Roux-en-Y gastric bypass (RYGB) induces weight loss and improves insulin sensitivity when evaluated by the hyperinsulinemic-euglycemic clamp (HEC). Surrogate indices of insulin sensitivity calculated from insulin and glucose concentrations at fasting or after an oral glucose tolerance test (OGTT) are frequently used, but have not been validated after RYGB. Our aim was to evaluate whether surrogate indices reliably estimate changes in insulin sensitivity after RYGB. Four fasting (inverse-HOMA-IR, HOMA2-%S, QUICKI, revised-QUICKI) and three OGTT-derived surrogates (Matsuda, Gutt, OGIS) were compared with HEC-estimated peripheral insulin sensitivity (Rd or Rd/I, depending on how the index was originally validated) and the tracer-determined hepatic insulin sensitivity index (HISI) in patients with preoperative type 2 diabetes (n=10) and normal glucose tolerance (n=10) 1 week, 3 months and 1 year postoperatively. Post-RYGB changes in inverse-HOMA-IR and HOMA2-%S did not correlate with changes in Rd at any visit, but were comparable to changes in HISI at 1 week. Changes in QUICKI and revised-QUICKI correlated with Rd/I after surgery. Changes in Matsuda and Gutt did not correlate with changes in Rd/I and Rd, respectively, whereas OGIS-changes correlated with Rd-changes at 1 year post-RYGB. In conclusion, surrogate measures of insulin sensitivity may not reflect results obtained with gold standard methodology after RYGB, underscoring the importance of critical reflection when surrogate endpoints are used. Fasting surrogate indices may be particularly affected by post-RYGB changes in insulin clearance, whereas the validity of OGTT-derived surrogates may be compromised by the surgical rearrangements of the gut.