Is There Good Evidence of Bacterial Infection?

The first step in empirical antimicrobial therapy is to critically ask whether there is good evidence of a bacterial infection. Too often, antimicrobials are prescribed on a 'just in case' basis or because an owner resists additional diagnostics. There are several very real drawbacks of 'just in case' antimicrobials:

 They add to the cost of the visit or hospitalisation, without contributing to a diagnosis.

 They may lead to diarrhoea, inappetence, or vomiting that obscure the underlying problem.

 They can cause adverse reactions or drug interactions.

 Importantly, they encourage the selection of resistant bacteria, both globally, in your hospital, and in that patient.

Fever alone is an inadequate criterion for prescribing an antimicrobial. If the probable source of fever cannot be localised during a physical examination, it is not possible to choose an appropriate spectrum of bacterial coverage or decide how long to treat. Physical examination findings that support the use of antibiotics without additional diagnostics are usually straightforward, such as a carnassial tooth root abscess, skin pustules or epidermal collarettes, or a traumatic wound with purulent exudate. Cytology should be performed on all exudates, particularly in areas endemic for systemic fungal infections, prior to considering antibacterials.

Leucocytosis alone is not a good justification for antimicrobials, since leucocytosis can result from stress, inflammation, glucocorticoid administration or hyperadrenocorticism. If a left shift and toxic change are present with leucocytosis, then a source of infection or significant inflammation should be pursued. Neutropenia with fever, however, is an established indication for empirical antimicrobials, at least in humans. Meta-analyses of studies in humans suggest that the benefit of antibacterials in neutropenic patients, even prior to fever, outweighs the negative effects of selecting for bacterial resistance. A beta lactam and fluroquinolone combination is recommended in humans, which provides coverage against gut flora to include anaerobes and Enterococcus (beta lactam) and gram-negatives (fluoroquinolone).

Antimicrobials are inappropriate in most cats with lower urinary tract signs. There is a < 5% incidence of positive urine cultures in cats with lower urinary tract disease. Clients' money is better spent on a urinalysis and bladder imaging for stones. Cats at higher risk for bacterial urinary tract infections are those with diabetes mellitus (13% prevalence), perineal urethrostomies or chronic renal failure (CRF). For these cats, urine culture is a good investment.

For cats with upper respiratory infections, there are no good studies comparing antimicrobials to placebo. Although Mycoplasma is commonly isolated from pharyngeal swabs in cats, underlying viral infection is typical, and cats have similar recovery rates whether antimicrobials with and without activity against Mycoplasmaare used. When active Mycoplasmaor Chlamydophila infection are documented or suspected, doxycycline remains the drug of choice. Suspension is preferable to capsules, to decrease the risk of oesophagitis.

Diarrhoeas in dogs and cats are usually not caused by pathogenic bacteria. For example, in dogs with acute diarrhoea, the prevalence of salmonellosis (2%), Campylobacter (5%), and Clostridium difficile toxin (10%) is low. Empirical antimicrobials, such as amoxicillin or fluoroquinolones, are not indicated for acute diarrhoeas. Fibre, probiotics, or Pepto-bismol, along with a short-term diet change, may be a better approach.

Finally, pancreatitis is usually sterile in dogs and cats. Antimicrobials are not indicated unless peritonitis, pancreatic abscess or loss of intestinal mucosal integrity (bloody diarrhoea with mucosal sloughing) develops. In humans, antimicrobials in severe (necrotising) pancreatitis do not reduce the incidence of secondary pancreatic infection or decrease mortality.

Helpful Diagnostics in Lieu of Culture

Pyuria with bacteriuria provides a strong indication for antimicrobials, although bacteriuria can be overdiagnosed. If cocci are frequently noted in your in-house urine sediments, be cautious; stain precipitates can mimic cocci. The majority of urinary pathogens in dogs are gram-negative rods. In cats, gram-positive and gram-negative pathogens occur at approximately the same rate; however, urinary infections are much less common in cats overall.

The finding of intracellular bacteria on cytology is a strong rationale for antimicrobials. A Gram stain, www.life.umd.edu/classroom/bsci424/LabMaterialsMethods/GramStain.htm , which is cheap but underutilised, narrows the spectrum to gram-positive, gram-negative, or mixed.

Deciding About Cultures

Cultures are not, in the practical sense, necessary for first-time empirical treatment of many routine bacterial infections, to include acute contaminated wounds, carnassial tooth abscess, infectious tracheobronchitis, superficial pyoderma, cat abscess or first-time bacterial cystitis.

Cultures are very important for any second-line antimicrobial treatment, to include lack of response to empirical treatment, relapse after treatment discontinuation, or waxing and waning signs. Examples of this include suspected bacterial bronchopneumonia, urinary tract infections, or non-responding pyoderma. Avoid antibiotic roulette in these cases! With recurrent urinary tract infections, serial cultures can be very helpful. Repeated culture of the same organism suggests inadequate clearance (e.g., immunosuppression, poor compliance, uroliths or accompanying prostatitis with inadequate drug penetration). Repeated culture of different organisms suggests new (usually ascending) infection (e.g., an anatomical defect such as an ectopic ureter or urethral incompetence, vulvar fold pyoderma, or poor perineal hygiene in debilitated or obese animals).

Cytology and cultures are also important for first-time treatment of long-standing infections, such as fistulous tracts, possible chronic foreign bodies, or non-healing wounds, as well as for serious or life-threatening infections, such as pyothorax, suspected endocarditis, osteomyelitis, joint sepsis, renal failure with suspected pyelonephritis, or suspected sepsis. Cultures are also recommended for suspected hospital-acquired infections (those developing > 72 hours after admission), since nosocomial bacteria may have multidrug resistance patterns, and hospitalised patients may be more susceptible to sepsis because of intravenous lines, urinary catheters, invasive procedures and immunocompromised states.

Practical Culture Techniques

Urine is ideally set up for culture within 15–30 minutes of collection, but this is impractical in many settings. Alternatively, a sterile syringe containing urine can be capped and refrigerated immediately, for up to 12 hours prior to culture. While some fastidious bacteria may not survive storage > 1 hour, this approach is adequate in most situations, and also allows quantitative culture. For aerobic, anaerobic and Mycoplasma cultures at other sites, our microbiologist at UW-Madison (Faye Hartmann) recommends A.C.T. II agar (Remel). These tubes contain a semisolid medium to which you can add fluid (or swabs if only scant fluid is available). Organisms should be stable for 24 hours for both aerobic and anaerobic culture set-up. Fluids for culture should not be placed in heparin or EDTA. If your available laboratory does not provide cultures at an affordable price or with a useful turnaround time, talk to the laboratory manager about your concerns.

Empirical First-Line Regimens

The first principle to follow for empirical antimicrobials is to use the narrowest spectrum possible (see Figures 1 and 2 overleaf). For example, choose amoxicillin or penicillin, instead of amoxicillin-clavulanate, for a cat abscess, and doxycycline, rather than a fluoroquinolone, for suspected Bordetellainfectious tracheobronchitis. Fluoroquinolones are over-used as first-line antimicrobials in both human and veterinary medicine. A common reason for misuse in human patients is the availability of narrower-spectrum first-line agents.

Second, consider tissue penetration. Beta lactams and aminoglycosides, which are relatively polar, have poor penetration into the prostate, eye, testes and bronchial secretions.

It should be assumed that urinary tract infections in male dogs involve the prostate, and antimicrobials with good prostatic penetration, such as fluoroquinolones, doxycycline, chloramphenicol or potentiated sulphonamides, should be chosen. For bronchitis without pneumonia, drugs that achieve high concentrations in bronchial secretions should be prescribed, to include fluoroquinolones, doxycycline, azithromycin or potentiated sulphonamides. For endophthalmitis or ocular trauma, azithromycin and fluoroquinolones show excellent ocular penetration.

Finally, treat for the shortest effective period possible. There is a trend towards shorter courses of antimicrobials in humans, with equivalent efficacy compared to longer regimens but with less emergence of resistance. For example, acute sinusitis, pneumonia or uncomplicated urinary tract infections are treated effectively with only a 3–7-day course of antibiotics in humans. Bacterial otitis in children can be treated with a single dose of azithromycin (as effective as multiple dosing over 7 days), and for community-acquired pneumonia, treatment regimens are continued for only 2–3 days beyond resolution of fever. In veterinary medicine, there has been little work to determine whether the longer courses that are recommended in textbooks are actually necessary. Consider using these shorter regimens, with a follow-up recheck 1 week after discontinuation. Shorter treatment regimens are less expensive for clients (allowing more resources for diagnostics and follow-up) and are associated with better compliance. Most importantly, shorter courses of antimicrobials lead to significantly less bacterial resistance.

Figure 1. Common isolates from bacterial infections in dogs and cats.

*Cultures strongly recommended at first presentation.

Figure 2. Typically effective antimicrobials for different microorganisms.

Initial concept for table courtesy of Dr Jill Maddison, Royal Veterinary College, University of London.

References

References are available upon request.