ISPAM Stakeholder Panel Meeting Book 3-14-17

Kernel-based gluten binary-like outcomes R. D. Fritz and Y. Chen 3

quantification curve of R-Biopharm R7001 ELISA kit) to obtain the exact gluten contents (Fig. 1 and Table 1). The 965 total servings were acquired to characterise the overall state of affairs of GF oatmeal labelling compliance at the serving size level and to gain insight into differences/biases that exist between laboratories, time frames and producers. In this latter regard, statis- tical comparisons were made via chi-square contin- gency tests between laboratories/time frames (‘7/14’ vs. ‘12/14’), between producers (‘Producer 1’ vs. ‘Producer 2’), between the two time frames for each producer (Producer 1 ‘7/14’ vs. Producer 1 ‘12/14’ and Producer 2 ‘7/14’ vs. Producer 2 ‘12/14’) and between date codes. No ‘statistically significant’ differences in ‘pro- portion of servings found noncompliant’ were found. Consequently, based on this data set, no evidence was found that the noncompliance rate at the serving size level differed between these two laboratories/time frames, producers or date codes. As Table 1 shows, about 95% of the 965 servings were found to have a gluten content below the limit of quantification (BLQ) of the gluten testing kit, R-Bio- pharm 7001 (i.e. < 5 ppm). It is worth mentioning that BLQ results do not equate to true zero results (and as gluten testing methodologies evolve, lower BLQ levels will no doubt be achieved). Continuing, 3.21% were found to be gluten positive but compliant, namely hav- ing a gluten content from 5 to < 20 ppm, and 1.76% were found noncompliant, with a gluten content of ≥ 20 ppm. Additionally, 0.62% of the 965 were found > 80 ppm, which is the quantification limit (ALQ) of the gluten testing kit. This equates to about one in every fifty-seven servings exceeding the FDA defined

Statistical analysis Data analysis was performed by custom macros using Excel, Microsoft Office 2013. Statistical significance was calculated at 95% confidence.

Results and discussion

In-market survey of GF oatmeal The ‘in-market’ assessment of GF labelled oatmeal was conducted from July through December, 2014. Nine hundred and sixty-five servings ( ~ 45 – 50 g per serving) were acquired from store shelves. These were from two well-known US producers and spanned twenty-one different date codes. The number of serv- ings obtained per date code varied from a low of ten up to a maximum of 108 servings. The survey had two parts, divided by time and laboratory. The first was comprised of 329 servings acquired in the summer of 2014. All of these sam- ples were tested at our internal analytical laboratory. This is referred to as the ‘7/14 Assessment’ (Table 1). The second part was comprised of 636 additional servings acquired late 2014. These samples were tested at a recognised external laboratory. This is referred to as the ‘12/14 Assessment’ (Table 1). The focus of our in-market survey was to examine whether the tested samples were in compliance with FDA gluten regulation ( < 20 mg kg –1 ). So, once we knew the gluten content of a sample was above or below 20 ppm, we did not do further analysis. That was the reason that we did not further dilute sample extracts with > 80 ppm readings (upper limit of the

Table 1 Oatmeal survey results

Observed non conforming ( ≥ 20 ppm)

# of Date codes

Qty. of servings tested

5 – 20 ppm

20 – 80 ppm

BLQ ( < 5 ppm)

Positives ( > 5 ppm)

> 80 ppm

Product

7/14 Assessment (Internal Laboratory tested)

Gluten-free oatmeal from Producer #1 Gluten-free oatmeal from Producer #2 Gluten-free oatmeal from both Gluten-free oatmeal from Producer #1 Gluten-free oatmeal from Producer #2 Gluten-free oatmeal from both Internal and external studies combined

5

228

Count

216

7

4

1

12

5

% of Total

94.74% 3.07% 1.75% 0.44% 5.26%

2.19%

2

101

Count

95

4

2

0

6

2

% of Total

94.06% 3.96% 1.98% 0.00% 5.94%

1.98%

7

329

Count

311

11

6

1

18

7

% of Total

94.53% 3.34% 1.82% 0.30% 5.47%

2.13%

12/14 Assessment (indep. Laboratory tested)

6

316

Count

302

6

5

3

14

8

% of Total

95.57% 1.90% 1.58% 0.95% 4.43%

2.53%

8

320

Count

304

14

0

2

16

2

% of Total

95.00% 4.38% 0.00% 0.63% 5.00%

0.63%

14

636

Count

606

20

5

5

30

10

% of Total

95.28% 3.14% 0.79% 0.79% 4.72%

1.57%

Both studies

21

965

Count

917

31

11

6

48

17

% of Total

95.03% 3.21% 1.14% 0.62% 4.97%

1.76%

Bold values indicate % servings noncompliant.

International Journal of Food Science and Technology 2016

© 2016 PepsiCo, Inc. International Journal of Food Science & Technology published by John Wiley & Sons, Ltd. on behalf of Institute of Food Science and Technology