Special Feature

Cutting G or GM-CSF Dose and Cost in Half?

By Thomas J. Smith, MD, FACP

I have seen the pharmacy budget at a great institution, one known for its specialty care and its community service. This center has always prided itself on having one set of standards—high—for care.

Recently, though, its leaders have begun to question if they can continue to provide the best level of care to everyone. The reason? Drug cost. It’s escalating rapidly and cancer drugs are the major player. I have listed some of them in Table 1. (See Table 1.) These are among the 50 highest. Let’s see—$360,000 for irinotecan and $500,000 for trastuzamab—where is that money going to come from? In their fixed budget, similar to many insurance companies or integrated health care programs, it must come from somewhere.

Note that almost all are supportive care or palliative chemotherapy drugs, indicating the importance that the healthcare system puts on symptom relief and prolongation of life. One of them, Trastuzamab (Herceptin), may prove to be the straw that broke the camel’s back—at $1000 a week additional—since it does not replace anything but adds a new treatment, for one-third of breast cancer patients. It may force the hospital to decide between it and other treatments. How ironic that the first really useful monoclonal antibody would help force a two-tier system that would not allow the uninsured to benefit.

    Table 1
    Drug Expense at a Major Hospital: 1998 Total Drug Budget 
    Rank  Item Cost($)
    1   Filgrastim (Neupogen) 1,042,000
    2   Immune globulin 705,000
    3   Paclitaxel (Taxol) 699,000
    4   Midazolam (Versed) 667,000
    5   Milrinone (Cordarone) 520,000
    11   Erythropoietin (Procrit, Epogen) 388,000
    New   Irinotecan (projected) (Camptosar) 369,000
    New   Trastuzamab (Herceptin) ?500,000
    15   Abciximab (Reopro) 340,000
    17   Ondansetron (Zofran) 322,000
    19   Carboplatin (Paraplatin) 312,000
    27   Granisetron (Kytril) 219,000
    38   Pamidronate (Aredia) 159,000

What if You Could Cut the Dose in Half?

For most drugs, that is not an option. We know that reducing the dose of most drugs does not produce the same clinical benefit, either curative or palliative.

But Toner and colleagues in Australia did just that for hematopoetic growth factors: cut the dose of lenograstim (glycosylated rHuG-CSF) for primary prophylaxis (to prevent febrile neutropenia) from 5 to 2 mcg/kg.1

Toner et al pointed out the dose-limiting toxicity of most chemotherapy is neutropenia, but that the high cost ($200-300/d) has led to hot debate and different guidelines across countries. They reasoned that reducing the dose would be one way to ensure access but cut cost.

The Trial

Only 52 patients were randomized, but 40 patients were sufficient to have an 80% chance to detect a one-day difference in duration of neutropenia or 90% to detect 1.2 days. About half the patients were naive to chemotherapy and/or radiation. The regimens (CHOP, CNEP, CNOP, PE/PEB/PEVB, AC/EC, DHAC, M-VAC) were all standard ones associated with a risk of febrile neutropenia (FN). There was no placebo arm. Patients received one dose in one cycle and then crossed over to receive the other dose the next cycle. After the study, patients could receive lenograstim at 2 mcg/kg for the duration of treatment. Lenograstim was continued each cycle until the ANC was more than 10,000/mm3.

Most patients never had neutropenia: 53% had no grade III neutropenia and 70% had no Grade IV neutropenia. The main results are shown in Table 2. (See Table 2.)

    Table 2
    Summary of Results____________________________________
    2 mcg/kg
    5 mcg/kg
    P value
    Mean days ANC < 1000 
    Mean days ANC < 500
    Mean days hospitalization for fever
    Days of treatment until ANC >10,000
    Total dose
    1.83 mg
    4.08 mg

There was no difference in the number of outpatient visits or incidence or duration of hospital admission.

What Does this Trial Mean?

This does NOT mean that we can automatically reduce the dose of CSF from 5 to 2 mcg/kg for every patient. These were moderate-risk patients receiving moderately intense chemotherapy, and they only got two cycles for evaluation, not six or eight of CHOP.

It does add to a growing body of evidence that CSF doses more than 2 mcg/kg are active enough to prevent important neutropenia and that the lower doses may be as safe and effective. A dose-finding study for lenograstim in Europe showed improvement in ANC at all doses more than 2 mcg/kg, and Seymour and associates concluded that the 2 mcg/kg/daily dose was sufficient for moderately myelosuppressive therapy.2 McQuaker and colleagues in the UK showed that even low-dose filgrastim, 300 mcg/kg/day, enhanced ANC recovery after stem cell transplantation with chemotherapy and filgrastim mobilized stem cells.3 In addition, Martinez and colleagues showed that 50 mcg/kg gave similar results to 200 mcg/kg G-CSF in stem cell mobilization.4 They did note that the higher dose gave better yields of CD-34+ cells on a second round of mobilization, if needed, suggesting the dangers of switching to low doses without good clinical data.

Does this mean we should be nihilistic about CSFs? Not at all, and I use them regularly in mobilization and to support dose-intensity when it’s important. They have an important role, and can be lifesaving and reduce toxicity. But, to date, I am not aware of a single study that unequivocally shows better disease free or overall survival with the modest increase in dose possible with CSFs alone. If they did not cost $300 per day, the decision to use them for most patients would be a no-brainer.

This also means that lower doses could save cost without worsening care. In this study, the low-dose arm gave savings of 55%. At this dose, CSF use would be cost-saving with any regimen with a 16% or greater risk of febrile neutropenia. It might be even more if the CSF were stopped before ANC is more than 10,000/mm3.

Where Do We Go From Here?

This study needs prompt replication in the United States with our available drugs, filgrastim or sargramostim. If confirmed, the total amount of CSF used could be decreased by half.

A study of low vs. higher doses in seriously ill patients with FN would also be welcomed. The available studies show minimal or no benefit to CSF use in the treatment of FN, but many of us continue to use it. At $50 a day, minimally effective therapy is less an issue than at $300 a day. And a large trial of both doses, or low vs. high vs. placebo, might establish whether there is a benefit or not.

Take Home Message

Low doses of CSFs may be just as effective as higher doses for moderately intense routine chemotherapy. Confirmation of this trial is critically important, and if confirmed, appropriate use of CSFs could be expanded at lower total cost.


1. Toner GC, et al. J Clin Oncol 1998;16(12):3874-3879.

2. Seymour AM, et al. Eur J Cancer 1995;31A:319-324.

3. McQuaker IG, et al. J Clin Oncol 1997;15:451-457.

4. Martinez C, et al. Bone Marrow Transplantation 1997; 20:855-858.