Previous Page Table of Contents Next Page


APPENDIX C

RISK ASSESSMENT


The development of “what-if” scenarios using a simple risk assessment model that compares the relative risk reduction potential of control measures for Enterobacter sakazakii and Salmonella enterica in powdered infant formula.

A simplified risk assessment was developed to consider the relative impact of different potential control measures on the microbiological safety of powdered infant formula. The purpose of the risk assessment was to demonstrate a number of concepts that need to be taken into account when quantitatively considering the relative efficacy of different risk reduction strategies. The scope of the risk assessment was restricted to the hospital environment, but the general principles considered should also be pertinent to the home environment. Figure A1 depicts a flow chart illustrating the scope of the powdered formula use encompassed in the risk assessment. The risk assessment was restricted to consideration of two of the identified microbiological hazards, Enterobacter sakazakii and Salmonella enterica, with a particular focus on the former. The risk estimates described in the risk assessments are based on the expected increase and decrease in risk as measured against a baseline level. That baseline was based on the following assumptions:

· The relative contribution of inherent contamination vs environmental sources to the presence of E. sakazakii in reconstituted powdered infant formula is 4:1 (i.e. 80% inherent contamination).

- The probability of powdered infant formula being contaminated is 0.025.

- The probability of infant formula being contaminated as result of environmental exposure during rehydration and preparation is 0.00625.

· The “time to consumption” from preparation to complete consumption is 4 hours.

· The temperature of the reconstituted infant formula during feeding is 25°C.

Figure A1. A flow chart depicting the scope of the powdered formula use encompassed in the risk assessment.

PIF, Powdered Infant Formula; PE, Contamination from the preparation environment; PC, Contamination from the infant formula; TTC, Time to consumption

The basic calculations employed in the risk assessment were to assume that E. sakazakii had an approximate 2-hour lag phase prior to initiating growth and that the powdered infant formula was rehydrated just prior to use (i.e. refrigeration and storage of the reconstituted product after preparation but before use). By then either manipulating the input data underlying the risk assessment or including additional models of potential treatments, a variety of “what-if” scenarios were developed, described below.

ENTEROBACTER SAKAZAKII

Scenario 1

The purpose of this scenario was to determine what could be achieved as a result of decreasing the frequency/extent of contamination of powdered infant formula with E. sakazakii. This decrease could be the result of any of a series of control measures performed at the manufacturing level prior to the sealing of the final product package. Examples of potential interventions could include enhancing plant sanitation, the successful implementation of microbiological criteria, or the adoption of some intervention technology. Five different contamination rate levels in intact packages were assumed, 0.025 (baseline), 0.0100, 0.0010, 0.0001 and 0.0000 (assumed level if able to treat the product after it was packaged). The calculated risk reductions are depicted in Table A1. The limiting factor regardin the degree of risk reduction that can be achieved by this approach is determined by the percent of the total servings of infant formula that become contaminated due to environmental contamination instead of contamination initially present in the powdered infant formula. Thus, the greatest reduction achieved would be limited by the assumed 20% that were environmentally contaminated.

Table A1. Effect of reducing the frequency/extent of contamination of powdered infant formula on the relative risk of Enterobacter sakazakii contamination.

Frequency of contamination of
powdered infant formula (PC)

Relative
risk

0.0250

1.00 (baseline)

0.0100

-1.92-fold

0.0010

-4.29-fold

0.0001

-4.90-fold

0.0000

-4.99-fold

Scenario 2

The purpose of this scenario was to determine the impact that holding the product at room temperature until it is consumed has on the relative risk compared to an assumed norm of a 4-hour “time to consumption” (TTC). The results for TTC values of 0, 2, 4, 6, 8 and 10 hours were compared (Table A2). An additional scenario was examined, with 1% of the feeds at 10-hour TTC, and the remaining 99% of feeds at 4-hour TTC. It is apparent that the relative risk at the longer TTC values increases exponentially, reflecting the exponential growth of the pathogen. It is also important to note that the relative risk at 0 and 2 hours is the same because the assumed lag of 2 hours would have resulted in the microorganism not having grown in these samples. It is apparent that extended holding times can greatly increase the relative risk if E. sakazakii is present.

Table A2. Effect of “time to consumption” duration between rehydration and the completion of feeding on the relative risk of associated with Enterobacter sakazakii.

Time to consumption (h)

Relative risk

0

-30-fold

2

-30-fold

4

1.00 (baseline)

6

+30-fold

8

+1 000-fold

10

+30 000-fold

10*

+300-fold

Note: * If only 1% of the servings were held for 10 hours.

Scenario 3

In this scenario, the effect of decreasing the rate of contamination due to environmental sources due to enhanced hygiene in the hospital environment is considered. The baseline frequency of contamination was assumed to be 0.00625, and the reduced levels considered included 0.00100, 0.00010 and 0.000000. This scenario (Table A3) is very similar to scenario 1, in that the risk reduction achieved will be limited to the percentage of the servings that are contaminated by E. sakazakii present in the formula when the product is unsealed.

Table A3. The effect of decreasing the rate of environmental contamination on the relative risk of E. sakazakii in rehydrated infant formula.

Frequency of contamination of
rehydrated infant formula due
to environmental contamination (PE)

Relative risk

0.00625

1.00 (baseline)

0.00100

-1.20-fold

0.00010

-1.24-fold

0.00000

-1.24-fold

Scenario 4

The purpose of this scenario was to explore the impact of having different susceptibilities among different age populations. In this scenario, a 28-day neonate (newborn up to 4 weeks) was assumed to represent a baseline susceptibility and it was assumed that very low-birth-weight (VLBW, i.e. under 1 500 g) neonates were 10 times more susceptible, 6-month olds were tenfold less susceptible, and 12-month olds were 100 times less susceptible. (These numbers are totally hypothetical at this point and information of relative susceptibilities is a research data requirement.)

Due to the essentially linear nature of the dose-response relationship at these levels of contamination, the increases and decreases in risk are proportional to the increases and decreases in susceptibility (Table A4). This is a highly simplified consideration of the scenario, since there was no attempt to consider the differential in consumption rates at the various ages. However, one additional factor that was considered was the interaction with environmental temperature. An elevated environmental temperature (32ºC) was considered for the VLBW infants. Likewise, the same elevated temperature was considered for the neonates to determine the impact of being in a tropical climate with a higher ambient temperature. It is apparent that the higher environmental temperature increases the relative risk due to the increased growth of E. sakazakii.

Table A4. The hypothetical example of the impact that varying degrees of susceptibility could have on the relative risk associated with the presence of Enterobacter sakazakii in rehydrated infant formula. Also considered was the impact of an elevated environmental temperature (25°C vs 32°C).

Age group

Assumed relative susceptibility

Ambient temperature (°C)

Relative risk

VLBW

+10-fold

25

+10-fold

VLBW

+10-fold

32

+66-fold

Neonate

1.00 (baseline)

25

1.00 (baseline)

Neonate

1.00 (baseline)

32

+6.6-fold

6 months

-10-fold

25

-10-fold

12 months

-100-fold

25

-100-fold

Scenario 5

The effect of subjecting formula to a post-preparation treatment capable of decreasing the level of E. sakazakii by 4-log cycles (99.99%) was evaluated for its impact on relative risk. This type of treatment (e.g. rehydration of powdered infant formula with hot water, the subsequent heating of filled bottles) would decrease contamination due to both initial contamination of the powdered infant formula during manufacture and the environmental recontamination of the formula during preparation. The baseline in this instance would be formula prepared with room temperature water and not subjected to any post-preparation treatment vs formula that was treated (Table A5). As a means of demonstrating the impact of only performing such treatments some of the time, a third situation where only 80% of the servings of formula were subjected to post-preparation was also considered (Table A5). It is apparent that such post-preparation treatments can significantly reduce the relative risk; however, the total risk reduction would be dependent on using such treatments on a consistent basis.

Table A5. Reduction of relative risk associated with Enterobacter sakazakii in rehydrated infant formula if subjected to a post-preparation treatment that effectively reduced the levels of the pathogen by 4-log cycles (99.99%).

Treatment

Relative risk

Not treated

1.00 (baseline)

4-log treatment

-10 000-fold

80% of servings receive a

-5-fold

4-log treatment


Scenario 6

This scenario explores the impact of combining control measures on the relative risk. Three subscenarios were considered:

It is apparent that combining a series of treatments can effectively increase the degree of risk reduction achieved (Table A6). These results should be compared with the results achieved with individual treatments in Tables A1-A3.

Table A6. Effect of combining risk reduction activities on the relative risk of Enterobacter sakazakii associated with powdered infant formula.

Control measures

Relative risk

No control measures

1.00 (baseline)

Reduce PC to 0.001 and PE to 0.001

-16-fold

Reduce PC to 0.001 and TTC to 2 hours

-132-fold

Reduce PC to 0.001, PE to 0.001 and TTC to 2 hours

-480-fold

The basic risk control principles demonstrated in the above scenarios for E. sakazakii would hold true for S. enterica, but the specific risk reductions achieved would vary to some degree based on the mode and sources of Salmonella contamination and its growth and survival characteristics. As a means of demonstrating this, two of the above scenarios were rerun for S. enterica so that the relative risk reductions expected with the microorganisms could be compared. In these scenarios the assumed TTC and environmental temperatures were kept the same, but the growth rate for S. enterica is somewhat slower than E. sakazakii and since the frequency of contamination with salmonellae, both initially in powdered infant formula and environmentally, appears to be substantially lower, the assumed PE and PC were both set at 0.0025.

SALMONELLA ENTERICA

Scenario 1

This is a repetition of E. sakazakii scenario 2 where the impact of TTC was examined. While the same pattern of relative risk reductions are observed for the two microorganisms, the somewhat slower growth rate of S. enterica results in the relative risk reductions being smaller (Table A7).

Table A7. Comparison of the impact of “time to consumption” durations on the relative risks associated with Enterobacter sakazakii and Salmonella enterica in powdered infant formula.

Time to consumption (h)

Relative risk - E. sakazakii

Relative risk - S. enterica

0

-31-fold

-12-fold

2

-31-fold

-12-fold

4

1.00 (baseline)

1.00 (baseline)

6

+31-fold

+12-fold

8

+1 000-fold

+150-fold

10

+30 000-fold

+1 100-fold

Scenario 2

This is a repetition of E. sakazakii scenario 6 that examined the effect of combining risk reduction strategies. In this scenario, the changes in PC, PE and TTC for E. sakazakii were kept the same as in E. sakazakii scenario 6. The changed conditions for S. enterica were a decrease in PC from 0.0025 to 0.0001, a decrease in PE from 0.0025 to 0.0010, and a decrease in TTC to 2 hours. Again, the pattern of relative risk reductions were similar for the two microorganisms, but the specific level of control achieved was dependent on the frequency and mode of contamination and the characteristics of the two microorganisms (Table A8).

Table A8. Effect of combining risk reduction activities on the relative risks of Enterobacter sakazakii and Salmonella enterica associated with powdered infant formula.

Control measures

Relative risk - S. enterica

Relative risk - E. sakazakii

No control measures

1.00 (baseline)

1.00 (baseline)

Reduce PC and PE

-4.5-fold

-16-fold

Reduce PC and TTC

-23-fold

-132-fold

Reduce PC, PE and TTC

-56-fold

-480-fold

BACKGROUND TO ASSUMPTIONS

Estimation of the ratio of Enterobacter sakazakii to other Enterobacteriaceae in neonatal and infant bloodstream infection (BSI) and meningitis:

Best guess

=

0.002

Upper bound

=

0.004

Lower bound

=

0.001

The proportion of E. sakazakii to all Enterobacter spp. was derived from data provided to the consultation by Dr Martin Cole (Australia) (Appendix B).

All age groups:

Reports of Enterobacter infections from blood/CSF
specimens from 1990-2003

E. sakazakii

21

All other Enterobacterspp.

1 676

Estimated proportion

~1%

The proportion of Enterobacter spp. among all Enterobacteriaceae was derived from the publications listed below. It was estimated to be about 20%. The ratio of E. sakazakii to all Enterobacteriaceae is then 1% of 20% or 0.002 (best estimate). The upper and lower estimates are based on expert opinion.

The values below are rough estimates based on data extracted from a number of scientifically heterogeneous publications from several countries:

Country

n*

% Enterobacter/ Enterobacteriaceae

Philippinesa

All infants up to 3 months

17 BSI
8 meningitis

18% in BSI
50% in meningitis

USAb

1st late onset sepsis BSI in VLBW infants

196

17%

USAc

Fulminant late onset neonatal sepsis

13

31%

Cameroond

Neonatal BSI and meningitis

53

21%

Taiwane

Neonatal meningitis

31

16%

Jordanf

Neonatal meningitis

53

19%

Mexicog

Neonatal meningitis

22

9%

* Best estimate of total Enterobacteriaceae from reference.
a Gatchalian, S.R., Quiambao, B.P., Morelos, A.M., Abraham, L., Gepanayao, C.P., Sombrero, L.T., Paladin, J.F., Soriano, V.C., Obach, M., & Sunico, E.S. 1999. Bacterial and viral etiology of serious infections in very young Filipino infants. Pediatric Infectious Disease Journal, 18(10): S50-55.

b Karlowicz, M.G., Buescher, E.S., & Surka, A.E. 2000. Fulminant late-onset sepsis in a neonatal intensive care unit, 1988-1997, and the impact of avoiding empiric vancomycin therapy. Pediatrics 2000, 106: 1387-1390.

c Kago, I., Wouafo Ndayo, M., Tchokoteu, P.F., Koki Ndombo, P., Ekoe, T., Doumbe, P., Tietche, F., & N’Koulou, H. 1991. [Neonatal septicaemia and meningitis caused by gram-negative bacilli in Yaounde: clinical bacteriological and prognostic aspects.] Bulletin de la Societe des Sciences Medicales du Grand-duche de Luxembourg, 84(5 pt 5): 573-581.

d Stoll, B.J., Hansen, N., Fanaroff, A.A., Wright, L.L., Carlo, W.A., Ehrenkranz, R.A., Lemons, J.A., Donovan, E.F., Stark, A.R., Tyson, J.E., Oh, W., Bauer, C.R., Korones, S.B., Shankaran, S., Laptook, A.R., Stevenson, D.K., Papile, L.A., & Poole, W.K. 2002. Late-onset sepsis in very low birth weight neonates: the experience of the NICHD neonatal research network. Pediatrics, 110: 285-291.

e Chang Chien, H.Y., Chiu, N.C., Li, W.C., Huang, F.Y. 2000. Characteristics of neonatal bacterial meningitis in a teaching hospital in Taiwan from 1984-1997. Journal of Microbiology, Immunology and Infection, 33: 100-104.

f Daoud, A.S., Al-Sheyyab, M., Abu-Ekteish, F., Obeidat, A., Ali, A.A., & El-Shanti, H. 1996. Neonatal meningitis in northern Jordan. Journal of Tropical Pediatrics, 42: 267-270.

g Sanchez-Saucedo, L.U., Gonzalez-Yunez, R.A., Avila-Figueroa, C., & Santos, J.I. 1990. Neonatal meningitis: observations on its etiology, mortality and sequelae (Spanish). Boletin medico del Hospital Infantil de Mexico, 47(11): 772-776.

Estimation of the ratio of E. sakazakii susceptibility between high risk groups and other infants

Based on the United States of America 2002 FoodNet laboratory survey for E. sakazakii isolates and United States of America census population data:

It was not attempted to estimate the relative susceptibilities among different susceptible subpopulations: e.g. immunocompromised non-neonate infants including HIV-infected, LBW neonates and immunocompromised neonates.

Estimates of population sizes

Estimates of population sizes based on data from the 2000 United States census:a

Age

Estimated population

28 days or younger

336 130

29 days to less than 1 year

3 697 589

<1

4 033 719

1-4

15 575 428

5-9

19 900 837

10-19

41 512 600

20-29

39 185 524

30-39

42 871 294

40-49

44 303 788

50-59

33 772 415

60-69

21 192 111

³ 70

26 020 990

Total

288 368 706

a Available from the U.S. Census Bureau (www.census.gov).

Estimates of case numbers, population and rates in United States FoodNet sites of Salmonella infections by age group, 2002


£ 28 days

29 days -<1 year

<1 year

1-4 years

5-9 years

10-19 years

Rate

52.0

147.3

139.4

59.29

18.52

10.92

Number

23

717

740

1 210

480

590

Population

44 237.2

486 611.8

530 849

2 040 976

2 592 406

5 403 341



20-29
years

30-39
years

40-49
years

50-59
years

60-69
years

>70
years

Total

Rate

12.66

11.94

10.25

9.81

10.23

11.56

15.99

Number

643

708

618

447

270

365

6 071

Population

5 078 736

5 931 766

6 030 705

4 555 746

2 638 867

3 158 296

37 961 688

FAO/WHO MICROBIOLOGICAL RISK ASSESSMENT SERIES

1 Risk assessments of Salmonella in eggs and broiler chickens: Interpretative Summary, 2002

2 Risk assessments of Salmonella in eggs and broiler chickens, 2002

3 Hazard characterization for pathogens in food and water: Guidelines, 2003

4 Risk assessment of Listeria monocytogenes in ready-to-eat foods: Interpretative Summary, 2004

5 Risk assessment of Listeria monocytogenes in ready-to-eat foods: Technical Report, 2004

6 Enterobacter sakazakii and microorganisms in powdered infant formula, Meeting Report, 2004


Previous Page Top of Page Next Page