Background: Isolated limb infusion (ILI) is a minimally invasive technique delivering regional chemotherapy to treat in-transit extremity melanoma. Determining perioperative factors that could predict toxicity is important to optimize strategies to improve clinical outcomes after regional chemotherapy in melanoma.
Methods: Perioperative factors from 171 ILI patients performed at eight centers from 2001 to 2008 were reviewed. The Wieberdink limb toxicity scale and creatine phosphokinase (CK) levels were used to measure toxicity. Logistic regression analysis was used to estimate the association between toxicity and perioperative parameters.
Results: Mild (grades I-II) and severe (grades >or=III) limb toxicity developed in 68% and 32% of patients, respectively. Melphalan adjusted for ideal body weight (aIBW) and papaverine were used in 47% and 63% of patients, respectively. Median peak CK for all patients was 563 U/l, and median peak occurred at postoperative day 4. On univariate analysis, papaverine and high CK levels (>563 U/l) were significantly associated with higher toxicity. On the contrary, aIBW was significantly associated with a lower risk of severe toxicity. Perfusate blood gas at 30 min [pH, PaO(2), and base excess (BE) ], limb temperature, and ischemia time were not predictive of limb toxicity. On multivariate analysis, severe toxicity was associated with female sex (P = 0.01), papaverine (P = 0.01), and high peak CK levels (P < 0.01). Independent predictors of high CK levels included younger age, unadjusted melphalan dose, and low PaO(2) at 30 min.
Conclusions: ILI can be performed with an acceptable morbidity. Papaverine use, female gender, and high peak CK were associated with higher limb toxicity. CK levels can be diminished significantly with aIBW.