In the first study to examine a full set of urinary analytes after administration of oral STS, neither healthy controls, nor hypercalciuric calcium stone formers, experienced statistically significant increases in urine calcium excretion or calcium oxalate or calcium phosphate supersaturation. In stone-formers, there were decreases in urine pH and citrate, and increases in ammonium excretion without a change in serum [HCO3
]. Similar results occurred in the control non-stone-formers, with the exception that citrate did not fall. These effects on urine chemistry, with no net change in lithogenicity, would not explain remarkable reductions in stone activity reported in a non-controlled, non-randomized clinical study 
. Of note, that study also demonstrated no significant difference in calciuria pre- and post-STS.
Our urine results are also not consistent with the findings of Yatzidis et al 
. In that paper, participants receiving the same 20 mmol dose of STS that we studied excreted about 4 mmol/d of thiosulfate, with only 0.5 mmol/d of sulfate excretion, an unlikely, below-normal, value. We note that we used a similar barium-precipitation method for sulfate measurement and therefore cannot account for the peculiarly low sulfate values reported in that paper. In contrast, we noted less than 1 mmol/d of thiosulfate in the urine of our study participants, and found that excreted sulfate was approximately equimolar with the amount of thiosulfate administered. Similarly, in the rat study, only 6% of orally administered thiosulfate appeared in the urine, considerably less than the excreted sulfate 
The present study is consistent with thiosulfate presenting a net acid load, suggested by increases in urine ammonium excretion, and decreases in citrate excretion (in stone-formers) and urine pH. The etiology of this acid load is not clear. Stool losses of bicarbonate or potential base could be an explanation, though most patients had no diarrhea and those who did, experienced it only transiently. Previous reports of metabolic acidosis after parenteral administration of STS for cyanide poisoning and calcific uremic arteriolopathy exist 
; parenteral administration of STS is not associated with diarrhea. Although only 3 patients had serum [HCO3
] measurements while taking STS, all 3 had clear evidence of having received an acid load based on urine chemistry while maintaining normal serum [HCO3
]. The lack of change in serum [HCO3
] in the stone formers and the lack of change in urine citrate in the controls (in whom serum [HCO3
] was not measured) suggest that ammoniagenesis was adequate to compensate for the presumed acid load. The difference in citrate excretion between controls and stone-formers despite similar increases in urine ammonium excretion, is interesting but not explained and suggests that perhaps stone-formers have greater renal citrate reabsorption in response to acid loads, contributing to relative hypocitraturia. The decrease in urine pH caused a significant increase in supersaturation of uric acid and could be associated with an increase in uric acid stone formation in some susceptible patients.
Some recent literature has claimed that sodium thiosulfate is a strong acid 
, when the molecule clearly lacks protons to donate. Others suggest that dissolution of sodium thiosulfate leads to formation of thiosulfuric acid 
, which cannot occur given that the pKa
’s of that strong acid are 0.6 and 1.7 and the common knowledge that the sodium salt of a strong acid is not itself an acid. More likely, the acid load is due to the oxidation of thiosulfate to sulfate by the liver, producing protons 
. An alternative explanation would be the oxidation of thiosulfate to sulfate by intestinal bacteria with absorption of sulfate and protons by the intestine 
. Either of these mechanisms is consistent with our finding that oral thiosulfate administration was followed by excretion of sulfate, not thiosulfate, in the urine. However, the possibility that this conversion occurs as the result of metabolism by intestinal bacteria remains open, since intravenously administered thiosulfate is rapidly excreted with only a small portion being metabolized 
. On the other hand the occurrence of acidosis in dialysis patients treated with parenteral STS for calcific uremic arteriolopathy might be attributable to these reactions occurring endogenously at a slow rate but one sufficient to produce acidosis when excretion is limited by low glomerular filtration rate.
The urine findings are similar to those noted in GHS rats given STS 
. These animals generally form calcium phosphate stones if fed normal rat chow. STS led to an increase in urine calcium excretion as well as increased urine ammonium, and decreased pH and citrate excretion. Therefore the significant reduction in stone formation was not explained by measurement of conventional urine chemistry and calculation of calcium phosphate or calcium oxalate supersaturations. Though acidification of urine could be expected to reduce formation of calcium phosphate stones, this possibility was discounted because in previous studies, administration of the acidifying agent ammonium chloride was not associated with reduction in stone formation 
. Increased urine calcium in the rats could be attributable to increased urine sodium excretion, given the sodium content of STS, or due to the metabolic acid load. Neither a significant increase in sodium nor calcium excretion occurred in our participants. In another rat study, ethylene glycol and ammonium chloride-induced nephrocalcinosis was not benefitted by administration of STS. 
One limitation of the study is the small sample size of the 2 groups and that the groups are not matched for age. Since the primary outcome was not a comparison of the effects of the drug on the two populations we do not consider the age difference to be important. In fact, combining the results of the 2 groups demonstrated that the results were highly consistent between them with the only difference being a fall in urine citrate in hypercalciuric stone-formers but not in the non-stone-forming controls. Finding healthy people willing to take an investigational drug and do multiple urine collections among a hospital population was more difficult than finding stone formers interested in advancing the knowledge of their disorder and its potential treatment. Only after the results in the healthy group did we realize that it would be desirable to measure serum chemistry after drug administration and because of logistical issues could not get blood while participants were still taking the drug in 2 of the 5 stone formers. Since serum [HCO3] did not change in the 3 in whom it was measured despite their significant changes in urine chemistry, we do not believe this small sample is an important limitation to knowing whether an acid load was presented. The urine findings clearly show that it occurred. We cannot speculate regarding the generalizability of these findings to a broader population of calcium stone formers with these or other urinary risk factors for stone formation.
The absence of a significant change in serum [HCO3
] does not necessarily confirm the safety of long-term STS administration, as the finding is potentially related to two limitations of the study: its small sample size and relatively short duration of STS administration. STS therapy for the prevention of nephrolithiasis would likely need to continue for many years, and as such, studies of at least months to years would be necessary to sufficiently rule out development of sustained metabolic acidosis. A larger study with a longer duration more similar in scale and length to that of Yatzidis 
would be necessary for a more concrete assessment of the safety of the drug. Yatzidis however did not report any serum or urine variables that would allow determination of whether there was evidence for the metabolic acidosis we show here. The implication of long-standing metabolic acidosis might be to reduce BMD. Despite the lack of an increase in urine calcium excretion in this and previous studies, the long-term effects of protein ingestion and net acid loads to reduce BMD are well known 
. Patients with hypercalciuria are particularly susceptible to these effects and have lower BMD compared with age- and gender-matched controls 
. Treatment with STS reduced the load needed to fracture femurs of nonuremic rats when compared to untreated animals; those animals also experienced hypercalciuria 
The appropriate therapy for prevention of recurrent calcium phosphate stones is unclear as no randomized controlled trials with this outcome have been completed 
. The role of administration of citrate is controversial because its effect to alkalinize the urine and increase calcium phosphate supersaturation may, to an extent, negate the effect of alkali to reduce calcium excretion and increased urine citrate to inhibit calcium salt crystallization 
. Therefore it is worth considering whether an agent that leads to urinary acidification without significant acidosis would be effective for such stones. STS did inhibit calcium phosphate stone formation in GHS rats despite increasing urine calcium, a finding not observed in these human studies 
. It might therefore be worthwhile to study the effects of STS specifically in this setting. Whether administering it with bisphosphonates or thiazides to diminish potential loss of BMD and reduce calciuria 
, would improve its safety profile and enhance efficacy can only be determined by appropriate clinical trials. However, overall, given the acid-loading effects of STS, as evidenced by the changes in urine chemistry, and the effects on bone characteristics demonstrated in animal studies, we are not confident that STS would be a safe and effective therapy for prevention of recurrent calcium oxalate or calcium phosphate stones.