Hyperthyroidism is a commonly encountered disease of domestic cats. The nature of the disease means that frequent monitoring of thyroid values (typically the “total thyroxine” or “TT4”) is necessary in order to monitor success of therapy and adjust doses of medications. Traditionally, quantification of thyroid values has required the submission of blood samples to external reference labs. In recent years several companies have developed in house assays for TT4, which may in theory allow for faster diagnosis of thyroidal disease, and more convenient trending of values for owners. This may be of significant benefit when dealing with critically ill patients who have no known history of thyroid disease, where rapid results may be very helpful.
The purpose of this study was to determine the accuracy of a particular in-house T4 test (The IDEXX Catalyst Total T4 immunoassay) and to compare its results with those preformed at a commercial reference lab. The first portion of the study was intended to validate the performance of the test in feline sera, and the second was a prospective cross sectional study utilizing cats presenting to a private referral practice.
In the first portion of this paper, the precision of the test was determined by the analysis of multiple replicates of various serum pools to determine within run and intra-assay variability. Serial dilutions of known amounts of thyroxine in feline serum were analyzed to examine linearity.
The cross-sectional portion investigated 157 cats over a 5 month period at a private referral practice in New York. 127 were hyperthyroid and 30 were previously treated with I131. Blood samples from each cat were collected, centrifuged to produce serum, and either analyzed in house or held at 4C until analysis at a reference lab. Cats were classified as hyperthyroid or euthyroid based on scintigraphy, and as “low”, “hi”, or “within reference interval” based on in house and laboratory TT4 results.
The intra- and inter- assay coefficients of variation for the in-house machine were found to be very low (means of 2.8% and 6.6% respectively). The assay was linear from 0-150nmol/L (the range most cats are likely to be found in) and within 9% variation from 150-250nmol/L.
There was no significant difference found in the median, range, and IQR between the values calculated by the in house and reference lab machines. Fixed or proportional biases between testing methods were not found, however the difference in measured values tended to increase as the TT4 increased. The 95% limits of agreement for the percent difference in normalized values was +/- 30%. What this amounts to is that for any given value on one analysis method, there is a 95% chance that the other method will be within +/- 30%. For example, a cat measuring a value of 25nmol/L in house will have a 95% chance of measuring between 17.5nmol/L and 32.5nmol/L.
When classifying cats as low, normal, or high T4, there was a 96.8% agreement between analysis methods. Values near the decision thresholds on either machine were most likely to result in discordant classification.
The authors concluded that the in house analysis of TT4 provided reliable and accurate values for TT4, comparable to the reference lab. It appears that the in house testing is largely equivalent to the reference lab for classification of cats as hypo- eu- or hyper- thyroid, however borderline cases may need to be re-tested or confirmed by other methods. While there was a relatively good correlation between in house and reference lab values, the range may be too large for practical trending between machines (ie a value of 50nmol/L in house may read between 35nmol/L and 65nmol/L at the reference lab).
One concern in this study was the inclusion of employees of the reference lab among the authors, which may be a source of potential bias. Future, independent validation studies would be recommended to confirm results. Further work on various sample types would also be recommended, as analysis on plasma is also possible. (MRK)