Background: Knowledge of eGFR outcomes after chronic renal impairment (CRI) in HIV-positive persons and its relationship to antiretroviral drug (ARVs) use is limited. Prior studies have shown increased tenofovir (TDF) discontinuation at eGFR<70 mL/min/1.73m2 limiting the value of eGFR<60 as a CRI cut-off. Methodology: D:A:D participants were followed from eGFR>80 (Cockcroft-Gault) after 1/1/2004 to CRI, defined as 2 consecutive eGFR<70 (>3 months apart), thus ensuring an eGFR decline of >10 before CRI. The median of all (and minimum 2) eGFRs at 12-24 months after CRI was compared to the median eGFR defining CRI, and eGFR changes were grouped into: improvement (increase >+10), stabilisation (-10 to +10) and progression (decrease >-10). Multinomial regression assessed odds of better eGFR outcomes (assuming improvement is better than stabilisation, which in turn is better than progression) adjusting for use and discontinuation of TDF, atazanavir (+/- ritonavir, ATV/r), lopinavir (LPV/r), other boosted protease inhibitors (PI/r) and abacavir (ABC), as well as demographics, HIV-related factors, HCV status and traditional renal risk factors (figure). Results: Of 1237 persons developing CRI, 23.2% improved eGFR, 69.0% stabilised and 7.8% progressed at 12-24 months after CRI. Those on TDF at time of CRI had lower odds of better eGFR outcomes compared to those never exposed to TDF, while those off TDF for >12 months at time of CRI had higher odds of better outcomes (figure). Similar trends were seen for ATV/r, LPV/r and other PI/r, but not ATV or ABC. Censoring follow-up for nephrotoxic ARVs used concomitantly (i.e. those on TDF for concomitant ATV/r use) showed similar results. Older persons (adjusted odds 0.61/10 yrs [95%CI 0.50-0.70]) and those with slowly declining eGFR prior to CRI (0.72 [0.55-0.96], <10 vs. >10/yr) had significantly lower odds of better eGFR outcomes, while those with diabetes for >5 yrs had marginally significant lower odds compared to non-diabetics (0.61 [0.36- 1.05]). No HIV-related factors were associated with better eGFR outcomes. Conclusions: Use of TDF, ATV/r, LPV/r and other PI/r, older age, diabetes and slowly declining eGFR were associated with decreased odds of better eGFR outcomes in HIV-positive persons after CRI. TDF discontinuation prior to CRI was associated with better eGFR outcomes, suggesting TDF associated eGFR decline may be halted or reversed with early cessation. There was some suggestion that this may also be true for ATV/r, LPV/r and other PI/r.