We are seeing a larger number of patients with underlying diseases who survive for a longer period of time and, therefore, get into the kinds of trouble that they were not able to survive and get into even up to 5 to 7 years ago. And so it is with neutropenic patients. We are seeing larger numbers of neutropenic patients. We are seeing patients with more severe neutropenia, and we are seeing patients with more prolonged neutropenia.

Just to sort of set the background, these data haven't changed in the last 30 to 35 years when it was first recognized that there was a link between neutropenia and infection. And the risk of infection is related to several factors related to neutropenia: the degree or depth of neutropenia, (the lower the neutropenia is the greater the chances of getting severe infection); also the duration of neutropenia, and there's some softer data indicating that if neutropenia develops very rapidly you have a greater chance of getting infected than if it takes a long time to develop. So there are several factors about neutropenia that are important--the two most important being the degree and duration of neutropenia.

When we see neutropenic patients there are several manifestations of neutropenic fever that we encounter, and we sort of try and pigeonhole these patients into what they might have at the onset of the febrile episode so that we might adequately take care of them. A large number of patients, most patients, will have what's called unexplained fever or FUO ("fever of unknown origin"). The percentages may differ depending on whether you use prophylactic antibiotics or not, but here are patients who have a febrile episode, you know that they're infected, but you don't have positive cultures and you don't have a clinical site of infection such as [in] cellulitis or pneumonia.

A certain percentage, 15%, 20%, 25% of patients will either have a clinical site such as pneumonia or cellulitis or will have positive cultures most often either from the blood or the urinary tract either as a sole cause of fever or, more often, superimposed on true infectious fever you may have noninfectious causes of fever such as neoplastic fever, tumor lysis syndrome, and then in patients receiving multiple drugs or receiving multiple transfusions you also will have fever associated with these. So this is sort of the spectrum or the nature of febrile episodes in neutropenic patients.

The playing field keeps changing, and so what you might have read in the 70s and 80s may not necessarily apply in your institution and mine anymore. And this is just to point out that as time has gone along the incidence of frequency of unexplained fever in my institution has increased. So we're seeing more patients who have cultures that are not positive or clinical sites that are not positive.

One of the reasons for this may be that they have been given more intensive chemotherapy; and then these patients are receiving prophylactic antibiotics. So even though they develop fever, we're not getting positive cultures in them.

Microbiologically documented infections? Again the playing field has changed dramatically. Gram-positives used to cause about a fifth of our documented infections back in the 70s. Currently, this is slightly older data, currently we're up to 70% as far as documented bacterial infections. And there are many institutions across the world where 70% of documented bacterial infections in this population are caused by Gram-positives, predominantly arising from catheter-associated infections. And, in response, Gram-negatives that used to cause 60%, 65%, 70% of our documented infections now are causing less than 30%.

I want to tell you these are documented infections, but Gram-positives are always in the equation as we will stress to you. What's also important is that polymicrobial infections, where you have multiple species of Gram-positives, Gram-negatives, anaerobes, even bacteria and fungi or bacteria and viruses coexisting together have increased in frequency and have, at least in my institution, doubled over the last few years.

A large variety of organisms cause infections in neutropenic patients, this is why when empiric therapy is given it necessarily has to be broad spectrum. We're not only documenting Gram-positives more often, but we're seeing a much bigger spectrum of Gram-positive infections. Coag[ulase]-negative Staphylococcus are isolated most frequently, but they're actually the organisms associated with the least amount of morbidity and mortality. They're a nuisance, but they don't necessarily always have to be treated with the most potent Gram-positive drugs such as glycopeptides.

On the other hand S aureus and viridans streptococci are associated with considerable morbidity and mortality-sometimes up to 30% to 40% in very accelerated or aggressive syndromes. Enterococcus has become an increasing pain, particularly VRE VRE (vancomycin-resistant Enterococcus). And then we have a wide variety of other Gram-positives -- Bacillus, Corynebacterium, Streptococcus species other than viridans, Micrococcus, Stomatococcus, etc.

Although we're isolating Gram-negatives less frequently, they're still part of the equation; and the big three still include E coli, Klebsiella, and Pseudomonas. And I want to point out just as again that Pseudomonas is a very common pathogen. There are data from some cancer centers saying that we don't see P aeruginosa any more, but you can extrapolate that to your population. You need to know what's going on in your hospital, and I'll show you in the next slide Pseudomonas is the second most commonly isolated Gram-negative from my institution. So I cannot use empiric regimens that do not have adequate anti-pseudomonal coverage, and there are those sorts of regimens out there. And then we see a wide variety of other Gram-negatives too, including the dreaded Acinetobacter and Stenotrophomonas. Fortunately, less than 4% of our infections are caused by these organisms; but they are increasing and are increasingly resistant to antibiotics.

This is just to show you similar data over time in terms of Gram-negatives. E coli has consistently been our most commonly isolated pathogen, and Pseudomonas has consistently been there, item number 2 or number 3. So it's very, very important to know what's happening in your institution because it could be totally different from mine or it could be totally different from what's happening right across the street. And if you're going to do three different hospitals and you get that call at 3 in the morning, you need to know which hospital it is, what microflora you're dealing with before you say such and such is the empiric regimen that you want to use.

What are the treatment options in the year 2000 for febrile neutropenic patients? These options are based on the guidelines that we created for the Infectious Diseases Society of America (IDSA). These are guidelines and recommendations. They don't apply to every institution. Local factors, again, are very important; but as a generalization, these are the choices that are available to us.

I call these a standard one-size-fits-all option because they are not based on risk. The newer way of treating febrile neutropenic patients, which we will discuss briefly, is risk-based therapy. But here you treat all patients as if they have the same risk. They're all given hospitalized therapy, and they're all treated with these options.

There are two major options. You can either use combinations or you can use single agent or monotherapy. And among the various combination regimens, again, there are different options which we will discuss. And, again, among the single agent or monotherapy regimens, there are different options--third-generation and fourth-generation cephalosporins and carbapenems. So you see the options are quite similar to what the situation is in other settings such as pneumonia, but the pathogens and the principles are slightly different.

Aminoglycoside-based regimens have been used for longer than any other kind of regimen, and you can combine an aminoglycoside in my institution. It happens to be amikacin simply because we have less resistance to amikacin than we do to gentamicin or tobramycin. But if you have institutions where those drugs have adequate susceptibility patterns, they are very effective and cheaper.

One can combine amikacin with either penicillins, cephalosporins (third- or fourth-generation), and even drugs like imipenem and meropenem. Now I put a question mark against ceftriaxone because this is the kind of the regimen I'm talking about which does not have adequate initial pseudomonal coverage. Aminoglycosides by themselves are not effective in Gram-negative infections in neutropenic patients. Ceftriaxone does not have adequate pseudomonal coverage. So in this sort of combination, one would be relying only on amikacin for anti-pseudomonal coverage which in a severely neutropenic patient is not adequate. So although there are studies out there talking about ceftriaxone plus amikacin, remember that if in your institution Pseudomonas is a problem, you might want to use something else.

These are standard regimens and are associated with certain advantages and disadvantages. You do have a broad initial antimicrobial spectrum, and there is moderate Gram-positive activity. Although clearly they're not optimum for all the resistant Gram-positives that we've talked about including penicillin-resistant viridans Streptococci and many of the Enterococci and methicillin-resistant S aureus and coag[ulase]-negative Staphylococcus. But the efficacy in many, many, many studies conducted over several decades for these sorts of combinations are between 65% to 85%. This is to the original regimen without modification of the regimen.

Perhaps because you're using drugs with different mechanisms of action there is a reduced potential for selecting resistant organisms, although the data here are soft as sometimes it happens, sometimes it doesn't. The downside is that particularly with aminoglycosides in elderly patients who've also received multiple other insults to their kidneys, you have the problem of nephrotoxicity. We generally don't measure ototoxicity, but it is a substantial problem. Again, in the elderly you need to monitor levels. This adds to the cost and discomfort. Aminoglycosides, as I said, must always be given in combination and then extended-interval therapy with aminoglycoside--the so-called once daily might reduce the level of toxicity. So these are standard regimens, but certainly there is the need for other alternatives.

One of the alternatives that became popular back in the 1980s was the so-called double beta-lactam combination. Here you generally combine an anti-pseudomonal penicillin with an extended-spectrum cephalosporin. So piperacillin or pip-tazo plus ceftazidime, aztreonam plus ceftazidime, and again in order to get better staphylococcal activity or Gram-positive activity. Even regimens like nafcillin plus ceftazidime have been used.

These are much less popular today than they used to be, again, because of the emergence of extended-spectrum beta-lactamase and Bush type 1 beta-lactamase producing organisms. Nevertheless, in studies done back in the 80s, clearly the efficacy was similar to aminoglycoside-based regimens. There was much less nephro and ototoxicity. You did have an initial broad spectrum; but, as we said, there's a greater potential for the emergence of resistance. The activity against many Gram-positives was suboptimal, and the toxicity that you did see to an increased level was the occurrence of rashes: 30% to 40% of patients got into trouble with cutaneous rashes. So an alternative, but perhaps not the greatest alternative in today in the year 2000.

Because of the tremendous increase in the frequency of Gram-positive infections, in my institution again from 20% up to 70%, and in many other institutions the same, basically vancomycin or glycopeptide therapy, teicoplanin in those countries where it's available has been combined very successfully again with a variety of drugs that have Gram-negative activity. The third-generation, fourth-generation cephalosporins, aztreonam in patients who have severe allergic problems to beta-lactams, pip-tazo, imipenem, quinolones. All of these have been used successfully in combination with vancomycin.

The advantages clearly here are that not only do you have a broad initial spectrum, but you have potentially the best spectrum available against Gram-positives including methicillin-resistant Gram-positives which still account for a substantial number of our coag[ulase]-negative Staphylococcus and some S aureus.

The efficacy here may be a little bit higher actually than with other regimens, to the original regimen, because one of the most common modifications that you make with other regimens is to add vancomycin. Here you have vancomycin up front already, so although the overall mortality at the end of therapy remains the same for most of these regimens, the response to the original regimen here might be a little bit higher.

There may be some added toxicity with vancomycin, but it isn't huge; but you still, in some patients, particularly with unstable renal function, have to monitor levels. But the big problem is the selection of resistant or tolerant organisms. Everybody knows the story of VRE (vancomycin-resistant Enterococcus) and that the use of both parenteral and oral vancomycin is to some extent responsible for the increase in VRE. But we're also seeing organisms that many of you may not be very familiar with. We see maybe 5 or 10 cases a year, but they're all vancomycin resistant: Leuconostoc, Pediococcus, and Lactobacillus.

With Ron Jones we just did a 10-cancer center surveillance study, and 11% of our Bacillus species and some of our Rhodococcus species now have MICs that are greater than 8 for vancomycin. So it's coming, and we need to be careful to use vancomycin only when it's necessary. It should clearly not be used on a routine basis even though vancomycin-containing regimens might give you a higher initial response; 2 or 3 days of fever is not going to hurt the patient as much as developing VRE or a resistant Gram-positive [infection] that's going to kill them.

We have less experience with some of the newer regimens that are coming out. Beta-lactams have been combined with quinolones. Both penicillins and cephalosporins have been combined with a quinolone, most commonly ciprofloxacin because of its good Gram-negative activity. And they might actually be a little bit better for tissue-based infections such as pneumonia because the quinolones have better penetration into lung tissue than do aminoglycosides. So my preference in somebody who has Gram-negative pneumonia in whom I need to give combination therapy may be a beta-lactam plus a quinolone rather than a beta-lactam plus some aminoglycoside. I do occasionally aerosolize the aminoglycoside.

They should clearly not be given, at least empirically, in patients who are receiving quinolone prophylaxis because if you've already developed an infection on quinolone prophylaxis, the likelihood is that it's not going to be very effective for that infection. There is less experience with these regimens, but in some studies that have been done, if they used in combination--clearly not quinolones alone but combination therapy with quinolones--that they are effective regimens.

We don't know the role of the newer agents--gatifloxacin, moxifloxacin, and gemifloxacin. They are broader spectrum, and they actually certainly have enhanced activity against Gram-positives. But we don't have clinical data....

The use of third-generation cephalosporins such as ceftazidime; I have put a question mark here now because in my institution, based on current susceptibility patterns, I would not be able to use ceftazidime as monotherapy because of ESBL-producing organisms, because the MIC ₉₀ for P aeruginosa is greater than 256. Fourth-generation cephalosporins such as cefepime do have better activity, both against Gram-positives and Gram-negatives than does ceftazidime. So they do represent a viable option for monotherapy, and clearly the carbapenems--imipenem and meropenem--also have a broad enough spectrum that they represent a viable option for monotherapy.

I want to show you some susceptibility data, again picking the three drugs that I believe are being used most widely for monotherapy or single-agent therapy and just running them against a variety of Gram-positives here. Dan Diekema out at Ron Jones lab published these data last year in Antimicrobial Agents in Chemotherapy. These isolates were collected from 10 different cancer centers in various different geographic areas of the United States so they give you an idea of what's going on countrywide and not just one particular center.

S aureus and these obviously are methicillin-susceptible isolates--98%, 99%, and 100% susceptible. But ceftazidime only 14%, 15%. Methicillin-susceptible, coag[ulase]-negative Staphylococcus similar story. A little bit higher susceptibility but clearly not as good as cefepime and imipenem. Viridans streptococci, which are sort of pen[icillin]-resistant pneumococcal equivalent, we don't see too much pen-resistant pneumococcus in neutropenic patients; but we certainly see a lot of viridans streptococcal infections, and clearly ceftazidime doesn't cut it against viridans streptococci anymore whereas cefepime and imipenem are better. And down the line you can see beta-hemolytic Streptococcus, Micrococcus. Bacillus is a problem; and, as I said, 8% to 12% of Bacillus now are resistant to vancomycin due, in this particular study, because the carbapenems were better than the cephalosporins. So overall, particularly for Gram-positives, we don't have good activity with the third-generation cephalosporins anymore. We do have better activity with the fourth-generation cephalosporins and the carbapenems.

These are data against Gram-negatives from my institution, and these data were generated back in 1995. So they're a little bit dated. We're getting ready actually to do our year 2000 susceptibility surveillance, but again you'll see imipenem, cefepime, meropenem, and in this study pip-tazo, being the 4 drugs that have the greatest Gram-negative activity. And the resistance rates among all Gram-negatives--we tested about 750 isolates--was less than 10% for these whereas for all the other drugs that we tested it was greater than 10%. And as I showed you in my institution, unfortunately, tobramycin has a high level of resistance among Gram-negatives. So I can't use it.

In terms of single-agent therapy, we do have, with some selected drugs--and in my opinion the fourth-generation cephalosporin cefepime or the carbapenems--you do have a broad initial antimicrobial spectrum including many Gram-positives and Gram-negatives. The carbapenems have much greater anaerobic activity. Do we necessarily need it? Probably not, but there are certain situations like perirectal abscesses or neutropenic enterocolitis where having that additional kick might be good; and if you're using a cephalosporin in that situation, you might want to add a drug like clindamycin or metronidazole.

We've had discussions about the stability of these agents against ESBLs and Bush type 1s, so I won't belabor the point. But there is an acceptable toxicity profile and certainly, although I haven't shown you multiple studies, established equivalent efficacy with monotherapy compared to either aminoglycoside-based on vancomycin-based combination regimens. Is there an increased likelihood for resistance? Probably. But we need to monitor these patients carefully, and there is certainly less likelihood for the emergence of resistance when we're using fourth-generation cephalosporins and carbapenems than we are using other monotherapy regimens.

I wanted to turn to what I call risk-based therapy in febrile neutropenic patients. We know that not all cancer patients with neutropenia have the same risk for developing infection or for getting into trouble once they have developed infection in terms of getting complications.

There are now what we call risk prediction strategies, either using clinical criteria, which my friend Anne Rubenstein and I have used for 10 years at M. D. Anderson [Cancer Center], or the two articles that were published in last year's New England Journal where there are patients who are stable hemodynamically who don't have microbiologically documented infections and in whom duration of neutropenia is short lived; 7 to 10 days.

There are statistically derived risk-prediction rules. The first one was published by Jim Talcott way back in 1989 and the more recent MASCC (Multinational Association for Supportive Care in Cancer) risk Index which was published last month in the Journal of Clinical Oncology, all enabling you to prescribe risk to patients--either high risk, moderate risk, or low risk very early on in the febrile episode.

This is a very simplified version of what risk assessment can do for you, but in my institution we consider patients to be high-risk patients when they have fever and neutropenia, if they have an underlying hematological malignancy such as leukemia or aplastic anemia if they received allogeneic bone marrow transplants or are getting therapy for allogeneic bone marrow transplants because in these patients they will have prolonged neutropenia that's generally greater than 14 days. These patients are given standard hospital-based therapy that we've just spent 15 minutes discussing.

There are patients who are at little bit lower risk than these, and I sort of call them moderate risk patients. These are patients with solid tumors such as breast cancer or even patients with lymphoma who are getting intensive chemotherapy and then are getting autologous bone marrow transplants or peripheral stem cell support. There the risk is not quite as great as this group. They have maybe an intermediate duration of neutropenia, 7 to 14 days; and in these patients what we do is we admit them to hospital initially, stabilize them over a period of 48 to 72 hours, and then if they're stable, we send them home on outpatient regimens.

The reason we do this is that it takes 72 hours or more for these patients to get colonized with hospital flora. So if we can actually get them out of the hospital before that time, we then don't encounter the same level of nosocomial pathogens such as Acinetobacter and Stenotrophomonas and others. And so it's still an advantage. If they're hemodynamically stable, etc., to get them out sooner.

And then there is the truly low risk patient. This is the patient with solid tumor who's receiving conventional chemotherapy. There are no other comorbid medical conditions. They have short-lived neutropenia--roughly equivalent to 7 days or less. In these you can give them hospital-based oral therapy, but more often you can give them outpatient parenteral antibiotic therapy or sequential therapy. If they don't have nausea/vomiting, they can tolerate oral therapy. And we've been doing that now for 10 years.

Basically these are the advantages and disadvantages associated with risk-based therapy. It's more cost effective. You encounter fewer nosocomial infections, and although this doesn't get a lot of press, quality of life, particularly for cancer patients, is very important; and it's clearly enhanced in this situation. There is more appropriate utilization of resources overall.

The disadvantages perhaps include noncompliance with oral regimens, although with cancer patients again we have not found these to be a major problem.

There is the potential for developing serious complications, so you really must have adequate follow-up and good patient selection. But again, we've not seen too many patients who have gotten into trouble at home; and it is, we only treat them if they live within a certain geographic radius of our hospital so that if they do get into trouble, they can come back in.

This is my summary slide. In the year 2000 the current treatment options for febrile neutropenia should actually be risk-based therapy. High-risk patients should receive standard hospital-based therapy according to the IDSA guidelines that we've talked about. Low-risk patients can safely be treated by outpatient therapy--parenteral or oral. Most of these are quinolone-based regimens because the quinolones still are the best oral agents with Gram-negative activity. And moderate-risk patients can be stabilized in the hospital and discharged early so that you can save them again exposure to nosocomial pathogens.