Risk analysis of the thermal sterilization process.

International Journal of Food Microbiology 47 (1999) 51–57

Risk analysis of the thermal sterilization process. Analysis of factors affecting the thermal resistance of microorganisms Stepan G. Akterian a , Pablo S. Fernandez b , Marc E. Hendrickx c , Paul P. Tobback c , b d, Paula M. Periago , Antonio Martinez * a

Higher Institute for Food Industry, Food Engineering Deptartment, 26 Maritza Blvd., BG-4002 Plovdiv, Bulgaria b ´ , E.P.S.O., Ctra. Beniel km 3.2, 03312 Orihuela, Spain Universidad Miguel Hernandez c Katholieke Universiteit-Leuven, Food Technology Laboratory, Kardinaal Mercierlaan 92, B-3001 Heverlee, Belgium d ´ ´ de Alimentos, Food Preservation, Apdo Correos 73, 46100 Burjassot, Valencia, Spain Instituto de Agroquımica y Tecnologıa Received 20 July 1998; received in revised form 6 November 1998; accepted 30 January 1999

Abstract A risk analysis was applied to experimental heat resistance data. This analysis is an approach for processing experimental thermo-bacteriological data in order to study the variability of D and z values of target microorganisms depending on the deviations range of environmental factors, to determine the critical factors and to specify their critical tolerance. This analysis is based on sets of sensitivity functions applied to a specific case of experimental data related to the thermoresistance of Clostridium sporogenes and Bacillus stearothermophilus spores. The effect of the following factors was analyzed: the type of target microorganism; nature of the heating substrate; pH, temperature; type of acid employed and NaCl concentration. The type of target microorganism to be inactivated, the nature of the substrate (reference or real food) and the heating temperature were identified as critical factors, determining about 90% of the alteration of the microbiological risk. The effect of the type of acid used for the acidification of products and the concentration of NaCl can be assumed to be negligible factors for the purposes of engineering calculations. The critical non-uniformity in temperature during thermobacteriological studies was set as 0.5% and the critical tolerances of pH value and NaCl concentration were 5%. These results are related to a specific case study, for that reason their direct generalization is not correct.  1999 Elsevier Science B.V. All rights reserved. Keywords: Critical factors; Critical tolerances; HACCP; Ranked order of factors; Variability of D and z values

1. Introduction Thermal sterilization is one of the most widely used industrial procedures for the microbiological stabilization of foods. Usually, food processes are *Corresponding author. Tel.: 134-6-3900022; fax: 134-63636301. E-mail address: [email protected] (A. Martinez)

designed to meet a pre-established probability of a non-sterile unit (PNSU). The level required varies with the target microorganism: for Clostridium botulinum spores a PNSU of 10 29 is considered to be necessary, whereas for other mesophilic spores the PNSU is 10 26 , while for thermophilic spores the PNSU varies between 10 22 and 10 26 (Dignan et al., 1989). The microbiological risk alteration, varies for a given sterilization process (time-temperature

0168-1605 / 99 / $ – see front matter  1999 Elsevier Science B.V. All rights reserved. PII: S0168-1605( 99 )00005-7

52

S.G. Akterian et al. / International Journal of Food Microbiology 47 (1999) 51 – 57

combination) as a function of the initial quantity of the microorganism in question and the thermal resistance parameter D value, which is directly related to certain environmental factors, such as pH level, NaCl concentration, and type of substrate ´ (reference, food) (Fernandez et al., 1994, Periago et al., 1998). Consequently, variations in these factors could originate alterations in the microbiological risk. It is generally accepted that the thermal process is a Critical Point that has to be controlled in order to guarantee the microbiological safety of processed foods. The Hazard Analysis Critical Control Points (HACCP) system is the main technique recognized worldwide for optimizing the safety of sterilized foods. A general problem in the development of an HACCP system is the identification of critical factors, establishment of critical levels and specification of the critical tolerance of the factors. The safety of a sterilization process can be evaluated according to the lethality achieved and the microbiological risk alteration (Akterian et al., 1997) of the target microorganisms that survive the thermal treatment. Akterian (1996) used sensitivity functions to study and control the sterilization process in canned foods. Simple approaches for controlling the sterilization process were proposed in the cases of fluctuations, systematic deviations and static errors of process variables. The sensitivity function method is adaptable and universal, and could also be applied for studying and controlling other complex food processes. In the present work, sensitivity functions were developed and applied to experimental kinetic data to: (1) rank the factors (pH, NaCl, substrate, target microorganism, etc.) that affected the D and z value according to their importance, (2) evaluate the risk alteration after the sterilization process according to deviations in these factors, and (3) identify factors and critical tolerances in the framework of an HACCP system for thermal sterilization processes.

values) (Rodrigo et al., 1993; Fernandez et al., 1994; Ocio et al., 1994; Periago et al., 1998) on the following factors: • Target microorganisms (M): Clostridium sporogenes (CS) and Bacillus stearothermophilus (BS) • Nature of the heating substrate in which the microorganisms were heated (S): either mushroom extract (ME) with a natural pH of 6.65, or reference medium (RM): phosphate buffer at pH 7.3 (for Clostridium sporogenes) or bidistilled water, pH 7 (for Bacillus stearothermophilus) • pH value of the heating substrate (pH): pH 4.65, pH 5.34, pH 6.22 • Type of acid employed (A): citric acid (CA) and glucono-d-lactone acid (GDL) • NaCl concentration (C): (from 0.5 to 3.5% w / v) • Temperature of the heating medium (T): in the range of 115–1258C for BS and 110–1408C for CS. The ranges for these factors include values that are commonly used for the food industry These experimental data were obtained by using two thermobacteriological methods: (a) capillarytubes (CT) as described by Stern and Proctor (1954), (Fernandez et al., 1994; Ocio et al., 1994; Periago et al., 1998), applied to the temperature range 110– 1258C, and (b) thermo-resistometer (TR) (Rodrigo et al., 1993) for the temperature range 121–1408C.

2.2. Identification of critical levels The selection of critical factors and the specification of critical tolerances of the factors are manually linked in accordance with the procedure of Akterian et al. (1997). In the first step of this procedure, the factors studied were classified in terms of the following features: A. Measurement (the capability of the factors to be measured)

2. Materials and methods

2.1. Microbiological data An analysis was made of the dependence of experimental thermobacteriological data (D and z

1. Measurable factors: temperature, pH value and NaCl concentration. 2. Non-measurable factors: the type of microorganism, the nature of the heating substrate and the type of acid employed.

S.G. Akterian et al. / International Journal of Food Microbiology 47 (1999) 51 – 57

Controllability

pH ? ≠z C ? ≠z F zpH 5 ]] 5 b 3 ? pH; F zC 5 ]] 5 b 4 ? C z ? ≠pH z ? ≠C

1. Controllable and on-line measurable factors: temperature. 2. Non-controlled but easily measurable: pH value and NaCl concentration. The second step in the procedure is related to the identification of critical factors as being the factors which have the highest partial effects (JP , Eq. 4.2) and whose deviations cause about 90% of the integrated risk alteration. The third step involves identification of the initial critical tolerances in accordance with the following criteria: 1. The critical tolerance for the critical controllable factors was assigned to be equal to the minimum deviation’s level of these factors, and the critical tolerance for the non-controllable factors was assigned to be equal to the average deviation’s levels of those factors. 2. For non-critical non-controllable factors, the critical tolerance was taken to be the average deviation level of those factors.

2.3. Risk analysis The risk analysis was based on the sets of sensitivity functions and the total differential. The sensitivity function FP 5 (≠F /F ) /(≠P/P) shows the change of an objective function (F, D or z value) caused by a unit of variation in a process factor P. The effect of temperature on D values was evaluated by Eq. (1), obtained by differentiation of the analytical relationship between D and z values T ? ≠D T F DT 5 ]] 5 2 ln(10)] D ? ≠T z

(1)

The effect of the measurable factors, pH value (pH) and NaCl concentration (C) were estimated by the corresponding sensitivity functions of D and z values:

F

D pH

53

pH ? ≠D C ? ≠D 5 ]]] 5 b 1 ? pH; F DC 5 ]] 5 b 2 ? C; D ? ≠pH D ? ≠C (2.1)

(2.2) where a 1 , a 2 , a 3 , a 4 , b 1 , b 2 , b 3 , b 4 are empirical coefficients of the following correlation equations fitted to the experimental data: D5a 1 ?exp(b 1 ? pH ); D5a 2 ?exp(b 2 ?C); z5a 3 ?exp(b 3 ? pH ); z5a 4 ? exp(b 4 ?C). The effect of the non-measurable factors, the type of microorganism (M), the nature of the heating substrate (S) and the type of acid used (A), were assessed by means of the following sensitivity functions: DBS 2 DCS DME 2 DRM F DM 5 ]]];F DS 5 ]]]]; DCS DRM DGDL 2 DCA F AD 5 ]]]] DCA z BS 2 z CS z ME 2 z RM F zM 5 ]]]; F zS 5 ]]]; z CS z RM z 2 z GDL CA F Az 5 ]]] z CA

(3.1)

(3.2)

where the superscripts are related to the D and z values and their meaning is: BS5B. stearothermophilus, CS5C. sporogenes, ME5mushroom extract, RM5reference medium, GDL5glucono-d-lactone, CA5citric acid.

2.4. Analysis of the variability of D and z values In the literature on thermobacteriology (Hersom and Hulland, 1980, Stumbo, 1973), D and z values are assumed to be constants for a particular microorganism and a fixed temperature, and for a particular microorganism, respectively. However, D and z values are affected by some environmental factors, such as pH, NaCl concentration and type of substrate (reference or food), (Fernandez et al. 1994, Periago et al., 1998), and these values vary in intervals (DP/P) var . The analysis of the effect of factors was carried out by using the following weight indexes: (1) The absolute effect of the variation (DP/P) var I DP 5 uF DP u aver ? (DP/P) var ; I PZ 5 uF PZ u aver ? (DP/P) var (4.1) The absolute effect IP represents the absolute value

S.G. Akterian et al. / International Journal of Food Microbiology 47 (1999) 51 – 57

54

of the change DD/D and Dz /z, respectively, when a deviation DP/P (for instance DpH /pH ) occurs (2) The partial effect of the variation (DP/P) var I DPi J 5 ]] ; n sI PjD d D Pi

O

j51

Z I Pi J 5 ]] n sI ZPjd Z Pi

(4.2)

O

The partial effect JP represents the relative value of the change change DD/D and Dz /z, respectively, caused by the deviation DP/P of a single factor P (as pH). The relative value JP is referred to the total change of DD/D caused by all factors studied. (3) The integrated effect of the variation (DP/P) var K DPs 5

OsJ d; D Pi

i 51

OsJ d s

K ZPs 5

Z Pi

(4.3)

i 51

The integrated effect KP represents a relative value of the change DD/D and Dz /z, respectively, caused by the deviation DP/P of several (s number) factors P. These factors (of s number) are selected among all factors studied having the highest partial effect. The physical interpretation of KP index is an accumulated relative effect of the first (s number) of factors which have the highest partial effect. The equation subscripts s and n are, respectively, the index related to the actual studied factor P and the total number of factors studied and aver subscript is the index related to the average value of the corresponding sensitivity function. The change in D and z values as a function of the deviations in the factors studied was evaluated by means of the following relationships derived on the basis from the total differential set:

U

U U U

U

DD DT DpH ] ¯ F DT ? ] 1 F DpH ? ]] D T pH DC 1 F DC ? ] 1uF DM ? 1u 1uF DS ? 1u 1uF DA ? 1u C (5)

U

U

U U

U

Dz DpH DC ] ¯ F zpH ? ]] 1 F zC ? ] 1uF zM ? 1u z pH C 1uF zS ? 1u 1uF zA ? 1u

UU

U

j 51

s

type of substrate and the type of target microorganism) intervene. This risk alteration Db /b was evaluated by the following relationship (Akterian et al., 1997) depending on the deviations in D and z values:

(6)

The risk alteration Db /b is a change of the risk b for the survival of a target microorganism after a sterilization process, when deviations of process characteristics (as DT and z values) and process factors (as temperature, pH, NaCl concentration, the

FD DD (FO /FT ) Db Dz ]¯ ] ? ] 1 ]]]Fz ? ] b Fb D Fb z

U U SU U

Dz U] U z Dz U] UD z

DD 5 27.8 ? ] 1 34.7 ? D DD 5 27.8 ] 1 1.25 ? D

U (7)

where b is the final concentration of a target microorganism after a sterilization process being used to quantify for an evaluation of the microbiological risk of consumer and FD 51, Fz 5041.548.5 and Fb 5 20.036 are sensitivity functions which show the change in F value caused by a unit of variation in one of the parameters D, z or b, respectively. The average value (underlined) of Fz was used in the next calculations (Akterian et al., 1997). The ‘worst case’ was taken into account by means of (i) the relation between the target (T) and obtained (O) sterilization values (F ); FT /FO 51.2; (ii) the absolute value of each term in Eq. 7.

3. Results The variation interval of the functions of sensitivity and the results related to the variability in the value of the thermal resistance parameters D and z can be seen in Tables 1 and 2. Figs. 1 and 2 give examples that show the effect of some of the factors analysed on the D value. The main factors that affect the thermal resistance parameters (D and z) were ranked according to the average absolute value of their functions of sensitivity. Temperature, pH and target microorganism were found as the most affecting factors, which explained about 95% of the total variation in D value (Table 1), and type of target microorganism, pH and substrate, explaining about 94% of the total variability in the z value (Table 2). The results related to the factors and critical tolerances are shown in Table 3. As can be seen, the main critical factors that affect the risk alteration (Db /b) are the target microorganism, the nature of

S.G. Akterian et al. / International Journal of Food Microbiology 47 (1999) 51 – 57

55

Table 1 Variability of D value depending on factor variation range s

Factor P

F DP min4average4max

(DP/P) var

1 2 3 4 5 6

Temperature (1104140) pH value (4.547) Type of microorganism Nature of heating substrate NaCl concentration (043.5%) Type of acid employed

2544 2354 223 20.442.346.6 20.3840.846.5 20.574 20.424 20.29 21.94 20.840 20.2740.0140.29

0.27 0.50 1 1 0.035 1

Total

–

–

I DP aver 9.45 1.15 0.80 0.42 0.03 0.01 11.86

J DP ,% aver

K DP , % aver

79.7 9.7 6.7 3.5 0.3 0.1

79.7 89.4 96.1 99.6 99.9 100

100

–

s5Order of the factor by its influence. F DP 5Sensitivity function for D. (DP/P) var 5Factor variation. I PD 5Absolute effect of the factor variation. J DP 5Partial effect of the factor variation. K PD 5Integrated effect of the factor variation.

Table 2 Variability of z value depending on factor variation range s

Factor P

F ZP min4average4max

(DP/P) var

I ZP aver

J ZP , % aver

K ZP , % aver

1 2 3 4 5

Type of microorganism pH value (4.547) Nature of heating substrate Type of acid employed NaCl concentration (043.5%)

20.46420.324 20.2 ]] 20.7420.240.6 ]] 0.0340.0840.12 ] 20.07420.0340.04 ]] 20.08420.0440 ]]

1 0.50 1 1 0.035

0.32 0.10 0.08 0.03 0.001

60.3 18.8 15.1 5.6 0.2

60.3 79.1 94.2 99.8 100

Total

–

–

0.531

100

–

s5Order of the factor by its influence. F ZP 5Sensitivity function for z. (DP/P) var 5Factor variation. I PD 5Absolute effect of the factor variation. J DP 5Partial effect of the factor variation. K PD 5Integrated effect of the factor variation.

the heating substrate, temperature and pH, explaining almost 96% of the total variability in the risk alteration. The critical tolerances considered for the critical and non-critical measurable factors (temperature, pH and NaCl concentration) as indicated in Section 2 would be 0.5% for the deviations in the heating temperature and 5% for the variation in pH and NaCl (Table 3). These suggested critical tolerances gener-

ate about 16.3% of the risk alteration caused by physical and chemical factors affecting D and z values.

4. Discussion The results of the present analysis of the experimental data by means of the sensitivity functions

S.G. Akterian et al. / International Journal of Food Microbiology 47 (1999) 51 – 57

56

Fig. 1. Effect of pH value and glucona-d-lactone on D value for Bacillus stearothermophilus.

Fig. 2. Effect of NaCl concentration on D value for Bacillus stearothermophilus.

´ coincide with those described by Fernandez et al. (1994), Ocio et al. (1994), Rodrigo et al. (1993) and Periago et al. (1998). Although it is well-known that temperature is the main factor affecting the D value of microorganisms, the results obtained in this work indicate that its effect is ten times higher than the other factors studied. Sodium chloride and pH also affect the thermal resistance of bacterial spores, although in the case of pH the effect seems to depend on the target microorganism. The nature of the substrate (reference or real food) also influences the D value. However, the type of acid used seems to produce no effect, and from the thermobacteriological point of view either of the two acids, citric or glucono-d-lactone, could be used. The specified critical tolerance of 0.5% for the non-uniformity in the heating temperature of samples, is a high requirement which can be only met by means of new thermobacteriological methods using volumetric (non-surface) heating of the studied sample and consequently quick cooling. Even this restrictive critical tolerance of 0.5% causes an average error in the D value of 17% calculated by Eq. 5. The specified critical tolerance of 5% for the other measurable and non-critical factors (pH and NaCl concentration) is reasonable for the practice in food technology. These tolerances can cause a negligible risk alteration Db /b of 4.7, calculated by Eqs. 5 to 7, which is about 3.5% of the total risk alteration for the sterilization process due to deviations of all parameters of the process (Akterian et al. 1997).

Table 3 The risk alteration Db /b depending on deviations in the factors a s

Factor P

uF DP u aver

uF ZP u aver

(DP/P) min4aver4max

I bP aver

J bP , % aver

K bP , % aver

1 2 3 4 5 6

Type of microorganism Nature of heating substrate Temperature pH value Type of acid employed NaCl concentration

0.8 0.42 35 2.3 0.01 0.8

0.32 0.08 – 0.2 0.03 0.04

1 1 0.54142% ] 245410% 1 245410 %

33.36 14.46 4.86 3.54 1.32 1.18

56.8 24.6 8.3 6.0 2.3 2.0

56.8 81.4 89.7 95.7 98.0 100

Total

–

–

–

58.72

100

–

a

The values of initial critical tolerances for critical factors are in bold. s5Order of the factor by its influence. F DP 5Sensitivity function for D. F ZP 5Sensitivity function for z. (DP/P) var 5Factor variation. I PD 5Absolute effect of the factor variation. J DP 5Partial effect of the factor variation. K PD 5Integrated effect of the factor variation.

S.G. Akterian et al. / International Journal of Food Microbiology 47 (1999) 51 – 57

The proposed approach based on the sets of sensitivity functions and total differentials is an objective tool for the development of HACCP systems and in particular the identification of critical factors and the specification of critical tolerances for measurable factors (pH, NaCl concentration, temperature). On the basis of the experimental heat resistance data analyzed, the type of target microorganism to be inactivated, the nature of the substrate and the heating temperature were identified as critical factors. The results presented were obtained on the basis of a specific case, and one must assume that each product formulation would be studied separately. However, these results depict a general tendency of the problem, and indicate that the development of the methodology used in this work could help to implement the sterilization process as a critical control point in the canning industry, contributing to an increase in the safety of processed foods. It could also be used to obtain an estimate of the risk alteration when important factors, such as pH, deviate from their standard value for a particular product. This could make it easier to take appropriate decisions about a batch of a processed product that has suffered a certain deviation in one of the factors.

Acknowledgements This research was undertaken in the framework of the post-doctoral research of Dr. Akterian sponsored by the Belgium Office for Scientific, Technical and Cultural Affairs.

57

References Akterian, S.G., 1996. Studying and controlling thermal sterilization of convection-heated canned foods using functions of sensitivity. J. Food Engin. 29, 329–338. Akterian, S., Smout, C., Tobback, P. and Hendrickx, M. (1997) Identification of CCP-limit levels for HACCP systems of thermal sterilization process using sensitivity functions. Proceedings of the conference, ‘Modeling of thermal properties and behavior of foods during production, storage and distribution’, pp. 6–12, 23–25 June 1997, Prague, (Czech Republic). Dignan, D.M., Berry, M.R., Pflug, I.J., Gardine, T.D., 1989. Safety considerations in establishing aseptic processes for low-acid foods containing particles. Food Technol. 3 (1989), 118–121. ´ Fernandez, P.S., Ocio, M.J., Sanchez, T., Martınez, A., 1994. Thermal resistance of Bacillus stearothermophilus spores heated in acidified mushroom extract. J. Food Prot. 57, 37–41. Hersom, A.C., Hulland, E.D., 1980. Canned Foods. Thermal Processing and Microbiology, 7th ed., Churchill Livingstone, Edinburgh. ´ ´ Ocio, M.J., Sanchez, T., Fernandez, P.S., Rodrigo, M., Martınez, A., 1994. Thermal resistance characteristics of PA 3679 in the temperature range 110–1218C as affected by pH, type of acidulant and substrate. Int. J. Food Microbiol. 22, 239–247. ´ ´ Periago, P.M., Fernandez, P.S., Ocio, M.J., Martınez, A., 1998. A predictive model to describe sensitization of heat treated Bacillus stearothermophilus spores to NaCl. Z. Lebensm. Unters. Forsch. A 206, 58–62. Rodrigo, M., Martinez, A., Sanchez, T., Peris, M.J., Safon, J., 1993. Kinetics of Clostridium sporogenes PA3679 spore destruction using computer-controlled thermo-resistometer. J. Food Sci. 58, 649–652. Stern, J.A., Proctor, B.E., 1954. A micro-method and apparatus for the multiple determination of rates of destruction of bacteria and bacterial spores subjected to heat. Food Technol. 8, 139– 143. Stumbo, C.R., 1973. Producing, harvesting and cleaning sores. In: Thermobacteriology in food processing. 2nd ed., pp. 61–68. Academic Press, New York, USA.