The use of micro-computed tomography (micro-CT) to study bone microstructure is continuously increasing. Thus, it is important to ensure that micro-CT can differentiate healthy and pathological bone. This study aimed to determine whether the reproducibility of bone histomorphometry and micro-CT, and agreement between the techniques, vary in bone samples with different metabolic status. Iliac crest biopsies (n = 36) were obtained from healthy subjects (n = 10) and from patients with osteoporosis (OP) (n = 15) or renal osteodystrophy (ROD) (n = 11). Micro-CT and histomorphometry analyses were repeated twice. Results were analyzed in separate groups and after pooling the data. Bone histomorphometry detected generally known differences between the diseases, whereas micro-CT did not detect differences between normal and ROD samples as effectively. Repeated measurements for BV/TV, Tb.Th, Tb.N, and Tb.Sp exhibited linear correlation coefficients (ρ) of 0.87-0.92 [coefficients of variations (CV), 8.3-27.2%] for histomorphometry and of 0.66-0.94 (CV, 4.4-23.4%) for micro-CT. There were no significant differences in reproducibility among samples from different study groups. Correlations between BV/TV (micro-CT) and mineralized bone volume (Md.V/TV, histomorphometry) were weaker than between BV/TV (micro-CT) and BV/TV (histomorphometry). When comparing the techniques, BV/TV, Tb.Th, and Tb.N displayed moderate correlations (ρ = 0.39-0.62, P < 0.05), and the agreement for BV/TV was highest in OP samples. The agreement between the techniques using clinical bone samples was moderate. Especially, micro-CT was less effective than bone histomorphometry in differentiating ROD from normal samples. The reproducibility was not affected by the health status of bone. Histomorphometry is still needed in clinical practice to study the remodeling balance in bone, and the methods are complementary.