298, 31–36.
Roberts, T., Baird-Parker, A., & Tompkin, R. (1996). Micro-organisms
in Foods. 5. Characteristics of microbial pathogens. London,
Weinheim, New York, Tokio, Melbourne, Madras: ICMSF, Blackie
Academic & Professional.
Roberts, T., Pitt, J., Farkas, J., & Grau, F. (1998). Micro-organisms in
foods. 6. Microbial ecology of food commodities. London, Wein-
heim, New York, Tokyo, Melbourne-Madras: ICMSF, Blackie
Academic & Professional.
Rocco, R. (1990). Fluorometric analysis of alkaline phosphatase in
fluid dairy products. Journal of Food Protection, 53, 588–591.
Saint-Denis, T., Humbert, G., & Gaillard, J. (2001). Heat inactivation
of native plasmin, plamsinogen and plasminogen activators in
bovine milk: a revisited study. Lait, 81, 715–729.
Sanders, G., & Sager, O. (1948). Heat inactivation of milk phos-
phatase in dairy products. Journal of Dairy Science, 31, 845–857.
Sandu, C., & Singh, R. (1991). Energy increase in operation and
cleaning due to heat-exchanger fouling in milk pasteurization.
Food Technology, 45, 84–91.
Sato, K., Dosako, S., Nakajima, I., & Ido, K. (1992). Effects of ionic
strength on thermostability of lactoperoxidase. Bioscience, Bio-
technology & Biochemistry, 56, 2054–2055.
Schaafsma, G. (1989). Effects of heat treatment on the nutritional
value of milk. Bulletin of the International Dairy Federation, 238,
68–70.
Schlimme, E., & Thiemann, A. (1992). Studies on alkaline phospha-
tase in bovine milk as a function of the stage of lactation. Kieler
Milchwirtschaftliche Forschungsberichte, 44, 371–382.
Schlimme, E., Buchheim, W., & Heeschen, W. (1994). Evaluation of
different heat treatments and heat indicators for consumption
milk. DMZ-Lebensmittelindustrie und Milchwirtschaft, 115, 64–69.
Schlimme, E., Clawin-Rädecker, I., Einhoff, K., Kiesner, C., Lorenzen,
P., Martin, D., Meisel, H., Molkentin, J., & Precht, D. (1996). Stu-
dies on distinguishing features for evaluating heat treatment of
milk. Kieler Milchwirtschaftliche Forschungsberichte, 48, 5–36.
Schlimme, E., Kiesner, C., Lorenzen, P., & Martin, D. (1997). Chemical
process parameters for thermal inactivation of alkaline
 W.L. Claeys et al. / Trends in Food Science & Technology 13 (2002) 293–311 311

phosphatase in milk. Kieler Milchwirtschaftliche For-
schungsberichte, 49, 207–219.
Sciancalepore, V., De Stefano, G., & Piacquado, P. (1996). Influence
of mono- and divalent cations on thermostability of lactoper-
oxidase in model systems. Milchwissenschaft, 51, 512–514.
Spreer, E. (1998). Milk and dairy product technology. Dresden,
Germany: Marcel Dekker, Inc.
Stabel, J., Pearce, L., Chandler, R., Hammer, P., Klijn, N., Cerf, O.,
Collins, M., Heggum, C., & Murphy, P. (2001). Destruction by
heat of Mycobacterium paratuberculosis in milk and milk pro-
ducts. Bulletin of the International Dairy Federation, 362, 53–61.
Straume, M., & Johnson, M. (1992). Analysis of residuals: criteria
for determining goodness-of-fit. Methods in Enzymology, 210,
87–105.
Taoukis, P., & Labuza, T. (1989). Applicability of time temperature
integrators as shelf life monitors of food products. Journal of
Food Science, 54, 783–788.
Thorne, J., Hardin, L., Spain, J., & Hardin, D. (1995). Factors influen-
cing milk composition from lactating cows. Journal of Dairy
Science, 78(Suppl.1), 179.
Tirelli, A., & Pellegrino, L. (1995). Determination of furosine in dairy
products by capillary zone electrophoresis : a comparison with
the HPLC method. Italian Journal of Food Science, 4, 379–385.
van Boekel, M., & Walstra, P. (1989). Physical changes in the fat
globules in unhomogenized & homogenized milk. Bulletin of the
International Dairy Federation, 238, 13–16.
van Boekel, M., & Walstra, P. (1995). Use of kinetics in studying
heat-induced changes in foods. Bulletin of the International Dairy
Federation (Special Issue), 9501, 22–50.
van Boekel, M. (1996). Statistical aspects of kinetic modeling for
food science problems. Journal of Food Science, 61, 477–485,
489.
van Boekel, M. (1998). Effect of heating on Maillard reactions in
milk. Food Chemistry, 62, 403–414.
van den Berg, L. (1984). The thermization of milk. Bulletin of the
International Dairy Federation, 182, 3–12.
Van Loey, A. (1996). Enzymic time temperature integrators for the
quantification of thermal processes in terms of food safety.
PhD thesis, K.U.Leuven, Laboratory of Food Technology,
Leuven, Belgium
Verheul, M., Roefs, S., & de Kruif, K. (1998). Kinetics of heat-induced
aggregation of b-lactoglobulin. Journal of Agricultural and Food
Chemistry, 46, 896–903.
Walstra, P., & Jeness, R. (1984). Dairy chemistry and physics. New
York: John Wiley & Sons.
Walstra, P., Geurts, T., Noomen, A., Jellema, A., & van Boekel, M.
(1999). Dairy Technology. Principles of milk—Properties and
processes. New York-Basel: Marcel Dekker, Inc.
Wilbey, A. (1996). Estimating the degree of heat treatment
given to milk. Journal of the Society of Dairy Technology, 49,
109–112.
Zehetner, G., Bareuther, C., Henle, T., & Klostermeyer, H. (1995).
Inactivation kinetics of g-glutamyltransferase during the heating
of milk. Zeitschrift für Lebensmittel-Untersuchung und-Forschung,
201, 336–338.
Zehetner, G., Bareuther, C., Henle, T., & Klostermeyer, H. (1996).
Inactivation of endogenous enzymes during heat treatment of
milk. Netherlands Milk & Dairy Journal, 50, 215–226.

Any Suggestions?
Articles published in TIFS are usually specially invited by the Editors, with assistance from our International
Advisory Editorial Board. However, we welcome ideas from readers for articles on exciting new and developing
areas of food research. A brief synopsis of the proposal should first be sent to the Editors, who can provide
detailed guidelines on manuscript preparation.

Mini-reviews focus on promising areas of food research that are advancing rapidly, or are in need of re-review
in the light of recent advances or changing priorities within the food industry. Thus they are shorter than
conventional reviews, focusing on the latest developments and discussing likely future applications and research
needs.

Features are similar in style to mini-reviews, highlighting specific topics of broad appeal to the food science
community.

The Viewpoint section provides a forum to express personal options, observations or hypotheses, to present new
perspectives, and to help advance understanding of controversial issues by provoking debate and comment.

Conference Reports highlight and assess important developments presented at relevent conferences worldwide.

TIFS also welcomes Letters to the Editor concerned with issues raised by published articles or by recent
developments in the food sciences.

All Review-style articles are subject to editorial and independent peer review by international experts in the
appropriate field.