In the this study we have investigated the threshold plasma concentration of lidocaine for reversal of mechanical 'allodynia' in a neuropathic pain model in the rat, defined the concentration-dependent limits of that reversal and compared the acute reversal during intravenous drug infusion with the persistent relief of allodynia assayed 48 h later. Actions of i.v. lidocaine on ipsilateral and contralateral legs were also assessed. Forty rats were sorted into five groups (n = 7-10) and underwent spinal root (L5-6) ligation to produce allodynia, as quantified by a lower force of von Frey hairs at the plantar hindpaw required to elicit paw withdrawal (PWT, paw withdrawal threshold). During surgery, intravenous catheters were placed for programmed lidocaine infusion and in some animals arterial catheters were also inserted for assaying lidocaine blood levels. PWTs were measured in ipsilateral and contralateral paws before and after ligation and during infusions which, beginning at 5 days after surgery, were conducted every other day to incrementing levels (1.1-9.7 microg/ml plasma). Ligation produced allodynia in ipsilateral paws (PWT = 1.22 +/- 0.42 g (+/-SEM)) and in contralateral paws (PWT = 4.99 +/- 0.61 g), both markedly lower than pre-operative control values for either paw (11.31 +/- 0.41 g). The ipsilateral allodynia was partially, but significantly and permanently reversed (to PWT = 6-8 g) after a lidocaine infusion to 2.1 microg/ml in two separate groups (n = 7, 8). Lower concentrations resulted in elevation of PWT during infusion but no sustained relief. The elevation of PWT during infusion at this threshold level among individual animals was positively correlated with the relief measured 48 h later, but higher lidocaine concentrations infused in subsequent dosings could exact no further sustained relief. The residual PWT level, after reversal by threshold lidocaine and greater, was constant for each individual rat tested over the next 14 days but varied substantially among individuals; some were restored to pre-operative PWTs and some were totally unresponsive to drug. Retrospective analysis revealed a significant and unanticipated correlation between the incidence of low pre-operative PWTs (< 10 g) and a lack of sustained reversal of post-operative allodynia by lidocaine. Contralateral allodynia, despite its acute reversal during infusion to 2.1 microg/ml and higher, was not persistently relieved after infusion of lidocaine to any concentration. Repeated infusions to subthreshold levels (<2 microg/ml) did not provide persisting relief of allodynia on either side, and infusions of saline were impotent. These findings show that experimental allodynia results from multiple factors, only some of which are sensitive to lidocaine treatment, and that prolonged reversal of allodynia is limited in extent and likely influenced by pre-existing factors.