The paper investigates the output-feedback control problem for a more general class of stochastic nonlinear systems with stochastic integral input-state stability (SiISS) inverse dynamics driven by noise of unknown covariance. With the aid of stochastic LaSalle theorem, small-gain type conditions and backstepping technique, an adaptive output-feedback controller is successfully constructed to render the closed-loop system globally stable in probability and ensure that all the signals in the closed-loop system are bounded almost surely.