While the relationships between energy requirements and age(Reference Burger1, Reference Pérez-Comago2) and body weight (BW)(Reference Earle and Smith3) have been determined in domestic cats, these reports utilised the outcomes of a series of studies in only a single colony of cats. Given the increase in obesity levels in companion animals and the number of recent studies that have investigated this phenomenon, meta-analysis may be an appropriate tool to accurately determine the energy requirements of adult cats. The use of meta-analysis is a powerful method by which the results from multiple independent experiments can be combined to produce more robust results than can be obtained from single experiments alone(Reference Phillips4). Combining results from different experiments using statistical meta-analytical methods can reduce the extent that differences in experimental conditions may have an impact on the calculated prediction equations and cause bias. Additionally, meta-analysis can identify factors that may influence these relationships.
Maintenance energy requirements for adult cats have been calculated to be 100 kcal/kg BW− 0 ·67 for normal-weight, and 130 kcal/kg BW− 0 ·4 for heavy cats(5). The National Research Council (NRC)(5) highlighted that the literature reporting maintenance energy requirements in adult cats is highly variable. For example, the effects of age on maintenance energy requirements differ markedly among publications, as does the effects of neutering. These effects may be explained by other factors that were not taken into consideration among studies, such as age of the cat or methodology used to determine the cats' energy requirements.
The objective of the present review was to conduct a meta-analysis on the energy requirements for maintaining BW in adult domestic cats in order to (a) determine predicted changes in energy requirements with weight, and (b) account for factors that may influence these relationships. Initial results have been reported in abstract format in Bermingham et al. (Reference Bermingham, Morris and Hawthorne6). For ease of comparison with the NRC(5), we have reported energy requirements in kcal/kg BW.
Materials and methods
Data classification
Data for the present review on the energy requirements of adult cats were limited to publications reported in English, and published from 1933 to 2009. These search criteria identified forty-two publications (Table 1) with data reporting the amount of energy required to maintain BW of adult cats. A ‘publication’ was defined as a distinct piece of published work, while ‘treatment’ represented a treatment group of cats within the publication (for example, a control or obese group).
Table 1 Literature used for the determination of allometric equations for maintenance energy requirements in domestic cats
BW, body weight; ME, metabolisable energy; FE, feeding experiments; IC, indirect calorimetry; DLW, doubly-labelled water; F, female; M, male; NF, neutered female; NM, neutered male.
* Where data are reported in the literature for feeding studies.
† Data obtained from Table 3.16 in the National Research Council (2006)(5).
‡ Pet cats.
Cats were classed according to sex and neuter status (entire male and female, neutered male and female). Due to an absence of body condition score or body composition assessment in the majority of publications, it was assumed that cat weight was an accurate estimate of whether the cat was lean, normal or overweight. Therefore, cats less than 3 kg in weight were classed as ‘light’, cats between 3·0 and 5·5 kg were classed as ‘normal’, and cats with BW greater than 5·5 kg were classed as ‘heavy’. Thirteen publications reported maintenance energy requirements on a fat-free basis (Table 1). This smaller dataset (thirty-four treatment groups) was used to investigate the effects of age, neuter status and methodology on maintenance energy requirements, when expressed on the basis of fat-free (lean) body mass. Cats were also grouped according to age, with cats less than 2 years (but older than 6 months) classed as ‘young’. Cats were classed as ‘adult’ between the ages of 2 and 7 years, and classed as ‘senior’ if they were older than 7 years of age.
In all cases, cats were domestic shorthaired from either private dwellings studied in a clinical setting, or from research colonies. For the purpose of the present review, energy requirements were synonymous with metabolisable energy (ME) intake (measured from feeding studies where BW was constant ( < 10 % change) over the feeding period), and total energy expenditure was assessed by doubly-labelled water (DLW). Additionally, resting energy expenditure as measured by indirect calorimetry (IC) was used as an estimate of energy requirements. Energy requirements for maintenance were therefore either classed as feeding experiments (FE), DLW or IC, and, where reported, the length of the study was also included (Table 1). Only data from FE that were more than 7 d were included in the present analysis.
The ME content of diets is determined by predictive equations using the ‘Atwater factors’ of 4·0, 9·0 and 4·0 kcal/g for protein, fat and carbohydrate, respectively(Reference Kienzle7). These Atwater factors were modified in 1997(8) to 3·5, 8·5 and 3·5 kcal/g for protein, fat and carbohydrate, respectively. Therefore, in order to compare energy intake in the FE studies pre- and post- the Atwater modifications, the energy intakes reported in the studies published before and including 1997 (n 7; Table 1) were recalculated using the modified Atwater equations unless: (1) the energy content of the diet was measured by bomb calorimeter or (2) the publication stated that modified Atwater equations were used to determine the energy content of the diet.
Statistical analysis
Energy requirements are provided as kcal/kg, rather than as kJ/kg, since the former units are used by the NRC for companion animals.
The initial dataset consisted of forty-two publications, representing 141 treatment groups of cats. Problems associated with fluctuating BW over the experimental period or missing data on BW led to twenty-six treatment groups being removed (Table 1). The revised dataset of 115 treatment groups was used for further analysis, and tested for normality and the presence of outliers. The dataset was found to be within normal statistical limits, and all 115 treatment groups were included in the meta-analysis.
A general linear model (Minitab® statistical software, version 15, 2006; Minitab Inc., State College, PA, USA) was used to predict average energy requirements and determine the influence of factors such as age, sex and neuter status and methodology used to determine energy requirements. The allometric equation of Y = aBWb was used to determine the relationship between BW and ME requirements, where Y is the dependent variable (ME requirement; kcal), a is the constant coefficient, BW is body weight (kg) and b is the allometric exponent(Reference Earle and Smith9). Additionally, a regression model was used to determine the allometric relationship between BW and energy intake by performing a regression of the logarithms of energy intake (kcal/d) and BW (kg)(Reference Earle and Smith9). In order to report data on a ‘kcal/kg BW’ basis, the log data generated by Minitab were back-transformed using the inverse of log-base10. Results, unless otherwise stated, are reported as mean and standard error of the mean.
Results and discussion
The literature review identified forty-two publications, detailing energy requirements on the basis of BW from 141 treatment groups containing 1933 cats. Some publications did not report BW of the cats used(Reference Caldwell10–Reference Taylor, Adams and Neville15) (twenty-six treatment groups), but of those that did, cats identified as having a normal weight predominated with seventy-nine treatment groups. Light (nineteen treatment groups) and heavy cats (seventeen treatment groups) had similar numbers of treatment groups (Table 2). Only 57 % of the treatment groups identified the age of the cats used in their study, with young cats (thirty-four treatment groups) and adult cats (twenty-seven treatment groups) most studied (Table 2). The energy requirement of neutered cats was reported more frequently than those of entire cats (thirty-four v. twenty-six treatment groups, respectively). Feeding studies were the most common method of investigating energy requirements for maintenance of BW (fifty-nine treatment groups), followed by IC (thirty-eight treatment groups), and finally DLW techniques (eighteen treatment groups; see Table 2).
Table 2 Maintenance energy intakes and body weight (BW) for domestic cats
(Mean values, standard errors and ranges)
ME, metabolisable energy; n, number of treatment groups.
Average energy intake required for the maintenance of BW was 222·1 (se 5·3) kcal/d, or 55·6 (se 1·1) kcal/kg BW (115 treatment groups; Table 2). Overall, maintenance energy requirements were significantly affected by weight, sex and neuter status, age and methodology (P < 0·05). As indicated in Table 2, there were large amounts of variation in the reported maintenance energy data, ranging from 122·5 to 401·0 kcal/d, which equated to 29·0–85·5 kcal/kg BW. The large amount of variation observed among studies has previously been attributed to the use of unmodified Atwater factors(5). However, the present dataset was adjusted for differences in Atwater factors, making it likely that other factors, such as sex/neuter status of the cat or age were the cause of the variation in the reported data.
The allometric equation for maintenance energy requirements in all cats was calculated with a back-transformed equation of 77·6 kcal/kg BW− 0·711 (Table 3; adjusted R 2 0·448; 115 treatment groups).
Table 3 Allometric regression equations (Y=aBWb) for maintenance requirements for cats based on the log-log regression of metabolisable energy (ME) intake (kcal/d) and body weight (kg)
Y, energy requirement; a, allometric coefficient; BW, body weight; b, allometric exponent; n, number of treatment groups.
* Back-transformed using the inverse of log-base10.
Effects of body weight
BW ranged from 2·6 to 6·6 kg (Table 2), and averaged 2·8 (se 0·02) kg in light cats (nineteen treatment groups), 4·0 (se 0·07) kg in normal cats (seventy-nine treatment groups) and 6·0 (se 0·07) kg in the heavy cats (seventeen treatment groups) (Table 2). When expressed as a proportion of BW, the energy requirements of light cats (62·8 (se 2·8) kcal/kg BW; nineteen treatment groups) were greater compared with both normal (56·3 (se 1·2) kcal/kg BW; seventy-nine treatment groups) and heavy cats (43·9 (se 1·7) kcal/kg BW; P < 0·05; seventeen treatment groups). Heavy cats have a lower maintenance energy requirement than normal and light cats (P < 0·01; Table 2).
The effects of BW on the allometric equations are indicated in Table 3. When both light- and normal-weight cats were incorporated into the same group to allow direct comparison with the normal and light equation identified by the NRC(5), the allometric equation was 56·2 kcal/kg BW− 0·966 (adjusted R 2 0·479; ninety-eight treatment groups). This value is lower than the NRC recommendations for light and normal cats (100 kcal/BW− 0·67). Both the predicted allometric equations for light- and heavy-weight cats have an adjusted R 2 of 0·000. This may be due to the low number of treatment groups, but more likely the high variation among cats at these BW. This suggests that more data are required in these classes of cats.
Figure 1 shows the actual data reported for domestic cats (kcal/d) compared with the predicted energy requirements using the allometric equations for light and normal cats and heavy cats from the present study and those of the NRC for light and normal cats and heavy cats. The prediction of energy requirements for maintenance of BW from the present study for light and normal cats and heavy cats is consistent with the actual ME requirements reported in the literature (Fig. 1) and suggests that the current NRC(5) recommendations over-predict the energy requirements for light- and normal-weight domestic cats. Over time this additional ME intake could lead to weight gain that may exacerbate specific orthopaedic, endocrine, cardiovascular and neoplastic diseases associated with obesity(Reference German16). However, when using predicted allometric equations either from the present study or those by the NRC, care should be taken to feed an individual cat to its specific requirements based on body condition and exercise levels due to the individual variation that exists among cats.
Fig. 1 Effect of body weight (BW) on the maintenance energy requirements in domestic cats (
; —), compared with the predicted requirements from the present study (– – –) and those predicted by the National Research Council(5) (······). Allometric equations predicted from the present study were 56·2 kcal/kg BW− 0·966 and 131·8 kcal/kg BW− 0·366 for light and normal cats and heavy cats, respectively, in the present study. The corresponding predictive equations from the National Research Council(5) were 100 kcal/kg BW− 0·667 and 130 kcal/kg BW− 0·40 for light and normal cats and heavy cats, respectively.
The amount of variation explained in both the overall, normal and combined light and normal cat groups was low, with an adjusted R 2 < 0·479, which prompted the investigation of factors that may have influenced the results. These factors included sex and neuter status, and methodology used to determine maintenance energy requirements.
The NRC(5) highlighted that cats classed as ‘normal’ may indeed be overweight and that energy requirements for maintenance would be better expressed on the basis of fat-free mass. On average, lean body mass was 3·2 (se 0·1) kg, ranging from 2·0 to 4·2 kg (Table 4). Mean maintenance energy requirements, when expressed on the basis of fat-free (lean) mass, were 68·6 (se 1·9) kcal/kg lean mass, ranging from 42·8 to 90·3 kcal/kg lean mass (thirty-nine treatment groups). The allometric equation generated for this dataset equated to 58·4 kcal/kg lean mass− 1·140 (adjusted R 2 0·690; thirty-four treatment groups). The amount of variation explained by this model was nearly twice that of the allometric equations generated from BW data.
Table 4 Mean maintenance energy intakes for domestic cats based on fat-free (lean) mass
(Mean values, standard errors and ranges)
ME, metabolisable energy; n, number of treatment groups.
* No data were reported in young cats (>2 years) on the basis of fat-free mass.
Age effects
Data showing maintenance energy requirements according to age are presented in Table 2. Forty-seven treatment groups did not have any information on the age of the cats so data from the remaining sixty-eight groups were used to determine maintenance energy requirements (Table 1). Despite this, age was seen to have a significant effect on the daily energy required for maintenance (P < 0·05; Table 2). When expressed as a proportion of BW, daily energy requirements were higher (P < 0·05) in young cats (59·4 (se 2·1) kcal/kg BW; thirty-four treatment groups), compared with adult cats (48·4 (se 1·8) kcal/kg BW; twenty-seven treatment groups). However, senior cats (48·4 (se 4·2) kcal/kg BW; seven treatment groups) had similar energy requirements to both young and adult cats (P>0·05).
Back-transformed allometric equations identified that young cats had daily maintenance energy requirements of 66·1 kcal/BW− 0·878, while adult cats needed 72·4 kcal/BW− 0·703 and senior cats 67·6 kcal/BW− 0·781 (Table 3). The allometric equations determined in the present study for young cats (Table 3) imply that young cats have higher energy requirements than older cats. For young cats, the regression model explained 59·6 % of the variation of the data; however, less variation was explained in adult (adjusted R 2 0·457) and senior (adjusted R 2 0·173) groups. Kienzle et al. (Reference Kienzle, Edtstadtler-Pietsch and Rudnick17) suggested that low exponentials of BW may be due to higher incidences of overweight cats in the population.
When expressed on the basis of fat-free mass, age had a significant influence on maintenance energy requirements (P < 0·05; Table 4). However, on analysis of the data there was no difference (P>0·05) between adult (69·5 (se 2·2) kcal/kg lean mass per d; twenty-seven treatment groups) and senior cats (60·2 (se 4·3) kcal/kg lean mass per d; seven treatment groups), with cats of unknown age status having the highest energy requirements (75·9 (se 3·2) kcal/kg lean mass per d; five treatment groups).
Published data on the effects of age on energy requirements are extremely variable. Burger(Reference Burger1), Harper et al. (Reference Harper, Stack and Watson18), and Peachey et al. (Reference Peachey, Harper and Dawson19) have all shown that energy requirement does not change with age. Similarly, Taylor et al. (Reference Taylor, Adams and Neville15) found no effect of age on ME intake in cats up to 10 years old. However, 12- to 14-year-old cats have been reported as having significantly higher daily energy requirements(Reference Taylor, Adams and Neville15). A retrospective analysis of colony cats showed that daily energy requirements decreased from 74 kcal/kg BW in young adult cats (>2–6 years of age) to 51 kcal/kg BW in older adults (7–15 years of age)(Reference Pérez-Comago2). Similarly, Kienzle et al. (Reference Kienzle, Edtstadtler-Pietsch and Rudnick17) suggested that the daily requirements of cats over 6 years did not change, averaging 130 kcal/kg BW− 0·40 (equivalent to 67 kcal/kg BW in a 3 kg cat). Nguyen et al. (Reference Nguyen, Mariot and Martin20) observed an age effect on energy requirements in castrated male cats. Scarlett et al. (Reference Scarlett, Donoghue and Saidla21) suggested that age effects may be masked by weight effects, as older cats tend to be less overweight than adult and younger cats.
Sex and neuter effects
Neuter status significantly affected daily maintenance energy requirements in cats (Tables 2 and 4; P < 0·001). When expressed as energy intake per d, reproductively entire male cats (280·3 (se 18·3) kcal/d; twelve treatment groups) tended to have higher (P < 0·10) energy requirements than reproductively entire female cats (215·0 (se 12·4) kcal/d; twelve treatment groups). However, when expressed as a proportion of BW this effect was not apparent. This trend was observed in neutered male (55·2 (se 3·2) kcal/kg BW; fourteen treatment groups) and female cats (58·2 (se 2·4) kcal/kg BW; twenty treatment groups). Overall, neutered cats (56·6 (se 2·0) kcal/kg BW; thirty-four treatment groups) required 10·4 % less energy than their reproductively entire counterparts (63·2 (se 2·2) kcal/kg BW; twenty-four treatment groups; P < 0·05).
When expressed on the basis of fat-free mass (kcal/kg fat-free mass), sex had no influence (P>0·05) on maintenance energy requirements (Table 4).
As previously observed for age, the effects of neutering on energy requirements also differed among studies. Kanchuk et al. (Reference Kanchuk, Backus and Calvert22) found that neutering increased daily energy expenditure, while Flynn et al. (Reference Flynn, Hardie and Armstrong23), Martin et al. (Reference Martin, Siliart and Dumon24) and Nguyen et al. (Reference Nguyen, Mariot and Martin20) all observed a decrease in energy expenditure and ME intake, in neutered animals. In contrast, both Fettman et al. (Reference Fettman, Stanton and Banks25) and Nguyen et al. (Reference Nguyen, Dumon and Siliart26) showed no difference in RMR and energy expenditure after neutering. Harper et al. (Reference Harper, Stack and Watson18) observed that the average daily energy requirement for maintaining BW in neutered females was 45 kcal/kg BW. Differences between male and female cats in their response to neutering may also exist. Hoenig & Ferguson(Reference Hoenig and Ferguson27) observed no effect of neutering on male cats; however, there was a significant reduction in daily energy requirements in female cats post-neutering. Additionally, Fettman et al. (Reference Fettman, Stanton and Banks25) showed that, while there were no differences due to neutering, neutered females had higher RMR than neutered males.
Methodology effects
Table 2 indicates the average values for the different methods of determining maintenance energy requirements. The method used to determine maintenance energy requirements had no effect on daily energy requirement; however, there was a significant effect (P < 0·01) on energy requirements when expressed as a proportion of BW. IC (50·0 (se 1·5) kcal/kg BW/d; thirty-eight treatment groups) gave slightly lower values than FE methodology (58·6 (se 1·6) kcal/kg BW/d; fifty-nine treatment groups; P < 0·05) and DLW (59·2 (se 2·4) kcal/kg BW/d; eighteen treatment groups; P < 0·10). However, as it is likely that cats measured using IC would be less active (i.e. in calorimetric chambers), this is not surprising. The allometric equations reported in Table 3 differed slightly from those reported by Nguyen et al. (Reference Nguyen, Dumon and Frenais28) for FE (150 kcal × BW− 0·40), DLW (83 kcal × BW− 0·89) and IC (54 kcal × BW− 0·65).
When expressed on the basis of fat-free mass, methodology influenced maintenance energy requirements (Table 4), with DLW (73·9 (se 2·6) kcal/kg lean mass per d; seventeen treatment groups) values higher (P < 0·10) than both IC (65·4 (se 2·2) kcal/kg lean mass per d; fifteen treatment groups) and FE (60·7 (se 6·0) kcal/kg lean mass per d; six treatment groups).
Points for reflection
Maintenance energy requirements for domestic cats in the present study were determined by a retrospective analysis of the data in the literature, and were affected by factors such as BW, age and the sex of the cat. The equations for normal and light cats (56·2 kcal/kg BW− 0 ·966) and heavy cats (131·8 kcal/kg BW− 0 ·366) in the present study were lower than those reported by the NRC(5). We recommend that three allometric equations should ideally be used when predicting the energy requirements for the maintenance of BW in the cat, namely light (53·7 kcal/kg BW− 1·061), normal (46·8 kcal/kg BW− 1·115) and heavy (131·8 kcal/kg BW− 0 ·366). However, the large variation among treatment groups may partly explain the relatively low relationships (as indicated by R 2) between BW and ME requirements. Therefore, care must be taken when applying these equations to individual animals. Ideally, allometric equations based on body condition score would be more informative than those based on BW. However, with the present dataset there are not enough publications reporting body condition scores and corresponding energy requirements for these to be determined. Authors and reviewers of future publications could look at including a more detailed set of background data in their publications so that these equations can be determined.
The use of lean mass to assess the maintenance energy requirements for domestic cats seemed to be less influenced by factors such as the weight and sex of the animal. While age had an effect on the maintenance energy requirements when expressed on the basis of lean mass, this may be a function of the lack of information in most of the age groups. A limitation of the present study is the assumption that normal cats weigh between 3·0 and 5·5 kg and cats with BW greater than 5·5 kg are heavy. It appears from the present study that lean body mass is the best predictor of maintenance energy requirements. If this is the case, there is a need to report background details concerning the body composition of the cats studied in such trials, so that a more accurate prediction of ME requirements can be made. Determining lean body mass is expensive and may not always be practical for determining the ME requirements for a particular cat population. Therefore the use of other tools (for example, The Feline Body Mass Index™) or using body condition score to assess leanness may be more practical.
For the purpose of the present review, it was assumed that physical activity of cats in colony settings would be relatively low, and resting energy expenditure as measured by IC was used as an accurate estimate of ME requirements. Unpublished results at the Waltham Centre of Pet Nutrition supported this claim. However, from the results of our analysis there was an effect of the methodology used when energy requirements were expressed as a proportion of BW, suggesting that the physical activity of colony cats may be higher than originally suspected.
In the literature, there were little data from older cats, especially those greater than 14 years of age. Similarly, 120 out of 123 treatment groups were colony cats (Table 1), with only the remaining three groups being pet cats. Colony cats included those in normal colony populations(Reference Taylor, Adams and Neville15, Reference Peachey, Harper and Dawson19), or those used to elucidate a wide range of dietary effects(Reference Nguyen, Dumon and Frenais28–Reference Prola, Dobenecker and Kienzle31). Pet cats were adult cats sourced from private households and monitored in the laboratory(Reference Parkman, Michel and Erswell32, Reference Riond, Stiefel and Wenk33). There are no published reports describing the energy expenditure of cats within a domestic home environment. It is likely that cats in the home environment may have even higher energy requirements than their colony counterparts, as they have the freedom to roam, and may require higher energy intake to maintain BW. Finally, all studies were carried out in laboratory environments, thus excluding any possible effects of environmental temperature fluctuation on energy requirements. This forms an important area of investigation for the future. A recent review by Hill(Reference Hill34) highlighted problems trying to extrapolate energy requirement data from healthy colony animals to pets at home which may have more or less exercise, be kept in a warmer or more controlled environment, are more likely to be overweight, and/or are a different breed to the standard colony cat.
Conclusions
The ME requirements reported in the present study were lower than those reported by the NRC, and this may have a significant impact on the feeding levels of cats and longer-term issues with obesity and related diseases. However, there was large variation among cats in the energy requirements for maintenance. Therefore care must be taken when devising the ME requirements for an individual cat, and other measures such as body condition score should be used to determine a more accurate ME requirement. Finally, the review highlighted the lack of data at the ‘extreme’ ends of the spectrum, namely, the lightest and oldest cats and those maintained in outdoor environments. The amount of variation in cat populations, and the scarcity of information available for young, senior and entire cats may mask any true differences in the energy requirements for maintenance of BW.
Acknowledgements
The present study received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
The statistical advice and discussions with John Koolaard (AgResearch Ltd, New Zealand) are gratefully appreciated. The authors acknowledge the insightful conversations with Dr Richard Hill (University of Florida, USA).
The literature review, meta-analysis and writing were conducted by E. N. B. at the Waltham Centre for Pet Nutrition. D. G. T., P. J. M and A. J. H. all contributed equally to data evaluation and writing of the manuscript.
The authors state that there is no conflict of interest.
