Our study describes for the first time in a controlled study, the role of TDF and the concomitant use of TDF and PI in HIV-infected paediatric patients. The long term impact of ARV, specifically of drugs associated with renal damage on adults, is a major topic in infected children, since they are expected to have a longer duration of exposure. Moreover registration studies of antiretrovirals in children are few, with limited time of observation, and are often not focused on drugs interaction.
We reported a significant variation in the eGFR after 2 years in the entire cohort (p = 0.006) with a mean net loss of 12.1 ml/min per 1.73 m2, confirming what was reported by other studies on HIV infected adults [24, 25]. We didn't find any significant eGFR variation in patients taking TDF+PI or TDF, even when compared to the group that was not treated with nor TDF nor PI, or TDF respectively. However, all three of the cases who developed a mild reduction in eGFR were on a regimen containing both TDF and PI.
This is consistent with two other studies on ARV-experienced children treated with TDF-containing regimens which didn't find evidence of impaired filtration rate, even if their sample sizes were very small (27 and 26 children respectively) and were not controlled [13, 14], and with other pediatric cohort studies reporting some cases of renal toxicity associated with TDF [26–28]. Furthermore, as reported in other studies, estimated eGFR cannot be considered a sensitive indicator of early renal damage related to TDF , and our observation could have been affected by limited sample size. In fact, other studies on adults, with a large cohort and a higher power, found that patients treated with TDF and PI based regimens had a significantly greater decline in renal function than those taking TDF with Non-NRTIs (NNRTIs) or non-TDF-based regimens [25, 30]. The role of co-administration of TDF and PI on renal function was also recently demonstrated in a large randomized clinical trial on 1858 ART-naïve adults, the ACTG 5202 study, comparing four different regimens (TDF+NNRTI; TDF+PI; other NRTIs+NNRTI; otherNRTIs+PI). In this study the only regimen with a significant reduction in estimated creatinine clearance was in the group on both TDF and PI .
The detection of early or mild TDF-related nephrotoxicity could be evaluated with other more specific tests for proximal tubule injury, such as serum phosphate and potassium levels.
In our study we didn't find any significant variation in potassium in all therapy groups, but there was a trend in reduction in Group C (TDF+PI) after 2 years, and this variation resulted significantly different if compared to the Group D (no TDF+PI). This seems to confirm what has been already demonstrated by observations in adults and in pediatric population on TDF [12, 26].
In healthy population serum phosphorus is slightly higher in the prepubertal age, decreasing to adult values after puberty, as also considered by the age-specific DAIDS event scale. A recent case-control study in children confirmed an association between the use of TDF and hypophosphatemia, but it didn't analyze the role of concomitant use of PI . In our analysis the levels of phosphatemia changed earlier (after 1 year) and significantly in the entire cohort. Moreover phosphate reduction resulted significant and worse in the group treated with TDF, and in the group treated with TDF and PI; notably four of the five patients who developed clinical hypophosphataemia were on this regimen.
The observation of a tubular damage associated with concomitant use of TDF and PI is consistent with pathogenetic cellular studies and PK analysis, showing that PI inhibition of MRP2 and MRP4 results in an increase of tubular concentration of TDF [22, 23, 32], and TDF clearance is slower in subjects receiving PI, with an AUC increased of 22-37%, and possible increase of mitochondrial damage .
However our analysis has some limitations: it is an observational study, and although the groups analyzed resulted not different for age, gender, HIV or ethnicity, other variables like immune status and disease progression, could have had a role. Furthermore we assessed renal impairment without analyzing urine tests and other more sensitive indicators such as β-2 microglobulinuria that may have shown some milder renal abnormalities. Moreover, we didn't assess the association of TDF and PI with lower bone mineral density as reported by other studies [13, 17, 33], and the clinical significance of kidney abnormalities observed remains not fully known. Further studies with more sensitive urine markers, bone density assays, and longer follow-up are warranted.