Ability of some food preservation processes to modify the overall nutritional value of food

Journal of Food Engineering 100 (2010) 613–621

Contents lists available at ScienceDirect

Journal of Food Engineering journal homepage: www.elsevier.com/locate/jfoodeng

Ability of some food preservation processes to modify the overall nutritional value of food Nawel Achir a,*, Janvier Kindossi b, Philippe Bohuon a, Antoine Collignan a, Gilles Trystram b a b

UMR 95 QualiSud (CIRAD, Montpellier SupAgro, Montpellier 1, Montpellier 2), F34398 Montpellier, France UMR 1145 Genial (INRA, Cemagref, AgroParistech), F91744 Massy, France

a r t i c l e

i n f o

Article history: Received 28 January 2010 Received in revised form 30 April 2010 Accepted 13 May 2010 Available online 24 May 2010 Keywords: Food processing Air drying Osmotic treatment Nutritional scores SAIN LIM Apple Pork belly

a b s t r a c t The impact of food transformations on final nutritional quality of food products is a major question that has been studied only partially. This paper propose an assessment of different preservation processes with the light of two overall nutritional scores, positive (SAIN) and negative (LIM), developed in 2008, in response to a European regulation made to improve nutritional information on processed food. With the example of two different products, apple and pork meat, the objective is to monitor the nutritional scores as a function of the preservation operation chosen, isolated or combined, traditional or innovative. The results show that nutritional scores are very different as a function of the mass transfers involved during the preservation operation. The results give insights in the ability of a technological choice to modify the nutritional quality of a food item and potentially orientate a food label. This work also shows the limitations of using global nutritional scores as a tool to evaluate a food process. Ó 2010 Elsevier Ltd. All rights reserved.

1. Introduction The recent reports of the World Health Organization underlined the severe impact of poor diets on health conditions. Studies estimate that around 30% of coronary heart diseases and 20% of strokes are caused by high fat, salt and free sugar consumption at the expense of fruits and vegetables (WHO, 2002, 2003). The severity of the problem encouraged policymakers to undertake a global strategy to promote physical activity associated to a balanced diet (Lee, 2005). In this context, the European Commission adopted a proposal of regulation on food labelling to allow only nutritional claims based on solid scientific data enabling consumers to have reliable information on the nutritional quality of a food product (European Commission, 2006). In response to this proposal, methodologies to profile food items in an objective manner, according to their nutritional composition, were proposed (Arambepola et al., 2008; Visioli et al., 2007; Scarborough et al., 2006; Darmon and Darmon, 2008). A nutrient profile model consists in classifying foods in more or less healthy groups and eventually ranks them as a function of their nutritional characteristics (Scarborough et al., 2006). The outputs of these models are food classes (positive or negative), allowance to carry logo, partial or global scores repre* Corresponding author. Address: 1101, Avenue Agropolis, BP 5098, F34093 Montpellier Cedex 5, France. Tel.: +33 467 87 40 91; fax: +33 467 61 70 55. E-mail address: [email protected] (N. Achir). 0260-8774/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.jfoodeng.2010.05.009

senting the food nutritional quality (Scarborough et al., 2007). The nutrient profile model of Darmon and Darmon (2008) deliver two independent scores presenting the positive and negative nutritional aspects of a food product. The first score, called SAIN for nutrient adequacy score for individual foods, is a nutrient density in an energy base. The second indicator, called LIM for limited nutrients, is calculated with percentages of excess of undesirable components in a mass base. To represent the nutritional quality of a product, Darmon and Darmon use these two independent scores as (x,y) coordinates to place raw and processed foods in a two-dimension frame. This surface representation of foods allows a high degree of freedom in their localization and enables a convenient visual classification of them. Along with comparing the nutritional quality of raw and processed foods, this tool could be very interesting to visualize the evolution of the product during its transformation. Indeed, during the life of a food product, from the harvest to its consumption, the human action during food processing is known to modify drastically its quality attributes, and therefore its nutritional characteristics (van Boekel, 2008). Some experimental works have already studied the impact of processes on foods nutritional quality but they usually analysed them by monitoring a single nutritional indicator such as a vitamin, a mineral content or a global energy density (Marty and Berset, 1986; Mattila et al., 1999; Bedi et al., 2006; Morgan et al., 1988; Moreira et al., 1997). However, for a same product, depending on the indicator chosen,

614

N. Achir et al. / Journal of Food Engineering 100 (2010) 613–621

recommendations to optimize the process conditions can be contradictory. For instance, frying potato chips at a high temperature is recommended to lower oil uptake but this condition enhances acrylamide formation (Pedreshi, 2005). In addition, optimization modelling outputs are usually based on the maximisation of the retention of one nutrient. However the concentration reached is not pondered with the nutritional needs and do not integrate other nutritional characteristics of the product. Therefore, the use of an overall nutritional assessment provided by nutrient model profiles could be very interesting during food processing to criticize each operation or their succession in order to identify the conditions or technological route to achieve the best nutritional trade-off. This work is not easy firstly because the knowledge of the nutritional composition evolution during food processes supposes a sufficient characterisation of the heat and mass transfers and coupled reactions. Secondly, the optimization amplitude of food processes is limited by stability, technological, sensory and cost constraints which must be respected (Oliveira et al., 1999). Therefore, the aim of this paper is to evaluate the sensitivity of a nutrient profile model, to estimate the nutritional impact of some food preservation processes. This work is done by analysing the nutritional trajectory of two different products: apple and pork belly during a single or a combined isothermal preservation process in the two-dimension plan made by the SAIN, LIM coordinates proposed by Darmon and Darmon (2008). The model responses are calculated and discussed for different process itineraries and choices, i.e., operating conditions, succession of food transformations, traditional or innovative technologies. These results are discussed as regard to the consequences on food labelling and the ability of the referential SAIN, LIM to describe appropriately the evolution of food during processing.

2. Materials and methods The scores SAIN and LIM were previously developed by Darmon and Darmon (2008) and in Darmon et al. (2009). They evaluate respectively the capacity of a food product to cover the nutritional needs and not to exceed the recommended intake in disqualifying nutrients. A sweet fruit product, apple, and a salted meat product,

Table 1 List of the 23 positive nutrients, their daily recommended values (RV) and the contents in 100 g of raw apple and pork belly used for the calculation of the nutrient density score SAIN. Nutrients

Unit

RV

Apple

Pork belly

Protein Fiber Monounsaturated fatty acids (MFAs) Linolenic acid (ALA) Linoleic acid (LA) Vitamin E Magnesium Potassium Calcium Iron Copper Zinc Selenium Iodine Vitamin A Vitamin D Ascorbic acid Thiamine (vitamin B1) Riboflavin (vitamin B2) Niacin (vitamin B3) Vitamin B-6 Folates (vitamin B-9) Vitamin B-12

g g g g g mg mg mg mg mg mg mg lg lg lg lg mg mg mg mg mg lg lg

65.0 30.0 44.4 1.80 9.00 12 390 3100 900 12.5 1.8 11 55 150 700 5.0 110 1.2 1.6 13.0 1.7 315 2.4

0.3 2.1 0.007 0.043 0.01 0.32 6.0 120 2.5 0.2 0.03 0.04 0.3 0.2 54.0 0.0 3.3 0.02 0.02 0.10 0.07 13.00 0

16.0 0.0 9.42 0.48 5.03 0.10 18.8 293 6.68 0.52 0.05 1.84 40 1.95 3.0 0.0 0.3 0.59 0.15 4.65 2.17 2.17 0.37

pork belly, are used to assess the following preservation processes: air drying, osmotic treatment, and smoking operation. Air drying and osmotic treatment have the same objective to stabilize food products by a decrease of the water activity by lowering water content and/or by adding salts or sugars. During the smoking operation (for pork belly only), volatile compounds which have antiseptic properties covers the meat. These unit operations can be combined. The choice of these operations is motivated by the involvement of mass transfers which can modify the energy density and micronutrient contents of the product. These mass transfers will be taken into account dynamically at the scale of each unitary operation in the case of apple processing and at the scale of the whole process till the domestic use for pork belly transformation. 2.1. The SAIN, LIM system The SAIN score is a nutrient density calculated by the unweighted arithmetic mean of the percentage adequacy for the food positive nutrients. The selection of the number of nutrient, from 5 to 23, is a trade-off between the need to have a complete model including nutrients of importance to public health and the need for a manageable number of nutrients (Darmon et al., 2009). Because apple and pork meat are very different in terms of composition, the definition of the SAIN score including 23 positive nutrient will be used (Darmon and Darmon, 2008) (Table 1): 23 P

Nut

 RV i i

i¼1

23

SAIN ¼

 100  100 E

ð1Þ

Where Nuti, is the quantity (g, mg or lg) of the positive nutrient i in 100 g of food, RVi, is the daily recommended value for nutrient i and E is the energy (in kcal) in100 g of food. The second score LIM calculates the mean content in disqualifying nutrients in 100 g of foods. The number of disqualifying nutrient is three in Darmon and Darmon (2008), i.e., sodium, saturated fatty acids and added sugars (Table 2): 3 P

LIM ¼

Nutj MRVj

j¼1

3

 100

ð2Þ

where Nutj, is the quantity (g, mg or lg) of disqualifying nutrient j in 100 g of food and MRVj, the daily maximal recommended value for nutrient j. Darmon et al. (2009) established thresholds values for SAIN and LIM scores. For the SAIN score, they defined an optimum value of 100% in 2000 kcal of food (reference daily energy intake). This optimum was therefore equivalent to 5% for 100 kcal of food. As a consequence, a SAIN value superior to five indicates a good nutrient density. Unlike the SAIN score, the LIM score is calculated for 100 g of food and the maximal recommended values are based on food intake rather than energy intake (Table 2). Thus, Darmon et al. (2009) defined a limit value for the LIM score of 100% in 1330 g (mean daily food intake of the French population). This threshold corresponds to 7.5% for 100 g of food. A LIM value inferior to 7.5 indicated a low content of disqualifying nutrient. Theses

Table 2 List of the three negative nutrients, their daily maximal recommended values (MRV) and the contents in 100 g of raw apple and pork used for the calculation of the limited nutrient score LIM. Nutrients

Unit

MRV

Apple

Pork belly

Saturated fatty acids (SFAs) Added sugars Sodium

g g mg

22 50 2 365

0.06 0 1.6

12.4 0 106.2

615

N. Achir et al. / Journal of Food Engineering 100 (2010) 613–621

2009). Experimental mass transfer data were chosen in previous works of Vega-Galvez et al. (2008) and Thémelin (1994) to reach an apple final water activity which range between 0.64 and 0.65 (i.e., stable products at ambient temperature). In the first technological choice (a), apple cubes were dried in a conventional air dryer at 60 °C during 524 min. In the second case (b), apple cubes underwent a previous osmotic treatment at 59 °C during 284 min (i.e., immersion in a concentrated sugar solution) followed by air drying in a conventional dryer at 69 °C during 300 min. The objective of this preliminary operation was to both partly dehydrate and impregnate the apples with sugars. The consequence of these two opposite mass transfers is the decrease of water activity by dehydration and sugar impregnation. 2.3. Pork belly processing

Fig. 1. Graphic representation of the two-dimension plan SAIN, LIM and the thresholds delimiting the four nutrient profile classes (log scale).

two limits enabled the representation of the SAIN, LIM coordinates of a food item into four nutrient profile classes from ‘‘recommended for health” to ‘‘consumption should be limited” which are represented in the log–log plot of Fig. 1. 2.2. Apple processing Fig. 2 presents the flowchart of apple processing by (a) air drying and (b) a combined osmotic treatment and air drying. Apples were of variety Granny Smith (Malus sylvestris Miller) and their composition is given in Table 1 and 2 (Favier et al., 1995; USDA,

Three distinct technological routes are compared in Fig. 3 for the production of pork belly meat (initial composition in Tables 1 and 2 (Favier et al., 1995; USDA, 2009)). The (a) and (b) routes are used in the traditional fabrication of smoked pork belly (Poligné et al., 2001a,b) and work out successive salting, and smoking-drying operations. The final product is dehydrated (residual water content of 28.9%) and salted (5.6% salt content) and is therefore stable (aw 6 0.83). The salting operation consists in a very short immersion (5 min) in a salted solution. The successive traditional drying-smoking operation is carried out during 22 h in a smoke room where meat pieces are hung up over a hearth. Therefore, this operation enables simultaneous cooking, drying, and smoking of the pork bellies. However, the direct contact between meat and hearth can lead to a product with a high content in a cancerigenic molecule resulting from wood pyrolysis which is benzo[a]pyrene (BaP). For instance, Poligné et al. (2001a) found a BaP final content of 6.9 lg/kg in the end products of the traditional

RAW APPLE

WL(%) = 86.5

100 kg

SuG(%) = 11.6 (b) Osmotic treatment – air drying

(a) Air drying

Air drying 60°C

542 min

Osmotic treatment

WL(%) = 83.0

59°C 59% sugar

WL(%) = 63.0 SuG(%) = 10.9

284 min

Formulated product 47.8 kg

W(%) = 49.0 Su(%) = 49.9

Air drying 69°C

WL(%) = 80.4

300 min

End product 17.0 kg

End product

W(%) = 20.5

Aw =0.64

30.4 kg Aw =0.65

W(%) = 20.0 Su(%) = 73.8

WL: water loss (kg/100 kg initial apple) SuG: sugar gain (kg/100 kg initial apple) W, Su: water and sugar content (kg/100 kg wet basis) Fig. 2. Flow sheet of two dehydration processes of apple cube: (a) traditional air drying process from Vega-Galvez et al. (2008) and (b) combined process of osmotic treatment following with air drying from Thémelin (1994).

616

N. Achir et al. / Journal of Food Engineering 100 (2010) 613–621

W(%) = 51.7 St(%) = 0.3 F(%) = 32.4

RAW PORK MEAT 100 kg

(c) Osmotic treatment Dry salting 27°C, 5 min

WL(%) = -0.5 StG(%) = 5.3

Soaking in a

FL(%) = 0.1

complex

Salted product

solution

105.7 kg

40°C

WL(%) = 24.3 StG(%) = 5.4 SuG(%) = 0.8

18 h (a) Traditional proces

Drying–smoking Ts: 50-120°C 22 h

End product 67.1 kg Aw = 0.83

(b) Innovative smoking

WL(%)= 31 StG(%)= −1.4

Drying–smoking

WL(%)= 31

Ts: 50-120°C

StG(%)= 0

8h

FL(%)= 0

FL(%) = 6.2

W(%) = 28.5 St(%) = 5.6

W(%) = 25.9

End product

FL(%) = 38.0

74.7 kg Aw = 0.80

BaP= 6.9µg/kg

W(%) = 32.0

End product

St(%) = 7.5

79.5 kg

F(%) = 43.1

Aw =0.81

BaP(%) = 0.3 µg/kg

St(%) = 6.0 Su(%) = 1.0 F(%) = 41.2

WL: water loss (kg/100 kg initial meat) FL: fat loss (kg/100 kg initial meat) SuG: sugar gain (kg/100 kg initial meat) StG: salt gain (kg/100 kg initial meat) W, F, Su, St: water, fat, sugar, and NaCl content Fig. 3. Flow sheet of three stabilisation processes of pork belly: (a) traditional process from Poligné et al. (2001a,b), (b) innovative smoking equipment from Bruneau et al. (2005) and (c) osmotic treatment in a complex solution at 40 °C from Poligné et al. (2005).

process (Fig. 3) which is much higher than the maximal daily recommended value (OFSP, 2008). To overcome this food safety problem, technological solutions were developed. The first one, presented in Fig. 3b (innovative smoking), is an innovative smoke device developed by Bruneau et al. (2005). In this case, the surface temperature of the product Ts is similar during the drying-smoking operation but contrary to the traditional process, the product is not in a direct contact with the hearth. In addition, the improvement of heat transfer in this device shortens the process duration to 8 h. The sensory properties of the meat are ensured by an aromatising step with cold fumes at a low content in BaP. Indeed, the resulting final BaP content was significantly lower in the final product (0.3 lg/kg). The second innovation, presented in Poligné et al. (2005) and in Fig. 3c, is a direct formulation of pork meat by osmotic treatment in a complex solution (consisting of water, salt, glucose syrup and liquid smoke flavouring) at 40 °C during 18 h. This operation enables a simultaneous salting, drying, cooking and aromatising of the meat.

2.4. Integration of mass transfers during processes in the SAIN and LIM scores 2.4.1. Dynamic integration of mass transfers during apple processing During air-drying and osmotic treatment, concentration of the ð0Þ initial positive nutrient i, Nuti , is multiplied by a factor

ðM ð0Þ =M ðtÞ Þ, where M ð0Þ and M ðtÞ are the food weight at t = 0 and time t: ðtÞ

Nuti ¼

M ð0Þ M ðtÞ

ð0Þ

 Nuti

ð3Þ ðtÞ

Because of water loss, which is the main mass transfer, Nuti is assumed to increase as a function of the process time. The Eq. (1) was modified to take into account the composition modifications during air-drying and osmotic treatment and generate a value of SAIN(t) at each time t of the process according to the experimental mass transfer data of Vega-Galvez et al. (2008) and Thémelin (1994): 23 P Mð0Þ i¼1

SAINðtÞ ¼

MðtÞ

ð0Þ i RVi

Nut



23 ðtÞ

E

 100

 100

ð4Þ

The energy in 100 g of apple, noted EðtÞ , is dependent on the only water loss during drying. In the case of osmotic treatment, EðtÞ is dependant on water loss but also on sugar gain. It is important to ðtÞ notice that Eqs. (3) and (4) only consider the evolution of Nuti because of the desired mass transfers of both operations. This situation, where no nutrient loss occurs, can be considered as the most optimistic scenario. Indeed, during processing, undesired loss of nutrients can happen because of two reasons: reactions and transport phenomena. For a same final product, the importance of a nutrient loss is very variable as a function of the raw material, the

N. Achir et al. / Journal of Food Engineering 100 (2010) 613–621

type of process, its conditions etc., and all these variables can affect the nutrient retention in a very important manner (Leskova et al., 2006). Therefore, we also consider a pessimistic scenario according to the following hypotheses. The chemical reactions affecting the nutritional quality of food are mainly caused or enhanced by temperature, oxygen and light. They affect vitamins, and especially vitamins A and C which are considered the most sensitive. Transport of nutrients can occur when the product is soaked in water and affect mainly hydrosoluble nutrients such as minerals or hydrosoluble vitamins (Leskova et al., 2006). As a consequence, vitamin loss are frequently reported during air-drying because of oxygen and temperature (Marfil et al., 2008; Timoumia et al., 2007).On the opposite, because of oxygen limitation, osmotic dehydration was proven to preserve micro-nutrient (Dermesonlouogloua et al., 2007; Rastogi et al., 2005) but, because of the immersion of the food product, this operation can induce a loss of water-soluble vitamins and minerals from the product to the concentrated solution (Peiro et al., 2006; Osorio et al., 2007). Therefore, in the most pessimistic case, we will consider the total destruction of vitamin A and C during drying and the total loss of minerals and hydrosoluble vitamins in the case of osmotic treatment.As well as positive nutrients, disqualifying nutrið0Þ ents Nutj concentrates by a factor (Mð0Þ =M ðtÞ ) during both processes. However, the nutrient ‘‘added sugar” is particular because by definition absent at t = 0. Therefore, in this case, the concentration was assessed at each time of the process thanks to the data of sugar gain of Thémelin (1994). Eq. (2) was modified in order to have a value LIMðtÞ at each time t of the process: ðtÞ

LIMðtÞ ¼

Nut1 MRV1

þ

3 P j¼2

M ð0Þ MðtÞ

3

ð0Þ



Nutj

MRVj

 100

ð5Þ

where indice 1 stands for added sugar while 2 and 3 stand for saturated fatty acids and sodium, respectively. 2.4.2. Integration of mass transfers at each step of pork belly processing Because of the complexity of the pork belly flowchart and the numerous nutrient transfers involved (water, fat, salt, sugar, BaP), integration of mass transfers is done at the scale of the whole process. Therefore, the SAIN and LIM scores are calculated with Eqs. (1) and (2) at each step of the process till the domestic use thanks to the nutrient analysis of pork belly available in Poligné et al. (2001a,b) and Bruneau et al. (2005).

617

after processing 100 kg of raw apples. For the combined process (b), water loss occurred in both osmotic treatment and air drying, but this loss was partly compensated by a sugar gain during the first osmotic dehydration step. Therefore, for the same final water activity, 30.4 kg of end product were obtained in this case. Fig. 4 presents the dynamic evolutions of SAIN and LIM scores as a function of processing time for the two technological routes (a) and (b). During air drying (Fig. 4a), the SAIN score remained equal to 3.6. This stability was due to the proportionality of nutrient and energy concentrations as a function of dehydration and the fact that the two values are on the numerator and the denominator of the SAIN score. The LIM score slightly increased from 0.13 to 0.78 (Fig. 4). This increase is due to the concentration of the initial saturated fatty acids in apple cubes because of dehydration and the fact that the energy density is not taken into account in the case of the LIM score. As a consequence, the air drying operation has a low impact on the SAIN, LIM scores of apple cubes and the end products also belonged to the neutral nutrient profile class of Fig. 1. However, the SAIN score might be over evaluated. Indeed, air drying lasted nearly 550 min. This important contact time between the air oxygen and the product at a temperature of 60 °C is assumed to degrade thermo-sensitive molecules (i.e., vitamin A and C). Therefore, supposing that the two vitamins were totally destroyed, the final SAIN was calculated and was found to drop from 3.6 to 2. This scenario was foreseen in Fig. 5. Regarding Fig. 4b, the consequences of mass transfers during osmotic treatment had a marked impact on the SAIN but also on the LIM score. Indeed, during this treatment, the SAIN score decreased from 3.6 to 2. This decrease was due to the increase of sugar content of the apple increasing its caloric-density. Therefore, the nutrient concentration because of water loss did not compensate in an equal manner the increase of caloric-density. The sugar impregnation during osmotic treatment had a higher consequence on the LIM score. Indeed, the LIM value increased at constant speed as a function of the osmotic process time and rise from an initial value of 0.13–15.4 after 284 min. The sugar implementation during this operation had a dramatic impact on the numerator value of the LIM score. During the following air drying, the SAIN score remained stable. Because of the water loss only, the increase of the SAIN denominator due to the sugar concentration was compensated by the positive nutrient concentration (numerator of SAIN). On the opposite, LIM score increased during air drying to reach a final value of 24.2. In Fig. 5, nutritional trajectories of the apple cubes are presented in the SAIN, LIM frame during air drying and during the

3. Results 3.1. Evolution of SAIN and LIM scores during apple processing The values of the 23 positive and negative nutrients (Tables 1 and 2, respectively) of apple lead to an initial SAIN score of 3.6 and a LIM score of 0.13. Apple belongs to the neutral class according to the classification of Fig. 1. However, it is important to notice that if the SAIN score was calculated with five positive nutrients as in Darmon et al. (2009), the denominator in Eq. (1) would be smaller and the resulting value would be 5.8, therefore placing the fruit in the first class of food recommended for health. The more important positive nutrients in apple are fiber and vitamin A (7% of the daily recommended values in 100 g). The little value of LIM is due to the low saturated fatty acids content of apple (0.3% of the daily maximal recommended values in 100 g). The experimental mass transfers during the two technological routes (a) and (b) presented in Fig. 2 were used to calculate the SAIN and LIM scores during processing with Eqs. (4) and (5). In the case of a single air drying, a water loss of 83 g for 100 g of initial apple occurred and finally, 17.0 kg of end products were obtained

Fig. 4. Evolution of SAIN (s) and LIM (*) scores as a function of processing time for (a) air drying (data from Vega-Galvez et al., 2008) and (b) combined osmotic treatment and air drying (data from Thémelin, 1994).

618

N. Achir et al. / Journal of Food Engineering 100 (2010) 613–621

Fig. 5. Trajectory of apple cubes in the SAIN, LIM frame during air drying and combined osmotic treatment and air drying. Prediction of the end products in case of vitamin destruction during air drying and of water soluble molecule leak in the combined process.

combination of osmotic treatment and air drying. As observed previously, during air drying, apple cubes remained in the neutral class 2 of Fig. 1 during the whole treatment. Indeed, the SAIN score was mathematically stable and the LIM score slightly increased (log scale) because of the concentration of few saturated fatty acids initially present in the apple. In the case of the combined process, apple cubes passed from nutrient profile class 2–4 after 94 min of osmotic treatment. Therefore, they turned into products of which consumption should be limited. In addition, during osmotic treatment, nutrient loss can occur with water leak. This scenario was also foreseen in Fig. 5. In this case, the SAIN, LIM coordinates would pass out of the frame defined by Darmon and Darmon (2008) that is to say with a SAIN score <1. Thus, while improving the stabilisation, the yield, and the sensory properties of apples, osmotic dehydration has a dramatic negative impact on the SAIN and LIM values. 3.2. Evolution of SAIN and LIM scores during pork belly processing Initial SAIN and LIM scores were calculated for pork belly according to the values of the 23 positive and negative nutrients (Tables 1 and 2, respectively). The scores were of 5.9 and 20.3 for SAIN and LIM, respectively. Both values are high and consequently set the raw meat in the class of products recommended in small quantities or occasionally (class 3 of Fig. 1). The high SAIN score is due the high content in protein and unsaturated fatty acids. Indeed, 100 g of raw pork meat provide 25% of the daily recommended value (RV) in protein, 21% in monounsaturated fatty acid, 27% in linolenic acid and 56% in linoleic acid. This meat is also rich in selenium (72% of daily RV) and vitamin B1, B3, B6 (49%, 36% and 128% of daily RV). If these micro-nutrients were not considered, i.e., calculated with five positive nutrients (Darmon et al., 2009), the SAIN score would be 1.7 therefore placing the meat in the fourth class of food to avoid or limit. The high LIM value is due to the high saturated fatty acid content of the meat. Indeed, 100 g provide 56% of the daily maximal recommended value.

The experimental evolution of pork belly compositions, presented in Fig. 3, were used to calculated the SAIN and LIM scores during the different steps of the three processes a, b and c. These values are synthesized in Table 3. The first operation of salting was identical for the traditional process (Fig. 3a) and the process involving the innovative device (Fig. 3b). After salting, the LIM value increased from 20.3 to 47.0. This marked increase was due to a salt gain of 5.3% impacting the LIM score. This operation did not affect the SAIN score as it did not modify significantly the caloric density of the product. Despite of the increase of the LIM value, both raw and salted products were in the same nutritional class. In the traditional route (a), the SAIN score slightly increased from 5.9 to 7.4 during the simultaneous process of drying, cooking and smoking. This evolution was linked to the significant fat loss by smelting of 6.2% (Fig. 3). This matter loss decreased the value of the SAIN score denominator (i.e., calories in 100 g of product) but only a part of the numerator (i.e., only the lipid nutrients). Therefore, the SAIN mathematically increased. Because of a simultaneous important water loss during the traditional process, the final salt and fat contents were much higher in the end products than in the raw meat. Therefore, the LIM value was high (53.3). This end product is still in class 3 of Fig. 1. During the innovative smoking process (Fig. 3b), the SAIN score remained the same because of a negligible fat loss. Therefore the only mass transfer was water loss which has no consequence on the SAIN score. Because of the absence of fat loss, the LIM value was higher for the final end products obtained with the innovative equipment (66) than with the traditional device. Regarding the values of SAIN and LIM during the osmotic treatment of Fig. 3c), the SAIN score remained equal to that of the raw meat. Indeed, sugar gain was negligible (0.8%) and thus did not impact significantly the caloric density of the meat. Salt gain was higher (5.4%) but salt does not impact the caloric density. Therefore, only water loss occurred and the simultaneous concentration of nutrients and caloric density of meat did not change significantly the SAIN value. Because of fat concentration and sugar gain, the LIM score also increase to a value of 57.3 comparable to the other treatments. This value was slightly higher than that obtained with the traditional process because of the additional sugar gain. It was lower than the value obtained with the first innovation because of a lower water loss to reach the same stability. In a general pattern, a strong increase of the LIM value was obvious for the three processed because of drying, salting and in the case of the second innovation, because of a slight sweetening. SAIN values were close as well for the three processes but penalizing for the two innovations because of the absence of fat loss during pork transformation. However, both innovations were made to improve the product safety. Poligné et al. (2001a) found a final benzo[a]pyrene (BaP) content of 6.9 lg/kg which is much higher than the maximal daily recommended value of 0.235 lg published for this molecule (OFSP, 2008). Therefore, it would be of interest to take into account this harmful constituent in cured pork belly. BaP could be added in the list of disqualifying nutrient to calculate a

Table 3 SAIN and LIM scores for raw and salted pork meat and for end products obtained after the processes presented in Fig. 3. Product

SAIN

LIM

Raw material Salted product End product from the traditional process (a) End Product from the innovative smoking equipment (b) End Product from the osmotic treatment (c)

5.9 5.9 7.4 7.4 5.8

20.3 47.0 53.3 66.0 57.3

N. Achir et al. / Journal of Food Engineering 100 (2010) 613–621 Table 4 SAIN and LIM4 scores integrating benzo[a]pyrene as the fourth disqualifying nutrient for raw and salted pork meat and for end products obtains after the processes presented in Fig. 3. Effect of domestic desalting on SAIN and LIM4 values. Product

SAIN

LIM4

Raw material Salted product End product from the traditional process End product from the innovative smoking equipment End product from the osmotic treatment Desalted product from traditional process Desalted product from innovative equipment Desalted product from osmotic treatment

5.9 5.9 7.4 7.4 5.8 7.4 7.4 5.8

15.2 35.2 115.9 52.8 43.2 84.9 40.2 35.2

new LIM score. This calculation was done with a maximal recommended value of 0.235 lg (OFSP, 2008) and a total number of disqualifying nutrients of four. The new LIM4 values are presented in Table 4 for the three pork processes. For the raw meat, LIM4 value is 15.3, which is lower the previous value of 20.3 of Table 3. This fact is inherent to the calculation of the LIM value (Eq. (2)) whose denominator is the number of disqualifying nutrients. The increase of this number from three (Table 3) to four (Table 4) decrease the LIM value. As well, for the same reason, the LIM4 value was lower for the final products obtained with the two innovations (b) and (c). In the case of the traditional drying-smoking, this decrease is hidden by a very marked increase of the LIM4 value (115.9) due to the very high concentration of BaP in the traditional end products. Values of Table 4 were used to draw the nutritional trajectories of pork belly during the three processes in the SAIN, LIM frame in Fig. 6. The three processes trigger off the increase of the LIM4 value because of the salt gains and the dehydration resulting in the concentration of both salt and fat. However, a very marked increase was due to the deposition of BaP in case of the traditional process. Indeed, while the two innovations enable end products to remain in the nutrient profile class 3 of Fig. 1 (recommended occasionally or in small quantities), the traditional process, with a LIM4 value of 115.9, position end products out of the frame of the SAIN, LIM4 system. They are thus assumed to be not recommendable even in small quantities. Therefore the consideration of an additional disqualifying nutrient can have important consequences on the final nutritional scores and on the classification of food items. These results must be moderated since end products are necessarily desalted before consumption for sensory needs. This operation at the domestic scale carried out by immersing products in

Fig. 6. Trajectories of pork belly in the SAIN, LIM frame during the processes presented in Fig. 3 and domestic preparation.

619

hot water lead to a final salt content of 4%. This operation is assumed to trigger off a BaP leak from the product to the water in the same proportion of the salt leak. From this hypothesis, the nutritional score SAIN and LIM4 were calculated and presented in Table 4 and Fig. 6. The LIM4 after domestic desalting was actually found to decrease for all processes. Particularly, for traditional process, the LIM value decreased from 115.9 to 84.9 bringing back the end products to the nutrient profile class 3. Therefore, the preparation of food products at the domestic scale can have a major impact of the nutritional score and can in this case decrease the toxicity. 4. Discussion The results of this work show the impacts of mass transfers and reactions involved in different preservation processes on the nutritional quality of products assessed with the SAIN, LIM model proposed by Darmon and Darmon (2008). In the case of apple or pork meat, the two scores have proven to be easy to manage and to well discriminate the different operations or processes. The resulting nutritional trajectories in the two-dimension frame allowed a visual and convenient analysis of the marked impact of the processes on the nutritional quality of foods. This plot is an interesting way to identify the main nutritional critical points of an operation or a process till the domestic use and may help to propose improvements such as a change of operation, process, and device but also to better communicate for an optimal use of the product at home. On a quantitative aspect, the use of SAIN and LIM scores suppose the knowledge of mass transfers but also nutrients evolutions during processing. Regarding the number of couple product/process in food industry, this supposes a tremendous data base. Though, realistic hypothesis can be done as regard to existing literature. Besides, the most common unitary operations of food processing are well described and robust models of heat, mass transfers and reactions are available in literature. The results of this paper are based on experimental data of mass transfers during air drying and osmotic treatment and on hypothesis about the reactions or diffusions associated. They showed that providing conditions used are favourable to nutrient retention, dehydration has a poor impact because of the simultaneous concentration of nutrients and energy. However, the concentration of penalising nutrient is taken into account in the LIM score because of the weight base used in its calculation. Therefore, a dehydrated food will be placed less favourably than the initial raw product in the SAIN, LIM system, which seems justified. Therefore, the two calculation bases of SAIN, LIM scores are complementary. Osmotic treatment has more important influence especially on the LIM score because the solutes used in this operation usually belong to the group of disqualifying nutrients: i.e., salts and sugars. Therefore, a sweetening operation can be very penalising which seems justified. One thing to question in the comparison of airdrying and osmotic dehydration is the discrimination between initial and added sugars. Indeed, dried apples are very sweet but have a low LIM because sugars come from the concentration of the initial ones. On the opposite, semi-candied apple cubes which may have a comparable sugar content, have a much more penalising position because the sugars are extrinsic. Therefore, this work of process analysis with the SAIN, LIM scores also highlights limitations in the use of a nutrient profile model to assess a food process. The limitations are important to consider as outputs of nutrient profile models are foods classification which may have consequences on consumer health or on products sells. Firstly, the characterisation of raw product is important and must be reliable as it represents the beginning point of a process. Initial calculation of SAIN and LIM scores showed that for a same product, the food class could change as a func-

620

N. Achir et al. / Journal of Food Engineering 100 (2010) 613–621

tion of the number of nutrients considered. For instance, by taking into account five and 23 nutrients, SAIN scores evolved from 5.8 to 3.6 and from 1.7 to 5.9 for raw pork meat and apple, respectively. Both evolutions had the consequence to differ the initial classification of the product. This fact was already noticed by model profile makers and users (Drewnowski et al., 2009). A higher flexibility of the calculation of the SAIN scores was proposed by taking into account only few base nutrients and by adding optional nutrients, in order to limit the ‘‘diluting” effect of a high number of nutrients while still representing the nutritional advantages of a product (Darmon et al., 2009). However, the appropriate and judicious selection of nutrients can be difficult for a non nutritionist. Other thing to question is the nutrient variability of the raw product as a function of the variety, origin, season etc., which is not considered in nutritional tables. This variability may also have important consequences on the positioning of raw products especially for those who have a high richness in one or two specific micro-nutrients such as vitamins which are not standardized at the industrial scale. Initial variability of a raw material may also have consequences on the final product. For instance, a variation of initial reducing sugars in potatoes or plantains has no influence on its initial positions. However it may have a direct consequence on the final concentration of acrylamide which is a toxic molecule (Taeymans et al., 2004). Actually, other limitation of SAIN LIM scoring of processed food is the absence of the consideration of toxic molecules. If this aspect is not considered, the risk is the possible promotion of products which can have a high content in toxic compounds. Indeed, in the example of pork processing, the resulting products obtained with the traditional process had a lower LIM than those obtained with the two innovative processes despite of the fact they contained more benzo[a]pyrene. Many other potentially harmful molecules can be found in common foods because of a high processing temperature during roasting, frying or smoking etc., Darmon and Darmon (2008) already thought about this problem and proposed for instance the consideration of trans-fatty acids which can be considered as a neoformed compound during oil refining or frying processes. However, to dissociate mass transfers and neoformed compounds during processes which both affect the LIM value, it would better to consider a new score which could represent the toxicological impact of a process. Lastly, food processes do not have only negative nutritional consequences on foods. Indeed, a food process can improve the digestibility or favour the destruction or neutralisation of toxic compounds. These particular improvements can also be difficult to take into account in the SAIN, LIM system.

5. Conclusion The nutrient profile model SAIN, LIM is an interesting tool to assess the nutritional quality of a food process according to the mass transfer involved. This approach gives insight in the main critical points of a process and may help improving technologies to optimize the nutritional quality of processed foods. This work also showed the limitation of such tools to assess the complexity of mass transfers and reactions during a food process and the resulting outputs can be questioned. Therefore, further work is needed to test nutrient profile models and to adapt them to the evaluation of food processes.

References Arambepola, C., Scarborough, P., Rayner, M., 2008. Validating a nutrient model. Public Health Nutrition 11, 371–378.

Bedi, R., Sadana, B., Verma, S., 2006. Effect of cooking methods on the mineral availability from commonly consumed vegetables. Journal of Food Science and Technology 43, 512–515. Bruneau, D.M., Sebastian, P., Lecompte, J.Y., Collignan, A., Rochery, V., 2005. Sizing of an innovative and improved meat smoking system. International Journal of Food Engineering 1. Darmon, N., Darmon, M., 2008. L’équilibre nutritionnel. Concepts de base et nouveaux indicateurs : le SAIN et le LIM. Lavoisier, Paris, France. Darmon, N., Vieux, F., Maillot, M., Volatier, J.-L., Martin, A., 2009. Nutrient profiles discriminate foods according to their contribution to nutritionally adequate diets: a validation study using linear programming and the SAIN, LIM system. American Journal of Clinical Nutrition 89, 1–10. Dermesonlouogloua, E.K., Giannakouroua, M.C., Taoukis, P., 2007. Stability of dehydro-frozen tomatoes pretreated with alternative osmotic solutes. Journal of Food Engineering 78, 272–280. Drewnowski, A., Maillot, M., Darmon, N., 2009. Testing nutrient profile models in relation to energy density and energy cost. European Journal of Clinical Nutrition 63, 674–683. European Commission, 2006. Regulation No 1924/2006 of the European Parliament and of the Council on Nutrition and Health Claims Made on Foods. Favier, J.C., Ireland-Ripert, J., Toque, C., Feinberg, M., 1995. Répertoire général des aliments. Tec & Doc Lavoisier, Paris, France. Lee, E., 2005. World Health Organization’s global strategy on diet, physical activity, and health: turning strategy into action. Food and Drug 60, 569. Leskova, E., Kubıkova, J., Kovacikova, E., Kosicka, M., Porubska, J., Holcıkova, K., 2006. Vitamin losses: retention during heat treatment and continual changes expressed by mathematical models. Journal of Food Composition and Analysis 19, 252–276. Marfil, P.H.M., Santos, E.M., Telis, V.R.N., 2008. Ascorbic acid degradation kinetics in tomatoes at different drying conditions. LWT - Food Science and Technology 41, 1642–1647. Marty, C., Berset, C., 1986. Degradation products of trans-b-carotene during heating in sealed glass tubes and extrusion cooking. Journal of Food Science 51, 698–702. Mattila, P., Ronkainen, R., Lehikoinen, K., 1999. Effect of household cooking on the vitamin D content in fish, eggs, and wild mushrooms. Journal of Food Composition and Analysis 12, 153–160. Moreira, R.G., Sun, X., Chen, Y., 1997. Factors affecting oil uptake in tortilla chips in deep-fat frying. Journal of Food Engineering 31, 485–498. Morgan, J.B., Calkins, C.R., Mandigo, R.W., 1988. Effect of trim level, cooking method, and chop type on lipid retention, caloric content, and cholesterol level in cooked pork. Journal of Food Science 53, 1602–1604. OFSP, Office fédéral de la santé publique Unité de direction protection des consommateurs. 2008. Fiche d’information, les hydrocarbures aromatiques polycyclique (HAP), . Oliveira, F.A.R., Lino, I.B.M., Oliveira, J.C., 1999. Chapter one: process optimization and minimal processing of foods: an international collaborative approach in the framework of the EU Copernicus program. In: Oliveira, F.A.R., Oliveira, J.C. (Eds.), Processing Foods: Quality Optimization and Process Assessment. CRC Press, Boca Raton, pp. 1–11. Osorio, C., Franco, M.S., Castaño, M.P., González-Miret, M.L., Heredia, F.J., Morales, A.L., 2007. Colour and flavour changes during osmotic dehydration of fruits. Innovative Food Science and Emerging Technology 8, 353–359. Pedreshi, F., 2005. Acrylamide occurence and oil uptake in potato chips. IFT Annual Meeting, New Orleans, Louisiana, 15–20 July. Peiro, R., Dias, V.M.C., Camacho, M.M., Martinez-Navarrete, N., 2006. Micronutrient flow to the osmotic solution during grapefruit osmotic dehydration. Journal of Food Engineering 74, 299–307. Poligné, I., Collignan, A., Trystram, G., 2001a. Characterization of traditional processing of pork meat into boucané. Meat Science 59, 377–389. Poligné, I., Collignan, A., Trystram, G., Pieribattesti, J.P., 2001b. Traditional techniques for processing boucané, a salted/dried/smoked meat product from Réunion. Tropical Science 41, 90–94. Poligné, I., Collignan, A., Trystram, G., 2005. Processing smoked pork belly by immersion in a complex solution at high temperature. Journal of Food Engineering 66, 155–169. Rastogi, N.K., Raghavarao, K.S.M.S., Niranjan, K., 2005. Developments in osmotic dehydration. In: Sun, D.W. (Ed.), Emerging Technologies for Food Processing. Elsevier, London, UK, pp. 221–243. Scarborough, P., Rayner, M., Stockley, L., 2006. Developing nutrient profile models: a systematic approach. Public Health Nutrition 10, 330–336. Scarborough, P., Boxer, A., Rayner, M., Stockley, L., 2007. Testing nutrient profile models using data from a surbey of nutrition professionals. Public Health Nutrition 10, 337–345. Taeymans, D., Wood, J., Ashby, P., Blank, I., Studer, A., Stadler, R.H., Gondé, P., Van Eijck, P., Lalljie, S., Lingnert, H., Lindblom, M., Matissek, R., Muller, D., Tallmadge, D., O’Brien, J., Thompson, S., Silvani, D., Whitmomre, T., 2004. A review of acrylamide: an industry perspective on research, analysis, formation, and control. Critical Review in Food Science and Nutrition 44, 323–347. Thémelin, A., 1994. Optimisation des procédés combinés de déshydratationimprégnation par immersion dans des solutions concentrées et de séchage par entrainement à l’air chaud. Thesis, école nationale supérieure des industries agricoles et alimentaires, p. 147 Timoumia, S., Mihoubia, D., Zagroubab, F., 2007. Shrinkage, vitamin C degradation and aroma losses during infra-red drying of apple slices. LWT - Food Science and Technology 40, 1648–1654.

N. Achir et al. / Journal of Food Engineering 100 (2010) 613–621 USDA, United States Department of Agriculture, 2009. National Nutrient Database for Standard Reference. van Boekel, M.A.J.S., 2008. Kinetic Modeling of Reactions in Foods. CRC Press. Vega-Galvez, A., Miranda, M., Bilbao-Sainz, C., Uribe, E., Lemus-Mondaca, R., 2008. Empirical modeling of drying process for apple (CV Granny Smith) slice at different air temperatures. Journal of Food Processing and Preservation 32, 972– 986.

621

Visioli, F., Poli, A., Peracino, A., Luzi, L., Cannella, C., Paoletti, R., 2007. Assessment of nutritional profiles: a novel system based on a comprehensive approach. British Journal of Nutrition 98, 1101–1107. WHO, 2003. Diet, Nutrition and the Prevention of Chronic Diseases. Report of a Joint WHO/FAO Expert Consultation, Geneva. WHO, 2002. The World Health Report 2002. Reducing Risks, Promoting Healthy Life, Geneva.