Table 1.

Examples of studies demonstrating the obesity paradox in patients with cancer, where BMI was determined either pre-, peri-, or post-diagnosis of cancer

Study	Time of BMI determination	Cancer	Number and country (% male)	BMI categories kg/m2	Results	Adjusted for	Comment
Reichle et al. (2015) [32]	Pre-diagnosis	Prostate (n = 1958), colorectal (n = 1013), breast (n = 1410), gynecological (n = 572), upper digestive tract (n = 635), urogenital (n = 607), lymphatic (n = 508), others (n = 1970)	n = 8673 Austria (58 %)	<18.5 18.5–24.9 25.0–29.9 ≥30.0	1.28 (1.02–1.60) 1.00 (referent) 0.93 (0.87–0.99) 1.06 (0.97–1.15)	Age at cancer diagnosis, sex, smoking status, primary location, stage
Tsang et al. (2016) [15]	Peri-diagnosis (date of receiving radiation therapy)	Distant metastases (bone, brain, others) with primary tumors (lung, breast, others)	n = 4010 Taiwan (55 %)	≤18.5 18.5–25.0 25.0–29.9 ≥30.0	1.41 (1.26–1.58) 1.00 (referent) 0.84 (0.78–0.91) 0.68 (0.57–0.81)	Age current, sex, performance status, primary tumor site, site of metastasis, multiple, onset of metastasis. EQD, chemotherapy, comorbidities, employment, alcohol, smoking, betel quid chewing, rural town	Alcohol, smoking, age, and comorbidities have a P value greater than 0.05
Schlesinger et al. (2014) [11]	Post-diagnosis (average 4 years after diagnosis)	Colorectal	n = 2143 (and n = 7565 in a meta-analysis) Germany (57 %)	25 ≤ 25.0–29.9 ≥30.0	2.12 (1.18–3.80) 1.00 (referent) 0.79 (0.71–0.89) 0.91 (0.80–1.04)	Age current, sex, alcohol, smoking status, tumor location, family history of CRC, metastases, other cancers (initially)	Prospective cohort study then meta-analysis with 7565 CRC patients

BMI body mass index, EQD equieffective dose (of radiotherapy), CRC colorectal cancer