METHODS AND RESULTS

Bacterial isolates were obtained from a variety of sources (table 1⇓). Each infection was of an invasive nature which would commonly merit prolonged antibiotic treatment. Bacterial isolates, identified initially only as CNS, were examined for purity before being subjected to susceptibility testing. Subsequent speciation with APIStaph (bioMerieux, France) revealed that all were Staphylococcus epidermidis except for one that was Staphylococcus lugdunensis (patient 6). Initial assessment included critical agar dilution testing with oxacillin (6 mg/l oxacillin in Mueller–Hinton agar with 2% NaCl supplementation and 24 hours incubation). This assessment was undertaken with standard methods and control organisms.¹⁰ All the isolates were susceptible in this assay. For confirmatory testing, the isolates were assessed by Etest (on Mueller–Hinton agar with 2% added NaCl) according to the manufacturer's instructions (AB BIODISK); minimum inhibitory concentrations (MICs) were determined according to standard criteria as recommended by the manufacturer. Oxacillin MICs ranged from 0.2 to 2.0 mg/l. In order to determine the presence of oxacillin heteroresistance, the zone of bacterial inhibition from Etest was examined carefully at 24 and 48 hours. Any suspect colony was transferred to confirm bacterial growth and then reassessed by Etest.⁷ Resistant CNS colonies are scant and usually pin-point; the repeat Etest thereafter can identify whether the isolate has an MIC greater than 6 mg/l. No such colonies were found after Etest screening of the isolates from these patients, thus reconfirming the susceptibility as initially determined by critical agar dilution.

Table 1⇑ details some characteristics of these six patients and their infections. All patients had serious and invasive infections, and all patients had preceding or existing illnesses which either complicated the treatment course or put them at risk of CNS infection. Given the clinical presentations, these patients all received an initial course of antibiotics, usually as an empirical regimen for presumed infection. Each went on to complete long term treatment with cloxacillin. Patient 1 continued to yield positive cultures for CNS from the external ventricular drain until intravenous cloxacillin was initiated. Patient 3 was treated with intravenous and subsequently oral cloxacillin; oral use was followed by demonstration of peak serum bactericidal activity of 1/256. Patient 4 continued to have positive blood cultures throughout the time that intravenous vancomycin was being administered despite adequate serum peak and trough levels (34–35 mg/l and 10–17 mg/l respectively). Blood cultures became negative only after intravenous cloxacillin was administered. Patient 5 also had consistently positive blood cultures despite receiving intravenous vancomycin, but therapy was complicated by the intermittent dosing required for intermittent haemodialysis and dose adjustment during treatment of active and advanced renal failure. A specific lumen of a multilumen vascular access device was thought to be colonised by and acting as the source of CNS infection. Blood cultures became negative within two days of commencing intravenous cloxacillin treatment despite leaving the central venous catheter in situ throughout the treatment period.

DISCUSSION

The mechanisms of oxacillin resistance in CNS are similar to those in methicillin resistant Staphylococcus aureus (MRSA)^11,12 and involve alteration in penicillin binding protein structure. Most resistant CNS have a resistance gene (mecA) which is analogous to the most common resistance gene among MRSA.¹³ When either CNS or MRSA are isolated, the bacteria may be heteroresistant, and the phenotypic expression of such resistance may be variable. Various improvements have been made in laboratory methods for detecting such heteroresistance, especially among MRSA.¹⁴

The frequency of oxacillin resistance in CNS is high (up to 60–80% or more), especially among CNS, causing hospital infections.⁷ The empirical use of vancomycin for suspected CNS infections is commonplace, and most physicians are hesitant to use β lactam agents for such infections because of the treatment failures that occur when β lactam agents are used, even when conventional susceptibility testing indicates that the bacteria are susceptible. Many CNS infections may be complicated by factors which promote the infection and its persistence, such as underlying illnesses and implanted foreign medical devices. Many CNS infections are treated with short term antibiotics and removal of any medical device or vascular access line. Vancomycin may be appropriate for such therapy. CNS infection that is susceptible to β lactam agents and which requires lengthy treatment is relatively uncommon. Vancomycin is expensive, relatively more toxic, and requires drug level monitoring. It should be used as little as possible to discourage emergence of resistance. β Lactam therapy appeared to be more efficacious in several of these cases, and oral therapy may be used in some circumstances, as illustrated by patient 3, allowing for more ambulatory care.

Although recognised as a dilemma for many years, the accurate determination of oxacillin resistance in CNS has received special interest in the past decade. US authorities have modified guidelines for susceptibility testing of CNS, decreasing the MIC breakpoint for definition of resistance considerably.¹⁵ This strategy will accurately define susceptible bacteria, but may exclude some CNS that are actually susceptible.¹⁵ Similar to MRSA, multiple resistance to other antibiotics is a marker for CNS that are likely to be genuinely oxacillin resistant, but this association is not consistent.¹⁶ Thus, a reliable direct determination of oxacillin resistance is still desirable.

Given that the major dilemma with oxacillin resistance determination among CNS is the phenomenon of heteroresistance, we have used a laboratory method which will determine heteroresistance with a high degree of accuracy. The number of patients in our study is small, and we believe that a larger study is warranted to confirm these findings. The application of a reliable technique which is associated with a reasonable frequency of cure would allow physicians and medical microbiologists to be more confident treating chronic CNS infections.