To determine if there are variables linked to responsiveness to pulse cyclophosphamide/methylprednisolone therapy in progressive Multiple Scerosis (MS). MS is a presumed autoimmune disease of the CNS in which immunosuppressive and immunomodulatory treatments are being used. We have treated patient with the progressive form of MS using a regimen consisting of pulse cyclophosphamide/methylprednisolone that is given as an outpatient at 4-8 week intervals similar to lupus nephritis protocols. We investigated a series of 95 consecutive progressive MS patient treated in an open label fashion in an effort to identify factors linked to response to treatment. Clinical outcome measures included status at 12 months and time to failure determined by EDSS change and global physician impression. For each endpoint associations were examined between outcome and patient characteristics including gender age at onset of disease and treatment, EDSS 1 year previously and at start of treatment, duration of MS, previous treatment, age at onset and duration of progression, and primary vs secondary progressive MS. Of the variables studied, age, gender, age at onset, and age at treatment did not correlate with response to therapy. The most significant variable that correlated with response was length of time the patient was in the progressive phase (P=0.048, 12 month change in EDSS; P=0.017, risk for time to failure). Patient that improved on therapy at 12 months had progressive disease for an average of 2.1 years prior to treatment, whereas those stable or worse had progressive disease for 5.0 and 4.1 years respectively. There was a trend (P=0.08) favoring positive clinical responses in secondary progressive as opposed to primary progressive patients. Our data suggest that progressive MS may become refractory to immunosuppressive therapy with time and early intervention when patient enter the progressive stage should be considered. Furthermore, in trials of immunosuppressive agent for progressive MS, duration of progression should be considered as a randomization and analysis variable.