Bone Mineral Density Decline According to Renal Tubular Dysfunction and Phosphaturia in Tenofovir-exposed HIV-infected Patients

José L. Casado; Carmen Santiuste; Monica Vazquez; Sara Bañón; Marta Rosillo; Ana Gomez; María J. Perez-Elías; Carmen Caballero; José M. Rey; Santiago Moreno


AIDS. 2016;30(9):1423-1431. 

In This Article


To date, loss of BMD has been observed after cART initiation, but in a greater proportion with the use of TDF, with stabilization after 48–96 weeks on therapy. This has been explained as a continued role of HIV infection or as part of an immune reconstitution syndrome.[18] Nevertheless, we found a lineal correlation between time on therapy and lower BMD in patients receiving TDF, a continuation of bone decline previously described in the continuation of the AIDS Clinical Trial Group (ACTG) 5224 s study.[19] Thus, the role of TDF on bone loss seems to be significant, as confirmed by small or no changes in BMD, in studies of dual therapies without TDF.[20] Furthermore, in HIV negative patients, preexposure prophylaxis studies found that initiation of TDF is also associated with mild but significant bone loss.[21]

However, the mechanisms underlying the effect of TDF have not been clearly established. In-vitro studies suggest that TDF may alter gene expression in both osteoblasts and osteoclasts,[22] but there have been several case reports and case series demonstrating the appearance of hypophosphatemic osteomalacia in patients receiving TDF.[23–26] Our study, in a well controlled population of HIV-infected patients receiving TDF in the clinical setting, show that chronic phosphaturia plays a key role in bone outcome. We observed that patients with increased phosphaturia had a reduced BMD, and that previous changes in the rate of phosphaturia during therapy were associated with BMD at inclusion. As confirmation, hyperphosphaturia was an independent factor for BMD decline in a median of around 3.5 years, after adjusting by age, sex, and BMI.

Mounting evidence supports that tubular abnormalities should be included between the toxicity of TDF during follow up.[27] Among them, hyperphosphaturia was observed in 50% of our patients. Previous in-vitro studies have demonstrated that phosphate transport from the ultrafiltrate across the proximal tubule epithelium is an energy-dependent process, therefore particularly sensitive to mitochondrial toxicity, and increased urinary phosphate excretion has been consistently associated with TDF use,[28] although few studies analyse simultaneously phosphataemia and phosphaturia, and its changes. Indeed, we found a significant inverse relation between both parameters. Furthermore, tubular dysfunction as cause of phosphaturia was confirmed, at least in part, by the significant correlation with other tubular parameters such as tubular proteins.

Therefore, our data confirm the importance of increased urinary loss of phosphate. That is, a phenomenon similar in part to the hypophosphatemic osteomalacia observed in the so-called tumour-induced osteomalacia or Fanconi syndrome. However, these two clinical entities are characterized by massive phosphaturia, secondary hypophosphataemia, and bone loss, and we did not observe high phosphaturia levels or maintained hypophosphataemia. By fact, hypophosphataemia was partly compensated with a significant increase in the tubular reabsorption of phosphate during follow-up. Thus, serum phosphate level was not useful to directly identify BMD loss, suggesting the influence of compensatory factors.

Indeed, maintenance of serum phosphate within a relatively narrow range is crucial for several important cellular processes, and it depends on a complex interplay between intestinal absorption of phosphate, bone resorption, shifts between intracellular and intravascular compartments, and renal excretion, influenced by levels of vitamin D, PTH, or other molecules such as phosphatins.[29] Bone plays a major role in phosphate homeostasis, as the hydroxyapatite matrix serves as the critical pool of phosphate, and serum phosphate levels are mostly determined by the rate of renal excretion.[29] Therefore, our data suggest that the maintenance of phosphataemia could lead to increased bone resorption in case of lack of, or insufficient, resolution of hyperphosphaturia, as confirmed by the association between fractional excretion of phosphate and reduced BMD at inclusion. Even more, a fractional excretion of phosphate below 20% could not be indicative of bone protection, and a higher level could be necessary to the maintenance of phosphataemia and to avoid bone resorption after a prolonged time on TDF.

These data are not surprising. It is already known the presence of a bone–kidney axis that coordinates bone mineralization and renal handling of phosphate and vitamin D metabolism. The use of TDF and secondary hyperphosphaturia could alter this interplay, resulting in different expression of bone loss, as result of the intensity of the injury and probably the mechanisms of response. Furthermore, bone loss was higher in patients with greater BMD at inclusion and it was a factor involved in the rate of decline, a finding previously described in the Simplification with Tenofovir-Emtricitabine or Abacavir-Lamivudine (STEAL) study,[30] emphasizing the complex regulation of bone–kidney and phosphate metabolism in case of tubular dysfunction. However, several questions remain to be elucidated, and unfortunately we did not include other molecules involved in phosphate metabolism, such as calcitriol or fibroblast growth factor-23, or more bone turnover markers in addition to BALP, limiting our capacity of an accurate and complete explanation.

In our study, classical factors for BMD decline, such as age, sex, BMI, or HIV-related factors play an important role in BMD at inclusion, but with the exception of the use of protease inhibitor, they were not related with bone loss in the relatively short time between both DXA scans. Although data are controversial, the use of protease inhibitor has been associated with bone loss in several studies, in relation with more severe TDF toxicity, and/or enhanced osteoclast activity.[31,32] Of note, higher 25OHD levels were associated with a better evolution of BMD, albeit it was not significant. Despite the mechanisms are not fully understood, the use of TDF has been associated with increased PTH levels, and this, in turn, stimulates the kidneys to produce 1,25OHD in the presence of optimal 25OHD levels, that enhances tubular reabsorption of phosphate.[33,34] Therefore, 25OHD can reduce the effects of TDF on phosphate homeostasis alteration.[35] Indeed, a recent study showed attenuation of bone loss with the use of vitamin D and calcium at cART initiation.[36]

We found no significant correlation between BMD decline and B2M or RBP, as previously described,[37] probably because of our inclusion criteria selected those patients able to remain longer on TDF, more than 6 years, and therefore with low tubular and renal toxicity. In any case, the number of tubular abnormalities was associated with BMD loss. In one recent cross-sectional study, RBP levels, but not phosphaturia, were associated with BMD decline.[38] In comparison with our data, the authors did not evaluate changes in phosphataemia or phosphaturia. Also, there was a significant correlation between phosphaturia and tubular proteinuria, suggesting that RBP or B2M could act as a marker of severity of tubular dysfunction. Even, it could be expected that tubular parameters could be useful in cross-sectional studies, since phosphataemia and phosphaturia could remain normal or close to normal during a long time on TDF. In support of this, patients switching from TDF to a novel prodrug, tenofovir alafenamide, TAF, had concomitant improvements in urinary biomarkers and BMD, even in patients with previous renal impairment.[39]

Taken together, our data clarify previous controversial issues. As hypothesis, tubular alteration and hyperphosphaturia are associated with an increased renal reabsorption of phosphate to maintain phosphataemia within normal range, and the imbalance is compensated through bone resorption in some patients, especially those with higher BMD. Even, this fact could explain initial stabilization, because of improvements in renal phosphate handling, or because of the adjustment secondary to the magnitude of the initial BMD decline, an equilibrium that could be altered with longer TDF exposure or more severe tubular involvement. Adequate vitamin D status could improve renal reabsorption of phosphate, or at least, reduces its consequences.[40] Moreover, the rapid improvement reported in small studies of TDF switch may be the result, at least in part, of mineralization of previously undermineralized bone matrix and osteoid, as observed with other causes of osteomalacia.[41]

Our study has limitations. First, our patients were a selected population since TDF could have been withdrawn with more severe renal toxicity, and therefore bone loss could be even worse with more advanced tubular dysfunction. In this line, our cohort was at low overall risk for bone-related harm as demonstrated by small BMD changes during follow-up. Also, this relatively short duration of the study and the small sample did not allow for investigation of fractures and the possible role of the factors described. Finally, we did not observe this association with BMD at spine, because of mild changes at this localization, and to the fact that BMD changes for the hip and spine measurements indicates different kinetics of response.

In conclusion, the emergence of tubular dysfunction and chronic phosphaturia correlated with reduced serum phosphate, and the altered equilibrium between phosphataemia and phosphaturia could be one of the main causes of BMD decline in patients on TDF. Also, serum phosphate concentrations may not reliably reflect total body phosphate depletion, neither even tubular phosphate reabsorption. Thus, cross-sectional serum phosphate measurements, and usual cutoffs for phosphaturia, may be of limited use in identifying patients at greatest risk of bone loss, and longitudinal evaluation is necessary. From a clinical point of view, patients on TDF are at increased risk for osteomalacia and it should be considered before indication of bisphosphonate therapy. Future studies should identify accurately the mediators involved in phosphate metabolism in case of tubular dysfunction or the changes in phosphaturia, phosphataemia, and BMD after TDF initiation or interruption.