Hot dogs: Fever of unknown origin (Proceedings)

In human medicine, fever of unknown origin (FUO) is defined as pyrexia of greater than two to three weeks duration (i.e. sufficient time for self-limiting infections to resolve) during which repeat physical examinations and standard diagnostic testing have failed to reveal an underlying cause. Additional criteria are also occasionally suggested, such as presence of non-specific signs of illness, or hospitalization to ensure patient compliance with antibiotic administration. Similar specific criteria have not been accepted by veterinary internists; however, using similar guidelines, most internists will define patients as having a fever of unknown origin if pyrexia persists beyond one to two weeks, physical examination prior to referral does not reveal any abnormalities, and the minimum data base fails to localize an underlying disease process to a particular organ system.

Differential Diagnosis

In theory, any inflammatory disease, including infectious, immune-mediated, and neoplastic diseases, may stimulate sufficient tumor necrosis factor-α and interleukin-1 and -6 production to result in persistent fevers. However, diseases which result in the official diagnosis of FUO are typically indolent, slowly progressive, or relatively occult. Two retrospective studies, published ten years apart, have reviewed the final diagnoses ultimately made in dogs referred to veterinary teaching hospitals for fever of unknown origin using approximately the same criteria as those mentioned above. Selected results in these populations of dogs included:

It is important to mention that both of these studies were performed in the United Kingdom, so the relative prevalence of various infectious diseases would be expected to alter results of similar studies in the United States; in truth, results from different sites even within North America would likely vary widely as well. Nevertheless, despite the variation in time and geography between these sites and the dog and cat populations most of the audience members here evaluate, it is clear that when a diagnosis can be made, the most likely causes of fever of unknown origin are most likely immune mediated diseases (with joint and CNS diseases predominating) or infectious diseases (with no single occult site of infection being more likely. Neoplasia accounts for less than 10% of fever of unknown origin, with leukemia being most likely.

Diagnostic Approach

The most important diagnostic test in the work-up of an animal with a fever of unknown origin is the methodical, thorough physical examination. Diagnostic testing to rule-in or –out even a small portion of the possible causes of fever is usually beyond the financial means of most owners; therefore, targeting diagnostic tests to exclude differentials for a higher-yield problem if at all possible is highly desirable. A full physical examination should include fundic examination, deep palpation of long bones and the spine to see if pain can be elicited, neurologic and orthopedic examinations, rectal palpation, and palpation of all joints. If a specific abnormality is found then a more refined problem list can be generated and work-up proceeds from there.

If no additional abnormal clinical findings other than fever are identified, then diagnostic tests should be selected which will help exclude large numbers of diseases, exclude diseases which are common causes of FUO, or which are over-represented in a given geographic area. A full minimum database (i.e. complete blood count, serum chemistry panel, and urinalysis) will help in many cases to target a specific organ system and as with the physical examination, hopefully lead to a higher-yield problem. The CBC should include a manual differential cell count and a cytologist's assessment of cell morphology; this is important because in some cases intra-cellular organisms (bacteria, rickettsial organisms) may be noted, or a 'leukocytosis' may actually be a chronic leukemia. Urinalysis should include both urine dipstick and sediment examination. Next-tier diagnostic testing most commonly includes survey thoracic and abdominal radiographs, urine culture, and FeLV and FIV testing in cats. I prefer radiographs over ultrasonography for evaluation of the abdomen when no abnormality has been noted on physical examination or bloodwork because the former allow for a 'one-shot,' global assessment; with abdominal ultrasound, the quality of the study is much more dependent on operator skill, the entire abdomen cannot be assessed at once, and there is a much higher likelihood that a focal image will never make it into the region being evaluated. Urine culture should be performed in all patients regardless of urine sediment or specific gravity results—chronic pyelonephritis or prostatitis may result in intermittent shedding of organisms into urine without any changes to the urinalysis, and in some cases identical bacterial isolates may be found in urine and blood of septic animals. FeLV and FIV are rare causes of FUO, but a positive cat with a fever may be considered more likely to be immunocompromised, and therefore secondary infections should be investigated, as well as an increased likelihood of lymphoproliferative neoplasia and autoimmune hematologic disease.

Unfortunately, if these first- and second-tier diagnostic tests fail to identify any additional problems to work-up and as a clinician you are still faced only with FUO on your problem list, then the third-tier list of possible diagnostic tests becomes much less clear-cut. Choice of tests should rely on patient signalment (young vs old; breed-related diseases) and geographic location and season (tick-borne diseases, fungal diseases, leptospirosis). Tests which should be considered based on those results reported in the two retrospectives discussed earlier include arthrocentesis (whether or not joint effusion or pain is present), abdominal ultrasound, blood cultures (even in the absence of a left-shift or toxic changes, and especially if a heart murmur is present; optimally, blood should be collected at times the patient is febrile), aspiration of lymph nodes or other parenchymal organs (perform if enlargement is suspected at all or if imaging suggests any change in echogenicity), infectious disease testing (e.g. Lyme, Ehrlichia spp., Anaplasma spp., Leptospira sp.). The final tier of diagnostic tests would include bone marrow aspiration or biopsy, 'autoimmune panels' even though other tests would not support IMHA, ITP, SLE, etc., echocardiogram, general anesthesia with CSF fluid collection, etc. It cannot be stressed enough that testing should be logical and proceed in a step-wise fashion in order to prevent exhaustion of owner funds, and the most important diagnostic test—the physical examination—cannot be repeated often enough in FUO patients.

Treatment

Treatment of fever because of the fear that high body temperature itself will result in secondary organ damage is controversial. In people, cooling is usually instituted when core body temperature is 104 F or greater. Because normal cat and dog body temperature is typically higher, most internists usually advocate withholding cooling therapy until temperature reaches 105 F. Optimal temperature-lowing therapy is unclear, but if the underlying cause of pyrexia is unknown, intravenous fluids and mechanical cooling (i.e. fans) should be chosen over antipyretic drugs. In order to increase the cooling effect of intravenous fluids, a portion of the lines can be run through ice water. Antipyretic drugs, which may blunt the beneficial and necessary effects of some fevers, should only be considered once other methods are insufficient to lower body temperature to a 'safe' range. Non-steroidal anti-inflammatory drugs are effective at lowering temperature, but carry the risk of renal and hepatic toxicity in patients that are already at increased risk due to fever-induced dehydration and hyperthermia-associated organ damage.

Many owners will also inquire at some point during diagnostic evaluation of patients with FUO as to whether empiric therapy with antibiotics or glucocorticoids can be attempted instead of spending additional money on diagnostic tests. I believe that for many pets owners this option truly should be given serious consideration. Arguments against empiric therapy include:

     • Fever in many patients with FUO is waxing and waning, and therefore what may be perceived as a 'response to therapy' actually represents the natural rhythm of the fever. Repeat trials of the same drug may therefore be inappropriately attempted based on false observations.

     • Without knowing what infectious agent is being treated or where the focus of infection truly is, suboptimal drug choice or inappropriate length of treatment of time may result in development of resistance.

     • Even worse than a lack of response to treatment is the possibility that further irreversible damage is occurring while time passes that could have been used for further diagnostic evaluation.

     • In the more recent FUO retrospective described above (Battersby et al), treatment with any drug prior to referral prolonged the median time until a diagnosis was made in dogs with FUO.

     • In this same study, treatment significantly decreases the likelihood of fever still being present at the time of referral, again making it unclear whether animals are truly improved or this represents a natural nadir in body temperature.

     • In the event that the underlying cause of FUO is a lymphoproliferative neoplasia (i.e. lymphoma), treatment may put these dogs into remission. Although this may be an apparent advantage, prednisone alone is not the optimal therapy for lymphoma—dogs with lymphoma treated with prednisone alone have a median survival of only 3 months, as opposed to longer survival rates with combination chemotherapy. Additionally, administration of prednisone prior to the diagnosis of lymphoma worsens long-term prognosis, as glucocorticoids induce multi-drug resistance against many other chemotherapeutic agents.

Ultimately, if owners still would like to attempt empiric therapy despite being informed of the risks outlined above, I will select a broad-spectrum drug or combination of drugs. If I note an improvement in the expected length of time (i.e. 3-5 days), then I will treat with this drug under the assumption that the patient has a nidus/complicated infection: treatment should be for a minimum of 6 weeks. If fever recurs after antibiotic therapy ends then I assume that there is likely a nidus of infection, but now have a clearer idea of what type of organism I am searching for based on the drug used. Only once patients fail an antibiotic trial will I consider a glucocorticoid therapy trial. Immune-mediated fever is rare; however, fevers will often resolve within 24-48 hours of starting therapy.