A prospective study of multiple needle biopsies versus a single open biopsy for testicular sperm extraction in men with non-obstructive azoospermia

To identify the predictive factors for testicular sperm extraction (TESE) and to understand the pathology associated with TESE, we carried out a prospective study in 40 consecutive men with azoospermia due to primary gonadal failure. The main outcome measure was the retrieval of at least one testicular spermatozoon. Endocrine and biophysical profiles, testicular histology, Johnsen score and testicular spermatids were used as predictors of sperm extraction. Spermatogenesis was quantified with the Johnsen score. A variable pattern of spermatogenesis was common, being present in 20 (50%) patients. Visualisation of testicular spermatids on testicular histology showed a strong association with TESE (P < 0.0001). Statistically significant differences were detected in plasma follicle stimulating hormone (FSH) and testicular volume between patients who had hypospermatogenesis and Sertoli cell-only or maturation arrest. There were no significant differences in Johnsen score, biophysical and endocrine profiles between the groups with successful and failed TESE. However, a statistically significant trend occurred with changes in histological pattern [chi2 for trend, P = 0.001; Pearson's coefficient (r) = 0.6], Johnsen score (P = 0.022; r = 0.5), testicular volume (P = 0.01; r = 0.5) and plasma FSH concentrations (P = 0.044; r = 0.4), albeit to a limited degree. Difference in the interpretation of histological patterns with different assessors was observed. The type of occupation or risk factors for azoospermia showed no association with testicular pathology or TESE. Variable histological patterns in different tubules in the same individual may explain the poor correlation of TESE with endocrine and biophysical profiles, Johnsen score and histological pattern. Differences in the amount of tissue used for TESE and histopathology, and misinterpretation of testicular histology rather than failure to quantify spermatogenesis may explain the poor correlation between histological patterns and TESE. Testicular spermatids predicted TESE. However, considerable overlap in values means that no single variable can provide a perfect discrimination between the groups with successful and failed TESE.