Twenty-five years after the first paper on etidronate in Paget's disease, there are few published papers that address bisphosphonate resistance as a specific clinical phenomenon. We report our data from two studies, Study 1 is a retrospective study of 20 patients,vith moderate to severe disease who were treated with intravenous (iv) pamidronate (221 +/- 18 mg [SEM]; range 60-360 mg), and after biochemical remission and relapse were retreated with generally larger iv dosage (293 +/- 28 mg; range 180-600 mg), The nadir bone turnover values were similar: plasma alkaline phosphatase (pAP) in 20 patients was 243 +/- 40 IU/l (mean +/- SEM) after the first course, and 267 +/- 44 IU/l after the second (reference range [RR] 35-135 IU/l). Likewise, fasting urinary hydroxyproline excretion (Hyp,) in 14 of the 20 patients was 4.5 +/- 1.1 mu mol/LGF and 4.1 +/- 0.9 mu mol/LGF, respectively (RR 0.40-1.92 mu mol/LGF). However the minimum duration of biochemical remission was significantly shorter after the second course-10.9 +/- 1.7 months (first) and 5.6 +/- 0.9 months (second) (p <0.03; Friedman's ANOVA n = 17). A subgroup of 10 patients who were followed for three courses showed a significantly higher pAP nadir in the third course. Study 2 is a prospective study of 40 patients, 23 previously untreated (NILPREV) and 17 previously treated with iv pamidronate (PAMPREV) and in biochemical relapse, who were randomly allocated to either oral alendronate 40 mg daily in 3 month units, or iv pamidronate 60 mg every 3 months. Treatment was continued until PAP and fasting urinary deoxypyridinoline/creatinine (Dpy/Cr) ratios (RR 5-27 mu mol/mol) were both in the reference range, or a clear plateau in each marker developed. At baseline, there were no significant differences in either marker between the two NILPREV groups and between the two PAMPREV groups. Using log-transformed data, in NILPREV the PAP reductions were significant and similar over the first 6 months, However, although each Dpy/Cr reduction was also significant, the difference in responses favored alendronate (p <0.015). In PAMPREV both markers showed no significant response to pamidronate; comparison showed a significantly greater response to alendronate (pAP p <0.02; Dpy/Cr p <0.002). Using two-way ANOVA, the pAP responses to alendronate in NILPREV and PAMPREV were similar and those to pamidronate were different (p = 0.034). The percentage of patients with both markers in the RR at 6 months or earlier were identical in NILPREV patients: alendronate 87% and pamidronate 87%, However in PAMPREV they were different: alendronate 83% and pamidronate 0% (p = 0.003). These data indicate: 1) patients treated with the same aminobisphosphonates for two courses show similar nadir values of bone turnover markers but a shorter remission time after the second course. In a third course the nadirs are significantly higher; and 2) in the alendronate/pamidronate comparison, NILPREV and PAMPREV patients showed similar PAP responses to alendronate, but significantly different responses to pamidronate. Thus, patients showing acquired partial resistance to one aminobisphosphonate (usually after two or more previous courses) are still capable of remission after exposure to another compound of the same class.