The posterior adiposity index: a quantitative selection tool for adrenalectomy approach

Academic Article


  • Background: Objective criteria are lacking to determine whether a laparoscopic transabdominal (LA) or retroperitoneoscopic (RP) approach to adrenalectomy is optimal. We hypothesized that imaging characteristics could predict patients for whom RP adrenalectomy is the optimal approach. Materials and methods: Retrospective cohort study of all patients undergoing minimally invasive adrenalectomy between 2014 and 2016 (n = 113) at one institution. Imaging measurements included distances between the skin and Gerota's fascia (S-GF), upper borders of adrenal and kidney (A-K), adrenal and 12th rib (A-R), 12th rib and iliac crest (R-IC), and perinephric fat (PNF). These characteristics plus patient body mass index, gender, age, tumor size, and diagnosis were compared with operative time and estimated blood loss using Pearson's correlation or ANOVA. Multivariable linear regression also identified independent predictors of operative time. Results: Half of patients underwent LA (n = 57) and RP adrenalectomy (n = 56). Median age was 57 y; 60% were female. Mean tumor size was 3.2 cm. Higher body mass index patients were more likely to undergo LA (P = 0.03). Increasing lesion size modestly correlated with longer operative time (r = 0.341). On bivariate analysis, S-GF and PNF distances moderately correlated with operative time (r = 0.464 and 0.494) for RP procedures. The sum of S-GF and PNF generated a Posterior Adiposity Index (PAI). The PAI strongly correlated with operative time for RP (r = 0.590). Nothing was significantly associated with estimated blood loss. Multivariate analysis revealed larger lesions (P = 0.025) and increasing PAI (P = 0.019) were predictive of longer operative time, with PAI ≥9 conferring the greatest risk (P = 0.004). Conclusions: Smaller tumors and PAI <9 are associated with shorter operative times in RP adrenalectomy. Surgeons can utilize preoperative images to calculate the PAI and determine whether an RP approach would be favorable.
  • Published In

    Digital Object Identifier (doi)

    Author List

  • Lindeman B; Gawande AA; Moore FD; Cho NL; Doherty GM; Nehs MA
  • Start Page

  • 26
  • End Page

  • 31
  • Volume

  • 233