Table 2 Summary of expert recommendations

I. ICU admission and prognosis

RI-1—Neutropenia should probably not be used as triage criteria in cancer patients considered for ICU admission. Performance status, comorbidities, and potentially life-prolonging treatment available are more relevant in this regard (Grade 2−, strong agreement)

RI-2—Neutropenia should probably not be considered as a prognostic factor in critically ill cancer patients (Grade 2-, weak agreement)

RI-3−Intensive care unit admission should probably not be delayed if ICU admission is deemed necessary in critically ill cancer patients (Grade 2-, strong agreement)

II. Prophylaxis and protective isolation

RII-1—Protective isolation should probably be considered in patients with profound (neutrophil count less than 500/mm³) and prolonged (expected neutropenia duration more than 7 days) neutropenia (Grade 2+, strong agreement)

RII-2—Protective isolation should not be considered as a sterile isolation (Grade 1-, strong agreement)

RII-3—Protective isolation should not delay ICU admission or limit patients’ clinical monitoring or access to patients’ rooms in cases of emergency (Grade 1-, strong agreement).

RII-4—Antibacterial prophylaxis should probably not be performed in critically patients with neutropenia (Grade 2-, strong agreement)

RII-5—Anti-Aspergillus prophylaxis should probably be used in critically ill neutropenic patients with acute myeloid leukemia or myelodysplastic syndrome with both induction and consolidation therapy used when neutropenia is expected to be profound (neutrophil count less than 500/mm³) and with an expected duration of at least 15 days (Grade 2+, weak agreement)

RII-6—Anti-Aspergillus prophylaxis should probably be used in high-risk critically ill neutropenic patients (myeloablative conditioning regimens, older patients, transplant in patients with active disease, umbilical/placental cord blood transplant) (Grade 2+, weak agreement)

RII-7−Anti-Aspergillus prophylaxis should probably be used in critically ill neutropenic patients with severe idiopathic medullary aplasia (neutrophil count less than 500/mm³) (Grade 2+, weak agreement)

III. Acute respiratory failure management

RIII-1—Acute respiratory failure should be considered as a therapeutic emergency in critically ill patients with neutropenia (Grade 1+, strong agreement)

RIII-2—Etiological diagnosis of ARF should be considered as a primary objective in this setting (Grade 1+, strong agreement)

RIII-3—The diagnostic workup should include systematic analysis of the underlying condition, severity and duration of neutropenia, underlying immunosuppression, preexisting treatment and prophylaxis, clinical course of ARF, and clinical and radiological features (Grade 1+, strong agreement)

RIII-4—Invasive and non-invasive diagnostic tests should probably be prescribed according to pretest probability rather than being performed systematically. This should particularly be the case for bronchoscopy with bronchoalveolar lavage (Grade 2+, strong agreement)

RIII-5—Pulmonary biopsies should probably be performed only on a case-by-case basis by a multidisciplinary team after careful assessment of both clinical suspicion and the risk-to-benefit ratio (Grade 2+, strong agreement)

IV. Organ failure and organ support

RIV-1—Neutropenic enterocolitis (typhlitis) should probably be considered in critically ill neutropenic patients with fever and acute abdomen, particularly in cases of recent cancer chemotherapy known to be associated with a high rate of oral or gastrointestinal toxicity (Grade 2+, strong agreement)

RIV-2—In adult patients, a complete diagnostic workup, including an abdominal CT scan with contrast media, should probably be performed (Grade 2+, strong agreement). In the pediatric setting, abdominal ultrasonography should probably be performed as first-line imaging (Grade 2+, strong agreement)

RIV-3—First-line colonoscopy should probably be avoided in patients with high suspicion of typhlitis (expert opinion, strong agreement)

RIV-4—Management of typhlitis should include broad-spectrum antibiotic therapy along with multidisciplinary management, including consultation of a general or abdominal surgeon (Grade 1+, strong agreement)

RIV-5—Neutropenia and thrombocytopenia should not modify the timing of surgery in patients with suspicion of digestive tract perforation (Grade 1+, strong agreement)

RIV-6—Neutropenia in itself should probably not modify ventilatory support in critically ill cancer patients (Grade 2−, strong agreement)

RIV-7—Invasive mechanical ventilation should probably not be delayed only as a consequence of neutropenia, underlying malignancy, or immunocompromised status (Grade 2−, weak agreement)

RIV-8—An indication for renal replacement therapy should probably not be modified by neutropenia in itself (Grade 2−, strong agreement)

V. Antibacterial therapy and source control management

RV-1—Combination therapy with aminoglycoside should probably be used as initial antibiotic therapy in neutropenic patients with severe sepsis or septic shock (expert opinion, weak agreement)

RV-2—Glycopeptide antibiotic adjunctive agents (or other agents active against resistant aerobic gram-positive cocci) should probably be considered for the following specific clinical indications

V-2-a—Suspected catheter-related infection (Grade 2+, strong agreement)

V-2-b—Skin or soft tissue infection (Grade 2+, strong agreement)

V-2-c—Severe sepsis or septic shock (Grade 2+, weak agreement)

V-2-d—Use of antipseudomonal -lactam agent with insufficient anti-gram-positive activity (ceftazidime, for example) (Grade 2+, weak agreement)

V-2-e—Grade III or IV mucositis (Grade 2+, weak agreement)

V-2-f−Known colonization with methicillin-resistant Staphylococcus aureus (Grade 2+, weak agreement)

RV-3—If used empirically, glycopeptide antibiotics should probably be reconsidered and discontinued in the following situations

 After 72 h and if no resistant gram-positive cocci have been identified (expert opinion, weak agreement)

 If infection is related to bacteria susceptible to a -lactam agent (expert opinion, strong agreement)

RV-4—Antibiotic de-escalation should probably be considered in the following situations

 When infection is related to susceptible organism (expert opinion, strong agreement)

 In patients without documented bacterial infection and with stable clinical condition (expert opinion, weak agreement)

RV-5—Indwelling catheters should probably be removed immediately in neutropenic patients with septic shock and no identifiable clinical infection (Grade 2+, strong agreement)

VI. Hematological management

RVI-1—Prophylactic use of G-CSF should probably be initiated or resumed in critically ill patients with neutropenia or requiring cancer chemotherapy with expected medullary toxicity (Grade 2+, weak agreement)

RVI-2—G-CSF should probably be stopped when worsening of respiratory status during neutropenia recovery is suspected or before neutropenia recovery in patients at high risk of worsening of respiratory status during neutropenia recovery (preexisting respiratory failure or pulmonary infection) (Grade 2+, strong agreement)

RVI-3—In euvolemic patients, transfusion of packed red blood cells should probably be performed according to a restrictive strategy in such a way as to maintain hemoglobin above a 7 g/dl threshold (Grade 2+, strong agreement)

RVI-4—Granulocyte transfusion should probably not be performed as adjunct therapy in neutropenic critically ill patients with severe infection (Grade 2−, strong agreement)