Recently, metabolic syndrome (MS) has gained attention in human medicine given its associations with development of diabetes mellitus and cardiovascular diseases [1]. Central obesity (determined by waist circumference) is critical to its definition, in conjunction with dyslipidaemia, hypertension and glucose intolerance. This suggests that MS is a complex cluster of metabolic risk factors that together may predispose to development of secondary diseases [1]. In humans, MS doubles the risk of cardiovascular disease whilst the risk of diabetes mellitus is increased five-fold [2]. More recently, proteinuria (or microalbuminuria) has been observed in patients with MS, suggesting altered renal function [3–5].

In veterinary medicine, equine MS is well described [6–8], and known to be a risk factor for laminitis amongst other pathologies [9, 10]. Canine obesity is associated with the development of insulin resistance, altered lipid profiles, and mild hypertension, which are ameliorated by weight loss [11–13]. Furthermore, overweight dogs are more likely to suffer from diabetes mellitus [14], whilst lifelong overfeeding leads to being overweight, metabolic disturbances and decreased lifespan [15]. Although laboratory dogs, rendered obese by overfeeding, have many features that resemble human MS [13], the authors are not aware of any studies examining the existence of MS in obese dogs. Therefore, the objective of the present study was to determine whether canine obesity-related metabolic dysfunction (ORMD, using a definition modified from that used for human MS) could be identified in pet dogs with naturally-occurring obesity, and whether it correlated with specific patient characteristics (e.g. signalment, body fat content), metabolic (e.g. adiponectin, insulin) and inflammatory (e.g. high-sensitivity C-reactive protein; hsCRP) biomarkers. We also determined the effect of weight loss on MS and its consequences.

Human MS is now well recognized and predisposes to cardiovascular diseases and type 2 diabetes mellitus [1]. In equine MS, obesity, regional adiposity, insulin resistance, hypertriglyceridaemia and hyperleptinaemia are part of the definition, and it is a risk factor for laminitis, altered reproductive function, and seasonal alterations in arterial blood pressure [6, 9]. In the current study, we classified pet dogs with naturally-occurring obesity on the basis of presence of ORMD (using a modification of criteria used for human metabolic syndrome). We then examined the effect of weight loss on these metabolic criteria, and sought to identify factors that were potentially associated with the syndrome. The criteria for ORMD were met in approximately 20% of obese patients pre-weight loss. The presence of ORMD was not associated with total fat mass, as measured by DEXA, but was associated with increased and decreased plasma insulin and adiponectin concentrations, respectively. These findings suggest that defining obese dogs on the basis of their metabolic status may have some merit, although further work is now required to determine the true significance of ORMD in terms of disease risk and outcome. Weight loss was associated with decreased body fat mass, BCS, SBP, circulating lipid concentrations, and plasma insulin concentration, whilst circulating adiponectin increased. These findings are similar to those reported in many studies of canine obesity [11–13, 28], and suggest that the study population was representative. Nonetheless, the study was small and extending the work further studies would now be desirable.

We chose to base our ORMD definition on the guidelines of the International Diabetes Federation [22], although other guidelines are available including those of the World Health Organization [29], the European Group for the Study of Insulin Resistance [30], the National Cholesterol Education Program [31], and a combined statement from the American Heart Association and National Heart, Lung, and Blood Institute [32]. The main reason for this was that the criteria used could readily be adapted to produce a practical method in dogs. For each criterion used, and in a similar manner to humans, either the upper limit of the respective laboratory reference ranges (e.g. cholesterol, triglyceride, and glucose), or internationally accepted criteria, above which the parameter is borderline (e.g. SBP) [23] was chosen as the cut-off. Whilst many of the criteria were identical to those used in the human system, others were substituted for similar parameters thought to be more relevant for dogs. The main advantages of such an approach were that all chosen parameters are already in 2widespread clinical use and techniques for measurement are better validated, thereby making the whole system more accessible for practicing veterinarians. For example, we replaced waist circumference (a human measure of central obesity), with BCS, since the significance of central obesity has not been studied in dogs. Given differences in canine anatomy, the human measures of central adiposity are not likely to be appropriate for dogs, and developing a clinical measure of abdominal obesity in dogs (e.g. waist circumference) would be challenging given the wide variability in size and shape amongst breeds. In contrast, differences in BCS are associated with both disease risk and decreased longevity [14, 15]. Nonetheless, it may be necessary to modify these criteria in the future, if a practical method of measuring central obesity can be validated.

In a similar manner, total plasma cholesterol concentration was used in place of HDL-cholesterol. The main reason for this modification was the fact that lipid profiles differ between dogs and humans, with humans demonstrating an ‘LDL pattern’, whilst HDL is the dominant cholesterol type in dogs [33]. Further, alterations in lipoprotein profiles differ in human and canine obesity: obese humans display increased LDL-cholesterol and decreased HDL-cholesterol [1], whereas the circulating concentrations of both LDL and HDL-cholesterol are increased in obese dogs [11, 34]. Thus, the use of total cholesterol, rather than cholesterol fractions, is more logical for the latter species.

Based upon definitions of 4 (i.e. SBP, cholesterol, triglyceride and glucose) parameters in addition to BCS, approximately 20% of the obese dogs of the study were classified with ORMD, respectively. This suggests that the prevalence of MS in obese dogs is somewhat less than for humans where prevalence is typically 22-28% and 50-60% in overweight and obese patients, respectively [35]. The reasons for such a difference are not known, and further elucidation of the underlying mechanisms of ORMD is recommended for comparative purposes.

In order to determine the significance of ORMD, we chose to assess a variety of other parameters including fasting insulin concentration, plasma adiponectin concentration, CRP and adiposity, as determined by DEXA. When the obese dogs of the present study were subdivided on presence or absence of ORMD prior to weight loss, only plasma adiponectin and insulin concentrations differed. Adiponectin was approximately twofold less, and insulin approximately twofold greater, in the ORMD group, and both were negatively correlated with one another. This finding is similar to that described in the human literature on MS [36]. In man, one of the major obesity and MS outcomes is insulin resistance and type 2 diabetes mellitus [1], and the risk of this condition is increased fivefold when MS is present [2]. The exact pathogenetic link between obesity and insulin resistance has not yet been fully elucidated [37]. However, adiponectin is known to have an insulin-sensitizing effect, acting through the AMP-activated protein kinase [38], so that the association with MS may (at least in part) be explained by the decreased adiponectin concentration that accompanies obesity. Adiponectin may also have anti-inflammatory effects, such that decreased adiponectin concentrations are associated with the increased risk of inflammation [39]. Thus, hypoadiponectinaemia observed in human MS may be responsible for development of secondary diseases due to increased susceptibility to inflammation and insulin resistance. Interestingly, no difference was noted in hsCRP, either when comparisons were made before and after weight loss, or when obese dogs were categorized as either having or nor having concurrent ORMD. This may suggest that associations amongst adiponectin, MS and obesity are independent of the function of this particular acute phase protein. Further, these findings are different from some previous studies examining CRP concentrations in obese dogs, where increased [12, 40] or decreased [13] concentrations have been seen. The reasons for such differences are not entirely clear, but may have resulted from differences in the test populations and assay used. Most notably, our work utilized a high-sensitivity assay recently validated for dogs [26], and such assays are thought to be more reliable in humans [41]. The findings regarding adiponectin concentrations in the current study are different from some [12, 42] but not other [43] studies. Again, the reasons for this are not clear but similar explanations would be feasible, namely that this related to population differences or assay type. In the present study, we used a high sensitivity human adiponectin ELISA assay previously validated for use in dogs [34]. In this assay, human calibrators were substituted for species-specific standards, in order to achieve similar affinity of antiserum against standards; this ensures that better differentiation between samples with greater and lesser adiponectin concentrations [34]. Nonetheless, high and low molecular weight adiponectin species were not measured in the current study, and these may differ in importance is development of obesity-associated consequences [44]. Thus, the true significance of these findings requires further study.

The presence of increased plasma insulin concentration, and decreased plasma adiponectin concentration implies physiological consequences to ORMD. However, in order to determine its true significance, further investigations would be needed examining other biomarkers and also clinical consequences. For example human MS is associated with dysregulated fatty acid metabolism [45], cardiac and vascular functional derangements [46], and hepatic manifestations such as non-alcoholic fatty liver disease [47]. Therefore, future studies could assess alterations in a variety of biomarkers in a prospective population of dogs. Epidemiological studies could also be considered, as a means of determining potential disease associations and, ultimately, risk of death in ORMD.

Another interesting observation from the current study was the fact that no differences in fat mass were identified between dogs classified with or without MS. This implies that canine obesity does not inevitably lead to metabolic dysfunction, and is similar to findings in man, where some obese individuals are determined to be metabolically healthy [48]. These patients are not insulin resistant, are normotensive, and have normal plasma concentrations of triglyceride, glucose, high-sensitivity C-reactive protein (hsCRP), and high-density lipoprotein (HDL) cholesterol. The reasons why some dogs may be protected from developing metabolic derangements of obesity are not known and require further study. Possible explanations, not examined in the current study include the age of onset obesity, time taken to become obese, duration of obesity prior to weight loss, and the type of diet fed during the obese stage. As a result, further work is required to determine the mechanisms involved in the development of ORMD.

As is often the case, this study has limitations that should be considered. The main study limitation was that client-owner out bred dogs were used, which undoubtedly added to study variability. For instance, environment, diet, exercise and husbandry were variable. Most notable is the issue of diet, since the amount of food eaten, and type of diet could have influenced the pre-weight-loss metabolic parameters, particularly cholesterol and triglycerides. Unfortunately, the information obtained from the owner regarding diet fed before enrolment was vague and incomplete. A range of foods was fed, including commercial pet food, treats, and human food. Further, owners frequently fed many different diets, rarely measured the amount fed out accurately, and did not record the extra food fed. As a result, it was not possible to generate a meaningful record of pre-weight-loss feeding that could be used in this study. Further, information was unclear as to the exact duration of obesity in many cases and, again, this may have influenced the results obtained. Moreover, ethical limitations meant that we were unable to perform more invasive but gold standard assessments of insulin sensitivity such as hyperinsulinaemic, euglycaemic clamps or minimal model analysis to determine insulin sensitivity [49, 50]. That said, the obesity had developed naturally and been longstanding (e.g. >12 months) in most cases, which is arguably more representative of the at-risk population of interest. Nonetheless, it would be sensible to consider further studies as a means of elucidating the underlying mechanisms of ORMD more precisely.

In conclusion, this study has described that up to a third of obese dogs do suffer from ORMD, which is characterized by hypoadiponectinaemia and hyperinsulinaemia. This study can form the basis of further investigations to determine pathogenetic mechanisms and the health significance for dogs, in terms of disease associations and outcomes of weight loss.