Cobalamin is a water-soluble vitamin with important intermediary metabolic functions. The importance of hypocobalaminaemia has long been recognised: deficiency can have neurological, haematological, cardiovascular and catabolic manifestations in man and encephalopathy, myelopathy, anaemia, anorexia, cold-intolerance and failure to thrive have been reported in cats [1],[2]. Hypocobalaminaemia has been reported in cats with gastrointestinal disease [3],[4], exocrine pancreatic insufficiency [5] and hyperthyroidism [6]. Decreased serum cobalamin can help to localise gastrointestinal disease [7] and is a negative prognostic indicator for alimentary lymphoma in cats [8].

Conversely, increased circulating cobalamin concentrations are generally attributed either to increased dietary intake or to previous supplementation and, as a consequence, are usually ignored. However, in the absence of supplementation, hypercobalaminaemia reflects altered metabolism [9]. In humans, hypercobalaminaemia is associated with myeloproliferative disorders [10]-[12] and with a wide range of solid neoplasms including liver, mammary, prostatic, pulmonary, gastric and pancreatic tumours [9],[13]. It has, therefore, been suggested to be a non-specific marker for cancer [13]. Associations have also been made with various non-neoplastic liver diseases [9],[14],[15] and with kidney disease in some studies [16],[17]. To date, its significance has not been investigated in veterinary species. Therefore, the aim of this study was to determine whether disease associations exist for hypercobalaminaemia in cats.

This study is the first to examine the significance of increased serum cobalamin in cats and the findings suggest that hypercobalaminaemic cats were more likely to be diagnosed with either a solid neoplasm or liver disease. Similar patterns are described in the human literature [9],[13]-[15],[19]-[21], and there is also a recent, as of yet, unpublished study evaluating cobalamin status in dogs with liver disease [22]. Therefore, the current assumption that hypercobalaminaemia does not have clinical significance in veterinary species should be questioned, and prospective studies should be considered to confirm these preliminary findings.

A similar association between hypercobalaminaemia and the presence of a solid neoplasm has been recognised in man [23], with the strongest associations for liver tumours and hepatic metastases [14],[15],[17],[24],[25]. Recently, associations with a wider variety of malignancies have been established [9],[13]. Unfortunately, in the current study, there were insufficient cases presenting with a solid neoplasm to enable associations with any particular tumour type. The reason for the hypercobalaminaemia in the cats with a tumour in the current study is not known. In humans, various mechanisms have been proposed, including the direct production of cobalamin binding proteins (transcobalamins) [9], upregulation of folate-centred metabolic pathways and production of cobalamin autoantibodies [13],[26]. In humans, the association between hypercobalaminaemia and neoplasia was first reported for haematopoietic malignancies, particularly chronic myeloid leukaemia [10]-[12]. Increased production of binding proteins (specifically, a transcobalamin derived from granulocyte membranes) is proposed to occur as a result of leucocytosis [21],[27]. No association with either haematogenous neoplasms or leucocytosis was detected in this study. This might be due to differences in human and feline transcobalamins, namely that cats lack transcobalamin I [28], or might reflect the type of haematogenous neoplasms encountered. The vast majority in this study were lymphoma (32/33) and there is conflicting evidence regarding an association with lymphoma in man. Some [17] but not other [11],[15],[16],[21] studies have suggested an association with hypercobalaminaemia. Additionally, as many of the cats with lymphoma had ileal involvement (the site of intestinal cobalamin absorption), this would likely be confounding, even if a true association with lymphoma does exist.

Hypercobalaminaemic cats in this study were also more likely to be diagnosed with liver disease, and similar associations are seen in humans [9],[14],[15]. Possible mechanisms include leakage into the circulation following hepatocellular injury and impaired hepatic uptake of cobalamin. An association with hepatocellular injury is less likely, given that there was no correlation between serum ALT activity and cobalamin concentration. The association between hypercobalaminaemia and hepatobiliary disease in cats is intriguing, since hepatobiliary disease is also associated with hypocobalaminaemia in this species [3],[4],[7],[28]-[30]. This discrepancy might reflect the tendency for concurrent gastrointestinal, pancreatic and hepatobiliary pathology in cats, having counteracting effects on serum cobalamin. Indeed, only 9 of the 17 hypercobalaminaemic cats with confirmed liver disease in this study had a disease confined to the hepatobiliary system. Alternatively, it could reflect differing effects of specific hepatobiliary diseases on cobalamin metabolism. A final observation, with regard to hepatic disease was the fact that one of the cats had portal vein hypoplasia. The reason why hypercobalaminaemia would be associated with a congenital disease is not known. One possibility might be that there are effects on cobalamin metabolism as a result of changes in hepatic function caused by the PVH. However, further work would be required to determine this.

Associations between kidney disease and hypercobalaminaemia have been identified in some human studies [16],[17], and decreased clearance of binding proteins from the circulation has been postulated [9]. Such an association has not been noted in all human studies [21], and there was no significant difference in serum cobalamin between dogs with chronic kidney disease and controls [31]. Although no such association was found in the current study, only six cats with kidney disease were included. Besides associations with disease, cats of pedigree breeds were more likely to be hypercobalaminaemic. In another study, hypocobalaminaemia was more frequently observed in pedigree breeds [28]. The reason for these findings is unknown.

This study has several limitations. First, methylmalonic acid (MMA) was not measured in any cat as part of the diagnostic investigations, and such results might have helped to clarify cellular cobalamin status. Although it might have been possible to measure MMA retrospectively, in cats where stored serum was available, variability in sample collection procedures (e.g. differing times between collection and freezing of samples), time and temperature of storage could have made the results less reliable. To address this limitation, we would recommend that a future study consider measuring both cobalamin and MMA in cats where methods of collection, handling and storage are standardised.

A second limitation was that data were collected retrospectively and, as a consequence, we were reliant upon the history taken at the time of presentation. Thus, the information acquired from the owner was variable, and this might have been incomplete, for instance information regarding prior cobalamin supplementation. A further limitation of the retrospective design was the fact that, despite a relatively large cohort of cats, relatively few were diagnosed with the diseases most strongly associated with hypercobalaminaemia in man. Further, even within disease categories, there was great diversity of the disease conditions; this was most marked for the two disease categories where an association with hypercobalaminaemia was noted, namely hepatobiliary diseases and solid neoplasms. To address these concerns, we would now recommend conducting a prospective study to assess cobalamin status in a more defined cohort of cats, i.e. with more uniform disease status. The results of the current study could be useful to determine an appropriate sample size to ensure that the project is not underpowered [32]. Further, a multicentre study or collaborative effort would be recommended to ensure that a larger enough cohort is recruited.

A final limitation is the issue of reference interval, not least given the wide variation in reference intervals amongst diagnostic laboratories for serum cobalamin in cats, despite widespread use of the same analytical assay. Reasons for such laboratory variability have previously been proposed in both cats and people, with regional variability, dietary factors, age and undiagnosed disease in control populations suggested as possible explanations [30]. The upper limit of the reference interval used in the current study is less than in some previously reported studies [33],[34] but comparable to others [3],[35]. This interval was based upon that used at other laboratories using the same assay, and then further validated by measuring serum cobalamin concentrations in at least 20 healthy cats over 1y age. Unfortunately, given the validation work was undertaken at least 10 years ago, full details are no longer available. For future studies, it would be advisable for all laboratories to review their reference ranges for cobalamin, and ensure that it is appropriate.

The current study has suggested possible associations between hypercobalaminaemia and the presence of either solid neoplasia or liver disease in cats. This suggests that veterinarians should pay greater attention to increased serum cobalamin concentrations when investigating sick cats and should consider thorough evaluation for neoplasia or hepatic disease where there is no evidence of prior supplementation.

^aImmulite 2000 Vitamin B12; Siemens Healthcare Diagnostics Inc., Deerfield, IL.

^bStatsDirect version 2.7.9, StatsDirect Ltd, Altrincham, UK.