Imprecision when using measuring cups to weigh out extruded dry kibbled food

Transcription

1 DOI: /j x ORIGINAL ARTICLE Imprecision when using measuring cups to weigh out extruded dry kibbled food A. J. German 1, S. L. Holden 1, S. L. Mason 1, C. Bryner 2, C. Bouldoires 2, P. J. Morris 3, M. Deboise 4 and V. Biourge 4 1 School of Veterinary Science, University of Liverpool, Neston, Wirral, UK, 2 Royal Canin (Schweiz) AG, Regensdorf, Switzerland, 3 The WALTHAM Centre for Pet Nutrition, Waltham-on-the-Wolds, Melton Mowbray, LE, UK, and 4 Royal Canin Research Center, Aimargues, France Keywords obesity, caloric restriction, feeding, commercial pet food Correspondence A. J. German, School of Veterinary Science, University of Liverpool, Leahurst Campus, Chester High Road, Neston, Wirral CH64 7TE, UK. Tel: ; Fax ; ajgerman@liv.ac.uk Received: 3 April 2010; accepted: 28 June 2010 Summary Many pet cats and dogs are fed dry extruded kibbled food by measuring cup, yet the precision and accuracy of this feeding strategy is not known. Over 12 studies, we assessed precision and accuracy of weighing out food portions, of various dry kibbled foods, by measuring cup. Poor precision was noted in all studies, with intra- and inter-subject coefficients of variation ranging from 2 to 13% and 2 to 28% respectively. Variable accuracy was also noted, which ranged from an 18% underestimate to an 80% over-estimate in portion size. No specific factors were associated with imprecision, but the degree of inaccuracy was negatively associated with portion size (R = )0.67, p = 0.022), and positively associated with the number of subjects participating in the study (R = 0.60, p = 0.048). This is the first study to document imprecision and inaccuracy of using measuring cups to estimate portions of extruded dry kibbled food. Over time, such errors could contribute to insidious weight gain in companion animals, potentially contributing to the development of obesity. Imprecision in measuring food portions could also contribute to failure of weight management programmes for obese animals. Introduction The majority of cats and dogs in developed countries are fed commercial pet foods (Robertson, 1996, 1999, 2003; Laflamme et al., 2008), and many are fed a dried extruded kibbled ration (Laflamme et al., 2008). Some authors have cited the widespread use of nutritionally complete and balanced commercial diets as a contributing factor in longer, healthier life spans in pets (Kraft, 1998). In contrast, other studies have implicated premium dry-expanded food as a risk factor for obesity in companion animals (Lund et al., 2005), which is increasing in prevalence and is now thought to be one of the most important medical diseases of companion animals (German, 2006). Although dry food is undoubtedly balanced in nutrients and easy to feed, it is energy dense and thus precision with feeding allocation is essential to avoid unwanted weight gain. Precision and accuracy is also important for weight management regimes, because most protocols use dietary caloric restriction as the principle means of weight loss and subsequent maintenance (German et al., 2007; Bissot et al., 2010). Most owners that feed dried food use a measuring cup, which enables the ration to be measured out quickly and conveniently. However, a recent human 368 Journal of Animal Physiology and Animal Nutrition 95 (2011) ª 2010 Blackwell Verlag GmbH

2 A. J. German et al. Imprecision when using measuring cups for dried pet food study suggested that there may be inaccuracy when using portion size aids (including measuring cups and spoons) to estimate portion size, with the potential both to over- and under-estimate amounts by over 20% (Godwin et al., 2004). Unfortunately, equivalent studies examining the use of portion size aids for companion animal foods are lacking. Therefore, the aim of the current study was to determine the precision and accuracy with which extruded dry kibbled food could be weighed out with a measuring cup. Materials and methods Diets and measuring cups Six different commercially available extruded dry kibbled diets were used in the study, representing products from three different manufacturers (Table 1). All diets are sold in bags of a variety of sizes, except for diet 2, which is sold in pre-weighed sachets of 20 g. The diets comprised various types (e.g. maintenance, growth, weight loss), macronutrient compositions, and had different kibble characteristics (size, shape and weight). For each of the foods, the measuring cup supplied by the manufacturer was used, and their dimensions varied (Table 1). Weights and measures For studies 1 3 and 5 12, electronic weigh scales (range g; Salter, Tonbridge, UK) were used as the reference method, which were calibrated with test weights (1 100 g; guaranteed accuracy 0.1%; Blake and Boughton, Thetford, UK) prior to each study to insure precision and accuracy. For study 4, a different set of calibrated electronic scales (PAGE 66100, Soehnle, Germany) was used as the reference method. Study participants Twelve different studies were conducted, all employing different diets, participants, and repetitions (Table 2). Four of the authors (AJG, SLH, CB and CB) participated in studies 1 3, whereas 10 different people participated in study 4, all of whom were Royal Canin employees involved in product technical support. Finally, three of the authors (AJG, SLH and SLM) participated in studies Table 1 Summary of study diets Diet Variable Type of diet Weight loss Weight loss Kitten growth Adult maintenance Adult maintenance Adult maintenance Species Feline Feline Feline Feline Canine Canine Kibble Shape Circular disc Circular disc Irregular pellet Pellet Circular disc Circular/cylinder Dimension (mm)* 11 x 9 x 4 11 x 11 x 4 8 x 6 x 4 7 x 7 x 7 20 x 20 x 9 14 x 14 x 9 Weight (g) Measuring cup dimensionsà Depth (mm) Top circumference (mm) Bottom circumference (mm) Diet analysis Metabolisable energy Moisture Protein Fat Crude fibre Ash Diet 1: Feline Obesity Management diet (Royal Canin); Diet 2: Feline Satiety Support diet (Royal Canin); Diet 3: Pro Plan Kitten Development (Nestle Purina, New Malden, UK); Diet 4: Science Plan Feline Mature Adult 7 + Active Longevity (Hill s Pet Nutrition, Watford, UK); Diet 5: Pro Plan Adult Large Dog (Nestle Purina, New Malden, UK); Diet 6: Canine Science Plan Adult 1-6 Advance Fitness (Hill s Pet Nutrition, Watford, UK). *Approximate dimensions, based upon measuring 3 5 representative kibbles. Dimensions measured by tape measure. àmean weight of a single kibble, based upon weighing kibbles, 3 5 times. Metabolisable energy expressed as metabolisable energy content (in kilocalories) per 1000 g as fed, and calculated using modified Atwater Factors (NRC, 2006); all remaining variables expressed as percentage on an as fed basis. Journal of Animal Physiology and Animal Nutrition. ª 2010 Blackwell Verlag GmbH 369

3 Imprecision when using measuring cups for dried pet food A. J. German et al. Table 2 Assessment of precision and accuracy by measuring cup Study Diet Subjects Repeats* Desired amount (g) Actual amount (g) Difference from actual Coefficient of variation (%) g % Intra-subject Inter-subject (49 72) 9 ()1 to 22) 18 ()2 to 44) 5 (2 6) (61 74) 7 (1 14) 14 (2 28) 5 (3 7) (23 43) 5 ()7 to 13) 10 ()14 to 26) 12 (3 13) (33 61) 8 ()7 to 21) 20 ()18 to 53) (38 48) 3 ()2 to8) 8()5 to 20) 5 (4 6) (35 63) 4 (0 28) 11 (0 80) 4 (4 4) (55 87) 1 ()5 to 27) 2 ()8 to 45) 2 (2 4) (87 119) 0 ()13 to 19) 0 ()13 to 19) 6 (2 15) ( ) 3 ()8 to 18) 2 ()5 to 12) 5 (2 6) (48 73) 8 ()2 to 23) 16 ()4 to 46) 9 (7 10) ( ) )2 ()18 to 9) )2 ()14 to 7) 5 (4 5) ( ) )10 ()24 to 13) )4 ()10 to 4) 4 (2 4) 4 Diet 1: Feline Obesity Management diet (Royal Canin); Diet 2: Feline Satiety Support diet (Royal Canin); Diet 3: ProPlan kitten (Purina); Diet 4: Feline Mature Adult Active Longevity (Hill s); Diet 5: ProPlan Dog (Purina); Diet 6: Canine Adult Advance Fitness (Hill s). For full details of diets, see Table 1. *No. repeats per subject. Study design Full details of the diets used, portion sizes and number of repetitions for each study is given in Table 2. The desired amount of food was first weighed on the calibrated electronic scales and then transferred to the correct cup. The cup was shaken to insure that the contents were level and the volume that the food occupied marked with an indelible pen. Each investigator then independently measured out either 1 (study 4), 5 (studies 2, 3 and 5 12) or 10 (study 1) portions of the desired amount of food (Table 1). Each measured portion was then placed in a clear plastic bag and, once all portions had been measured out in a study, they were then weighed on the electronic scales (without the plastic bag) and the actual amount that had been measured out was recorded. Data handling and statistics All continuous data are expressed as median (range) and all statistical analyses were performed with a computer software programme (Stats Direct, version 2.6.2; Stats Direct, Altrincham, UK). Data were first assessed for normality with the Shapiro-Wilk test. Precision data (intra- and inter-subject coefficients of variation, CV) were not normally distributed and associations with other parameters were assessed either using Kendall s tau (for continuous data) or the Mann Whitney test [where variables contained two categories e.g. number of study subjects and measuring cup size (see below)]. Given that accuracy data (difference between actual weights measured and desired weight) were normally distributed, possible associations with other parameters (e.g. portion size, kibble size, number of study subjects, measuring cup size) were assessed by linear regression. For the number of study subjects and measuring cup size, dummy variables were created for use in regression analysis. With regards to study subjects, those employing three subjects were assigned a value of 0, whereas studies with four or more subjects were assigned a value of 1. For measuring cup size, there was little variability amongst cups in either their circumference or diameter and the greatest variability was instead between short and tall cups. As a consequence, a dummy variable was created with cups <100 mm depth being assigned a value of 0, and cups >100 mm depth being assigned a value of 1. Univariable analysis was initially performed for all parameters. Multivariable analysis was then conducted by initially including variables identified as p < 0.2 on univariable analysis. Variables were then eliminated in a backwards stepwise fashion, with variables only being retained if their effect remained significant in the final model (p < 0.05), or if their removal substantially influenced (10% or greater change) the effect of other variables. The level of statistical significance was set at p < 0.05 for two-sided analyses. Results Poor precision was noted in all studies with intraand inter-subject CVs of 2 15% and 2 28% 370 Journal of Animal Physiology and Animal Nutrition. ª 2010 Blackwell Verlag GmbH

4 A. J. German et al. Imprecision when using measuring cups for dried pet food respectively (Table 2). Inaccuracy was also noted in the amount of food measured out (ranging from an 18% under-estimate to an 80% over-estimate). Intra-subject precision was not significantly associated with any factor (portion size, p = 0.36; number of subjects, p = 0.34; kibble weight, p = 0.74; cup size, p = 0.35). Similarly, no factor was significantly associated with inter-subject precision (portion size, p = 0.11; number of subjects, p = 0.59; kibble weight, p = 0.43; cup size, p = 0.85). In contrast, univariable analysis demonstrated that the number of study subjects was positively associated with the degree of inaccuracy (more variability when more subjects were used in the study, p = 0.014), whereas both the portion size (greater inaccuracy with smaller portions, p = ) and cup size (greater inaccuracy with smaller cups, p = 0.020) were negatively associated with accuracy (Table 3). However, on multivariable analysis, only the positive association with number of subjects (p = 0.048) and the negative association with portion size (p = 0.022) remained in the final model (Table 3). Discussion The current study has assessed accuracy and precision for using measuring cups to measure out various portions of dry cat and dog food. Four feline and two canine dry kibbled foods were chosen for the studies, such that a broad range of diet types were represented. Furthermore, the portion sizes that were chosen were designed to be equivalent to Table 3 The results of univariable and multivariable analysis for the association between inaccuracy in measurements and various factors Predictors R R 2 p-value Univariable Portion size ) Intra-subject CV Inter-subject CV No. subjects* Kibble weight ) Cup size ) Multivariable 0.84à 0.71à à Portion size ) No. subjects* CV, coefficient of variation. *Dummy variable created for no. subjects, where 0: 3 subjects; 1: 4 subjects. Dummy variable created for measuring cup data, with two groups: 0: short, i.e. <100 mm; 1: tall >100 mm. àresults for the final multivariable model, containing portion size and no. subjects as predictors. Values in bold are those where p < amounts approximating to the typical meal size of the companion animal for which it was designed. Considerable variability in both accuracy and precision was noted for all cups and diets tested; although some combinations had acceptable precision and accuracy, others demonstrated either marked inaccuracy (up to 80% over-estimate) or marked imprecision (up to 28% CV amongst subjects), or both. Inaccuracy was greatest when small portion sizes were measured, with a tendency for these to be over-estimated, that is, excessive kibbles weighed out. These findings raise concerns about the potential for overfeeding when using measuring cups as a feeding aid, especially with small dogs and cats. Over the long-term, such inaccuracies with the use of measuring cups could pre-dispose to gradual and insidious weight gain ultimately leading to obesity. Furthermore, as precision and accuracy are both required when feeding a weight loss diet, inaccuracy when using a measuring cup might lead to failure to lose weight. Nonetheless, it should be emphasised that development of obesity is typically multi-factorial, and thus use of measuring cups would likely be only one factor in the tendency to weight gain. First, energy requirements can vary markedly amongst dogs, ranging from 70 to 150% of textbook means (NRC, 2006), and such variability may be more significant than the variability encountered when using a measuring cup. Furthermore, energy expenditure is also important and some animals may balance variable intake by varying their energy expenditure over time. Moreover, other factors relating to feeding behaviour have also been implicated, such as excessive feeding of table scraps and treats (Robertson, 1996, 1999). Finally, the potential for either weight gain or weight loss as a result of inaccuracies in portion size, would not necessarily be a concern if body weight and condition were monitored regularly, because subtle weight gain would be noted and the diet adjusted accordingly. However, a recent UK study has demonstrated that many pet dogs are inconsistently weighed at veterinary practices and rarely have a condition score performed (German and Morgan, 2008). One limitation of the current study was the fact that all participants were either veterinarians or employed by a pet food company. Such individuals would be expected to have more experience in feeding companion animals than the general pet-owning population and, thus, the study may have underestimated the true imprecision and inaccuracy that exits. A further study limitation is the fact that the Journal of Animal Physiology and Animal Nutrition. ª 2010 Blackwell Verlag GmbH 371

5 Imprecision when using measuring cups for dried pet food A. J. German et al. study was short-term in that measurements were performed back to back, and variation over time (e.g. over a number of days or weeks) was not assessed. Moreover, although a range of kibble sizes and varieties were used, it was not possible to test all shapes of kibble. Thus, it is likely that real-life variability is likely to be greater than predicted in the current study. Therefore, additional longer-term studies would be advisable using participants from a variety of backgrounds and incorporating a wider range of kibble types. Although each cup was already marked with its own-graded scale, we were not certain whether these were actually accurate. As a result, we chose to calibrate each cup, prior to its use, with a portion of food weighed accurately on calibrated electronic scales, so that any resulting inaccuracies would be the result of errors in the measuring rather than errors in the graded scale on the cup. Interestingly, the graded scales of four of six cups were inaccurate, and all would have lead to over-estimates in the portion size, by up to as much as 40% (data not shown). Such inaccuracies are likely to add further imprecision for owners using measuring cups. Given the concerns that have been highlighted over measuring cups, alternative approaches may be necessary to insure accurate food dosing, especially when portion size is small. One approach would be for food to be weighed with kitchen scales and this was the reference method used for the current study. This would be the preferred approach of two of the authors (AJG, SLH) and is the method used for all cases undergoing weight management at their clinic. However, as some owners may find the use of scales laborious, alternative strategies should be considered. One approach would be to use pre-weighed portions, which have recently become commercially available for a feline weight management diet (current study, diet 2). Such sachets are demonstrably more accurate (median 0.5% of 10 measurements) and precise (between-portion CV 1%) (data not shown). Furthermore, owners of obese cats preferred pre-weighed sachets to using a measuring cup in a recent weight loss study (Bissot et al., 2010). Unfortunately, such a strategy is currently limited by its lack of availability and additional cost. A final solution, therefore, would be for owners to calibrate their own cups for the desired type of food and portion size. In this respect, the daily food portion could be initially weighed on scales and transferred to the cup. The walls of the cup could then be cut so that they are level with the top of the food portion. Although imprecision would likely still exist, errors would be reduced because it would be more difficult to over-fill. Such a strategy, unfortunately, requires more work on the part of the owner, and a new cup would need to be used, and individually calibrated, every time the amount of food in the daily ration changed. In conclusion, the current study has demonstrated imprecision and inaccuracy when using measuring cups for weighing out extruded dry food for companion animals. Such imprecision has the potential to lead to problems such as weight gain and failure with a weight programme. Conflict of interest statement Royal Canin manufactured the diets used in this study. Four of the authors (Bryner, Bouldoires, Deboise, Biourge) are employees of Royal Canin. Royal Canin also financially supports the posts of A. J. German and S. L. Holden, at the University of Liverpool. P. J. Morris is an employee of Mars All of these authors were involved in the study design, in the collection, analysis and interpretation of data, in the writing of the manuscript, and in the decision to submit the manuscript for publication. Acknowledgements AJG s senior lectureship is funded by Royal Canin. The study was funded by Royal Canin. The authors wish to thank the 10 participants of study 4. References Bissot, T.; Servet, E.; Vidal, S.; Deboise, M.; Sergheraert, R.; Egron, G.; Hugonnard, M.; Heath, S. E.; Biourge, V.; German, A. J., 2010: Novel dietary strategies can improve the outcome of weight loss programmes in obese client-owned cats. Journal of Feline Medicine and Surgery 12, German, A. J., 2006: The growing problem of obesity in dogs and cats. Journal of Nutrition 136, 1940S 1946S. German, A. J.; Morgan, L. E., 2008: How often do veterinary surgeons assess the bodyweight and body condition of dogs? Veterinary Record 163, German, A. J.; Holden, S. L.; Bissot, T.; Hackett, R. M.; Biourge, V., 2007: Dietary energy restriction and successful weight loss in obese client-owned dogs. Journal of Veterinary Internal Medicine 21, Godwin, S. L.; Chambers, S. L.; Clevland, L., 2004: Accuracy of reporting dietary intake using various 372 Journal of Animal Physiology and Animal Nutrition. ª 2010 Blackwell Verlag GmbH

6 A. J. German et al. Imprecision when using measuring cups for dried pet food portion-size aids in-person and via telephone. Journal of the American Dietetic Association 104, Kraft, W., 1998: Geriatrics in canine and feline internal medicine. European Journal of Medical Research 3, Laflamme, D. P.; Abood, S. K.; Fascetti, A. J.; Fleeman, L. M.; Freeman, L. M.; Michel, K. E.; Bauer, C.; Kemp, B. L. E.; Van Doren, J. R.; Willoughby, K. N., 2008: Pet feeding practices of dog and cat owners in the United States and Australia. Journal of the American Veterinary Medical Association 232, Lund, E. M.; Armstrong, P. J.; Kirk, C. A.; Klausner, J. S., 2005: Prevalence and risk factors for obesity in adult cats from private US veterinary practices. International Journal of Research in Veterinary Medicine 3, National Research Council, Nutrient Requirements of Dogs and Cats. National Academy Press, Washington DC, Robertson, I. D., 1996: A survey of diets offered to dogs in metropolitan Perth, Western Australia. Australian Veterinary Journal 73, Robertson, I. D., 1999: The influence of diet and other factors on owner-perceived obesity in privately owned cats from metropolitan Perth, Western Australia. Preventative Veterinary Medicine 40, Robertson, I. D., 2003: The association of exercise, diet and other factors with owner-perceived obesity in privately owned dogs from metropolitan Perth, WA. Preventative Veterinary Medicine 58, Journal of Animal Physiology and Animal Nutrition. ª 2010 Blackwell Verlag GmbH 373