Chemotherapy-Induced Neutropenia (CIN) — Comprehensive Research Report

Prepared 2026-08-31 for dismech knowledge-base curation.

A framing note before the data: this entry is not a disease in the "broken gene → broken protein → sick person" sense. It's an iatrogenic injury — we deliberately poison a fast-dividing tissue to kill another fast-dividing tissue, and the bone marrow is collateral. So several template sections (causal genes, inheritance, natural disease in other species) don't map cleanly, and I've said so plainly rather than manufacturing content. Where the section does have a real analogue — germline pharmacogenomics standing in for "causal genes," cytotoxic drug class standing in for "environmental exposure" — I've curated that instead and flagged the substitution.

Evidence-verification status, stated up front: PMIDs marked ✅ below were retrieved and read this session (via NCBI E-utilities or PMC). PMIDs marked ⚠️ are derived from PubMed URLs and search-result summaries and were not opened directly — treat every quoted string from those as a lead, not a snippet. Before any of these become snippet: values in a dismech entry, run just fetch-reference PMID:NNNNNNNN and match the exact substring against the cache. Ontology IDs marked ✅ were resolved live against the EBI OLS4 API this session; ones marked 🔍 need a local OAK lookup before binding.

1. Disease Information

What it is

Chemotherapy-induced neutropenia is the drop in circulating neutrophils that follows cytotoxic anticancer therapy, because the drugs that hit dividing cancer cells also hit the dividing granulocyte precursors in marrow. The marrow is basically a factory running three shifts — stem cells, a dividing (mitotic) pool, and a maturing (post-mitotic) storage pool — and chemotherapy shuts down the middle shift. The warehouse keeps shipping for a few days, then runs dry. That lag is the nadir.

Its clinically dangerous form is febrile neutropenia (FN): fever arriving while the neutrophil count is on the floor, which is a medical emergency because the usual sign of infection — pus, inflammation, a visible abscess — requires neutrophils to make, and there aren't any. The fever is often the only sign.

Identifiers

⚠️ MONDO gap — important curation finding. A direct OLS query for MONDO terms matching "chemotherapy-induced neutropenia," "drug-induced neutropenia," or "febrile neutropenia" returned nothing. MONDO carries MONDO:0001475 (Neutropenia) plus a long tail of congenital neutropenia syndromes, but no drug-induced or chemotherapy-induced child that I could resolve. Recommended handling in dismech: anchor disease_term on MONDO:0001475 with a mappings.mondo_mappings entry using mapping_predicate: skos:narrowMatch, per the design-decisions rule for promoted concepts lacking their own term, and record the gap in notes:. Do not invent a MONDO ID here — my own memory note (never guess MONDO ids) exists precisely because guessed IDs resolve to real-but-wrong concepts.

Synonyms

Chemotherapy-induced neutropenia; CIN; chemotherapy-induced myelosuppression (broader — includes anemia and thrombocytopenia); myelotoxicity; neutropenic fever; febrile neutropenia; FN; chemotherapy-induced febrile neutropenia; drug-induced agranulocytosis (when profound); bone marrow suppression.

Note the near-synonym trap: "chemotherapy-induced myelosuppression" (CIM) is the multilineage parent, of which CIN is the neutrophil lineage. Trilaciclib's label and the trilaciclib literature use CIM, not CIN. Keep the two distinguishable in the entry.

Data provenance

Both. Individual-patient EHR/claims data drive the burden and incidence figures (Boccia et al. pooled US claims and hospital datasets ✅ PMID:35552754); aggregated trial and registry data drive the regimen-specific risk estimates and the guideline thresholds.

2. Etiology

Primary cause

The cause is the drug. This is a dose-dependent, largely predictable, on-target-off-tissue toxicity, and the "genetic etiology" here is host pharmacogenomics modulating exposure, not a disease gene.

Highest-risk agent classes (mechanistic grouping):

High-FN-risk (>20%) regimen exemplars documented in the literature: TAC (docetaxel/doxorubicin/cyclophosphamide) is consistently classified >20% ⚠️; TC (docetaxel 75 + cyclophosphamide 600 q3w) has reported FN rates spanning 4–69% across studies, and a meta-analysis supports G-CSF primary prophylaxis for it ⚠️ PMID:26337685; TPF (docetaxel/cisplatin/5-FU) for head and neck ⚠️; dose-dense AC-T; R-CHOP in older lymphoma patients; topotecan and platinum-etoposide in SCLC.

Risk factors

Treatment-related (dominant): - Regimen myelotoxicity and dose intensity — the single largest determinant - Prior chemotherapy or radiotherapy (especially pelvic/vertebral fields — marrow reserve is a bank account you can overdraw) - Concurrent radiotherapy - Cycle number: FN clusters in cycle 1 — a prospective multinational MASCC study of 364 patients across 1,601 cycles found FN in 5% of first cycles, 3% in cycles 2–3, 1% in cycles 4–6, with 9% of patients having ≥1 FN episode and 59% of all FN events occurring in cycle 1 ⚠️ (PMC10570161). That front-loading is why primary prophylaxis beats secondary.

Patient-related: - Age ≥65 — the most consistently replicated host risk factor - Low baseline ANC / low pretreatment absolute lymphocyte count / low pretreatment monocyte count ⚠️ (PMC5922370) - Poor performance status, low albumin, low BMI (or, in some analyses, obesity via dosing practice) - Bone marrow involvement by tumor - Renal or hepatic impairment (reduced drug clearance → higher exposure) - Comorbidities: cardiovascular, hepatic, diabetes - Advanced disease stage; hematologic malignancy > solid tumor - Open wounds / active infection

Genetic risk factors — see §4, but in brief: DPYD (fluoropyrimidines), UGT1A1 (irinotecan), TPMT and NUDT15 (thiopurines), CYP2B6 (cyclophosphamide), CDA (gemcitabine/capecitabine), plus GWAS hits at SLC15A1, NR1I2(PXR), FMO3, and platinum loci at 2q24.3 and 17p12.

Environmental / lifestyle: malnutrition; poor dentition and periodontal disease (a portal of entry); crowded or high-exposure living or occupational settings; no strong evidence that smoking or alcohol independently drive CIN incidence, though both correlate with comorbidity burden.

Protective factors

• G-CSF prophylaxis — the dominant modifiable protective factor (§12)
• Trilaciclib — pharmacologic myeloprotection via transient G1 arrest of HSPCs
• Dose reduction / schedule modification — protective against FN but at the cost of relative dose intensity (§11); a genuine two-edged trade
• Genetic: the flip side of the pharmacogene story — normal-function DPYD, UGT1A1*1/*1, wild-type TPMT/NUDT15 are effectively protective alleles for the corresponding drugs
• Higher baseline ANC — and this is where the Duffy story sits (§4/§9), because a lower baseline ANC in Duffy-null individuals is NOT a risk factor for infection, but is systematically misread as one

Gene–environment interaction

This condition is almost the platonic case of GxE: the "environment" is a precisely dosed, precisely timed chemical exposure, and the genotype sets the exposure–response curve. A DPYD poor metabolizer given standard 5-FU experiences a pharmacokinetic exposure several-fold higher than intended; the same genotype with no fluoropyrimidine exposure is clinically silent. CPIC and DPWG both publish genotype-directed dose reductions, and DPYD/UGT1A1 pre-emptive genotyping is now implemented in multiple health systems ⚠️ (PMC8827955, PMC9262778).

3. Phenotypes

The defining feature of CIN's phenotype list is what's missing: neutropenia is largely asymptomatic until it isn't. Without neutrophils you cannot mount the local inflammatory response that produces the usual signs of infection, so a patient can go from "fine" to "septic" with almost no intermediate signal. Fever is often the only manifestation, and that's precisely why it's treated as an emergency.

Severity grading — and a 2025 change you must not miss.

CTCAE v5.0 "Neutrophil count decreased," ANC thresholds ⚠️: - Grade 1: <LLN – ≥1500/µL - Grade 2: <1500 – ≥1000/µL - Grade 3: <1000 – ≥500/µL - Grade 4: <500/µL

CTCAE v6.0 (released 2025) shifted every band down one notch ✅ PMID:41158990 (Merz LE, HemaSphere 2025;9(10):e70242): - Grade 1: <1500 – 1000/µL - Grade 2: <1000 – 500/µL - Grade 3: <500 – 100/µL - Grade 4: <100/µL

The paper's own framing, quoted from the article: "It is no longer accurate or appropriate to use neutropenia severity grading from the 1980s to assess treatments from the 2020s." Two drivers: modern agents aren't 1980s cytotoxics, and the old thresholds pathologized the normal ANC range of Duffy-null individuals. Any dismech entry citing a "grade 3/4 neutropenia" rate must state which CTCAE version the source used — the same patient reclassifies between v5 and v6.

Febrile neutropenia definition (IDSA/common clinical): single oral temperature ≥38.3 °C (101 °F), or ≥38.0 °C sustained over one hour, plus ANC <500/µL or <1000/µL with predicted decline to <500 ⚠️. Note that some sources cite ANC <1500 for the FN definition; the <500 threshold is the operational one.

Quality of life. Trilaciclib trials measured patient-reported outcomes alongside myeloprotection endpoints and reported improvements in fatigue and physical wellbeing domains ✅ PMID:34408488. Beyond the direct symptom burden, CIN drives hospitalization (average LOS 6–10 days ✅ PMID:35552754), social isolation during nadir periods, and the anxiety of the "call us if you have a fever" instruction that reorganizes a patient's whole month around a thermometer.

4. Genetic / Molecular Information

Causal genes: none. There is no germline or somatic mutation that causes CIN. What genetics contributes is susceptibility magnitude — how much drug exposure a given host generates from a given dose, and how readily their progenitors die from it.

Pharmacogenes with actionable, guideline-backed evidence

GWAS findings

• Biobank Japan, 13,122 cancer patients across regimens (antimicrotubule, paclitaxel-based, docetaxel-based) — GWAS of severe neutropenia/leucopenia ⚠️ PMID:23648065 (and a follow-up, PMC7657179)
• Platinum-based chemotherapy in NSCLC: rs13014982 (2q24.3) and rs9909179 (17p12) associated with myelosuppression risk ⚠️ PMID:25823687
• Docetaxel-induced myelosuppression (Chinese Han cohort): SLC15A1 rs2297322, PXR/NR1I2 rs3732359, FMO3 rs2266782 ⚠️
• CALGB 90401/60404 (Alliance) GWAS of docetaxel-induced neutropenia ⚠️

The Duffy / ACKR1 story — and why it belongs in this entry

ACKR1 (atypical chemokine receptor 1, the Duffy blood group gene) 🔍. Homozygosity for the -67T>C promoter variant (rs2814778) abolishes erythroid ACKR1 expression and produces the Duffy-null associated neutrophil count (DANC) phenotype: a genuinely lower circulating ANC with no increased infection risk, because the neutrophils are redistributed to tissue rather than absent. This is not a disease; it was historically mislabeled "benign ethnic neutropenia."

Numbers ✅ (from PMID:41158990 and Blood Advances 2023;7(3):317 ⚠️): - Normal ANC reference range, Duffy-null adults: 1200–1540/µL; Duffy non-null adults: 2000–7500/µL - ~66–67% of Black/African American individuals in the US carry the Duffy-null phenotype - Median ANC 2820/µL (Duffy-null) vs 5005/µL (non-null) among healthy Black adults ⚠️ - "Duffy null patients have ~10% lower eligibility for clinical trials due to ANC criteria alone." ✅

A 2024 Dana-Farber analysis found that more than half of studied cancer treatments would require dose reduction, delay, or discontinuation if a Duffy-null participant's ANC fell below a threshold that is normal for their phenotype ⚠️. This is a mechanism by which an ontology-adjacent measurement error becomes an undertreatment disparity, and it is the reason CTCAE v6 moved the grade boundaries.

Curation suggestion: model this as a genetic: entry with relationship_type: MODIFIER (or SUSCEPTIBILITY for the misclassification pathway), and consider a discussions: entry with kind: KNOWLEDGE_GAP on whether Duffy-null status modifies true CIN risk versus only its measurement.

Epigenetics, chromosomal abnormalities, somatic variants: not applicable to CIN itself. Relevant adjacent finding: clonal hematopoiesis (CH) — pre-existing PPM1D, TP53, DNMT3A, TET2 clones are selected for by cytotoxic therapy, and CH-derived therapy-related myeloid neoplasms occur in up to 2.3% of autologous-HSCT recipients at a median 2.6 years post-transplant, with worse outcomes than non-CH-derived tMN ⚠️ (Leukemia 2024, PMC11147764). This belongs in §11 as a long-term complication of the same cytotoxic exposure, not as a cause of CIN.

5. Environmental Information

For CIN, "environmental factor" is the drug exposure, which makes this section unusually well-defined.

Suggested environmental: entries with influences_mechanisms links (remember: target here is a bare node name, not an entity reference — the pathograph does not use <kind>#<name>):

A note on predicate choice: TRIGGERS and EXACERBATES are in qc_plugins.CAUSAL_PREDICATES and count toward compliance scoring. Chemotherapy genuinely triggers this, so TRIGGERS is honest here — but antibiotic prophylaxis's net effect on outcomes is contested, so MODULATES is the right non-committal predicate for that one.

Lifestyle factors: limited independent evidence. Nutritional status and oral/dental hygiene are the two with plausible mechanistic links (barrier integrity, portal of entry).

Infectious agents — as consequence, not cause. The organisms that matter in FN, worth curating with NCBITaxon bindings 🔍: - Gram-negative: Escherichia coli, Klebsiella pneumoniae, Pseudomonas aeruginosa (the one that kills fast) - Gram-positive: coagulase-negative staphylococci (most common isolate, often line-related), Staphylococcus aureus, viridans group streptococci (mucositis-associated, can cause ARDS) - Anaerobes: Clostridioides difficile - Fungi: Candida spp., Aspergillus spp. (risk rises sharply beyond ~7 days of neutropenia) - Viral reactivation: HSV, VZV, CMV

Critically: most FN episodes are culture-negative. Only a minority yield a microbiologic diagnosis, which is the empirical anchor for the "febrile mucositis" reframing in §6.

6. Mechanism / Pathophysiology

The causal chain

1. Cytotoxic drug is administered and distributes systemically, reaching bone marrow — leads to —
2. DNA damage / mitotic arrest / nucleotide starvation in the marrow mitotic pool. The mitotic pool is the committed granulocytic progenitor compartment (CFU-GM, myeloblasts, promyelocytes, myelocytes) that is "sensitive to myelosuppressive chemotherapeutic drugs" ⚠️ (PMC6472781) — leads to —
3. p53 stabilization → transcriptional induction of the BH3-only protein PUMA (BBC3) → BAX/BAK-dependent mitochondrial outer membrane permeabilization → intrinsic apoptosis of progenitors. Evidence: "PUMA is largely responsible for the apoptotic effect downstream of p53 in hematopoietic stem cells after irradiation" and "Many chemotherapeutic agents function, in part, by inducing apoptosis through p53-dependent up-regulation of proapoptotic BH3-only proteins such as PUMA" ⚠️. Puma-null mice are protected from HSPC depletion ⚠️ (Blood 2010;115:3472 and 115:4707). This step is inferred for most cytotoxics from irradiation and dyskeratosis-congenita models rather than demonstrated directly for each agent — mark directness: INDIRECT and evidence_source: MODEL_ORGANISM accordingly.
4. Branch A (acute, reversible): progenitor apoptosis halts new granulocyte production — results in —
5. Depletion of the marrow post-mitotic storage pool over the following days as it continues releasing without resupply — results in —
6. Circulating neutropenia. The kinetics fall straight out of neutrophil biology: neutrophils have a circulating half-life of only 6–8 hours and are produced at 5×10¹⁰–10×10¹¹ cells/day ⚠️ (PMC2930213). A cell type with that turnover has essentially no buffer; the count follows production with only the storage-pool transit time as delay. Hence the nadir at days 10–14 for most regimens, with docetaxel notably earlier (median day 7) ⚠️ (PMC6472781) — leads to —
7. Loss of innate first-line defense, plus, in parallel:
8. Branch B (the mucosal arm, running concurrently): the same cytotoxic hits the gut and oral epithelium — results in — mucosal barrier injury (mucositis) — results in — loss of epithelial barrier integrity plus antibiotic- and chemotherapy-driven dysbiosis of commensal microbiota, with commensals converting to pathobionts — results in — bacterial translocation into the bloodstream and, independently, cytokine release / inflammatory response even in the absence of proven infection ⚠️ PMID:25196917 ✅. The authors of that paper propose "febrile mucositis" as a more complete frame than "febrile neutropenia," on the grounds that MBI "creates a port-de-entrée for resident micro-organisms to cause blood stream infections and contributes directly to the occurrence of fever by disrupting highly regulated host-microbe interactions." This is the best available explanation for why most FN is culture-negative.
9. Branches A and B converge — result in — febrile neutropenia, and when uncontrolled, sepsis, organ failure, death.
10. Branch C (long-term): cytotoxic exposure also drives HSC stress-induced premature senescence (not just apoptosis) and bone marrow niche perturbation — mesenchymal stromal injury, elevated intracellular ROS ⚠️ (PMC6114852) — leads to — impaired self-renewal and long-term hematopoietic damage, distinct from the acute nadir. The literature draws the distinction explicitly: "If hematopoietic progenitor cells are induced apoptosis and depleted by chemotherapy, acute myelosuppression occurs, but if hematopoietic stem cells undergo senescence with impaired self-renewal ability, long-term damage to the hematopoietic system occurs" ⚠️.
11. Branch D (clinical-management feedback loop): neutropenia → dose delay/reduction → reduced relative dose intensity → potentially worse cancer outcomes (§11). The toxicity loops back onto the disease being treated.
12. Branch E (emerging, model-organism only): neutropenia may itself be permissive for metastasis. ✅ PMID:37538353 (Russo et al., Oncoimmunology 2023): cyclophosphamide and doxorubicin — but not cisplatin — increased lung metastatic burden in mice by inducing neutropenia; anti-Ly6G neutrophil depletion reproduced the effect, and G-CSF rescue prevented it. "CIN affected the early metastatic colonization of the lung, quite likely promoting the proliferation of tumor cells extravasated into the lung at 24–72 hours." Strictly evidence_source: MODEL_ORGANISM; this is a HUMAN_MODEL_MISMATCH candidate, not a human claim.

The counter-regulatory arm (and how the drugs exploit it)

G-CSF / CSF3R signaling. G-CSF (CSF3) is "the major regulator of neutrophil production under basal conditions of hematopoiesis" and is essential for "emergency" granulopoiesis in response to bacterial infection ⚠️. Mechanism: G-CSF binds CSF3R → conformational dimerization aligns Box1/Box2 motifs → recruitment and activation of JAK1/JAK2/TYK2 → tyrosine phosphorylation → downstream STAT3 (principal axis in granulocyte-monocyte progenitors), plus MAPK/ERK, PI3K-AKT-mTOR, SRC-family kinases, NF-κB ⚠️ PMID:41950547. STAT3 both drives myeloid differentiation and provides negative feedback via SOCS3. Therapeutic G-CSF shortcuts the depleted-storage-pool problem by expanding and accelerating the surviving progenitor compartment.

CDK4/6 dependence — the trilaciclib insight. HSPCs proliferate in a CDK4/6-dependent manner; many tumors (notably SCLC, which is almost universally RB1-deficient) do not. Transient pharmacologic CDK4/6 inhibition therefore "induces reversible G1-arrest in CDK4/6-dependent cells (such as HSPCs and lymphocytes)" ⚠️, parking the marrow safely outside S-phase for the duration of chemotherapy exposure while the RB-null tumor keeps cycling into the cytotoxic. It's a beautifully asymmetric trick — like closing the storm shutters on one house while the neighbors' windows stay open, on purpose.

Ontology suggestions for the pathograph

Cell types (CL) 🔍 — verify all before binding: - neutrophil; granulocyte monocyte progenitor cell; common myeloid progenitor; hematopoietic stem cell; promyelocyte; myeloblast; mesenchymal stem cell of bone marrow (niche); intestinal epithelial cell; oral mucosa epithelial cell

Biological processes (GO) 🔍: - granulocyte differentiation; myeloid cell differentiation; neutrophil chemotaxis; intrinsic apoptotic signaling pathway in response to DNA damage by p53 class mediator; cell cycle arrest / G1/S transition of mitotic cell cycle; cellular senescence; hematopoietic stem cell proliferation; response to oxidative stress; cytokine-mediated signaling pathway; JAK-STAT cascade; innate immune response; defense response to bacterium

Anatomy (UBERON) 🔍: bone marrow; blood; oral mucosa; intestinal mucosa; spleen

Chemicals (CHEBI) 🔍: docetaxel, doxorubicin, cyclophosphamide, carboplatin, 5-fluorouracil, irinotecan, gemcitabine, topotecan

biological_scale tagging suggestion: step 2–3 nodes → MOLECULAR; steps 4–5 → CELLULAR; step 8 mucosal barrier → TISSUE; step 9 febrile neutropenia / sepsis → ORGANISM. Keep one value per node; if a node wants two, split it.

Molecular profiling

• Transcriptomics: GEO holds numerous marrow and PBMC datasets under chemotherapy; also mouse 5-FU/cyclophosphamide myelosuppression series. Run just discover-datasets and just verify-datasets before committing any accession — and remember that verification proves existence, never relevance.
• Microbiome: several strong recent datasets. "Specific gut microbiota changes heralding bloodstream infection and neutropenic fever during intensive chemotherapy" (Leukemia 2019) ⚠️; gut diversity and resistome as FN outcome biomarkers in pediatric HSCT (Sci Rep 2024) ⚠️; AML neutropenic-fever microbiome profiles (PLOS ONE 2020) ⚠️.
• Proteomics / metabolomics / lipidomics: no established CIN-specific signature located. Record the absence rather than leaving it blank.
• Functional genomics: DepMap and CRISPR screens for chemosensitivity determinants exist but are tumor-focused, not marrow-focused.

7. Anatomical Structures Affected

Primary organ: bone marrow (UBERON 🔍) — specifically the hematopoietically active red marrow of the pelvis, sternum, vertebrae, ribs, and proximal long bones. Involvement is bilateral, diffuse, and systemic, not focal.

Secondary organ involvement (all consequences of failed defense or shared cytotoxic injury): - Blood — the compartment where the deficit is measured - Oral cavity and oropharynx — mucositis - Gastrointestinal tract — mucositis, typhlitis (classically cecum/ascending colon) - Perianal region — cellulitis - Lung — pneumonia, invasive aspergillosis - Skin and vascular access sites — cellulitis, line infection - Any organ, via hematogenous seeding once bacteremia is established

Body systems: hematopoietic and immune (primary); gastrointestinal, integumentary, respiratory (secondary).

Tissue level: hematopoietic tissue; stratified squamous and columnar mucosal epithelium; marrow stroma/vasculature (the niche).

Cell populations targeted: the mitotic granulocytic progenitor pool preferentially; HSCs at higher/repeated doses; lymphocytes; erythroid and megakaryocytic progenitors (multilineage CIM); intestinal crypt stem cells; marrow mesenchymal stromal cells.

Subcellular (GO Cellular Component) 🔍: nucleus (DNA damage); mitochondrion (MOMP, intrinsic apoptosis); spindle/microtubule cytoskeleton (taxane target); specific and azurophil granules (neutrophil function).

Lateralization: not applicable — systemic and symmetric.

8. Temporal Development

Onset: acute and predictable, keyed to the infusion. No congenital or age-of-onset concept in the usual sense; the "onset" is pharmacologic. Age distribution mirrors who gets chemotherapy — bimodal-ish, with pediatric oncology and a large adult/geriatric mass.

Within-cycle course (the canonical arc): 1. Day 0: chemotherapy administered 2. Days 1–4: ANC often transiently normal or elevated (storage pool still shipping; steroids and stress demargination can raise it) 3. Days 5–7: descent begins 4. Days 7–14: nadir — most regimens days 10–14; docetaxel monotherapy median day 7 ⚠️; some regimens (gemcitabine, carboplatin) run later or biphasic 5. Days 14–21: recovery as surviving progenitors repopulate 6. Day 21: next cycle — assuming recovery is complete, which is the crux of every dose-delay decision

Duration: self-limited per cycle in solid-tumor chemotherapy (typically ~5–10 days of grade 3–4). In AML induction and HSCT conditioning, profound neutropenia lasts weeks, and duration is the dominant risk multiplier for invasive fungal disease.

Course pattern: episodic and cycle-locked — the clearest example of a recurrent, treatment-entrained disease course in oncology. Model as temporality: RECURRENT (or TRANSIENT per episode) rather than CHRONIC.

Progression across cycles: cumulative marrow injury means nadirs often deepen and recovery slows over successive cycles, especially with alkylators and carboplatin.

Remission: treatment-induced (G-CSF) or spontaneous with time. Complete hematologic recovery is the norm; incomplete recovery signals marrow reserve exhaustion, marrow involvement by tumor, or evolving myelodysplasia.

Critical intervention windows: - Before cycle 1 — because 59% of FN happens there ⚠️. Primary prophylaxis decided before the first dose is the single highest-yield intervention in this whole entry. - 24–72 h after chemotherapy — the pegfilgrastim/filgrastim administration window (G-CSF given too close to the cytotoxic can theoretically recruit progenitors into cycle at the worst moment). - First hour of fever onset — "door-to-antibiotic time"; empiric broad-spectrum antibiotics within 60 minutes is the standard-of-care target.

9. Inheritance and Population

Inheritance: not applicable. CIN is acquired and iatrogenic. Do not populate an inheritance: block with a Mendelian mode. The heritable component is pharmacogenomic susceptibility — model with genetic: entries typed SUSCEPTIBILITY or MODIFIER, and if you want an inheritance concept at all, the honest one is polygenic/multifactorial (HP:0010982 🔍 — verify) with a description naming the pharmacogenes. Penetrance, anticipation, germline mosaicism, consanguinity: all not applicable.

Epidemiology

For dismech prevalence: records, use the structured slots — measure_type: ANNUAL_INCIDENCE for the per-million and per-1,000 figures, rate_per_100000 normalized (2.4–15.4/million → 0.24–1.54 per 100,000; 7.8/1,000 cancer patients → 780 per 100,000 cancer patients, and note the denominator is cancer patients, not general population, in population:). Never let a per-cycle FN rate and a per-patient FN rate sit in the same comparison.

Population demographics

• Sex ratio: no strong intrinsic sex effect on CIN; observed distributions track underlying cancer epidemiology and body-composition-based dosing.
• Age: ≥65 is the consistently replicated high-risk group; pediatric patients have the longest hospitalizations (~8 days, up to $65,000 per admission ✅).
• Ancestry: the Duffy-null issue is the dominant population-genetics finding — ~66–67% of Black/African American individuals in the US, and high frequency in Middle Eastern populations. It affects measured ANC, trial eligibility (~10% lower ✅), and dose-modification decisions, without affecting infection risk. NUDT15 variant frequencies are highest in East Asian populations; DPYD variant spectra differ substantially by ancestry (HapB3 is European-enriched; several African-ancestry-relevant variants are under-tested by standard panels).
• Geography: the burden distribution is a healthcare-access story, not a biology story — G-CSF availability and cost determine outcomes far more than any regional biological variation.

10. Diagnostics

Diagnosis is a lab value plus a thermometer. There's no imaging finding, no biopsy, no genetic test that makes the diagnosis.

Core tests

• Complete blood count with differential → absolute neutrophil count. ANC = WBC × (% segmented neutrophils + % bands). LOINC 🔍: search "Neutrophils [#/volume] in Blood by Automated count" — resolve before binding.
• Temperature measurement — oral, per the FN definition.
• Blood cultures — at least two sets, from peripheral vein and each lumen of any central line, before antibiotics.
• Site-directed cultures: urine, stool (if diarrhea), sputum, wound, CSF if indicated.
• CMP, LFTs, lactate, CRP/procalcitonin (adjunctive, not diagnostic).
• Chest imaging — CXR; high-resolution chest CT if prolonged neutropenia or fungal concern (the halo sign in invasive aspergillosis).
• Serum galactomannan / β-D-glucan for invasive fungal surveillance in high-risk hematologic patients.
• Bone marrow biopsy — not routine for CIN. Reserved for unexplained, prolonged, or non-recovering cytopenias where marrow involvement or a therapy-related myeloid neoplasm is the question.

Reference ranges (for a Biochemical entry with interpretation_bands)

Adult ANC, conventional: 1500–8000/µL (LLN commonly 1500). But the Duffy-null finding means a single population-wide interval is wrong for a third of some populations. If you curate reference_ranges, add a population:-stratified pair:

• Duffy non-null adults: 2000–7500/µL ✅ PMID:41158990
• Duffy-null adults: 1200–1540/µL ✅ PMID:41158990

and then interpretation_bands mapping ANC <1500 / <1000 / <500 / <100 to the CTCAE tiers — stating the version, since v5 and v6 disagree (§3). This is exactly the graded-category case the interpretation_bands slot exists for, and the CKD-MBD entry is the pattern to copy.

Genetic testing

Not diagnostic for CIN, but predictive and increasingly standard: - DPYD genotyping before fluoropyrimidines — mandated pre-treatment in the EU; CPIC guideline-backed - UGT1A1 before irinotecan (especially at high dose or in FOLFIRINOX) - TPMT and NUDT15 before thiopurines — CPIC guideline-backed - Pre-emptive multi-gene PGx panels ⚠️ (PMC8827955 describes implementation) - WGS/WES/CMA/karyotype/FISH/mtDNA/repeat-expansion testing: not applicable - ACKR1 (rs2814778) genotyping or Duffy phenotyping: increasingly proposed to avoid misclassifying DANC as pathologic neutropenia ⚠️ (Am J Hematol 2025, "Diagnosis of Duffy Null-Associated Neutropenia During Chemotherapy")

Clinical criteria and risk stratification

• FN definition (IDSA): as in §3
• MASCC Risk Index — 7 clinical factors at fever onset; score <21 = high risk. Internationally validated ⚠️
• CISNE (Clinical Index of Stable Febrile Neutropenia) — designed for apparently stable patients; in head-to-head work CISNE-I had sensitivity 0.22 / specificity 0.91 vs MASCC low-risk sensitivity 0.95 / specificity 0.17 ⚠️ PMID:31864874. The two are answering different questions: MASCC casts a wide net, CISNE is a stricter filter for who can genuinely go home. Validated in gynecologic oncology ⚠️ PMID:34504931. A sober JCO OP commentary is titled "Risk-Stratifying Treatment Strategies for Febrile Neutropenia—Tools, Tools Everywhere, and Not a Single One That Works?" ⚠️ — worth citing for honest uncertainty.
• Lyman model and similar multivariable models predict FN risk before cycle 1 to guide primary prophylaxis.

Differential diagnosis

Screening

• Pre-treatment CBC and risk assessment before every cycle — the operational screen
• Pre-treatment pharmacogenomic screening for DPYD/UGT1A1/TPMT/NUDT15
• No population-level screening program; not a newborn-screening or carrier-screening concept

11. Outcome / Prognosis

Mortality ✅ (all from PMID:35552754)

Morbidity and burden ✅

• Average hospital LOS: ~6 days (elderly), ~8 days (children), up to 10 days (adults)
• Per-hospitalization cost: up to $15,000 (elderly), $40,000 (adults), $65,000 (children)
• Total annual US cost (2012): >$2 billion for adults, up to $880 million for children
• 94% of FN emergency-department visits end in admission

Complications

Bacteremia and sepsis; septic shock; invasive fungal disease; typhlitis; ARDS (notably viridans streptococcal); prolonged hospitalization; central line loss; C. difficile colitis; and the treatment-level complication — dose delay and dose reduction.

Recovery

Complete hematologic recovery is the rule for a single cycle. The prognostically important question is whether it recovers in time for the next cycle.

The dose-intensity paradox — the most important prognostic content in this entry

Two findings sit in tension and both are real:

(a) Reduced relative dose intensity (RDI) is associated with worse cancer survival. RDI ≥85% is the conventional threshold for early breast cancer ⚠️; a systematic review and meta-analysis found RDI associated with survival in advanced solid tumors ⚠️ PMID:33973301; the early-breast-cancer/aggressive-lymphoma analysis is ⚠️ PMID:20227889. Older age, obesity, and FN itself are associated with reduced RDI ⚠️.

(b) Experiencing neutropenia is associated with better survival. A meta-analysis of 13 trials, 9,528 patients found a hazard ratio of death of 0.69 (95% CI 0.64–0.75) for patients with higher-grade neutropenia/leukopenia versus lower-grade or none ⚠️ PMID:20960191. Grade ≥3 neutropenia on day 8 of cycle 1 in mCRPC docetaxel: 1-year OS 81% vs 68% ⚠️.

These reconcile if you read neutropenia as a pharmacodynamic dosimeter — a readout that the patient actually received a biologically active exposure. As one gastric-cancer analysis puts it, "the absence of such toxicity indicates that the dosages of drugs are not pharmacologically adequate" ⚠️. But the inference has an obvious confounding hazard (patients who survive long enough to have cycle-1 toxicity, patients with better organ function, immortal-time bias in some designs). Curate this as SUPPORT + directness: INDIRECT, and put a discussions: entry with kind: KNOWLEDGE_GAP on whether the association is causal or a marker of exposure. Do not let the entry imply that inducing neutropenia is therapeutic.

Long-term

Therapy-related myeloid neoplasms: up to 2.3% of autologous-HSCT recipients at median 2.6 years; risk factors age ≥60, male sex, radiotherapy, ≥3 prior lines, graft cellularity; post-aHSCT tMN enriched for PPM1D and TP53 lesions with shorter latency and worse OS; CH-derived tMN does worse than non-CH-derived ⚠️ (Leukemia 2024, PMC11147764).

Prognostic factors

MASCC and CISNE scores; ANC nadir depth; duration of neutropenia (the strongest predictor of invasive fungal disease); documented bacteremia; hemodynamic instability; comorbidity burden; age; hematologic vs solid malignancy; time-to-antibiotic.

12. Treatment

Prophylaxis — the main event

Colony-stimulating factors (G-CSF). The current authority is ✅ PMID:41740078 — "WBC Growth Factors: ASCO Guideline Update," J Clin Oncol 2026, which replaces the 2015 version following a systematic review of RCTs and meta-analyses published September 2014 – August 2025. Core recommendation, quoted: "Prophylactic use of CSFs to reduce the risk of febrile neutropenia is warranted when the risk...is approximately 20% or higher." For regimens with <20% risk, primary prophylaxis may still be warranted based on age, medical history, or disease characteristics ⚠️. Named acceptable agents: filgrastim, pegfilgrastim, eflapegrastim, and biosimilars ✅.

Agents and suggested NCIT bindings:

Eflapegrastim efficacy signal: severe neutropenia in cycle 1 15.8% vs 24.3% for pegfilgrastim ⚠️.

Trilaciclib. First-in-class CDK4/6 inhibitor given before chemotherapy for myeloprotection in ES-SCLC. Pivotal evidence: ✅ PMID:31504118 (Ann Oncol 2019, phase Ib/randomized II, 94 patients, trilaciclib + etoposide/carboplatin) — grade ≥3 AEs 50% vs 83.8%, improvements across neutrophil, RBC, and lymphocyte measures, comparable antitumor efficacy, no trilaciclib-related grade ≥3 AEs. Pooled phase 2 analysis ✅ PMID:34408488 (123 trilaciclib vs 119 placebo; NCT02499770, NCT03041311, NCT02514447) — myeloprotection across age groups, greater in patients ≥65, with improved QoL and on-schedule chemotherapy delivery. target_mechanisms should point at the progenitor-apoptosis node with a MODULATES/preventive framing.

Plinabulin (investigational). PROTECTIVE-2 phase 3 (n=221, TAC q21d, plinabulin 40 mg day 1 + pegfilgrastim 6 mg day 2 vs pegfilgrastim alone): grade 4 CIN prevention in cycle 1 31.5% vs 13.6%, P = .0015 ⚠️. FDA breakthrough designation Sept 2020 ⚠️. Not FDA-approved — curate as investigational and do not imply availability.

Antimicrobial prophylaxis. Fluoroquinolone prophylaxis is considered for high-risk patients with expected prolonged profound neutropenia (classically ANC <100/µL for >7 days), balanced against resistance selection and C. difficile risk. Antifungal (posaconazole) and anti-PJP (TMP-SMX) prophylaxis per risk group. This area is genuinely contested — use MODULATES-style hedging, not causal predicates.

Treatment of established febrile neutropenia

• Empiric broad-spectrum IV antibiotics within 60 minutes, after cultures. IDSA monotherapy options: an antipseudomonal β-lactam (cefepime), a carbapenem (meropenem, imipenem-cilastatin), or piperacillin-tazobactam ⚠️ (IDSA 2010 update, Clin Infect Dis 2011;52:e56).
• Add vancomycin, aminoglycoside, and/or fluoroquinolone for hypotension, pneumonia, suspected line infection, or known resistance ⚠️.
• Continue "until ANC recovery is imminent" or until a specific infection dictates otherwise ⚠️.
• Empiric antifungal therapy for persistent fever beyond 4–7 days in high-risk patients.
• Outpatient oral management (ciprofloxacin + amoxicillin-clavulanate) for carefully selected low-risk patients — ASCO/IDSA outpatient guideline ⚠️ (JCO OP 2018).
• Therapeutic G-CSF in established FN: guidelines generally do not recommend routine addition, reserving it for high-risk features. In AML induction specifically, a JSCO systematic review found G-CSF "significantly shortened the duration of neutropenia" but "did not correlate with infection-related mortality and did not affect disease progression/recurrence or overall survival" ⚠️ (PMC11043120). That's an important negative and belongs in the entry.
• Granulocyte transfusion — salvage only; evidence weak.
• Supportive care: NCIT:C15747 Supportive Care; hydration, antipyretics, source control, line removal.

Treatment-strategy content

• Dose reduction / cycle delay — effective at reducing FN but see the RDI paradox (§11)
• Regimen substitution to a lower-myelotoxicity option
• Genotype-guided dosing: CPIC DPYD, UGT1A1, TPMT/NUDT15 — the personalized-medicine arm of this entry
• Duffy-informed ANC thresholds so that DANC patients are not needlessly dose-reduced ⚠️

Adverse effects of the treatments

G-CSF: bone pain (very common), splenic rupture (rare), ARDS (rare), capillary leak; theoretical concern about stimulating CSF3R-expressing myeloid malignancies. Trilaciclib: fatigue, hypocalcemia, hypokalemia, headache, injection-site reactions. Broad-spectrum antibiotics: C. difficile, resistance selection, microbiome collapse (which loops back to §6, branch B — the treatment feeds the mechanism).

13. Prevention

Primary prevention (stop the neutropenia happening): - FN risk assessment before cycle 1, using regimen risk + patient factors (Lyman-type model) - G-CSF primary prophylaxis at the ≥20% threshold ✅ PMID:41740078 - Trilaciclib in ES-SCLC receiving platinum/etoposide or topotecan - Regimen selection favoring lower myelotoxicity where efficacy-equivalent - Pre-emptive pharmacogenomic dosing (DPYD, UGT1A1, TPMT, NUDT15) - Dental evaluation and source control before starting therapy

Secondary prevention (stop the neutropenia becoming a catastrophe): - CBC monitoring around the expected nadir - Secondary G-CSF prophylaxis after a prior FN episode or dose-limiting neutropenia - Patient education: the thermometer protocol — take a temperature for any chills or malaise, call at ≥38 °C, do not take antipyretics that mask it - Antimicrobial prophylaxis in the high-risk subset

Tertiary prevention: - Door-to-antibiotic ≤60 minutes - Risk-stratified admission via MASCC/CISNE - Central line care bundles - Antifungal surveillance in prolonged neutropenia

Immunization: inactivated influenza and COVID-19 vaccines are recommended in chemotherapy patients (ideally timed away from nadir); pneumococcal per risk. Live vaccines contraindicated. These prevent infection, not neutropenia — keep the distinction in the entry.

Behavioral / public health: hand hygiene; food-safety precautions; avoiding sick contacts and crowds during nadir; the "neutropenic diet" is not supported by evidence and has been de-implemented in most centers — worth curating explicitly as a refuted intervention with supports: REFUTE if a suitable citation is obtained.

Genetic counseling: not applicable in the reproductive sense; pharmacogenomic counseling is the analogue.

14. Other Species / Natural Disease

There is no natural disease here — CIN cannot occur without an anthropogenic exposure. But there are two genuine cross-species angles:

Veterinary iatrogenic CIN. Dogs (NCBITaxon:9615 🔍) and cats (NCBITaxon:9685 🔍) receiving veterinary chemotherapy (doxorubicin, vincristine, cyclophosphamide, lomustine, carboplatin) develop dose-limiting neutropenia with nadir timing that differs by agent — lomustine and carboplatin notably later. Canine lymphoma CHOP protocols are the highest-volume setting. This is a real clinical entity in veterinary oncology and evidence from it grades as evidence_source: MODEL_ORGANISM per dismech rules (veterinary observations in non-human mammals). Breed effects are real and mechanistically interesting: ABCB1 (MDR1) nt230(del4) in Collies and related herding breeds (VBO 🔍) causes profound sensitivity to P-glycoprotein substrates including vincristine and doxorubicin — a genuine veterinary pharmacogenomic parallel to the human DPYD story, and worth a genetic: or comparative note. Search OMIA for the ABCB1 entry to get a citable identifier.

Evolutionary conservation. Granulopoiesis, CSF3/CSF3R signaling, and p53/PUMA-mediated progenitor apoptosis are conserved across vertebrates, which is why the mouse models in §15 work at all. Orthologs: Csf3r, Trp53, Bbc3 (mouse); csf3r, tp53 (zebrafish) — NCBI Gene IDs 🔍.

Zoonotic potential / cross-species transmission: not applicable.

15. Model Organisms

Because the initiating lesion is a drug, CIN is unusually easy to model — you give the animal the drug. That makes it one of the better-supported iatrogenic conditions for animal_models: curation.

Mammalian induced models

The exemplar link to curate — Russo et al. ✅ PMID:37538353 is the strongest single animal_models entry available here, because it has all four ModelMechanismLink components in one paper:

animal_models:
- name: Cyclophosphamide-induced neutropenia in Renca/LLC-bearing mice
  species: Mouse
  publication: PMID:37538353
  modeled_mechanisms:
  - target: <your neutropenia node's bare name>
    relationship: RECAPITULATES
    fidelity: MODERATE
    limitations: >-
      Single high-dose administration rather than cyclic human dosing; murine
      neutrophil biology and marrow reserve differ from human; the metastasis
      readout has no confirmed human counterpart.
    readouts:
    - name: Circulating neutrophil count
      target: <same bare name>
      direction: DECREASED
      ...

Note the anti-Ly6G arm and the G-CSF rescue arm (which prevented the metastasis phenotype) are natural PERTURBS and RESCUES links respectively — a rare chance to use three relationship values from one paper.

Other systems

• Zebrafish — transparent larvae with fluorescent neutrophil reporters (mpx:GFP, lyz:DsRed) allow live imaging of granulopoiesis and chemotoxic depletion. Useful for screening; poor for marrow-niche biology (no long-bone marrow).
• In vitro: human CD34+ HSPC colony-forming assays (CFU-GM) are the classic myelotoxicity readout; ex vivo bone marrow cultures; iPSC-derived hematopoietic differentiation. These are experimental_models: (non-animal), not animal_models:.
• Bone-marrow-on-a-chip — emerging NAM (new approach methodology) for chemotherapy myelotoxicity prediction; a good experimental_models candidate with experimental_model_type set appropriately.
• Computational: semi-mechanistic PK-PD models of the neutrophil lifecycle (Friberg-type myelosuppression models, with proliferation, transit compartments, and feedback) are widely used for dose optimization and are strong computational_models: candidates with modeled_mechanisms links to the progenitor and storage-pool nodes.

Model databases: MGI, IMPC, KOMP, IMSR, MMRRC (mouse); ZFIN (zebrafish); Cellosaurus/ATCC (lines); OMIA (veterinary).

Curation notes for the dismech entry

A few things I'd flag before anyone starts writing YAML:

1. The MONDO gap is the first decision. No MONDO term for chemotherapy-induced neutropenia resolved via OLS. Anchor on MONDO:0001475 with skos:narrowMatch, record the gap, and consider whether this is better modeled as a kb/modules/ mechanism module than a disease entry — "cytotoxic myelosuppression" recurs across every myelosuppressive drug toxicity entry the KB will ever have, and the treatment-toxicity module family already exists for exactly this shape of thing. Run just list-modules and grep kb/modules for "myelosuppression" and "neutropenia" before deciding; a module here would let anthracycline, taxane, and platinum toxicity entries all conforms_to one chain instead of re-deriving it.
2. Every "grade 3/4 neutropenia" number needs a CTCAE version stamp. v6 (2025) moved all four boundaries. A rate quoted from a 2019 trial and one from a 2026 trial are not the same measurement.
3. The dose-intensity paradox needs a discussions: entry, not a confident causal claim in either direction.
4. The Russo metastasis finding is MODEL_ORGANISM only and is a textbook HUMAN_MODEL_MISMATCH candidate.
5. Verify every ⚠️ PMID with just fetch-reference before quoting. The exact-substring rule is unforgiving, and several of the quotes above came from search-result summaries that may have paraphrased. My own memory file on this (preverify-snippets-before-writing-yaml) exists because a blocked Write is a silent no-op.
6. Duffy/ACKR1 deserves its own curated thread through genetic:, reference_ranges (population-stratified), and probably a notes: line in diagnostics. It's the most consequential recent development in this space and it's an equity finding as much as a biology one.

Sources

• Chemotherapy-Induced Neutropenia and Febrile Neutropenia in the US: A Beast of Burden That Needs to Be Tamed? — The Oncologist 2022 (PMID:35552754)
• WBC Growth Factors: ASCO Guideline Update — JCO 2026 (PMID:41740078)
• A paradigm shift in neutrophil adverse event grading: Why now? — HemaSphere 2025 (PMID:41158990)
• Chemotherapy-induced neutropenia elicits metastasis formation in mice — Oncoimmunology 2023 (PMID:37538353)
• Mucosal barrier injury, fever and infection in neutropenic patients with cancer: introducing the paradigm febrile mucositis — Br J Haematol 2014 (PMID:25196917)
• Myelopreservation with the CDK4/6 inhibitor trilaciclib — Ann Oncol 2019 (PMID:31504118)
• Myeloprotective Effects of Trilaciclib, pooled phase 2 — Cancer Manag Res 2021 (PMID:34408488)
• Chemotherapy-induced neutropenia and emerging agents for prevention and treatment — Cancer Treat Rev 2022
• Early neutropenia on day 8 with docetaxel: mechanisms within the neutrophil pool system
• Neutrophil kinetics in health and disease
• A prospective, real-world, multinational MASCC study of febrile neutropenia occurrence
• Meta-analysis of neutropenia or leukopenia as a prognostic factor (PMID:20960191)
• Relative Dose Intensity and Survival in Advanced Solid Tumors: Systematic Review and Meta-Analysis (PMID:33973301)
• CISNE versus MASCC: Identifying low risk febrile neutropenic patients (PMID:31864874)
• Risk-Stratifying Treatment Strategies for Febrile Neutropenia — JCO OP
• IDSA Clinical Practice Guideline for Antimicrobial Agents in Neutropenic Patients with Cancer: 2010 Update
• Outpatient Management of Fever and Neutropenia in Adults — ASCO/IDSA Guideline Update
• Efbemalenograstim Alfa: First Approval — Drugs 2023 (PMID:37368138)
• FDA Approves Trilaciclib to Reduce Chemotherapy-Induced Myelosuppression (Feb 2021)
• PROTECTIVE-2 meets primary endpoint of preventing grade 4 neutropenia (plinabulin)
• Absolute neutrophil count by Duffy status among healthy Black and African American adults — Blood Advances 2023
• Diagnosis of Duffy Null-Associated Neutropenia During Chemotherapy — Am J Hematol
• Dana-Farber: cancer trial criteria disadvantage patients of African and Middle Eastern ancestry (2024)
• Duffy-null Associated Neutrophil Count (DANC) — ASH
• GWAS of chemotherapeutic agent-induced severe neutropenia/leucopenia, Biobank Japan (PMID:23648065)
• GWAS of myelosuppression in NSCLC with platinum-based chemotherapy (PMID:25823687)
• Impact of UGT1A1 Polymorphisms on Febrile Neutropenia in FOLFIRINOX
• Modern Developments in Germline Pharmacogenomics for Oncology Prescribing
• Deletion of Puma protects hematopoietic stem cells — Blood 2010
• Chemotherapy-induced niche perturbs hematopoietic reconstitution
• CSF3R signalling beyond granulopoiesis (PMID:41950547)
• G-CSF and its receptor in myeloid malignancy — Blood
• Specific gut microbiota changes heralding bloodstream infection and neutropenic fever — Leukemia 2019
• Gut diversity and resistome as biomarkers of febrile neutropenia outcome — Sci Rep 2024
• Clonal hematopoiesis-derived therapy-related myeloid neoplasms after autologous HSCT — Leukemia 2024
• Primary prophylaxis with G-CSF after AML induction: JSCO systematic review and meta-analysis
• Risk of febrile neutropenia with docetaxel/cyclophosphamide: meta-analysis (PMID:26337685)
• Protective effects of BB-10010 on chemotherapy-induced neutropenia in mice — Exp Hematol
• Febrile Neutropenia — StatPearls
• EBI Ontology Lookup Service (OLS4) — used for all ✅-marked term resolutions

Reference Validation

Checked with linkml-reference-validator 0.2.1.

22 of 37 references resolved; the rest could not be looked up either way.