The study, led by L. Falchero of the Hopitaux Nord-Ouest respiratory medicine department in Villefranche-sur-Saone, drew on the prospective national cohorts KBP-2020-CPHG and ESCAP-2020-CPHG, which captured patients newly diagnosed with lung cancer in 2020 and treated in non-academic public hospitals across France. Because these cohorts were designed to enroll consecutive patients nationwide rather than selected referral cases, they offer an unusually faithful snapshot of everyday oncology, in contrast to the carefully curated populations of randomized trials. From a total of 527 patients with small-cell lung cancer, the investigators applied predefined eligibility criteria for G-CSF primary prophylaxis, based on the febrile neutropenia risk of each chemotherapy regimen and on patient-related risk factors, in line with guideline recommendations.

The headline finding is stark in its simplicity. Of the 527 patients, 447, or 84.8 percent, met the predefined eligibility criteria for G-CSF primary prophylaxis. Yet among those eligible patients, only 285, or 63.8 percent, actually received the prophylaxis. In other words, roughly one in three patients for whom guidelines would have supported preventive G-CSF treatment did not get it. A sensitivity analysis, designed to test whether the eligibility definition itself drove the result, confirmed the picture: 439 patients remained eligible under the alternative assumptions, and 281 of them, or 64.0 percent, received prophylaxis, a proportion essentially unchanged from the primary analysis.

Perhaps the most revealing part of the study lies in who did and did not receive the treatment. Patients given G-CSF primary prophylaxis were significantly younger than those who were not, with a mean age of 66.4 years versus 69.4 years, a difference that reached statistical significance at p less than 0.001. Performance status, a standardized measure of a patient’s functional fitness, also separated the groups sharply: 76.3 percent of the prophylaxis group had a good performance status of 0 or 1, compared with 56.0 percent of the non-prophylaxis group, a difference significant at p less than 0.0001. These patterns hint at a systematic bias in prescribing, in which fitter, younger patients are preferentially offered the protective therapy while older and frailer patients, who may in fact be at higher risk of severe complications from febrile neutropenia, are left without it.

The biology behind the recommendation is worth unpacking. G-CSF is a growth factor that stimulates the bone marrow to produce neutrophils, the white blood cells that form the first line of defense against bacterial and fungal infections. Intensive chemotherapy regimens such as carboplatin or cisplatin combined with etoposide, the backbone of small-cell lung cancer treatment, suppress the marrow profoundly, and the resulting nadir in neutrophil counts typically occurs one to two weeks after each cycle. When fever develops during this window of profound neutropenia, mortality and intensive-care admission rates rise substantially, and prior studies cited by the authors have documented both the incidence of febrile neutropenia in chemotherapy-naive lung cancer patients receiving etoposide plus platinum and the risk factors that predict it, including older age and poor performance status.

Prophylaxis given before the first chemotherapy cycle, so-called primary prophylaxis, is intended to blunt that nadir before it happens, rather than reacting after a first febrile episode. Beyond preventing infections, maintaining the neutrophil count also helps preserve chemotherapy dose intensity, the ability to deliver each planned cycle on schedule and at full dose, which a systematic review and meta-analysis by Lyman and colleagues linked to cancer outcomes. Japanese studies and a 2022 systematic review and meta-analysis published in International Journal of Clinical Oncology have specifically examined the effectiveness and safety of G-CSF primary prophylaxis in lung cancer, supporting its role in this population. French national thoracic oncology guidelines, most recently updated in 2025 by the Auvergne-Rhone-Alpes ARISTOT group, similarly codify when prophylaxis should be offered.

Against that backdrop, the descriptive survival data from the French cohorts are intriguing but must be read with caution, as the authors themselves emphasize. Among patients eligible for primary prophylaxis, median overall survival was 11.2 months in the group that received it, with 3-month and 6-month overall survival of 89.8 percent and 75.4 percent, compared with 8.4 months and 72.8 percent and 63.0 percent respectively in the group that did not. Progression-free survival showed a similar early separation, with 3-month and 6-month rates of 86.7 percent and 56.1 percent in the prophylaxis group versus 70.4 percent and 51.2 percent in the non-prophylaxis group. However, the authors are explicit that these estimates are descriptive only: there were baseline imbalances between the groups, and the survival curves cross at around 12 months, meaning the apparent early advantage cannot be interpreted causally. Sicker patients received less prophylaxis, and sicker patients also die sooner, a classic confounding pattern that no observational analysis of this design can untangle.

The finding of incomplete guideline implementation is not unique to France. A German representative survey published in Supportive Care in Cancer documented gaps in G-CSF guideline adherence among German oncology practices, and a Moroccan single-center experience likewise reported variable use of primary prophylaxis. Physician preference studies add another layer of complexity: a discrete-choice experiment published in BMJ Supportive and Palliative Care found that French physicians weigh factors beyond clinical risk when deciding on febrile neutropenia prophylaxis, and a 2024 questionnaire study on outpatient G-CSF prescribing concluded that considerations beyond clinical factors shape real-world decisions. Practical barriers, including injection logistics, cost concerns, and the perception that elderly or frail patients derive less benefit, likely all contribute to the shortfall observed in the French data.

The therapeutic landscape of small-cell lung cancer has also shifted in ways that raise the stakes of supportive care. The landmark IMpower133 and CASPIAN phase 3 trials established chemoimmunotherapy, combining platinum-etoposide with atezolizumab or durvalumab, as the new first-line standard for extensive-stage disease, and the KBP-ESCAP cohort program itself has documented the impact of immunotherapy on real-life survival in France. In this era, avoiding treatment delays and dose reductions caused by neutropenic complications matters more than ever, since interruptions may compromise the delivery of both cytotoxic and immunotherapeutic components. Safety questions around G-CSF use during concurrent chemoradiotherapy in limited-stage disease, explored in secondary analyses of the CONVERT trial and in retrospective cohort studies of pegylated G-CSF, add nuance but do not negate the case for prophylaxis in patients treated with chemotherapy alone.

What the French analysis ultimately delivers is a benchmark. By quantifying, in a nationwide prospective cohort of 527 consecutive small-cell lung cancer patients, that only about 64 percent of guideline-eligible individuals received G-CSF primary prophylaxis, and by showing that the shortfall concentrates among older patients and those with poorer performance status, the study gives clinicians, hospital administrators and guideline developers a concrete target for quality improvement. The authors suggest that interventions such as electronic medical record prompts, which a 2015 study showed could influence growth factor use, might help close the gap. Whether the survival differences observed in the early months reflect genuine benefit or merely the confounding of patient selection, the message for practice is unambiguous: the decision to withhold prophylaxis should be a deliberate, guideline-informed judgment, not a default driven by age or frailty, because the patients most often left unprotected may be precisely those least able to survive a febrile neutropenic episode.

Subject of Research: Real-world use of G-CSF primary prophylaxis against febrile neutropenia in small-cell lung cancer

Article Title: Real-world use of G-CSF primary prophylaxis in small cell lung cancer: secondary data analysis from the French national cohorts KBP-2020-CPHG and ESCAP-2020-CPHG

Article References: Real-world use of G-CSF primary prophylaxis in small cell lung cancer: secondary data analysis from the French national cohorts KBP-2020-CPHG and ESCAP-2020-CPHG. (n.d.). https://doi.org/10.1007/s00520-026-11223-7

Image Credits: AI Generated

DOI: 10.1007/s00520-026-11223-7

Keywords: small-cell lung cancer, G-CSF, primary prophylaxis, febrile neutropenia, chemotherapy, guideline adherence, real-world evidence, supportive care, etoposide-platinum, performance status, France, national cohorts

For years, clinical guidelines have drawn a bright line at twenty percent. If a chemotherapy regimen is expected to cause febrile neutropenia in at least one in five patients, guidelines from the American Society of Clinical Oncology, the European Organisation for Research and Treatment of Cancer, and the National Comprehensive Cancer Network recommend giving granulocyte colony-stimulating factor, or G-CSF, prophylactically from the first cycle. This injected growth factor stimulates the bone marrow to keep producing neutrophils through the nadir of each chemo cycle, and randomized evidence shows it reduces both febrile neutropenia and chemotherapy-associated mortality. The trouble is that the twenty percent figure for R-CHOP has rested largely on trial data, and a new systematic review argues that those data may systematically understate what actually happens in ordinary clinics.

The review, published in Supportive Care in Cancer by Jamiel Reyes and Alexander Reyes, took an unusually careful approach to this question. The authors searched MEDLINE, CENTRAL, ClinicalTrials.gov, Europe PMC, Epistemonikos, Lens.org, Google Scholar, and citation trails from inception through 8 September 2026, looking for studies that reported febrile neutropenia incidence in adults with diffuse large B-cell lymphoma receiving first-line R-CHOP or R-CHOP-like immunochemotherapy. The protocol was prospectively registered in PROSPERO, and the team assessed risk of bias with the ROBINS-I tool for observational studies and RoB 2 for randomized trials, then graded the certainty of the evidence using the GRADE framework.

Fifteen studies covering 5,440 patients and 885 febrile neutropenia events made the final analysis. But the authors did something that many meta-analyses of this kind skip: they stratified everything by study design, pooling observational cohorts separately from clinical trials. The rationale is technical but important. Trials and routine-practice cohorts measure the same complication under materially different conditions, and blending them can produce a single number that describes neither setting well. In trials, patients tend to be fitter, monitoring is more intensive, supportive care protocols are more standardized, and prophylactic G-CSF use is often higher and more consistent than in the community.

The observational stratum told a strikingly consistent story. Across ten cohorts enrolling 4,020 patients, the pooled incidence of febrile neutropenia was 19.0 percent, with a 95 percent confidence interval of 17.1 to 21.1 percent. Heterogeneity was low, with an I-squared statistic of just 34 percent, meaning the studies agreed with one another far more than is typical in meta-analyses of this kind. The prediction interval, which estimates where a future study would likely fall, ran from 15.3 to 23.3 percent, and the estimate proved stable across every sensitivity analysis and leave-one-out test the authors performed, including a prespecified analysis restricted to the nine cohorts using core R-CHOP. In other words, no single study was driving the result.

The trial stratum looked very different. Five clinical trials enrolling 1,420 patients yielded a pooled febrile neutropenia incidence of 11.7 percent, but with a confidence interval stretching from 5.7 to 22.7 percent and an I-squared of 79.4 percent, indicating severe heterogeneity. Individual trial estimates ranged from 2.4 to 20.0 percent, a nearly tenfold spread that makes the pooled figure difficult to interpret. The formal comparison between the two designs did not reach conventional statistical significance, with a p-value of 0.051, but the authors note that the difference depended substantially on a single trial. When the large 2013 trial by Cunningham and colleagues comparing 14-day and 21-day R-CHOP schedules was omitted, the trial-stratum estimate shifted to 15.5 percent and the gap between designs essentially disappeared.

What does this mean for the twenty percent threshold? The observational estimate of 19.0 percent sits squarely at, not below, the conventional trigger for primary G-CSF prophylaxis, and its prediction interval crosses well above it. The authors’ conclusion is blunt: in routine practice, approximately one in five adults with diffuse large B-cell lymphoma receiving first-line R-CHOP-based therapy develops febrile neutropenia, and trial-derived estimates should not be assumed to represent this real-world burden. For clinicians who have been reassured by trial figures suggesting the risk is closer to one in ten, the message is that the community setting may be considerably riskier than the pages of a phase 3 protocol suggest.

The mechanistic reasons for the discrepancy are not hard to reconstruct. Trial populations are enriched for patients who meet strict eligibility criteria, often excluding older adults with multiple comorbidities, patients with poor performance status, and those with the very risk factors, such as advanced age, elevated lactate dehydrogenase, bulky disease, and low baseline blood counts, that observational studies have repeatedly linked to febrile neutropenia. Trials also tend to enforce prophylactic growth factor use more rigorously and to monitor neutrophil counts on tight schedules, catching problems before they become fevers. Real-world cohorts, by contrast, capture the full spectrum of patients, including elderly and frail individuals who make up a large share of diffuse large B-cell lymphoma diagnoses, and they reflect the uneven adoption of prophylaxis in everyday oncology.

The stakes extend beyond the acute emergency itself. Febrile neutropenia is one of the leading causes of chemotherapy dose reductions and treatment delays, and for a curable disease like diffuse large B-cell lymphoma, compromised dose intensity can translate directly into worse survival. Earlier research, including the INC-EU prospective observational study and analyses of R-CHOP delivery by Pettengell and colleagues, documented how neutropenic complications disrupt treatment schedules and drive hospitalizations. Prophylactic G-CSF, particularly the long-acting pegfilgrastim given once per cycle, has been shown to preserve relative dose intensity, and studies in patients aged 75 and older suggest it reduces mortality in this vulnerable group. If roughly one in five unselected patients experiences a febrile neutropenia event without adequate prophylaxis, the case for applying the guideline threshold generously, rather than treating it as a ceiling, becomes considerably stronger.

The review is not without caveats, and the authors are transparent about them. The between-design difference was not statistically significant at the conventional threshold, and the analysis was reported as an observation rather than a hypothesis test. The registered protocol underwent three amendments, including replacing a comparative outcome with patient-level febrile neutropenia incidence as the primary endpoint and adopting the design stratification during revision, changes the authors documented in the PROSPERO record. All fifteen studies were pooled irrespective of design in a secondary analysis, yielding 17.0 percent with high heterogeneity of 83.3 percent, a figure that illustrates precisely why the stratified approach matters: pooling across designs produces a number with a confidence interval so wide and an I-squared so high that it obscures the two very different realities it averages together. Still, the core observational finding is robust, low-heterogeneity, and consistent with the biological logic of the regimen. As newer first-line options such as polatuzumab vedotin-based therapy enter wider use and add their own myelosuppressive profiles to the mix, the review’s central lesson will only grow in relevance: when deciding who needs prophylactic growth factors, the numbers that count are the ones from the clinic down the street, not just the ones from the trial protocol.

Subject of Research: Febrile neutropenia incidence in adults with diffuse large B-cell lymphoma receiving first-line R-CHOP-based therapy

Article Title: Febrile neutropenia in adults with diffuse large B-cell lymphoma receiving first-line R-CHOP-based therapy: a systematic review and meta-analysis of observational and trial evidence

Article References: Reyes, J., & Reyes, A. (2026). Febrile neutropenia in adults with diffuse large B-cell lymphoma receiving first-line R-CHOP-based therapy: a systematic review and meta-analysis of observational and trial evidence. Supportive Care in Cancer, 34(10), Article 1065. https://doi.org/10.1007/s00520-026-11301-w

Image Credits: AI Generated

DOI: 10.1007/s00520-026-11301-w

Keywords: diffuse large B-cell lymphoma, febrile neutropenia, R-CHOP, G-CSF, primary prophylaxis, meta-analysis, systematic review, real-world evidence, supportive care, neutropenia, rituximab, clinical trials