Table 1.

Studies included in meta-analysis (by year of publication)

First author, year and reference in brackets	Study design and country	Population (n), description of study sample	Outcome assessed	Significant patient factors, OR and confidence intervals	Factors assessed in multivariable model	Comments on quality
Fujiwara 2017 [28]	Retrospective study Observational Japan	n = 244 Lung	Febrile neutropenia	Male gender OR 4.26 (1.26–20.33) Radiotherapy pre-treatment. OR 6.09 (1.67–23.81)	Age, gender, ECOG PS, cancer type, stage, albumin, AST, total bilirubin, baseline neutrophil smoking, radiotherapy, surgery, chemotherapy treatment	High Authors failed to report missing data handling
Julius 2017 [7]	Retrospective study Observational US	n = 635 Gynaecological cancers	Chemotherapy-induced neutropenia including treatment delays and treatment dose reductions	Metabolic comorbidities** mainly diabetes mellitus Age-negative effect OR 0.865 (0.788–0.951)	Age, BMI, treatment, cancer type, stage, prior treatment (cycles received previously and regimen)	Moderate Authors failed to report missing data handling Population studied may influence generalisability
Naito 2017 [29]	Retrospective study Observational Japan	n = 66 Oesophageal cancer	Grade 3/4 neutropenia	Baseline platelet count OR 0.98 (0.97–0.99) ALT OR 1.15 (1.02–1.35) PPI administration OR 37.95 (3.53–1660.64)	Age, PPI treatment, baseline neutrophils and platelets. Albumin, ALT	Moderate Authors failed to report missing data handling Inadequately powered study.
Agiro 2016 [8]	Retrospective study Observational USA	n = 8745 Breast cancer TC n = 4815 TCH n = 2292 AC n = 1638	Neutropenic episodes Including febrile neutropenia	TC CSF OR 0.29 (0.22–0.39) age effect > 65 years (ref) 18–44 OR 0.52, 0.31 to 0.85 Comorbidity 0.65 0,42 to 1,00 CSF 0,29 0.22 to 0.39 TSH age effect > 65 years (ref) 18–44 OR 0.92 0.46 to 1.83 0.46 0.23 to 0.93 0.19 0.12 to 0.30 AC 1.21 0.75 to 1.93	Age, co-morbidity, stage, CSF use, prophylactic antibiotics	High Authors failed to report missing data handling
Kim 2016 [33]	Retrospective study Observational Korea	n = 610 Early breast cancer	Febrile neutropenia and relative dose intensity	GFR < 60 ml/min 2.806 (1.3–6.1) Age 2.804 (1.16–6.8) Co-morbidity 2.12 (1.078–4.536)	Age, co-morbidity, stage, renal function, WBC count, haemoglobin level, CSF use	High Authors failed to report missing data handling
Altwairgi 2013 [37]	Retrospective Observational Canada	n = 239 Early breast cancer	Febrile neutropenia	No patient-specific factors	Age, treatment, GCSF	High missing data not fully reported
Jiang 2013 [41]	Retrospective study Observational China	n = 141 Non-small cell lung	Grade 3–4 neutropenia	Age OR 3.819 (1.23–11.83) Albumin (g/dl) OR 3.3 (1.13–9.87) BSA > 2 OR 4.040 (1.45–11.22)	Age, weight, gender, PS, renal, diabetes, albumin, BSA	High missing data not fully reported
Lyman 2011 [16]	Prospective study Observational US	n = 3760 Cancers included colorectal (n = 521), small cell lung (n = 210), non-small cell lung (n = 697), ovarian (n = 312), breast (n = 1473), lymphoma (n = 547)	Neutropenic episodes FN or grade 3 or 4	Immunosuppressive medication OR 1.554 (1.105–2.187) AST > 35 u/L OR 1.422 (0.991–2.04) ALP > 120 u/L OR 1.469 (1.058–2.040) Bilirubin >1 mg/dL OR 2.152(1.235–3.747) Low baseline WBC OR 0.930 (0.8920–0.969) GFR OR 0.993 (0.989–0.997)	Age, prior chemotherapy, immunosuppressive medications, high AST, ALT or bilirubin, reduced white blood count or estimated GFR, patients with small-cell lung cancer, with planned RDI 85% CSFs	High All criteria met
Schwenkglenks 2010 [44]	Prospective study Observational Multinational	n = 444 Curative breast	Grade 4 neutropenia	Increasing age OR 1.35 (1.06–1.73) Increasing weight OR 3.85 (1.84–8.07) Vascular comorbidity OR 2.29 (1.25–4.20), baseline WBC OR 0.87 (0.76–0.99), higher baseline bilirubin OR 4.38 (1.25–15.33)	Age, weight, co-morbidity, liver, renal, FBC, RDI, CSF	High All criteria met in assessment
Pettengell 2008 [45]	Retrospective study Observational Belgium, France, Germany, Spain and the UK	n = 749 Non-Hodgkin’s lymphoma	Febrile neutropenia	Age OR 2.2 (1.21–4.01) Recent infection OR 3.07 (0.99–9.52) low baseline albumin < 35 g/l OR 4.76 (0.09–2.99)	Age, dose intensity, cycle 1 FN, CSF, renal, co-morbidity	High All criteria addressed
Dranitsaris 2008 [46]	Prospective study Observational Data from multicentre RCT	n = 509 Metastatic breast cancer	Any neutropenic complication > G2 NCI neutropenia or febrile neutropenia	Age > 59 OR 1.90 (0.95–3.78) PS WHO 2 OR 6.65 (2.54–17.39), cycle > 1 2.41 (1.32–4.39) Baseline neutrophil 2 × 109 cells/L 4.25(0.99–18.2)	Age, treatment, PS, stage, number of cycles	Moderate Study did not fully address confounders. Missing data not reported
Schwenkglenks 2006 [43]	Retrospective study Observational Luxenberg, Belgium, France, Germany, Spain and the UK	n = 2860 Early breast cancer	Neutropenic events Delays or hospitalisations	Higher age 1.02 (1.01–1.03) Higher BSA 3.85 (1.84–8.07) Lower BMI, 0.003 Body mass index 0.97 (0.94–0.99)	Age, weight, treatment, diarrhoea, regimen, cycles, radiotherapy	High All criteria addressed

ANC absolute neutrophil count, ALT alkaline transaminase, AST aspartate transaminases, BMI body mass index, BSA body surface area, CSF colony-stimulating factors, ECOG Eastern Cooperative Oncology Group, FBC full blood count, FN febrile neutropenia, GFR glomerular filtration rate, GCSF granulocyte colony-stimulating factor, N numbers of patients, PPI proton pump inhibitor, PS performance status, RCT randomised controlled trials, RDI relative dose intensity, UK United Kingdom, US United States, WBC white blood cells

**Authors did not detail statistics of finding