Fever of unknown origin (Proceedings)

Hyperthermia versus True Fever

Normal body temperature in the dog and cat ranges from approximately 100 F to 103 F. Not all animals that have a temperature above this range, however, necessarily have a fever. An increased body temperature may be due to true fever, or it may be due to hyperthermia. True fever occurs when the animal's own thermoregulatory mechanisms (controlled by the thermoregulatory center in the hypothalamus) have been ‘reset' to recognize a higher body temperature as desirable, and the animal actually actively tries to achieve and maintain this higher temperature. In contrast, in the hyperthermic patient, the animal's thermoregulatory ‘set point' is still normal, and the animal recognizes that a high body temperature is abnormal, and actively tries to reduce its body temperature back down to normal levels.

The two causes of a high body temperature have entirely different causes, entirely different presentations, and entirely different treatments. It is therefore important, when presented with an animal with a high body temperature, to first of all distinguish between hyperthermia and true fever. Fortunately, it is usually a straightforward process to differentiate fever from hyperthermia, both because the two conditions have different causes, and because they have different presentations.

Hyperthermia only has a limited number of causes. Healthy animals have a well-developed ability to maintain a normal body temperature, and hyperthermia therefore only develops in extreme conditions: conditions of extreme heat and humidity (for example, animals locked in a parked car), extreme exertion or muscle activity (such as over-exertion or excitement on a hot day, seizures due to any cause, or muscle activity associated with conditions such as malignant hyperthermia), increased metabolic activity (hyperthyroidism) or marked impairment of thermoregulatory mechanisms such as panting (for example, laryngeal paralysis or brachycephalic syndrome). Such causes can usually be readily identified based on history and physical examination. Fever, in contrast to hyperthermia, can be caused by a very wide range of infectious, inflammatory, neoplastic and iatrogenic (drug-induced) etiologies, many of which may not be immediately identifiable on initial examination. As a crude general rule, if initial history and physical examination do not identify an obvious cause of hyperthermia, then a patient with an elevated body temperature is far more likely to have true fever rather than hyperthermia.

Animals with hyperthermia also present very differently than animals with fever. Because hyperthermic dogs and cats have a normal thermoregulatory set point, and perceive that they are hotter than they should be, they will be exhibiting behaviors designed to cool body temperature, particularly panting, cutaneous vasodilation (mucus membranes will be bright pink, with a rapid capillary refill time) and cold-seeking (people and horses also sweat). Unless they have extremely elevated body temperatures, hyperthermic animals are typically also alert, although they may be excited, agitated or distressed.

Hyperthermic animals can occasionally have extremely elevated body temperatures, sometimes as high as 110° F, or even higher. Feverish animals, in contrast, despite having a high body temperature, perceive that they are colder than they should be, and therefore exhibit behavior designed to raise body temperature, such as shivering, huddling, cutaneous vasoconstriction, piloerection, heat-seeking, and release of hormones such as epinephrine (causing tachycardia) and thyroxine. Mucus membranes in feverish animals may be pale and cold with a sluggish capillary refill time (due to vasoconstriction), or reddish and congested with a rapid capillary refill time if fever is due to a process that causes vasodilation, such as sepsis. Because the various inflammatory cytokines that serve to raise the thermoregulatory set point also tend to be ‘feel bad' cytokines, animals with a fever also tend to be lethargic, depressed and miserable. In animals with a true fever, body temperature very rarely rises above 106° F or, at most, 107° F.

Hyperthermia versus Fever: Clinical Signs

Table 1

Hyperthermia Fever Demeanor: excited, agitated, distressed Demeanor: lethargic, depressed, anorectic Behavior: panting, vasodilation, sweating, cold-seeking Behavior: shivering, huddling, vasoconstriction, piloerection, heat-seeking Body temperature: up to 110 F or even more Body temperature: Rarely above 106 to 107 F

In true fever, the animal's thermoregulatory set point has been reset (raised) by the effects of a range of fever-inducing substances known as pyrogens. Pyrogens may either be exogenous (introduced from outside the body), or endogenous. The range of potential exogenous pyrogens is enormous, and includes many different organisms (viruses, bacteria, and fungi), antigens, foreign substances, and drugs. Exogenous pyrogens, however, do not usually induce fever directly. Instead, exogenous pyrogens typically induce production of endogenous pyrogens, a rather stereotypical range of substances (particularly inflammatory cytokines such as interleukin-1, interleukin-6, and tumor necrosis factor) produced in response to a wide range of different stimuli, including infection, non-infectious inflammation, immune-mediated disease, neoplasia and pyrogenic drugs. Endogenous pyrogens serve to reset the thermoregulatory set point via the local action of prostaglandins such as prostaglandin E.

Since a wide range of different disease processes and exogenous pyrogens trigger the release of the same set of endogenous pyrogens, the end result, fever due to resetting of the thermoregulatory center, is the same in for many diseases. Furthermore, since many endogenous pyrogens are also potent mediators of systemic inflammation, animals with a fever due to almost any cause will also often exhibit many of the typical features of the systemic inflammatory response, such as lethargy, anorexia, an inflammatory leukogram, anemia of chronic disease, induction of acute phase proteins such as serum amyloid A and C-reactive protein, and mild hypoalbuminemia. These systemic inflammatory changes, therefore, are to be expected in the feverish animal, and are not indicative of any one particular cause of fever.

Causes of Fever

Rather simplistically, some clinicians assume that any patient with a fever is suffering from some form of an infectious process. Although simplistic, this approach has some superficial validity, because undoubtedly the most common cause of fever in small animals is infection. For many patients, unless history and physical examination suggest otherwise, infection (usually bacterial or viral) will be the most likely cause of a fever of one or two days duration. In such patients, waiting a few days to see if fever resolves, or giving an empirical course of broad-spectrum antibiotics, can be clinically defensible. Many times, the patient will get better, either because of, or despite, therapeutic intervention.

In reality, however, the range of different disease processes that can cause fever is enormous. Clinicians should always, therefore, keep an open mind, and realize that the next patient that they see with a fever could very well have a non-infectious disease. While it is truly impossible for any clinician to remember all of the different infectious agents and non-infectious diseases that can cause a fever, it is practicable to characterize the different etiologies of fever into some broad and easy to remember categories:

Infectious

·         Viral

·         Bacterial

·         Fungal

·         Protozoal

·         Rickettsial

·         Mycoplasma

·         Other (algal, oomycetal)

Non-infectious inflammatory

Immunologic

·         Immune-mediated disorders

Neoplastic

Drug-associated

Within these broad categories, it is rational to focus on, and look for, the commonest diseases first, such as a viral infection in a puppy or kitten from a humane shelter, a cat bite abscess in an outdoor tom cat, a tick-borne infection in a dog covered in ticks, or cancer in a very old and thin dog or cat. However, in many instances where the cause of fever is not immediately obvious, it is dangerous to assume that the feverish patient must have one of a limited handful of common diseases. In reality, there are many hundreds of potential infectious and non-infectious causes of fever. In these circumstances, it often makes more sense to keep an open mind, to concentrate on localizing the source of the fever, and to then attempt to isolate or identify a specific cause, than it does to keep non-selectively plowing through an ever-lengthening list of potential causative infectious agents and disease processes.

Initial diagnostic approach:

Usually, in an animal presenting with a high temperature, it is a straightforward process to distinguish between hyperthermia and fever. In circumstances where it is more difficult, for example with a quiet dog or cat coming in for vaccination on a hot day, it is worth rechecking the temperature of the patient after it has been rested in an air-conditioned environment. As a general rule, however, if an animal presents for investigation of signs such as anorexia and lethargy, is quiet on physical examination, and has an elevated temperature, it is far more likely to have a fever than it is to have hyperthermia.

Fever of Unknown Origin

Fever of unknown origin (FUO), also known as pyrexia of unknown origin (PUO), is a relatively common and challenging problem in clinical practice. Interpreted literally, any animal that has a fever in which a cause has not yet been determined has a ‘fever of unknown origin'. In more practical terms, however, a more workable clinical definition for fever of unknown origin would be ‘any fever of greater than a few days duration in which the cause is not obvious on initial history and physical examination'.

History and Physical Examination

In any animal with a fever, history and physical examination should be as comprehensive as possible. The objective of performing the initial history and physical examination is to make every attempt to find any clue, however subtle, that helps to localize the disease process. Whenever a disease process can be localized in a pyrexic animal, the etiology of the disease process can usually be determined. On the other hand, whenever a disease process cannot be localized, finding a cause for a patient's fever can be time-consuming, difficult and expensive. History and physical examination represent the first, best and least expensive opportunity to localize the disease process causing the fever.

History

Owners should be very carefully questions for any historical data that might suggest a causative etiology, and any clues that might help localize the disease process:

Table 2

Historical Clue Might Suggest Vomiting or diarrhea Gastrointestinal disease Coughing or wheezing Respiratory disease Lameness Musculoskeletal disease Seizures Neurological disease Multiple animals affected Contagious diease Owner affected Contagious, zoonotic disease Tick exposure Tick-borne infectious agent Recentn boarding at kennels Infectious agent Recent travel overseas Exotic disease Recent breeding history Sexually transmitted disease (brucellosis) Recent estrus or whelping Uterine infection Recent medications (esp. tetracyclines) Drug-induced fever Apparent response to antiobiotic therapy Bacterial infection Apparent response to glucocorticoids Inflammatory/immune-mediated disease

Historical clues that suggest an etiology or disease localization should be aggressively pursued diagnostically. Even subtle clues may serve to direct clinicians along the correct diagnostic trail. While there is a methodical, step-by-step ‘screening' diagnostic approach that can be applied any patient with fever unknown origin, this process is only used as a ‘default' approach if initial investigation has not already suggested a more focused diagnostic approach. Any diagnostic clues that are uncovered along the way should as a general rule first be investigated before embarking on further ‘screening' tests.

Physical Examination

Physical examination in the pyrexic animal, like history collection, should be comprehensive. In addition to a standard thorough physical examination including thoracic auscultation and abdominal palpation, neurologic, orthopedic and ophthalmologic (including fundic) examinations should also be performed. Certain anatomic locations are notorious for harboring pyrogenic disease processes (particularly infectious and immune-mediated diseases), and a mental ‘check list' of prime locations to thoroughly examine will ensure that no disease process is missed.

Table 3

Prime Locations for Examination Disease Processes Uncovered Cardiac valves (auscultate for murmurs) Valvular endocarditis Cervical, thoracic, and lumbar spine (palpate) Discospondylitis, meningitis Joints (palpate and flex/extend thoroughly) Polyarthritis Long bones (palpate) Panosteitis, neoplasia, osteomyelitis Muscles (palpate) Polymyositis Prostate (palpate) Prostatitis, prostatic abscess Testicles (palpate) Orchitis Lymph nodes (palpate) Infection, neoplasia

The importance of a thorough physical examination in the patient with a fever cannot be overemphasized. Physical examination is, without doubt, the single diagnostic step that is most likely to lead to a definitive diagnosis in the patient with fever of unknown origin.

Performance of a single collection of history, and a single thorough physical examination, may not be sufficient in the patient with fever of unknown origin that is proving to be difficult to diagnose. Disease processes often change or progress over time and, even if the disease process remains static, a second or subsequent careful examination may reveal clues that were missed the first time. Regular, repeated physical examination is probably the most effective, and the most cost-effective, diagnostic tool that can be used in a patient with persistent fever of unknown origin. Physical examination clues that suggest an etiology or disease localization, as with historical clues, should be aggressively pursued diagnostically.

The importance of simple clues revealed by history and physical examination should not be minimized. Signalment alone, for example, can provide valuable diagnostic clues. Young animals are more likely to have infectious processes, middle-aged animals are more prone to immune-mediated diseases, and older animals are of course more susceptible to neoplastic processes. Species and breed may also provide useful clues. For example, young dogs from breeds that are prone to diseases that impair different aspects of neutrophil or immune function, such as Weimaraners, are more likely to have recurrent bacterial infections, middle-aged Cocker and Springer Spaniels are far more at risk for developing immune-mediated blood disorders, and Shar-Peis of any age are prone to Familial Shar-Pei Fever.

Minimum data base

History and physical examination will often reveal an obvious cause for a patient's fever. However, whenever fever has persisted for more than a few days, and no obvious disease process has been ascertained, further investigation is indicated. The objective of further investigation is to cast a wide diagnostic net as cost effectively and non-invasively as possible. Obviously, both results of previous history/physical examination and owner expectations/finances may have an impact on the exact nature of further testing, but a standard minimum data base for the fever of unknown origin patient should include routine hematology, a serum biochemical profile, urinalysis, heartworm testing and, in cats, retroviral testing and measurement of thyroid hormone levels. This fever of unknown origin minimum data base is non-invasive and cost effective in that it screens for a wide range of disease processes with a relatively limited number of tests.

Fever of Unknown Origin

Minimum Data Base

Routine Hematology (complete blood count, including platelets and smear examination)

Serum Biochemistry (including electrolytes, creatine phosphokinase, amylase, lipase)

Urinalysis (preferably by cystocentesis or, at least, via catheterization)

Heartworm Testing (antigen test in dogs, arguably antigen/antibody testing in cats)

Cats:    Retroviral Testing (feline leukemia/immunodeficiency viruses)

        Thyroid Hormone (cats older than 8 years old)

Results of the minimum data base in a patient with fever of unknown origin will almost never be entirely normal, because:

1.        Dogs and cats with a fever have predictable hematologic and biochemical abnormalities associated with activation of systemic inflammatory response mechanisms. Such abnormalities are common, and are not specific to any particular disease process. Patients with infection, immune-mediated disease or inflammation can all exhibit similar hematologic and biochemical abnormalities.

Changes may include

·         Inflammatory or stress leukogram, represented by various combinations of neutrophilia (with or without a left shift), monocytosis, eosinopenia and lymphopenia

·         Anemia of chronic disease, typically a mild non-regenerative anemia

·         Mild hypoalbuminemia (albumin is a ‘negative acute phase protein')

·         Mild to marked hyperglobulinemia, associated with both induction of acute phase proteins and increased production of immunoglobulins.

2.        Well over thirty different parameters are measured in a typical data base containing a hemogram, biochemistry and urinalysis. With this large number of test results evaluated at a single time, it is highly likely that one or more abnormal screening test results will be detected that are in reality ‘normal outliers' (that is, normal for the particular patient). Obviously, not all of these abnormal findings have any clinical relevance, and not all will have anything to do with the cause of the fever. Clinicians must try to distinguish between normal outliers and abnormal results that have real diagnostic meaning.

Abnormal results are more likely to be normal outliers if they are only just outside the reference range, or if they are typically not associated with disease processes (for example, no important disease process typically causes a low amylase). In contrast, abnormal results that are likely to have diagnostic significance are those results that are situated well outside normal reference ranges, and those results that correlate well with historical and physical examination findings. Mild to moderate thrombocytopenia, for example, is far more likely to be clinically significant in a dog with a history of recent tick exposure and painful joints on physical examination.

Sometimes, determining whether an abnormal result has clinical significance can be a major diagnostic challenge, since the cause of an animal's fever of unknown origin may be indicated by a very subtle abnormal test result. In these circumstances, it always preferable to investigate any suspicious abnormality (even if it is eventually proven to be a red herring) than it is to ignore it.

3.        With luck, the initial minimum data base will indicate the specific disease process underlying the fever of unknown origin, or at least serve to localize the disease process.

Test results that can be of diagnostic value include:

·         Moderate to marked anemia, which if regenerative may be seen with immune-mediated hemolytic anemia and with blood-borne parasites such as Mycoplasma hemofelis (hemobartonellosis) or Cytauxzoon felis in cats and Babesia canis or Babesia gibsoni in dogs, and if non-regenerative may be seen with bone marrow diseases such as leukemia or histoplasmosis

·         Thrombocytopenia, which can be seen with both acute and chronic tick-borne infections and immune-mediated thrombocytopenia

·         Extreme leukocytosis, which may sometimes be seen with acute and chronic leukemia

·         Severe neutropenia, which may be seen with retroviral infections and marrow diseases

·         Identification of blood-borne organisms, such as Mycoplasma hemofelis or Cytauxzoon felis in cats, or Babesia, Hepatozoon or Ehrlichia species in dogs

·         Identification of disease-specific blood cell changes on examination of blood smears, such as spherocytes (highly suggestive of immune-mediated hemolytic anemia)

·         Elevated serum biochemical values localizing the disease process to the pancreas (amylase, lipase), kidneys (urea, creatinine), liver (ALT, SAP) or muscles (CPK)

·         Hypoalbuminemia and proteinuria, which may be seen with protein-losing nephropathies such as immune-mediated glomerulonephritis or amyloidosis

·         Pyuria or bacteriuria, indicating the probable presence of a urinary tract infection.

·         Positive feline leukemia virus or feline immunodeficiency virus tests

·         Elevated thyroid hormone levels

Diagnostic clues indicated by the minimum data base should be aggressively and specifically pursued until a diagnosis is made. Minimal further diagnostic testing may be indicated when a specific etiologic diagnosis (for example, hepatozoonosis based on identification of the organism within neutrophils) has already been made. In contrast, it is not sufficient to refrain from further investigation if the minimum data base has served to localize a disease process, but has not as yet identified a specific etiology. For example, markedly elevated liver enzymes, while they serve to localize the disease process to the liver, do not in isolation provide sufficient clues to enable determination of the cause of the liver disease: further liver-specific investigation, such as abdominal radiography, abdominal ultrasonography, hepatic cytology or histopathology, Toxoplasma and Leptospira serology, and questioning owners about access to hepatotoxins, may be indicated in order to obtain a definitive diagnosis.

Specific diagnostic approaches based on the results of the initial history, physical examination and minimum data base will vary widely, and will be determined by the results obtained. In the vast majority of patients with fever of unknown origin, the initial diagnostic approach will either provide a definitive diagnosis, or at least serve to localize disease to a specific location or organ system. Once a disease process has been localized, a definitive diagnosis is usually only a few diagnostic steps away, and is typically based on cytology, biopsy or culture of the affected body system.

While it is beyond the scope of this presentation to cover the specific diagnostic investigation of every body system, such investigations are usually straightforward. Certainly, focused investigations are much more clear-cut than the kind of screening investigation that is required in a pyrexic patient with no localizing signs. When the ‘problem-orientated approach' to diagnosis is used, Fever is a very non-specific problem that leads to expensive and extensive diagnostic testing. In contrast, more specific problems (for example, Pancytopenia, Myopathy or Pyuria) have a much more focused approach that can usually narrow down a limited range of differentials to a single definitive diagnosis with only a handful of further tests.

Further diagnostic investigation:

Unfortunately, in a relatively small sub-set of fever of unknown origin patients, initial diagnostic testing will not provide a definitive diagnosis, and will not be able to localize the disease process. In such patients, there is a structured, step-by-step approach that can be applied in order to obtain a diagnosis. Owners should be warned, at this point, that further testing may on occasion need to be both extensive and extensive. However, with an exhaustive and methodical approach, a diagnosis can usually be made, a prognosis determined, and treatment instituted.

As a general rule, more advanced testing in a stable fever of unknown origin patient will be conducted in a sequential, step-wise fashion, with the results of one series of tests obtained before further testing is commenced. Testing will generally commence with tests that are the most likely to obtain a diagnosis for the least expense and invasiveness, and will then progress through to, if needed, tests that are of lower diagnostic yield and more costly and invasive. In the unstable or very unwell fever of unknown origin patient, however, a large battery of tests may need to be conducted simultaneously in order to obtain a diagnosis and commence treatment before the animal succumbs.

Diagnostic imaging

Thoracic and abdominal radiography is usually indicated as the next step in the diagnostic approach to a pyrexic patient. Thoracic radiographs are indicated even if the patient exhibits no history or clinical signs of respiratory disease, and even if thoracic auscultation is normal, since significant thoracic and respiratory diseases can be remarkably well-hidden. Thoracic radiographs should be carefully examined for evidence of disease processes such as pneumonia, focal abscesses, granulomatous disease, neoplasia or pleural effusion. Abdominal radiographs should also be collected and carefully examined. Furthermore, the clinician should take care to carefully examine any vertebrae, disc spaces, ribs, sternebrae, or pelvic or limb bones that may be visible on the radiographs for evidence of disease processes such as discospondylitis, osteomyelitis, bone neoplasia, hepatozoonosis or panosteitis. Specific spinal, bone or joint radiographs may also be indicated if pain can be localized to these regions.

Abdominal ultrasonography is now becoming more readily available to general practitioners and, for the patient with a fever, tends to provide more useful diagnostic clues than does abdominal radiography. Abdominal ultrasonography may identify such causes of fever as neoplasia or abscesses within organs or abdominal lymph nodes, pancreatitis, pyelonephritis, and low-grade peritonitis. The diagnostic yield of abdominal ultrasonography will to some extent depend on both the quality of the ultrasound machine used and the experience of the ultrasonographer.

Blood and Urine Cultures

Aerobic and anaerobic blood cultures and cystocentesis urine culture are indicated in the difficult to diagnose fever of origin patient. Ideally, samples should be collected either prior to the commencement of antibiotic therapy, or at least several days after antibiotic therapy is discontinued. However, if culture must be attempted despite the presence of concurrent antibiotic therapy, blood culture bottles containing an antibiotic-binding resin may increase the chances of obtaining a positive blood culture. Blood culture samples should be collected as aseptically as possible and, if possible, at least three samples should be collected from multiple different veins at time intervals of at least an hour apart.

A single positive blood or urine bacterial culture should be interpreted with caution, since inadvertent contamination of blood and urine culture specimens with skin bacteria is relatively common. Multiple positive blood sample cultures, in contrast, are highly suggestive of the presence of bacteremia and probable underlying endocarditis, particularly if the same organism is obtained from all samples. Positive urine cultures with the same organism are also common in patients with bacteremia and endocarditis. Bacteremia and endocarditis can occasionally occur in the absence of obvious clues such as a heart murmur, bounding pulses or congested mucous membranes.

Fecal cultures

The intestinal tract, particularly the large intestine, normally contains a large and diverse population of microorganisms, and non-specific culture of feces is therefore of no diagnostic value. The culture of specific enteric pathogens such as Campylobacter, Salmonella and Shigella, in contrast, can be of diagnostic significance, particularly in the feverish patient with signs of gastrointestinal disease. Selective culture for fecal pathogens is therefore indicated in the fever of unknown origin patient, particularly in the presence of diarrhea.

Advanced and Invasive Testing

Uncommonly, a thorough investigation of a patient with fever of unknown origin will fail to provide any specific diagnostic clues. In these circumstances, a diagnosis will occasionally be obtained by a final series of screening tests:

·         Echocardiography will, surprisingly, sometimes reveal either bacterial endocarditis or low grade pericarditis, even in the absence of characteristic clinical signs such as a cardiac murmur or muffled heart sounds. In the hands of a skilled ultrasonographer, vegetative endocarditis may be recognized by such subtle abnormalities as mildly thickened heart valves and aortic or mitral valvular insufficiency.

·         Lymph node fine needle aspiration cytology will occasionally reveal early neoplastic disease (such as lymphosarcoma) or indicators of lymph node inflammation or hyperplasia, even in the absence of obvious lymphadenopathy.

·         Despite an apparent lack of clinical signs attributable to orthopedic, neurologic or marrow disease, joint taps (multiple smaller and larger joints), cerebrospinal fluid taps and bone marrow aspirates/biopsies will occasionally reveal important diagnostic clues to diseases such as polyarthritis (which can be immune-mediated, bacterial or rickettsial), meningitis or marrow neoplasia (such as aleukemic leukemia) or infection (such as histoplasmosis or leishmaniasis). Since joint taps, cerebrospinal fluid taps and bone marrow collection are all easiest performed under general anesthesia, all three tests are often saved until last and then undertaken under the same anesthetic.

Extremely rarely, fever of unknown origin may be caused by a lesion in or near the hypothalamus (typically a brain tumor) altering the thermoregulatory center's ability to respond normally to homeostatic demands. In these rare instances, ‘end of the diagnostic line' advanced imaging (computerized tomography or magnetic resonance imaging) may reveal the presence of a hypothalamic tumor, even in the absence of clinical signs of neurologic disease.

Disease-specific testing

Numerous serologic tests are available that are designed to identify diseases caused by specific infectious agents, or by specific immune-mediated disorders. During the investigation of a frustrating fever of unknown origin case, it is certainly tempting to run a series of serologic tests designed to identify the more common infectious and immune-mediated diseases. However, performance of most of these tests in the absence of prior diagnostic clues derived from the results of history collection, physical examination, a minimum data base and diagnostic imaging is, unfortunately, highly likely to result in confusing test results and a high incidence of erroneous diagnoses based on false positive serologic tests. With a few specific exceptions, most serologic tests for infectious and immune-mediated diseases are neither sensitive nor specific enough to be trusted in the absence of supportive clinical evidence.

Specific testing for infectious agents or immune-mediated disorders is, on the other hand, often extremely valuable when the results of a prior work-up suggest that the disease being tested for is a highly likely differential. Since such disease-specific serologic tests have much more diagnostic strength when supported by the results of prior diagnostic investigation, such tests are usually best left until relatively late in the investigation of fever of unknown origin.

Infectious agents

Numerous serologic tests, usually based on the presence of antibodies against an infectious agent, are available to test for specific infectious disease. A positive test for antibodies against a specific infectious agent, however, often cannot distinguish between an active disease process and prior natural or vaccine-related exposure with subsequent immunity. In areas where sub-clinical exposure to the infectious agent is common, and actual disease caused by the agent is relatively uncommon, many animals will have a positive test result due to past exposure and subsequent immunity, and few will have a positive test result due to active disease. In these circumstances, a single positive test result will typically be a false positive, and consequently diagnostically misleading.

There are numerous examples of serologic tests that commonly will provide false positive results due to natural or vaccine-related exposure rather than active infectious disease:

·         Fungal serology (for example, Blastomyces, Coccidiodes, Histoplasma or Aspergillus)

·         Rickettsial serology (particularly, Rocky Mountain Spotted Fever but also, arguably, Ehrlichia)

·         Protozoal serology (Toxoplasma or Neospora)

·         Bacterial serology (arguably, Lyme disease and, in vaccinated dogs, potentially Leptospira)

·         Viral serology (feline corona virus as part of testing for Feline Infectious Peritonitis and, now that a vaccine is available, potentially also Feline Immunodeficiency Virus)

There are, however, instances when a test for an infectious agent is highly valuable, and less likely to provide a false positive results:

1.        When the test identifies an antigen rather than an antibody.  

Since the presence of significant amounts of antigen derived from an infectious agent strongly suggests the presence of the organism, a single positive test result can be diagnostic. Examples of antigen tests with a relatively high level of diagnostic specificity include feline leukemia virus testing, heartworm antigen tests, and tests for the capsular antigen of Cryptococcus. Although polymerase chain reaction (PCR) tests for most infectious agents are only just now becoming commercially available, PCR testing, because it detects antigen rather than antibody, also has the potential to be more diagnostically specific.

2.        When the presence of antibodies indicate the presence of disease.

With some infectious agents, exposure to the organism and subsequent seroconversion does not lead to clearance of the infection. When a disease cannot be cleared despite seroconversion, then a positive serologic test result also indicates the presence of disease. The best example of this is a positive antibody test against feline immunodeficiency virus. Other examples of organisms that cannot be easily cleared from the body include Leptospira, Brucella and, arguably, Borellia (Lyme disease) and Ehrlichia. In the absence of recent vaccination, a very high Leptospira titer, for example, is supportive of the presence of active disease or at least a chronic carrier state.

3.        When antibody levels are extremely high.

Some laboratories claim that extremely high titers of antibodies against certain infectious agents (for example, Toxoplasma or Leptospira) are strongly suggestive of active disease.

4.        When seroconversion can be documented.

Paired antibody tests that document seroconversion from negative to positive, or rising titers over a period of 2 to 4 weeks, strongly suggest the presence of active disease. While performing paired tests and documenting seroconversion is diagnostically specific, it has the disadvantage that a definitive diagnosis will always be delayed by at least 2 weeks.

5.     When IgM levels are high.

Since high levels of IgM antibody against an organism only occur during early exposure, high IgM titers usually suggest active or recent disease. A positive IgM titer for Toxoplasma or Neospora, for example, is more diagnostically specific than a positive IgG titer.

With the exception of the above specific circumstances, however, serologic testing for infectious organisms is best performed once prior work-up is highly suggestive of a particular infectious agent. Even in these circumstances, it is still preferable to attempt to obtain a diagnosis with a more specific means (for example, identifying a fungal organism on culture, cytology or histopathology) rather than to rely on serology alone.

Tick-borne infections

Serologic testing for the various Ehrlichia species, Rocky Mountain Spotted Fever, and Lyme disease, is indicated in patients with various combinations of:

·         A history of tick exposure or travel to areas with endemic tick-borne diseases

·         Thrombocytopenia or clinical signs suggestive of vasculitis

·         Polyarthritis

·         Diffuse or multifocal central nervous system disease

Fungal Serology

Serologic testing for fungal organisms such as Blastomyces, Aspergillus, Coccidiodes, or Histoplasma may be indicated (although direct organism identification is always preferable) in patients with various combinations of:

·         Pyogranulomatous lymph node disease

·         Granulomatous lung disease with perihilar lymphadenopathy

·         Pyogranulomatous osteomyelitis or arthritis

·         Uveitis or chorioretinitis

New antigen-based tests on serum or urine have recently become available for a number of the important fungal infectious diseases, although the diagnostic accuracy of such tests has not yet been established in dos and cats.

Coronavirus serology

Despite claims to the contrary, there is still no serologic test that has been definitively proven to differentiate pathogenic (causing Feline Infectious Peritonitis) from non-pathogenic coronaviruses. For this reason, a ‘positive FIP test' can only be diagnostically useful when interpreted in combination with a range of other clinical features consistent with a diagnosis of FIP:

·         Polyclonal hyperglobulinemia

·         High protein exudative body cavity effusions

·         Pyogranulomatous inflammation of organs such as the liver or kidneys

·         Anterior uveitis

·         Diffuse or multifocal central nervous system disease

·         Lymphopenia

Specialized tests that may have higher specificity for diagnosis of FIP include PCR for viral antigen in the blood or tissues, and immufluorescence or immunohistochemistry to detect viral antigen in cells (for example, within macrophages in an abdominocentesis sample)

Numerous other infectious agents may be potential considerations in particular circumstances. New organisms are being regularly identified and incriminated in various disease processes, and the number of serologic and PCR-based tests that are available is expanding yearly. Infectious diseases that may need to be considered and tested for in individual pyrexic dogs and cats include hemobartonellosis (mycoplasmosis), bartonellosis, toxoplasmosis, neosporosis, babesiosis, leishmaniasis, hepatozoonosis and trypanosomiasis.

Immunologic tests

A number of serologic tests for immune-mediated disease are commercially available, including antinuclear antibody (ANA), rheumatoid factor, and Coomb's testing. Unfortunately, however, such tests are overused by many clinicians, and both false positive and false negative tests are common. Antinuclear antibody tests, for example, are typically negative in most immune-mediated diseases except for systemic lupus erythematosis, and commonly positive in animals with sepsis or cancer. Rheumatoid factor is typically negative for most of the common forms of immune-mediated polyarthritis in small animals, and is only likely to be positive in the relatively rare patient with true rheumatoid arthritis. The Coomb's test has little meaning except when performed in patients with hemolytic anemia that is suspected to be immune-mediated, and is commonly positive in animals with other forms of hemolytic anemia (hemobartonellosis, for example), and negative in many animals with true immune-mediated hemolytic anemia.

The practice of running an ‘immune panel' (ANA, rheumatoid factor and Coomb's test) relatively early in the investigation of fever of unknown origin leads to a large number of false positive test results, and an unacceptably high rate of erroneous diagnoses of immune-mediated disease. Because of their relatively limited diagnostic utility, serologic tests for immune-mediated disease should therefore only be performed when strongly supported by specific clinical circumstances:

Antinuclear Antibody

Testing for antinuclear antibody should only be considered when a diagnosis of systemic lupus erythematosis is a consideration. Specifically, the ANA should be run in patients with two or more organ systems affected by suspected immune-mediated disease:

·         Immune-mediated hemolytic anemia

·         Immune-mediated thrombocytopenia

·         Glomerulonephritis

·         Immune-mediated skin disease

·         Polyarthritis

·         Polymyositis

Rheumatoid Factor

Testing for rheumatoid factor is only indicated in patients with:

·         Polyarthritis

·         Radiographic evidence of joint destruction (erosive polyarthritis)

·         No evidence (on culture or cytology) of an infectious polyarthritis

Coomb's Test

Coomb's testing is only indicated with patients with:

·         Hemolytic anemia (either regenerative or non-regenerative)

·         No evidence of a non-immune cause of hemolysis (for example, zinc toxicoses)

Therapeutic trials

It is always tempting, in an animal with a fever of unknown origin, to treat empirically and see if the animal gets better. Sometimes, this is appropriate, when a bacterial infection is strongly suspected but unproven (for example, a painful cat one day after a cat fight is probably brewing up a cat bite cellulitis), or when immune-mediated disease is likely but as yet not proven (for example, a young Cocker spaniel presenting with acute agglutinating hemolytic anemia). Sometimes, although neither desirable nor appropriate, empirical therapy will also become necessary because of limitations in owner budget and expectations. Regardless of why empirical therapy is instituted, some general guidelines apply:

Avoid therapy that interferes with diagnostic testing

Whenever possible, withhold

·         Antibiotic therapy until all bacterial cultures (such as blood and urine cultures) have been collected and submitted.

·         Corticosteroid therapy until a lesion has been localized, tests for immune-mediated disease (such as an ANA or Coomb's test) have been performed, and samples for cytology or histopathology have been collected. Prior steroid therapy can lead to false negative ANA and Coomb's tests, and can markedly impair the cytologic and histopathologic diagnosis of immune-mediated disorders and hematologic and round cell neoplasia (such as lymphosarcoma and leukemia).

·         Anti-inflammatory therapy until a lesion has been localized. Anti-inflammatory therapy can markedly reduce heat, pain and swelling, making it impossible to, for example, localize joint or spinal pain.

Treat with appropriately aggressive doses and drugs

A therapeutic trial is of limited value unless the clinician can be absolutely sure that the drug used will be effective if it is the appropriate therapeutic decision. A failure to respond to antibiotic therapy, for example, has little diagnostic value if the antibiotic was used inappropriately (for example, wrong spectrum of activity, or wrong dose rate).

Therapy must therefore be appropriately aggressive:

·         Antibiotic therapy should usually be appropriate to the organism suspected. When a bacterial infection is suspected, for example, antibiotic therapy should generally be broad-spectrum, bactericidal, and able to penetrate into locations that may be infected. When a tick-borne disease is suspected, an appropriate drug (usually doxycycline) should be given at an appropriate dose for an appropriate duration.

·         Glucocorticoids must be given at appropriate and aggressive immunosuppressive doses if immune-mediated disease is suspected.

Accurately record response to therapy

Owners must keep an accurate diary, and accurate measurements of apparent response to therapy must be made, in order to differentiate coincidental from real responses to therapy.

Antipyretic therapy

Treatment with antipyretic therapy, usually with non-steroidal anti-inflammatory drugs (NSAIDs) such as aspirin and related compounds, is often initiated in fever of unknown origin patients. Therapy with NSAIDs, however, is in reality uncommonly indicated. With rare exceptions, NSAID therapy does nothing to treat the primary disease process, and merely masks the underlying disease without aiding its resolution, making it more difficult to identify and diagnose. Veterinarians are, however, still tempted to give NSAIDs because they will often make a patient feel and look better, although such responses are rarely sustained unless the underlying disease is identified and treated. Clinicians should be aware that, unless a fever is extreme (greater than 106° F), the fever itself is rarely life-threatening. Usually, therefore, non-specific treatment of fever with NSAIDs is best withheld until appropriate diagnostic testing has been completed.

Antipyretic therapy is appropriate in patients where an underlying disease process has been identified but is not readily treatable with more specific therapy such as antibiotics, antifungals or immunosuppressive agents. Fever due to disease processes that lack a specific treatment, such as panosteitis, hypertrophic osteodystrophy or Familial Shar-Pei Fever, is therefore appropriately treated with NSAIDs.

Patient cooling strategies

Clinicians are often tempted to treat pyrexic patients with cooling strategies that are designed for animals with hyperthermia or heat stroke, such as cooling fans, cold water immersion, ice packs, cold fluids or cold water enemas and gastric lavages. Such cooling methods are, however, typically ineffective, and may actually often be detrimental to the patient. Since pyrexic animals have a raised thermoregulatory set point, and actually perceive that they are too cold despite a higher than normal body temperature, cooling techniques will only make the patient feel colder and more miserable, and exacerbate counterproductive warming behaviors such shivering and piloerection.

Specific therapy

Specific treatments for patients with a fever are many, and will vary widely depending on the ultimate final diagnosis. It is therefore imperative to attain an accurate definitive diagnosis. Discussion of the specific treatments of all of the diseases that can cause fever is beyond the scope of this presentation.