July 2009
Feature Story

William Check, PhD

Vancomycin and Staphylococcus aureus—long-time partners at the infectious disease dance. Monitoring and dosing guidelines for vancomycin to treat S. aureus infections were established decades ago. In the ever-changing world of laboratory medicine, it’s nice to know some practices are dependable. And yet—it seems all is not settled between these two. Witness the consensus review published in January titled “Therapeutic monitoring of vancomycin in adult patients” (Rybak M, et al. Am J Health-Syst Pharm. 2009;66:82–98), a joint document of the American Society of Health-System Pharmacists, the Infectious Diseases Society of America, and the Society of Infectious Diseases Pharmacists. The publication reviewed evidence regarding effective serum trough concentrations and dosing regimens for vancomycin amid suspicion that resistance in S. aureus strains has grown incrementally, as well as concern about an increase in therapeutic failures. The review contained a set of recommendations; also in the review was a discussion of potential toxicity and recommendations about when to monitor to avoid it. As a result, vancomycin and S. aureus look less like dancing partners and more like determined pugilistic adversaries in a long-running series of title fights. Like Ali and Frazier, they keep getting up off the canvas and returning for another round. As with much of infectious disease practice, it’s a real-time display of evolution in process.

Michael J. Rybak, PharmD, MPH, was first author on the consensus review. Evidence of decreasing susceptibility of S. aureus to vancomycin and increasing rates of therapeutic failure were the reasons that the vancomycin monitoring recommendations were reconsidered, explains Dr. Rybak, associate dean for research, professor of pharmacy and medicine, and director of the Anti-Infective Research Laboratory in the Eugene Applebaum College of Pharmacy and Health Sciences at Wayne State University. The guideline pertains only to S. aureus. “Most of the published information dealing with adjustment of vancomycin dose is applied to Staph aureus,” he says. He adds, “Several surveillance studies have demonstrated decreasing susceptibility,” sometimes called “MIC creep,” “and there is a mounting group of papers showing increasing rates of failure, particularly in patients with complicated bacteremias, such as infective endocarditis.”

As a result, the consensus group recommended raising the target for trough serum vancomycin concentration from the previous range, 5 to 10 mg/L, to a new range of 15 to 20 mg/L. Dr. Rybak emphasizes that daily dosing regimens to achieve this serum concentration will vary depending on several factors, including the patient’s body weight and renal function and the minimal inhibitory concentration of the infecting strain of S. aureus. “If the organism has an MIC of 2 micrograms per milliliter or greater, you are not likely to reach the target concentration with conventional dosing,” such as 15 mg/kg (1.2 g for an 80-kg person) every 12 hours, he says.

“In the past, clinicians would get a warning when the trough serum concentration exceeded 10 mg/L because these values exceeded the recommended range,” Dr. Rybak continues. “What we are doing is expanding the acceptable range of values. This is an attempt to take an old drug—it was approved for human use in 1958—and extend its use, to do the best we can given the susceptibility patterns we are seeing.” With regard to pathologists, Dr. Rybak says, “The bottom line is that if clinicians adopt these recommendations, you are likely to see more vancomycin levels drawn and more toxicology involvement. There will be vancomycin assays being run and more variability in serum concentrations, including some over and some under recommended levels.”

Atarget for trough serum concentration between 5 and 10 mg/L is “completely out of date,” agrees Yashpal D. Agrawal, MD, PhD, director of the central laboratory and point-of-care testing at New York Presbyterian Hospital-Cornell Campus and associate professor of clinical pathology and laboratory medicine. He calls the January 2009 review “a very nice consensus guiding document.” Not only is a range of trough levels of 15 to 20 mg/L more appropriate, but also conventional dosing using ideal body weight is outmoded, in Dr. Agrawal’s view. “In the older literature, they talk about a standard vancomycin dose of one gram per 12 hours,” says Dr. Agrawal, a member of the CAP Toxicology Resource Committee. “That is not very useful. Dosing should use actual body weight, to avoid the risk of subtherapeutic levels in overweight patients. Also, where practical, you must correlate dose with the MIC of the bacteria. Unless you combine these two, you will produce subtherapeutic levels of vancomycin and keep generating resistance.”

Sheila Dawling, PhD, CChem, FRSC, says her laboratory report on vancomycin monitoring gives a target range of 5 to 20 mg/L, adding that 5 to 15 is sufficient for most infections. “In practice, 20 applies to high-risk patients and those with a resistant strain,” says Dr. Dawling, director of the toxicology/TDM laboratory and associate director of clinical chemistry at Vanderbilt University Medical Center. Now the pharmacy and infectious disease departments at Vanderbilt University Hospital are proposing to increase the reference range to 10 to 25 mg/L. “There is some worry that we are breeding more resistant strains by not treating aggressively,” she says. In support of the change, the medical center pharmacy’s infectious disease specialist sent Dr. Dawling the January 2009 consensus review. Dosages will approximate 18 mg/kg instead of the 12 mg/kg typically used now, so that 1-g doses will be minimal and there will be many more doses of 1,250 mg, 1,500 mg, and 1,750 mg.

In addition to the widely quoted and influential consensus document, another indication that changes are needed in vancomycin monitoring and dosing was the question and answer published in the “Q&A” column in the November 2008 issue of CAP TODAY. Replying to a reader’s query about whether his lab’s reference range of 5 to 10 mg/L for vancomycin serum trough levels should be raised, Steven M. Johnson, MD, began by noting, “There is much that is controversial, uncertain, or evolving about vancomycin monitoring.” In the course of a wide-ranging discussion of this topic, Dr. Johnson, director of microbiology and serology at Christian Hospital in St. Louis, wrote that “[H]igher trough ranges of 15 to 20 or 15 to 25 µg/mL may be used by some physicians and institutions when more aggressive dosing is considered to be of value.” An example would be S. aureus pneumonia. He closed by suggesting that “Collaboration among the laboratory, the pharmacy, and infectious disease physicians may be of help in reaching a consensus on guidelines for an institution’s needs.”

“Many readers of CAP TODAY are general pathologists,” Dr. Johnson said in a recent interview. “They are not reading about vancomycin all the time and have questions like the one that this physician sent—Is there a current consensus about the range? People would like to have some short answers,” Dr. Johnson acknowledges. “However, we don’t have all the answers. We don’t have a perfect regimen. We have a regimen that is highly effective for many patients, but not all.” In this context, Dr. Johnson says, “The important question is what is reasonable to do based on the information we have at this time.”

In addition to the pathologist’s basic responsibilities—to ensure specimen integrity and labeling, to issue guidelines on how and when to obtain samples, and to make sure measurement is accurate, with vancomycin “we are going beyond the fundamentals and providing guidelines about whether initial dosing is appropriate or should be adjusted,” Dr. Johnson says. “It is appropriate for us to provide those guidelines for vancomycin.” A pathologist’s advice on this point might be more valuable in smaller hospitals, in a situation where either the attending physician, who might be selecting the dose and adjusting it, or the pharmacist is not sure about what target ranges to seek. In larger hospitals, pharmacy tends to be larger and “their intervention is very helpful,” Dr. Johnson says. “Ideally, pharmacy, ID, and pathology would follow the same guidelines.”

Vancomycin levels are monitored to get the most benefit out of the antibiotic, agrees Ronald N. Jones, MD, president and CEO of JMI Labs in North Liberty, Iowa, and adjunct professor of Medicine at Tufts Medical Center. “It puts a burden on laboratory medicine departments to provide those services. Physicians believe they have to order those tests to appropriately dose vancomycin,” he says. Dr. Jones has one concern with the new guidelines. Trying to maximize favorable clinical outcomes by striving for higher trough levels encounters the linear relation with increasing toxicity. Dr. Jones questions whether existing MIC data mandate the increase in serum trough levels and increased dosing. “It is fair to say that the information by which we have been led to make these decisions is, I believe, soft data.” In truth, the consensus group rated the evidence supporting its recommendations as the lowest category, level III.

Another specialist with doubts is Paul G. Ambrose, PharmD, FIDSA, director of the Institute for Clinical Pharmacodynamics in Latham, NY. Current vancomycin dosing recommendations don’t often achieve the recommended serum trough levels, Dr. Ambrose agrees. However, he doesn’t automatically conclude that increasing doses is the correct response. To Dr. Ambrose, the issue is, Should we raise daily dose or lower trough levels? “I remain to be convinced that we have a major vancomycin efficacy problem,” he says. “In clinical trials with vancomycin versus new drugs, we are not seeing major differences in efficacy. [See “Vancomycin alternatives.”] Even with higher doses, only 15 percent of those patients reach the target trough level. If vancomycin efficacy is decreasing, then this drug should be tanking [in the comparative trials].” But it’s not. Dr. Ambrose considers 15 to 20 mg/L to be a somewhat arbitrary value based on “soft science.” He and his colleagues are collaborating with the National Institutes of Health on a study design for a clinical trial that will define the exposure/response relationships for vancomycin for efficacy and for toxicity. “Then we can come up with a dosing algorithm to minimize toxicity and maximize efficacy,” he says. “It’s amazing,” Dr. Ambrose adds. “After 50 years we still haven’t figured out how to dose vancomycin to get optimal efficacy with lower toxicity. Once we dose it correctly, it might turn out still to be a darned good drug.”

Dr. Rybak highlights details of the consensus guideline. For instance, the most accurate measurement of a therapeutic dose is the ratio AUC/MIC (area under the curve divided by minimal inhibitory concentration). “The AUC/MIC ratio drives efficacy, but most people measure trough levels as a surrogate marker,” he says. Some evidence supports that an AUC/MIC of =400 would be a reasonable target for efficacy. “A trough level of 15 to 20 is essentially equivalent to an AUC/MIC of around 400,” Dr. Rybak says. The consensus report cites evidence from one article supporting the conclusion that “For pathogens with an MIC of 1 mg/L, an AUC/MIC of approximately 250 can be achieved in most patients with 1 g every 12 hours based on a patient with an actual body weight of 80 kg and normal renal function,” which is defined as creatinine clearance of 100 mL/min. Increased MIC or body weight or decreased renal function may require dose adjustment.

As more aggressive dosing is used, monitoring becomes necessary because of the heightened toxicity risk. “At the lower doses used to reach trough levels of 5 to 10, you probably don’t run into problems with toxicity,” Dr. Rybak says. Also, achieving a trough level between 5 and 20 mg/L, as was the previous standard for dosing at Detroit Medical Center, was “fairly easy,” he says. “We had a lot of room for error. At higher doses, however, you are more likely to have adverse reactions. And monitoring will be more aggressive as we shoot for a smaller, more defined range of trough levels in patients.” Treatment will more often walk the line between efficacy and toxicity. “Aiming for the narrow window between 15 and 20 will be more difficult,” Dr. Rybak says.

It follows that not all patients need to have their vancomycin trough levels measured. Because the potential for toxicity at higher doses is the main reason for monitoring, monitoring is not mandatory in patients with uncomplicated infections caused by organisms with lower MICs and who are getting short-term therapy.

Dr. Johnson agrees there is a potential tradeoff between toxicity and efficacy. “However,” he says, “I still don’t see a predictable trend in renal damage. To me it’s reasonable to accept an undefined small risk of renal toxicity in exchange for the chance to treat more people with serious infections effectively.” Among “serious infections” he includes MRSA pneumonia, meningitis, osteomyelitis, and endocarditis. “These are the patients who are more likely not to do well,” he says. “In our hospital these patients are not extremely common, but they do occur occasionally. Our guideline recommends considering higher levels for these more serious infections.” Infectious disease physicians at Christian Hospital tend to aim for a trough level of 15–20 mg/L in such patients.

Actual vancomycin dosing at Christian Hospital is a collaboration among the laboratory, pharmacy, and ID physicians. A guideline goes out with the laboratory report. “We occasionally have physicians who want to do things on their own, so they want the guideline on their report,” Dr. Johnson says. Otherwise, pharmacy handles vancomycin dosing and adjustments. “Usually the ID physician will see people with more serious infections, then specify a target range. Pharmacy makes the dosing adjustment accordingly. It’s a pretty good system. But also I think it is all right for labs to put an indication on their report that higher levels may be appropriate for more serious infections.”

Dr. Agrawal specifies several details of dose adjustment. First, creatinine clearance must be taken into account. He advocates using the classic Cockroft-Gault equation for this. A shorthand method now in general use reports a value greater than 60 for patients with normal clearance. “It doesn’t give a precise value,” Dr. Agrawal says. “If you use the classic equation, it gives an actual value for dosing calculations.”

Second, the therapeutic level being targeted must, where possible, be correlated with the MIC of the organism. While 15 mg/L might be acceptable in most cases, for an organism with an MIC greater than one, a concentration of 20 mg/L might be needed. “Not everyone appreciates that a standard dose of 15 to 20 mg/kg of antibiotic will not always give you an AUC/MIC =400,” Dr. Agrawal says. “There is a need for clinical pathologists to discuss these issues with pharmacy or the treating physician.”

He calls the notion that you don’t need to monitor vancomycin therapy a partial misconception. “You don’t need to monitor in a so-called uncomplicated patient—one who is otherwise healthy with normal kidney function and a highly susceptible organism,” he says. “In reality, if you are giving vancomycin, especially in the hospital, chances are that things are a bit more complex.” In particular, you need to monitor any time you aim for a higher dose, such as 15 to 20 mg/L.

Dr. Agrawal doesn’t have great fear about toxicity with higher doses. “The older literature talks about nephrotoxicity and organ toxicity based on impure preparations of vancomycin,” he says. (The 2009 consensus review says “Vancomycin was initially dubbed ‘Mississippi mud’ because of the brown color of early formulations, which were about 70% pure.”) “More recently, you need to give very, very high doses before you see toxicity.” Vancomycin concentration should always be monitored when the drug is combined with another nephrotoxic drug.

Dr. Dawling, too, notes that fears about toxicity are based on older data with impure preparations. “Now that preparations are much purer, I think there is very little evidence to support that,” she says.

Her lab’s vancomycin report goes into the hospital system, which automatically sends an alert to pharmacy for review. Pharmacy might make a recommendation—“hold dose,” “increase dose,” or “be on lookout for declining renal function”—that is appended to the patient’s chart.

“We have another automatic function in place,” Dr. Dawling says. An electronic alert is sent to the treating physician if a patient’s serum creatinine increases by more than 0.5 mg/dL or by more than 50 percent from the previous reading. This, of course, does not necessarily reflect vancomycin toxicity. “I think you are going to get a lot of sick patients who will fit these criteria anyway,” she says. Either way, it could be a signal that the dose must be reduced.

Dr. Dawling points out a serious technical consideration related to the new target range for serum trough vancomycin concentrations. “I was just reviewing CAP proficiency data,” she says. “It’s all very well for people to recommend a certain target as long as they all measure the same value on the same sample. But if intralaboratory accuracy is not good enough, it could make the whole thing a bit of a joke.” Dr. Dawling points to results in the first chemistry set for 2009. For the CHM-02 specimen, she says, agreement among mean values for three widely used methods is “not so bad”: FPIA (n=494), 48.3 mg/L; PETINIA (1,389), 50; immunoturbidimetric (616), 44.2. For two less commonly used methods, however, the means were outliers: homogeneous EIA (192), 53.4; chemiluminescence (242), 31.5. With this degree of difference among methods, does it make sense to try to discriminate between 20 and 25 mg/L?

Dr. Jones has more fundamental reservations: He doubts the literature about widespread MIC creep. “Our surveillance data, which include hundreds of thousands of data points, show no evidence of widespread MIC creep,” he says. A key to this discrepancy, in his view, is that he uses reference methodology. “Much of the data from individual medical centers has been generated by nonreference methods,” he says. These are automated or manual commercial methods, some not validated against the broth-based microdilution method, which Dr. Jones calls “the only true reference method.” Because commercial methods are used so widely, their data become “the rule of fact,” Dr. Jones says. “This gives rise to the perception that MICs are higher. We may need to address the potential adverse clinical outcomes of this perception by making MICs correct.”

Dr. Ambrose of the Institute for Clinical Pharmacodynamics identifies as one consequence of this perception the guideline issued in 2005 by the American Thoracic Society and IDSA for managing hospital-acquired, ventilator-associated, and health care-associated pneumonia (Am J Respir Crit Care Med. 2005;171:388–416). The guideline recommended an initial dose of 15 mg/kg, about 1 g/12 hours, then monitoring and maintaining a trough level of 15 to 20 mg/L. “Unfortunately, the possibility of actually attaining a trough level of 15 to 20 following that regimen is quite low,” Dr. Ambrose says. In patients with moderately impaired renal function (creatinine clearance of 40 to 80 mL/min), it is about 15 percent. “What happened,” Dr Ambrose says, “was that physicians started treating with higher and higher doses. In some institutions, they got up to 4 grams per day trying to meet the goal of 15 to 20 as a trough level. That really does raise the risk of vancomycin nephrotoxicity.

“I would treat patients based on how they are doing clinically. I don’t have a problem with doses of 1 gram twice daily or 15 to 20 mg/kg, but if the patient was doing well I wouldn’t adjust the dose upward to reach a recommended trough level.”

“Treat the patient, not the blood level,” Dr. Jones says. The situations in which you may want to consider blood levels and adjust for toxicity are the hospitalized patients with documented bacteremia that persists in throwing off positive cultures after a week of therapy or in any case with a secondary focus of infection, such as a heart valve, endocarditis, or osteomyelitis, he says. “But the usual patients getting vancomycin for mild to moderate MRSA, who make up the vast majority of those getting vancomycin, do not require any use of lab services for serum vancomycin levels.”

Related Links

• Vancomycin alternatives