Proficiency Testing of Feed Constituents: A Comparative Evaluation of

Subscriber access provided by Northern Illinois University

Article

Proficiency testing of feed constituents: a comparative evaluation of European and developing country laboratories and its implications for animal production H.P.S. Makkar, I. Strnad, and J. Mittendorfer J. Agric. Food Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jafc.6b02452 • Publication Date (Web): 20 Sep 2016 Downloaded from http://pubs.acs.org on September 21, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers

Journal of Agricultural and Food Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Subscriber access provided by Northern Illinois University

and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Agricultural and Food Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 30

Journal of Agricultural and Food Chemistry

Proficiency Testing of Analytical Chemistry Laboratories

z-score =

𝑥𝑙𝑎𝑏 −𝑥𝑅𝑇 𝑆𝐷𝑅𝑇

𝑥𝑙𝑎𝑏 , laboratory result 𝑥𝑅𝑇 , mean of all results of the ring test 𝑆𝐷𝑅𝑇 , standard deviation

| z | ≤ 2: satisfactory result 2 < | z | < 3: doubtful result

ACS Paragon Plus Environment

| z | > 3: unsatisfactory result


1 2 3

Proficiency testing of feed constituents: a comparative evaluation of European and developing country laboratories and its implications for animal production

4 H. P .S. MAKKAR1, I. STRNAD2, J. MITTENDORFER2

5 6 7 8 9 10

1,

Food and Agriculture Organization of the United Nations (FAO), Animal Production and Health division, Rom, Italy 2,

Austrian Agency for Health and Food Safety (AGES), Institute for Animal Nutrition and Feed, Wieningerstrasse 8, 4020 Linz

11 12

Corresponding author: Harinder P.S. Makkar

13

E-mail: [email protected]

14

Tel: +390657054944

15

Fax: +390657055749

16 17

Short title: Proficiency testing for feed analysis

18


Page 2 of 30

Page 3 of 30

19 20


Proficiency testing of feed constituents: a comparative evaluation of European and developing country laboratories and its implications for animal production

21 22

ABSTRACT

23

Proficiency tests, with two feed samples each year, for various constituents (proximate, macro-

24

and micro-minerals, feed additives and amino acids) were conducted in 2014 and 2015. A total

25

of 40 and 50 European and 73 and 63 developing country feed analysis laboratories participated

26

in the study in 2014 and 2015 respectively. The data obtained from these two sets of

27

laboratories in each year enabled a comparison of the performance of the European and

28

developing country laboratories. Higher standard deviation and several folds higher coefficient

29

of variation were obtained for the developing country laboratories. The coefficients of

30

variation for chemical composition parameters, macro-minerals, micro-minerals and amino

31

acids were higher by up to 9-fold, 14-fold, 10-fold and 14-fold respectively for the developing

32

country laboratories compared with the European laboratories in 2014, while the corresponding

33

values for the 2015 were 4.6-fold, 4.4-fold, 9-fold and 14-fold higher for developing county

34

laboratories. Also higher number of outliers were observed for developing countries (2014:

35

7.6–8.7% vs. 2.9–3.0%; 2015: 7.7–9.5% vs. 4.2–7.0%). The results suggest higher need for

36

developing country feed analysis laboratories to improve the quality of data being generated.

37

The likely impact of higher variability of the data generated in developing countries towards

38

safe and quality preparation of animal diets, their impact on animal productivity and possible

39

ways to improve the quality of data from developing countries are discussed.

40 41

Key words: Proficiency test, ring test, feed analysis, feed chemical composition, mineral

42

analysis

43 ACS Paragon Plus Environment


44

INTRODUCTION

45 46

Feed is the base of the livestock production systems, impacting almost every sector of the

47

livestock production – from animal reproduction, health and welfare – to farm economic

48

viability, environment, animal product safety and quality. The feed cost can account for up to

49

70% of the total costs for the production of an animal product. High feed costs can wipe out a

50

livestock rearing operation. Good nutrition (protein, energy and minerals in the right amounts)

51

increases animal production, reproductive efficiency (higher cyclicity, lower age at first calving

52

and lower calving interval) and higher productive life1,2, which result in higher profitability to

53

farmers. Feed nutrients (70 to 90% of nitrogen and phosphorus) are lost into the environment

54

through manure, which if not managed properly can lead to environmental pollution. A number

55

of welfare problems in animals are elicited by the feeding of poor quality or unsafe feeds.

56

Nutrient deficiency or excess in feed can lead to metabolic disorders in ruminants such as

57

acidosis and lameness causing welfare issues; whilst breeding animals of monogastric species,

58

which are restrict-fed to optimise health and production, may suffer from chronic hunger3. In

59

addition, these imbalanced nutrient situations also increase enteric methane emissions4. Society

60

demands production of animal products in a manner that is least damaging to the environment,

61

takes welfare needs of animals into account and brings benefits that are equitable.

62

Precision feeding or a properly balanced diet, free of undesirable substances, fed in the right

63

amount that meets the animal’s nutrient requirement avoids physical and psychological

64

suffering from hunger and thirst; and is crucial for optimal performance, economic viability of

65

livestock operation (by decreasing feed cost), decreasing environment pollutants (by decreasing

66

release of nitrogen and phosphorus in manure and enteric methane)5,6. Precision, or balanced

67

feeding, results in more animal products from less feed. It will also translate to less use of

68

energy, land and natural resources7,8. In order to practice precision or balanced feeding, a pre-


Page 4 of 30

Page 5 of 30


69

requisite is the availability of good quality data on chemical composition of feed ingredients

70

and feeds. Therefore the laboratories should conduct analyses using the right methods in the

71

right manner, using good laboratory practices including the use of internal quality control

72

samples.

73

A proficiency test is an inter-laboratory test that allows the evaluation of the performance of

74

laboratories, and is based on analysis of similar homogeneous samples. It is critical to ensuring

75

quality of analyses being performed in a laboratory. A proficiency test is an element of external

76

quality assurance (EQA). EQA promotes both quality improvement and standardisation of the

77

test procedures. Both EQA and internal quality control (IQC) are essential elements to good

78

laboratory practices. It is in the interest of the laboratories to assess their performance,

79

especially using proficiency tests, since it allows them to evaluate their performance vis-a-vis

80

their peers and is a valued step towards certification and accreditation. It also gives assurance to

81

the customers that the results they get are the right ones.

82

The International Analytical Group, Section Feeding Stuffs (IAG) has many years of

83

experience in conducting proficiency tests (also called ring test) for European feed analysis

84

laboratories. In 2014 and 2015 the Food and Agriculture Organization of the United Nations

85

(FAO) invited laboratories from developing countries for participation in this exercise. FAO’s

86

vast networks in different countries were used to deliver the samples to feed analysis

87

laboratories located in different parts of the world. The main goal of conducting the proficiency

88

test was to enable laboratories to assess and improve their feed analysis performance. The IAG

89

had also conducted the proficiency test in 2014 and 2015 for European countries using the

90

same test samples as used for developing countries laboratories. As a result of this proficiency

91

test, it was considered pertinent and prudent to compare the performance of the two groups of

92

countries: European and developing countries; and to use the results of the study to develop

93

strategies and means to enhance quality of data emerging from feed analysis laboratories.



94 95

Materials and Methods

96

Test material

97

The proficiency tests are generally conducted with one or two common samples, which are

98

used by all participating laboratories. The four samples (two per year) used for the proficiency

99

test were commercial products available in 20 kg paper bags.

100

Sample 1; 2014 and 2015: Green meal pellets. This feed belonged to the category of forages

101

and roughages and this matrix was especially selected for analysis of different fibre fraction

102

parameters. Analyses conducted on this sample were the classical proximate parameters

103

including mineral and trace elements.

104

Sample 2; 2014: Compound feed for pigs. This was a compound feed for weaned piglets (25%

105

inclusion rate). Besides analysis of classical proximate parameters this compound feed was

106

especially intended for the analysis of different feed additives (enzymes, vitamins, organic

107

acids and antioxidants). For the proficiency test this sample was spiked with formic acid (2%)

108

and butylated hydroxytoluene (BHT) (50 mg/kg). In developing countries there were too few

109

laboratories that analyzed these parameters; therefore, theses parameters were not taken for

110

comparisons in this paper.

111

Sample 2; 2015: Complete feed for poultry. Besides analysis of classical parameters this mixed

112

feed was especially selected for the analysis of different feed additives (enzymes, vitamins,

113

organic acids, coccidiostats and antioxidants).

114

The content of the 20 kg paper bags (7 in number) was thoroughly mixed, subsequently split in

115

portions of about 0.7 kg9, which were filled into plastic bags and sealed. The samples were sent

116

to the participating laboratories by express means and these were received in the laboratories


Page 6 of 30

Page 7 of 30


117

within 15 days of dispatch. The homogeneity of the samples was tested and validated according

118

to Thompson and Wood10 by determining protein, crude fibre and fat using a near infrared

119

spectroscopy (NIRS) apparatus (Foss NIR Spectrometer InfraXact Lab) and selected elements

120

using ICP-MS respectively. The laboratories stored the samples in a cool and dark place. The

121

analyses were completed within one month of receiving the samples. The analyses conducted

122

by the participants are given in tables, which were for general chemical constituents, macro

123

minerals, trace elements, feed additives and amino acids. Metabolizable energy estimation was

124

also included. Laboratories were asked to specify the analytical methods applied. Most

125

laboratories followed this requirement. Either the laboratories referred to the international

126

standard methods (e.g. ISO, EU or AOAC) or described the method.

127 128

Proficiency test

129

The proficiency test by the International Analytic Group (IAG) is conducted annually with two

130

samples (green meal and compound feed) and is organized by the Institute for Animal Nutrition

131

and Feed, Austria Agency for Health and Food Safety (AGES). This proficiency test is

132

conducted without prescribing any methods of analysis in order to evaluate the laboratories

133

competence. Public and private laboratories were invited to participate. All laboratories are

134

provided with excel-sheets for reporting the analytical results. Each result should be the mean

135

of such a number of replicates the laboratory normally uses in feed analysis, expressed on dry

136

matter basis and reported with 3 significant digits in the units indicated. The participants are

137

asked to give information about the methods used by them in a standardized format, comprising

138

of three parts. Part 1 seeks information on: Digestion and extraction techniques; Part 2 on

139

Separation, pre-concentration and clean up; and Part 3 on Method of detection. Alternatively



140

the applied international standard procedure could be reported. All laboratories followed the

141

standard published protocols.

142

This proficiency test is being conducted since 1977 and a number of European laboratories

143

have been participating in the test on a regular basis. Based on the results of the proficiency

144

tests for many years, it can be assumed that the participating laboratories have integrated

145

quality control systems and follow good laboratory practices and generate quality data.

146

In 2014 and 2015 proficiency tests was organized jointly by FAO and IAG and conducted by

147

AGES. The laboratories participated in the proficiency test in response to an open invitation

148

sent through various FAO networks and the FAO website. The main aim of this joint test

149

between European and developing country laboratories was to enable laboratories in

150

developing countries to assess and improve their feed analysis performance. Both groups of

151

laboratories, European and developing countries were provided with the same samples for the

152

proficiency test. The annual IAG proficiency test (European laboratories) was separately

153

evaluated and discussed. The data of European and developing countries were evaluated

154

together, in order to get a good population of laboratories and a holistic view of the results.

155

Additionally, an extra internal evaluation was done for the comparison of both groups of

156

laboratories.

157

In 2014, in total 113 laboratories from 42 countries participated in the study, the distribution of

158

which was: developing countries, 73 laboratories from 28 developing countries; European

159

countries, 40 laboratories from 14 European countries. These laboratories were both from

160

public and private sectors. Several laboratories which were neither from Europe nor from

161

developing countries (e.g. Australia) were excluded for the comparison. In 2015, in total 113

162

laboratories from 49 countries participated. In this year, for comparison of the two groups 63


Page 8 of 30

Page 9 of 30


163

laboratories from 27 developing countries and 50 laboratories from 22 European countries were

164

taken.

165 166

Statistical evaluation

167

As part of the proficiency test the following parameters were calculated: average, standard

168

deviation, coefficient of variation. Before calculating these parameters, Dixon outlier test11 with

169

critical values, at 5% error, as listed in Wernimont12 was used to identify outliers, which were

170

eliminated from the calculation.

171 172

The calculated parameters are described below:

173 ∑

174

Average of all results (x) as: x =

175

Standard deviation (SD) measures the amount of variation or dispersion from the

176

average.

177 178

Coefficient of variation is a relative measure of dispersion. It is the ratio of the

179

standard deviation to the mean (x) expressed as a percentage. It provides information

180

of variation in percentage comparable to other parameters.

181

=

∗ 100

182

Results which are outside the mean ± 2 x standard deviation boundaries were marked with an

183

"s" (so called straggler) but remain in the set of evaluation data. Values presented as ‘below the

184

limit of detection’ were not taken into account in the evaluation.

185

Average and standard deviation are accepted as assigned values for calculation of the z-value13.



186

An often used quantitative criterion for the evaluation of the laboratory performance is the z-

187

score14 and the International Harmonized Protocol of the Proficiency Testing of Analytical

188

Chemistry Laboratories15. This value was calculated by dividing the difference between the

189

laboratory result (xlab) and the mean of all results of the ring test (xRT) for a particular parameter

190

by the standard deviation.

191

z-score =

| |

192 193

The following internationally accepted classification is used:

194

| z | ≤ 2: satisfactory result

195

2 < | z | < 3: doubtful result

196

| z | > 3: unsatisfactory result

197

Using this approach, all doubtful or unsatisfactory results must be critically examined by the

198

laboratory staff for all possible sources of error. Measures should be taken to avoid errors and

199

improve results.

200

For the detection of the significant difference between the means and the standard deviations of

201

the two groups Students-t test and F-test were used respectively16. Both tests were applied with

202

a probability of 95%. Following the international protocol for proficiency tests17 only

203

parameters with equal or more than 8 participants were taken into account for the comparison

204

between the two groups. The parameters for which data were available from 6 and 7

205

participants have also been given the table but not taken for comparison.

206 207

RESULTS


Page 10 of 30

Page 11 of 30


208

The evaluation of the proficiency test results show that the number of laboratories of

209

developing countries analysed mainly proximate analysis of feed constituents and macro-

210

minerals. The laboratories that also analysed trace elements and amino acids were few.

211

Carotenoid, xanthophyll, sugar, selenium, mercury, nickel, fluoride, iodine, vitamin D3,

212

enzymes, antioxidants, coccidiostats were analysed only in 3 laboratories in developing

213

countries at a maximum; and hence not included in the proficiency tests..

214 215

Proximate analysis

216

Tables 1 and 2 present mean and standard deviation (SD) for chemical composition data for the

217

forage and the compound feed sample, respectively.

218

For the forage sample mean values differed significantly (P < 0.05) only for acid detergent

219

lignin (ADL) in 2014; and for crude ash and crude fibre in 2015. The CV for ADL observed

220

was as high as 89% (Table 1). Determination of ADL occurs in several steps and is therefore

221

prone to errors. Challenges associated with determination of fibre fractions and higher

222

deviations in determination of these parameters have also been observed in previous ring tests

223

conducted by IAG. High values of CV for ADL (up to 42%) have been reported18,19. Amongst

224

the fibre fractions, generally highest CV has also been observed for ADL in other proficiency

225

tests for feed analysis19,20. Points worthy to note are significant differences between SD of the

226

two groups and overall higher SD for the values obtained by developing country laboratories

227

(Table 1). Only SD of protein (Dumas) was comparable in 2015 between European and

228

developing country laboratories.

229

For the compound feed sample, means for crude fibre, crude ash and HCl insoluble ash (a

230

measure of soil/sand contamination) in 2014 and protein (Kjeldahl), crude ash, HCl insoluble

231

ash and, crude fat and metabolizable energy in 2015 differed significantly for the two groups



232

of laboratories. While for all parameters the SD values were significantly different between the

233

two groups and always higher for developing countries, suggesting higher spread of values for

234

these countries. This is also reflected in several fold higher coefficient of variation (CV) of

235

values obtained by developing country laboratories (Table 2). Only SD of protein (Dumas) was

236

comparable between the two groups of the laboratories in 2015.

237 238

Macro- and trace-minerals, feed additives and amino acids

239

A pattern similar to that reported for proximate analysis parameters was obtained for both

240

macro- and trace-minerals in both samples: mean values are statistically similar for most of the

241

analyses (except for iron, copper and two amino acids in 2014; potassium, iron, copper and

242

three amino acids in 2015) but SD values are significantly different. Standard deviation as well

243

as CV values are several folds higher for developing country laboratories (Tables 3–6).

244

Amino acid also followed the similar pattern (Table 7), measured only for the compound feed

245

sample. A number of feed additives such as Vitamin A, E and D3, phytase, BHT, organic acids,

246

coccidiostats were not or seldom analysed in developing countries.

247

A further indicator for the difference between European and developing country laboratories is

248

the percentage of outliers. In 2014 for the forage sample the outliers for European and

249

developing country laboratories were 2.9% and 7.6% respectively; and for the compound feed

250

sample the outliers were 3.0% and 8.7% respectively. In 2015 for the forage sample the outliers

251

for European and developing country laboratories were 4.2% and 7.7% respectively; and for

252

the compound feed sample the outliers were 7.0% and 9.5% respectively (data not shown in

253

tables). Greater numbers of outliers were produced by developing country laboratories.

254


Page 12 of 30

Page 13 of 30


255

DISCUSSION

256

This is the first time that a proficiency test for feed analysis has been conducted in a systematic

257

manner and at an extensive scale for developing country laboratories. The objective was to

258

evaluate the laboratory performance of participating laboratories, aid them in locating

259

systematic errors, which is often hard to achieve by other means, and identify assays that need

260

improvement in the participating laboratories. The basic elements of a proficiency test are that

261

the test samples sent through a proficiency testing programme should not be given any special

262

treatment and that the analyses on them should be conducted following the same set of

263

procedures including the number of replicates and the standard methods as for the samples

264

analysed in routine. These internationally accepted basic elements formed the basis of the

265

current study. Also following the internationally accepted procedure, the assays that had z

266

values of 2 < | z | < 3 and | z | > 3 were identified as the ones producing doubtful and

267

unsatisfactory results, which were informed to the laboratories to critically examine for all

268

possible errors and follow good laboratory practices and integrate quality control methods as

269

described in 13.

270

Higher SD, several fold higher CV of almost all analyses, higher percentage of outliers for

271

developing country laboratories, and a significant difference (P < 0.05) for SD between

272

European and developing country laboratories for a large number of analyses (Tables 1–7)

273

demonstrate that there is a greater need for developing country laboratories to improve the

274

quality of analysis. As some examples the CV of ADL, crude fat and acid insoluble ash (2014

275

forage sample) of 88%, 40% and 13% in developing countries are unacceptably high. Similarly

276

CV of micro-minerals of 40 to 98% in developing countries illustrate a big scope for

277

improvement in the manner these assays are being conducted (Tables 1–7). It may be noted that

278

in proficiency tests there is no obligation regarding methodology used in the assays since the

279

main issue is to test the performance of the procedures that are in routine use at each laboratory.



280

As for 2014, the data analysed from the proficiency testing in 2015 showed higher number of

281

outliers for developing countries. The data also illustrated higher standard deviation and several

282

folds higher coefficient of variation for assays conducted in developing country laboratories.

283

For developing country feed analysis laboratories the coefficients of variation for chemical

284

composition parameters, macro-minerals and micro-minerals were higher by up to 2.4-fold,

285

2.7-fold and 2.7-fold respectively, compared with those for European feed analysis laboratories

286

for the forage sample (the values for 2014 being 4.2-fold, 8.4-fold, and 4.1-fold respectively);

287

and for the compound feed the values for chemical composition parameters, macro-minerals,

288

micro-minerals and amino acids were higher by 2.6-fold, 2.3 fold, 2.5-fold, and 4.5-fold for

289

2015 and 3.4-fold, 3.3-fold, 1.8-fold, and 5.0-fold for 2014 respectively (Table 8). Unlike in

290

2014, the standard deviation of protein determined by Dumas method was comparable between

291

the two groups of the laboratories in 2015 (Tables 1 and 2). Although lower variations in the

292

results were obtained in 2015, but still the variations for developing countries are higher than

293

for European countries. Again highlighting that developing countries need to pay higher

294

attention towards integrating quality control systems and follow good laboratory practices in

295

feed analysis laboratories. A similar proficiency test is currently taking place and it is also

296

planned to follow the progress of the laboratories that participated in all three proficiency tests.

297

In the past, at a number of fora, experts visiting developing country laboratories have expressed

298

the need to introduce the use of internal control samples in the analyses and for them to follow

299

good laboratory practices such as regular standardization of spectrophotometers, frequent

300

calibration of balances, pipettes and thermometers, and use of properly washed glassware,

301

among others. The results of this study corroborate these views originated from field

302

observations, and illustrate the need to strengthen quality assurance systems in these

303

laboratories. Without a robust quality control system, the laboratory personnel are unable to

304

evaluate the quality of the results being generated.


Page 14 of 30

Page 15 of 30


305

Implications for developing countries

306

Developing countries need to address this issue seriously because it could adversely impact

307

their export and increase wastage of feed and food items for not meeting the quality and safety

308

standards. Contaminated feed has often resulted in food of animal origin being recalled and/or

309

destroyed, causing huge economic losses and negatively affecting food security.

310

Generation of sound data is fundamental to implementation of nutritional principles and for

311

getting benefits from them. Also for sustainable development of the livestock sector, generation

312

of sound chemical composition data of feed ingredients and mixed or compounded feed is vital.

313

Feed industries are neither able to resource good quality ingredient nor prepare balanced feeds

314

without having reliable chemical and nutritional value data on feed ingredients.

315

In addition, without sound data on chemical composition, precision feeding or balanced feeding

316

approaches that demand nutrient provision, as per the nutrient requirements of the animal

317

cannot be used. Unbalanced feeding results in lower profit to farmers, production below the

318

genetic potential of animals, reproductive problems for example longer first age of calving and

319

longer calving interval, animal being more prone to metabolic diseases such as milk fever and

320

ketosis, shorter productive life, poor animal health and welfare, and excessive amounts of

321

pollutants released to the environment6,7. Feeding of unbalanced diets could have severe

322

implications for developing countries and this could be avoided by generating good data.

323

Furthermore, ensuring good laboratory practices will enhance the health and safety of the

324

laboratory workers, protect the environment from laboratory-discharged pollutants and increase

325

efficiency of the laboratories in developing countries. Other spin-offs will be enhanced research

326

and education capabilities of students graduating from R&D institutions, promotion of a better

327

trading environment between developing and developed countries, increased quality of

328

research, and meeting of the requirements of international standards. It will also enhance



329

confidence of the customers towards analytical laboratories serving them that all technical,

330

administrative and human factors that influence the quality of the results being generated are

331

under continuous supervision with the aim to prevent non conformity and identify opportunities

332

for improvement.

333 334

Suggested measures to improve quality of the data

335

Governments should consider increasing investment for improving laboratory infrastructure

336

and laboratory proficiency. Development of sound training programmes and their effective

337

execution is warranted. Higher donors’ attention to this issue and provision of greater funds for

338

developing capacity of laboratory personnel are needed.

339

Many functional laboratories can improve the quality of the data without much investment. A

340

culture of generation of quality data needs to be practiced through a change of mindset that

341

generation of any data is not important. This can be achieved by integrating quality control

342

systems and following good laboratory practices. FAO has taken a number of steps in this

343

direction: produced various manuals on quality control systems in feed analysis

344

laboratories13,21,22; distributed them at no-cost; made them available on FAO website for free

345

downloading, arranged on-line courses23 on risk management and quality control enhancement;

346

and by organized proficiency tests every year. FAO has also been providing relevant

347

educational support to the laboratories through its network of experts24. However, these efforts

348

need to be translated into a movement within countries by formulating government policies, for

349

example setting up of a body overseeing quality of data being generated by laboratories and the

350

laboratory operations; and by supporting laboratories through investments and capacity

351

development. The efforts of International organizations will only be catalytic towards


Page 16 of 30

Page 17 of 30


352

furthering a culture of generation of quality data. Large impact can only be generated by

353

government actions and policies, including promoting public-private partnerships.

354

The donors, besides supporting efforts that enhance quality control systems in laboratories,

355

should also demand putting in place of a proper control mechanism for the data being generated

356

by the laboratories in the framework of their sponsored projects. Similarly, journals while

357

considering the work for publication in the journal should also seeking information from

358

authors on the quality control set up in their research laboratories. Certification and

359

accreditation of laboratories by an outside agency may be encouraged but most laboratories in

360

developing countries do not have funds to achieve this. Use of internal standards will enable the

361

laboratory personnel to evaluate if they are producing quality data. Also a creation of a network

362

of laboratories within a country and running of annual in-country proficiency tests would

363

contribute to furthering the proficiency of laboratories, at a relatively low cost.

364

The progress made in improving the quality of data generated could be assessed, both at the

365

levels of a country and participating laboratory, by annual monitoring of z-scores. External

366

quality assessment or proficiency testing and internal quality control are critical to ensuring

367

quality of test practice. These processes permit laboratories to monitor and assess their long-

368

term performance, which could be compared to those of peer laboratories for proficiency

369

testing.

370

The results discussed here have implications for both monogastric and ruminant production

371

systems. The unsound data could adversely impact trade, increase feed wastage, render

372

precision feeding ineffective, and make feed industries incapable to resource good quality

373

ingredients or prepare good feeds. Unbalanced feeding results in: decrease in profit to farmers,

374

production below the genetic potential of animals, reproductive problems, metabolic diseases,

375

shorter productive life, poor animal health and welfare, and excessive amounts of pollutants



376

released to the environment. Spin-offs of improved data quality will be enhanced research and

377

education capabilities of students. The study will enable putting in place mechanisms in

378

developing countries to improve quality of chemical composition data of feed resources.

379

In conclusion, the chemical composition values of animal feeds, which reflect their nutritional

380

value, have a wide variation around the mean and a higher percentage of outliers for developing

381

countries. The quality of data from feed analysis laboratories in developing country laboratories

382

needs improvement. Developing countries need to address this issue seriously because it could

383

adversely impact their export and increase wastage of feed and food items by not meeting the

384

required quality and safety standards. Investment in improving skills of laboratory staff and

385

laboratory infrastructure, coupled with putting in place a practice of following good laboratory

386

practices, are expected to improve the quality and reliability of the data. Also there is a need to

387

further strengthen already ongoing training programs and to establish formal training programs,

388

if they do not exist, for laboratory staff that provide effective training on methods for

389

determination of chemical composition of feeds and feed ingredients. Technical Without sound

390

data, countries will not be able to fully exploit the benefits, in terms of enhanced animal

391

productivity, decreased environmental pollution and other social benefits, of using sound

392

nutritional concepts and practices in the livestock sector.

393 394

ACKNOWLEDGEMENT

395

We thank FAO representative in the countries of the participating laboratories for their help in

396

making available the proficiency test samples to the participating laboratories. Also thanks to

397

Claudia Pecanka from AGES; Antonella Falcone and Enrico Masci from FAO Headquarter in

398

Rome; and FAO Representatives in countries in which the participating laboratories were


Page 18 of 30

Page 19 of 30


399

located for their excellent support in distribution of the proficiency samples. Authors also thank

400

Dr. Johan De Boever for useful suggestions.

401 402

REFERENCES

403

1. Hersom, M.J. Opportunities to enhance performance and efficiency through nutrient

404 405

synchrony in forage-fed ruminants. J. Anim. Sci. 2008, 86, E306–E317. 2. Delgadillo, J.A.; Martin; G.B. Alternative methods for control of reproduction in small

406

ruminants: A focus on the needs of grazing industries. Animal Frontiers 2015, 5, 57–65.

407

3. FAO. Impact of animal nutrition on animal welfare – Expert Consultation 26−30 September

408

2011 – FAO Headquarters, Rome, Italy. Animal Production and Health Report. No.1.

409

Rome 2012. (http://www.fao.org/docrep/017/i3148e/i3148e00.pdf) (30 May 2016).

410

4. Flachowsky, G.; Kamphues, J. Carbon footprints for food of animal origin: What are the

411 412 413

most preferable criteria to measure animal yields? Animals 2012, 2, 108–126. 5. Nahm, K.H. Efficient feed nutrient utilization to reduce pollutants in poultry and swine manure. Critical Rev. Environ. Sci.Technol. 2012, 32, 1–16.

414

6. Garg, M.R.; Sherasia, P.L.; Bhanderi, B.M.; Phondba, B.T.; Shelke, S.K,; Makkar, H.P.S.

415

Effects of feeding nutritionally balanced rations on animal productivity, feed conversion

416

efficiency, feed nitrogen use efficiency, rumen microbial protein supply, parasitic load,

417

immunity and enteric methane emissions of milking animals under field. Anim. Feed Sci.

418

Technol. 2013, 179, 24–35.

419 420

7. Makkar, H.P.S. Precision feeding – a developing country perspective, in Precision livestock farming ’13, Berckmans, D.; Vandermeulen, J. Eds. 2013, pp. 95–105.

421

8. Beever, D.E.; Drackley, J.K. Feeding for optimal rumen and animal health and optimal feed

422

conversion efficiency: The importance of physical nutrition, in Optimization of feed use

423

efficiency in ruminant production systems, Makkar, H.P.S.; Beever, D.E. Proceedings of

424

an FAO Symposium held in Bangkok, Thailand, 27 November 2012, FAO, Rome (2013)

425

(http://www.fao.org/3/a-i3331e.pdf) (30 May 2016)

426 427 428 429

9. Thompson, M.; Wood, R. The International Harmonized Protocol for the Proficiency Testing of (Chemical) Analytical Laboratories, Pure Appl. Chem. 1993, 65, 2123–2144. 10. Dixon, W.J. Processing Data for Outliers, Biometrics. International Organization for Standardization (ISO) document ISO 5725-1981, 1953, pp. 74−89.



430 431 432

11. Wernimont, G.T. Use of Statistics to Develop and Evaluate Analytical Methods (W. Spendley, editor), AOAC, Arlington, VA. 1985, Table A-9, p. 156. 12. FAO. Quality assurance for animal feed analysis laboratories, FAO Animal Production and

433

Health Manual No. 14. Rome, 2011.

434

(http://www.fao.org/docrep/014/i2441e/i2441e00.pdf) (30 May 2016).

435 436 437 438

13. ISO/IEC-17043. Conformity Assessment – General requirements for proficiency testing, International Organization for Standardization, Geneva, 2010. 14. ISO-13528. Statistical methods for use in proficiency testing by Inter-laboratory Comparison, International Organization for Standardization, Geneva (2005).

439

15. Thompson, M.; Ellison, S.R.; Wood, R. The international harmonized protocol for the

440

proficiency testing of analytical chemistry laboratories (IUPAC Technical Report). Pure

441

Appl. Chem. 2006, 78, 145–196.

442 443 444 445

16 Sachs, L. Angewandte Statistik, Anwendung statistischer Methoden, Springer; 7. Auflage (5. März 1992). 17. Horwitz, W. Protocol for the design, conduct and interpretation of collaborative studies. Pure Appl Chem 1988, 60, 855–864.

446

18. Lanari, D.; Pinosa, M.; Tibaldi, E.; D’Agaro, E. Risultati di due ring test condotti sulla

447

composizione chimica di alcuni alimenti e sulla digeribilità in vivo. Zootecnica e

448

Nutrizione Animale 1992, 17, 285–295.

449

19. Bovera, F.; Spanghero, M.; Galassi, G.; Masoero, F.; Buccioni A. Repeatability and

450

reproducibility of the Cornell Net Carbohydrate and Protein System analytical

451

determinations. Ital. J. Anim. Sci. 2003, 2, 41-50.

452 453

20. EGRAN. Technical note: Attempts to harmonize chemical analyses of feeds and faeces for rabbit feed evaluation. World Rabbit Sci. Assoc. 2001, 9, 57–64.

454

21. FAO. Quality assurance for microbiology in feed analysis laboratories, by RA Cowie, in

455

FAO Animal Production and Health Manual No. 16, Makkar, H.P.S. Ed. FAO, Rome,

456

2013.

457

22 de Jonge, L.H.; Jackson, F.S. The feed analysis laboratory: Establishment and quality

458

control. Setting up a feed analysis laboratory, and implementing a quality assurance

459

system compliant with ISO/IEC 17025:2005, in Animal Production and Health

460

Guidelines No. 15, Makkar, H.P.S., Rome, FAO, 2013 (http://www.fao.org/3/a-

461

i3535e.pdf) (30 May 2016)

462 463

23 FAO. An FAO-TAMU on-line course to enhance capacity of laboratory managers on quality control systems, 2014.


Page 20 of 30

Page 21 of 30


464

(http://www.fao.org/ag/againfo/home/en/news_archive/AGA_in_action/2014_Enhancing

465

_Feed_Quality_an_Safety.html) (30 May 2016).

466

24 FAO. Network of experts: Support for strengthening quality control systems in animal feed

467

analysis laboratory in developing countries, 2011

468

(http://www.fao.org/ag/againfo/home/documents/Network_Quality-control.pdf) (30 May

469

2016).

470



471 472

Page 22 of 30

Table 1 Chemical composition obtained for the forage samples in 2014 and 2015 by the two groups of laboratories (values without parentheses are for 2014, while values in square parentheses are for 2015) European countries

Protein Kjeldahl, g/kg

No. of labs 25 [30]

Mean

SD (CV)

112 [145]

2.50x (2.23) [2.38x (1.65)] x

Developing countries No. of labs 59 [49]

Mean

SD (CV)

113 [143]

9.05y (8.04) [8.54y (5.98)] y

Protein Dumas, g/kg

19 [27]

114 [148]

3.08 (2.70) [7.71 (5.21)]

11 [7]

115 [144]

5.00 (4.37) [6.50 (4.51)]

Crude fibre, g/kg

22 [30]

261 [250]

8.16x (3.13) [11.9x (4.77)]

49 [43]

256 [246]

27.4y (10.69) [21.9y (8.93)]

ADFOM, g/kg

20 [25]

318 [313]

17.8 (5.60) x [21.4 (6.83)]

37 [25]

321 [314]

42.3 (13.20) y [44.4 (14.16)]

NDFOM, g/kg

20 [26]

521 [488]

17.0x (3.26) [24.6x (5.03)]

36 [25]

540 [478]

52.2y (9.65) [44.7y (9.35)]

ADL, g/kg

20 [17]

50.7 [59.7]

a

5.39 (10.63) x [5.17 (8.67)]

27 [15]

79.8 [54.1]

70.6 (88.56) y [21.3 (39.44)]

Crude ash, g/kg

30 [37]

115a a [110 ]

3.94x (3.41) x [2.33 (2.11)]

65 [58]

111b b [105 ]

5.60y (5.03) y [5.68 (5.41)]

16

49.0

2.28 x (4.65)

23

49.2

6.51y (13.23)

[21]

[29.0]

[1.6 (5.53)]

[21]

[29.4]

[3.72 (12.66)]

25

26.1

2.79 x (10.71)

43

29.5

11.8 y (40.13)

[33]

[30.6a]

[3.8 x (12.4)]

[43]

[33.5b]

[7.82 y (23.31)]

HCl insoluble ash, g/kg

Crude fat (including hydrolysis), g/kg

x

x

x

b

y

y

y

473 474 475 476 477 478

ADFOM , acid detergent fibre corrected for organic matter; NDFOM, neutral detergent fibre corrected for organic matter; ADL, acid detergent fibre; values in round parentheses are coefficient of variation (CV) in % . n, number of observation Within a row, means with different superscripts a and b differ at P

Proficiency Testing of Feed Constituents: A Comparative Evaluation of

Recommend Documents