Nelson S, Heyder AM, Stone J, Bergeron MG, Daugherty S, Peterson G et al. A randomized controlled trial of filgrastim for the treatment of hospitalized patients with multilobar pneumonia.
Overall Study Question
To assess whether filgrastim (r-metHuG-CSF [recombinant human methionine granulocyte colony- stimulating-factor]) is superior to placebo in decreasing the therapeutic failure rate in patients with multilobar community-acquired pneumonia. Four hundred eighty (480) hospitalized adults (>=18 years old) receiving antibiotics for multilobar community-acquired pneumonia were randomized to receive either filgrastim or placebo in addition to their standard antibiotic therapy. Study drug was started within 24 hours of hospitalization and continued for <=10 days, or until white blood cell (WBC) count reached > 75 x 109 L, IV antibiotics were discontinued, or patient was discharged from hospital. The primary endpoint was the incidence of therapeutic failure through day 29. Therapeutic failure was defined as development of organ dysfunction (ARDS, DIC, acute renal failure, or shock), empyema, or death from any cause. Secondary outcomes included mortality at 28 days, number of days in ICU, time to stability or improvement as demonstrated on chest radiograph, and incidence of chest radiograph resolution. Safety endpoints were adverse events and clinical laboratory abnormalities.
Are the Results of the Study Valid?
Was assignment of patients randomized?
Yes. Patients were randomized to receive filgrastim or placebo.
Were all patients who entered the trial properly accounted for and attributed at its conclusion?
Yes. Sixteen (3%) of patients did not finish the study for various reasons (patient request, administrative decision, or loss to follow-up). Withdrawals for each respective reason were balanced across the groups. The authors stated that the primary endpoint was analyzed on an intention-to-treat basis, while the analysis of secondary outcomes were based on an evaluable subset of subjects who met all eligibility criteria and received >=3 doses of study drug.
Were patients, their clinicians, and study personnel ‘blind’ to treatment?
The methods section suggests that this study was double-blinded, however, details of this blinding are not provided. The only detail provided is that the manufacturer provided the study drug and vehicle (to act as placebo) in identical vials. Although the primary care provider making daily care decisions may have been blinded, we do not know if persons evaluating the chest radiographs or determining the degree of organ dysfunction were blinded, which could potentially result in bias.
Were the groups similar at the start of the trial?
Yes. Baseline demographic characteristics were similar across the groups, and baseline disease characteristics were similar.
Aside from the experimental intervention, were the groups treated equally?
Aside from the intervention, patient supportive care and monitoring (concomitant medications, microbiologic cultures etc.) were allowed as medically indicated. Assuming the primary care provider was blinded, the groups were most likely treated similarly.
Overall, are the results of the study valid?
What were the Results?
How large was the treatment effect?
The incidence of patients experiencing therapeutic failure through day 29 was 42 (17.3%) in the placebo group versus 40 (16.9%) in the filgrastim group. The difference between the groups was not statistically significant in the intent-to-treat analysis, nor the evaluable subset analysis. This study was adequately powered, and sufficient patients recruited, to detect an absolute decrease in the proportion of treatment failures of approximately 50%, although no alpha and beta values for this calculation were provided. As well, no statistically significant effect of treatment was seen in any of the secondary endpoints or safety endpoints.
How precise was the estimate of the treatment effect?
No treatment effect was demonstrated and confidence intervals were not reported.
Will the Results Help Me in Caring for My Patients?
Can the results be applied to my patient care?
Yes. Although no treatment effect was noted, the patients enrolled in the study are representative of patients with multilobar community-acquired pneumonia. As well, this study appears to confirm the results of a similar study done by these authors in which filgrastim was not superior to placebo in terms of time to resolution of morbidity, mortality, or length of stay in hospital.
Were all clinically important outcomes considered?
Are the likely treatment benefits worth the potential harms and costs?
Filgrastim was not shown to be superior to placebo in decreasing the incidence of therapeutic failure, morbidity, or mortality in patients with multilobar community-acquired pneumonia. Therefore, it is unlikely that a treatment “benefit” would outweigh any potential risks.
The premise behind this study is that endogenous granulocyte colony-stimulating factor concentrations are increased in a variety of infectious disease states, suggesting an important role for this growth factor in these circumstances. As animal models of infection have indicated filgrastim can augment host response, and filgrastim has been shown to be well tolerated in humans for other indications (non-neutropenic diabetic foot infections), the authors decided to investigate whether filgrastim would be of benefit in patients with severe pneumonia. Amgen (manufacturer of filgrastim) has now funded several studies to this effect, both with safety endpoints (Phase 1) and clinical endpoints (Phase 3). Although filgrastim appears to be well tolerated, any treatment benefit has not been demonstrated. At this time, recommending filgrastim for this patient population is not warranted.