Articles and Resources

Preventing Mixing Errors And Cross-Contamination Of Premixes And Feeds

The basic premise for manufacturing formula feeds is that by "least cost" formulation of ma ingredients, and additive feeds may be manufactured that provide optimum nutrition and health to animals, poultry and fish. It is assumed that feed must be mixed so that each serving or at least each day's consumption of feed should provide nutrients and additives at formulated levels.

The "level of scrutiny" is thereby critical, as a cow may consume 10 kilos of feed a day and a shri less than 1 gram. Even if the mixing of feeds is validated by analyzing for one or many nutrients or additives as tracers for all other ingredients, mixing may not be adequate if the particle sizes of specific ingredients or additives are not fine enough to provide a uniform dispersion at the "level of scrutiny¡ã An example would be adding a powdered vitamin that clumps on the end wall of a mixer. When it falls off the wall, it will be present in the feed as formulated but it will not be adequately dispersed even if other ingredients are.

With this as a caveat, how can one prevent mixing errors and cross-contamination of premixes and feeds? The answer is by careful design and selection of mixing equipment, conveyers, bucket elevators, holding bins, pelletmills and all other feedmill equipment with which feed is handled. Since most feedmills are designed with components from many vendors with specific expertise, the builder must rely on the knowledge of these suppliers. In the end, however, since equipment wears and changes over time the only way to know the capability of a feedmill and to thereby prevent errors in mixing and/or in cross-contamination of feeds is to study these specific issues.

Determining Mixing Of Premixes And Feeds
Knowing the capability of the feedmill to mix feeds will allow optimization of mixing parameters including: mixing time, batch size, and speed (revolutions per minute). Optimizing manufacturing conditions will: ensure the quality of feed manufactured, increase plant capacity while reducing labor, energy and equipment depreciation costs. The optimum mixing conditions will depend upon the bulk density of the feed, the shape of the feed ingredients and the size of feed particles. It is critical mixers not be overloaded and several studies have suggested it takes longer to mix large particles than it does to mix smaller ones.

Validating The Mixing Process
One must consider at least five issues before studying mixers to determine the adequacy of mixing and to optimize mixer performance.

· Selection of one or more tracers.
· Addition of the tracer to the test feed.
· Sampling the feed.
· Analyzing the samples.
· Interpreting the results.

A.Selection of the tracer
Whatever tracer or tracers are chosen for the mixer test, data from their analysis will be used to evidence the mixing for all other ingredients. If they yield results typical of a complete mix, one will assume all other ingredients are also mixed.

At least the following criteria should be considered in selecting one or more than one tracers for the test.

*The tracer should be contributed to the feed from only one source. If a feed is formulated with both corn and wheat with 12% protein, analyzing for protein is meaningless as even if no mixing occurred analytical results from a series of samples would yield a low coefficient of variation (CV).

*The tracer should be a microingredient. While it may be reasonable to assume that if a tracer added at 50 grams/tonne is completely mixed then macroingredients added at 1% or more will also be mixed, it is seemingly unreasonable to make the converse assumption. Drugs such as halofuginone (1) are added to feeds at 3ppm or less and selenium is added at 0.3ppm or less. Is it reasonable to assume these are mixed if the mixer study employs salt (sodium chloride) as the tracer with the salt added to the feeds at 2% of the total formula, especially if salt is also contributed to the feed from other ingredients such as fish meal?

*The analytical procedure to determine the tracer must be accurate and reliable with minimum analytical error.

*The analytical procedure should be inexpensive.

*The analytical procedure should provide quick results, ideally "on the spot" so additional batch of feed can be studied with mixing parameters changed based upon initial test results.

*One should be able to interpret the results objectively.

The most commonly used tracers for validating mixing are:

*Salt (sodium chloride)- in the Official ASAE (American Society of Agricultural Engineers) method. Assays cost for chloride or for sodium as low as USD $15/sample with analytical error for chloride as low as 2% or 3% coefficient of variation and for sodium 5%.

*Minerals such as zinc, manganese and cobalt. Determined by atomic absorption spectroscopy. Assay cost as low as $25/sample with analytical error as low as 5% to 7%.

*Amino acids- such as lysine and methionine. Determined by HPLC. Assay cost as low as $50/sample with analytical error as low as 5%.

*Vitamins such as Riboflavin. Determined chemically or by HPLC. Assay cost as low as $60/sample with analytical error as low as 7% to 10%.

*Medicated premixes- such as amprolium. Determined chemically. Assay cost as low as $50/sample with analytical error as low as 10% to 15%.

*Colored iron particles or colored fine iron powder (2). Retrieved magnetically from feed samples, with colored spots developed or with colorimetric readings made from dye dissolved from the colored iron powder. Assay cost as low as $10/sample or much less if performed by feedmill personnel with analytical error as low as 2% to 3%.

B.Addition of the tracer(s) to the test feed
This should be via a premix that can, if necessary, be prepared by mixing the tracer additive in ground corn or another diluent. If the tracer is to be added at 50 grams/tonne, the 50 grams might be mixed in 450 grams of diluent to make a 500-grams/tonne addition.

The location of the tracer premix addition is usually where a "hand add" premix would normally added. If one is studying the design capability of a mixer as required for paddle mixers that often have difficulty achieving an end-to-end mix, two tracers may be added to a batch one at each end. Alternately, if one wants not only to study mixing capability but also the capability of automated metering equipment to consistently add microingredients to a series of batches of feeds, the tracer may be added via such automated addition equipment with samples then taken not from one batch of feed but from a series of batches.

The tracer is normally added after the mixer is fully loaded. Starting and ending times for the test must be carefully controlled and addition of the tracer must be coordinated with the sampling plan.

C.Sampling the test feed
Ideally, one takes samples from within the mixer. All samples must be "grabs" not composites. samples are taken from within a mixer, it may be adequate to take as few as three samples- one sample from each end and one from the middle.

Often it is difficult to obtain samples from within a mixer. In such instances, one may take a series of ten or more "grab" samples from the discharge of a batch off a screw conveyer exiting a sur bin. These samples will reflect not only mixer performance but incomplete cleanout of feeds from the mixer and surge bin. Such cross-contamination if it is occurring can be documented by taking samples from the next following batch of feed.

A 500-gram sample for the mixer performance test is adequate though samples from following batches to be tested for cross-contamination should be larger, possibly 2-kilos. Samples must be marked and they should not be homogenized prior to analysis as this will weaken the power of the test. If the samples are homogenized, one will compare results for 500-gram samples whereas if the samples are not mixed one will compare results for the sample analyzed (i.e. 80 grams or less).

D.Analyzing the test feed
This depends upon the tracer(s) one employs. If one uses vitamins, minerals, drugs, salt or amino acids, samples must be boxed tightly and shipped to the applicable laboratory for analysis. If one tests for chlorides via test strips or colored iron particulates or very fine colored iron powder, the tests may be performed at the plant.

E.Interpreting the test results
For the mixer performance test, this should be done by comparing the coefficient of variation (CV) found from the test data with the CV inherent to the method. The method CV is what one expects from repeat analysis of the same homogenous sample and can be as little as 2% for chloride chemical analysis or as high as 20% for many feed drug assays.

Results follow for studies of two mixers where colored iron particulates were employed, one study yielding results evidencing a complete mix and one evidencing a statistically significant mixing error follow.

In both cases red colored iron particles were formulated at 50 grams/tonne with an expected count of 100 particles from analysis of 80 grams of feed. The feed tested in both was broiler mash with a mixer capacity of 3 tonnes and a mixing time of 3 minutes.

The complete mixes yielded CV's consistently less than the 10.1% expected from a complete mix This was better than one could reasonable expect. The incomplete mix yielded CV's consistentl greater than the 10.2% expected.

The Chance Probabilities for the complete mixes were consistently greater than 5%, likelihoods typical of a complete mix. The Chance Probabilities for the incomplete mixes were in 4 or 5 batches less than 1% and would be expected in less than 1 in 100 studies of a complete mix. Since this data was not typical of a complete mix, the mix is judged incomplete.

Correcting A Mixing Problem
Common causes of incomplete mixing that may be easily corrected include overloading the mixer and mixing for too short a time. One should visually inspect the mixer when loaded to be sure ribbons or paddles are visible at least 15 centimeters above the feed. If the mixer is overloaded, one can reduce the batch size and run another mixer test. If the mixer is not overloaded, one may run another test with the mixing time extended by possibly a minute. If results are improved, one should repeat the test to be sure the improvement in consistent.

One may also visually inspect the mixer blades or paddles to be sure they are not severely coated with fat that can distort their ability to mix. One should also inspect the discharge gates to be sure they have no dead spots and to be sure they are not leaking. Ribbons or paddles should also be inspected to be sure they are not worn or broken. Typical life expectancy for ribbons may be 2 to 4 years. If one cannot improve results by reducing the loading or increasing the mixing time, one can investigate increasing the speed (revolutions per minute) of the mixer or re-engineering the mixer blades or paddles. If all else fails, it may be necessary to replace the mixer.

Cross-contamination Of Feeds And Premixes.
Currently, the greatest concerns over cross-contamination of feeds relate to "carryover' of rumina by-product formulated in non-ruminant feeds to ruminant feeds where it could spread mad cow disease and "carryover' of medicated feeds into non-medicated feeds. Medicated feed carryover a concern where trace levels of the drug in feeds can lead to tissue residues in meat, poultry and fish and condemnations or import refusals by the European Union, Japan and other countries with demanding import standards.

One way to prevent cross-contamination from causing problems is to not use ruminant by-products or drugs in feeds or to make feeds for only one species at the feedmill. This may be effective but it can also be costly. An alternate approach is to study the capability of the feedmill and to develop procedures to control cross-contamination at levels sufficiently low they will cause no problems.

Determining Cross-contamination Of Feeds And Premixes
Analytical methodology for determining trace level contamination of ruminant by-products or of most drugs into withdrawal feeds is limited and in many cases not precise or accurate. Most medicated feed assays yield accurate results at formulated levels but do not yield meaningful results in feeds at 1% or less of the formulated level. Since drug assays of meat, poultry or fish tissue may have far lower levels of detection than in feeds, this creates a situation where one may find tissue residues but be unable to determine whether the drug was present in the feed or not.

One solution is to employ simple easy to detect tracers, one example being colored iron particulates and to interpret tracer results as being indicative of the drug. It has also been found that fine powders will "carryover" more in feeds than granulated products. Electrostatic and van der Waals forces m also cause a tracer to behave differently than a drug. The validity of using colored iron tracers as indicators for a medicated feed was addressed by studying cross-contamination at the premix level rather than at the feed level, utilizing a medication with a good assay with a low level of detection.

This study compared counts for particulate iron particles, colorimetric readings for very fine colored iron particles and chemical assays for the drug amprolium. The study was performed at a premix plant where the drug was formulated at 2.5% into the initial batch of premix and each of the two iron based tracers were formulated at 1.1-kilos per tonne.

The level of detection for the drug was 50 ppm or 0.2% of the formulated level. The level of detection for the red iron particle counts, the colorimetric readings from the red tracer and the colorimetric readings from the blue iron powder were all about 0.02% the formulated level. The tracer results were thus about 10 times more "sensitive" than the chemical assay results. In this study, the amproli and the two tracers appeared to cross-contaminate similarly.

References
Eisenberg, S and D. Eisenberg, 1992. Markers in Mixer Testing- Closer to Perfection, Feed Management, November, pgs. 8-11 and 20.
Corrigan, O.I, L. Wilkinson, J. Ryan and O.F. Corrigan, 1994. The Use of Microtracers (tm) in a Medicated Premix to Determine the Presence of Tiamulin in Final Feed, Drug Development and Industrial Pharmacy, 20(8). 1503-1509.
Eisenberg, D. Mix With Confidence, 1994. International Milling, June, pgs. 31-33.
N. Amornthewaphat, K.C. Behnke and J.D. Hancock, 1998. Effects of Particle Size and Mixing Time on Uniformity and Segregation in Pig Diets, Kansas State University Department of Grain Science and Industry, Swine Day.
Heidenreich, E., 1998. Quality Assurance by Avoiding Carry Over and Cross-Contamination in Feed Compounding, June Biomin Symposium- "From Quality Feed to Quality Food", Vienna, Austri IFF-Research Institute Feed Technology, Braunschweig-Thune, Germany.
Elaboration d'un guide technique pour la realisation des controles d'homogeneite et des contaminatio croisees dans la perspective de l'agrement 1999. Tecaliman (Centre technique de la Nutritio Animale), Nantes, France, December.
Eisenberg, D. The Use of Colored Iron Particles in Determining Cross Contamination of Medicated Feeds at Feedmills and Premix Plants, 2003. Zootecnica International, March pgs. 42-47.

Tags · Mixing Errors · Cross-Contamination · Quality · Premixes · Feeds

17.02.2008. 13:13