The prevalence of HIV-associated kidney disease has decreased in the era of antiretroviral therapy (ART), but drug-induced nephrotoxicity is emerging.1 Long-term exposure to nephrotoxic antiretroviral drugs, such as tenofovir disoproxil fumarate (TDF), may lead to progression of renal dysfunction and incomplete reversibility of toxicity, even after drug withdrawal.2,3 Moreover, some urinary biomarkers of early kidney injury were found to be associated with important clinical outcomes, such as mortality and severe heart disease.4,5 Therefore, it is important to detect early renal dysfunction in HIV-infected individuals, particularly in those exposed to TDF.

TDF is widely used as a preferential drug for ART according to worldwide guidelines, and it is also part of the only approved regimen for pre-exposure prophylaxis to HIV infection.6 The drug is excreted through glomerular filtration and proximal tubular secretion, and its use has been related to an increased risk of rapid decline in renal function and chronic kidney disease.2 Clinically, renal insufficiency is usually assessed by an estimated glomerular filtration rate (eGFR),7 but proximal tubular dysfunction is more common and specific for TDF-induced toxicity8 and can occur with the absence of glomerular defects.9

The evaluation of urinary biomarkers, such as α1-microglobulin (α1M), β2-microglobulin (β2M) and N-acetyl-β-d-glucosaminidase (NAG), may improve the identification of those at risk of tubular toxicity and may detect early stages of renal dysfunction.10,11 However, the evaluation criteria of these proteins in tubular dysfunction have not been established and little is known about the change of these new biomarkers during ART. The role of the aforementioned biomarkers in diagnosing and monitoring early tubular damage still needs to be explored.

HIV transmission is increasing markedly faster among MSM than other high-risk groups in China and some other countries.12,13 MSM are at risk of TDF toxicity due to its comprehensive role in treatment and pre-exposure prophylaxis. However, most biomarkers of early renal dysfunction have not been validated in HIV-infected patients let alone in specific key-populations, such as MSM. Our study aims to investigate the presence of early renal dysfunction among HIV-infected MSM and to find the indicators of early renal injury in MSM with TDF exposure.

Two-hundred-thirty-nine HIV-infected MSM were enrolled in the study and stratified by ART into the three groups: TDF group (n=97), non-TDF group (n=66), and non-ART group (n=76). The characteristics of the participants are given in Table 1. In the study sample, the frequency of hypertension, diabetes and hypophosphatemia were 2.9% (n=7), 2.5% (n=6), and 9.6% (n=23), respectively. Although almost half of the patients in the TDF group were treatment experienced with a non-TDF-containing regimen, the total ART time in both ART groups was similar (p>0.05).

In the cross-sectional study, an eGFR<90mL/min/1.73m2 was seen in eight patients (3.3%) and an eGFR<60mL/min/1.73m2 was seen in only two patients (0.8%). Albuminuria was present in 12.1% of our participants. As shown in Fig. 1 and Table 2, the TDF group had a higher frequency of α1M proteinuria (α1MCR >25.8mg/g) and a higher α1MCR value than the other two groups (all p<0.05). A higher frequency of β2M proteinuria (β2MCR >0.68mg/g) was also observed in the TDF group compared to the non-TDF group (p=0.024). No significant differences were found in β2MCR among the three groups. Statistical analysis revealed no significant differences in abnormal frequency or ratio-to-creatinine value of urinary NAG and albumin among the three groups.

Fig. 1.

Proportion of tubular and glomerular proteinuria among the study groups.

Albuminuria was defined as ACR >30mg/g. Other kinds of proteinuria were defined as β2MCR >0.68mg/g, α1MCR >25.8mg/g, and NAGCR >9.1IU/g, respectively. ACR, urinary albumin-to-creatinine ratio; β2MCR, beta2-microglobulin-to-creatinine ratio; α1MCR, alpha1-microglobulin-to-creatinine ratio; NAGCR, N-acetyl-β-d-glucosaminidase-to-creatinine ratio.

(0.09MB).

β2MCR and α1MCR (in urine) were both correlated with serum β2M (Fig. 2A and B, all p<0.001). In multivariate analyses, TDF exposure (OR 2.62, 95%CI 1.09, 6.30), body weight <60kg (OR 3.05, 95%CI 1.45, 6.40), and an eGFR <90mL/min/1.73m2 (OR 13.77, 95%CI 2.37, 80.03) were associated with UQ α1MCR; after adjusting for serum β2M, TDF exposure (OR 4.16, 95%CI 1.43, 12.10), body weight <60kg (OR 3.30, 95%CI 1.43, 7.63), and the cumulative time on ART (OR 1.02, 95%CI 1.00, 1.04) were related to UQ α1MCR (Table 3). Moreover, protease inhibitor exposure (OR 3.09, 95%CI 1.00, 9.55), CD4 cell count <200cells/mm3 (OR 2.96, 95%CI 1.40, 6.25), and eGFR <90mL/min/1.73m2 (OR 25.26, 95%CI 2.74, 232.71) were associated with UQ β2MCR; after adjusting for serum β2M, only cumulative time on ART (OR 1.02, 95%CI 1.00, 1.04) was related to UQ β2MCR (Table 4). Factors with too low prevalence, such as diabetes (2.5%) and hypertension (2.9%), were not included in the analysis.

Fig. 2.

Correlations between β2MCR and serum β2M (A), α1MCR and serum β2M (B). β2MCR, beta2-microglobulin-to-creatinine ratio; α1MCR, alpha1-microglobulin-to-creatinine ratio.

(0.15MB).

To further study the impact of TDF on kidney function, we followed-up 56 MSM from the non-ART group who started TDF/3TC/EFV. As shown in Table 5, α1MCR was the only urinary biomarker that dramatically increased after three and six months on TDF-containing therapy compared to baseline data (p=0.023 and 0.011, respectively). There was a slight increasing trend of β2MCR (in the urine) at three months, but it disappeared at six months (all p>0.05), concurrent with a significant decrease in β2M after six months of ART (p=0.028).

In an era in which early ART and pre-exposure prophylaxis are generally encouraged,17 TDF use and its nephrotoxicity are likely to draw increasing attention.18 In HIV-infected females, some urinary biomarkers were related to TDF use and could forecast declining kidney function.19,20 Among HIV-infected MSM, we observed a higher rate of α1M and β2M proteinuria in TDF users in a cross-sectional study and confirmed urinary α1M elevation after TDF exposure in a longitudinal study. HIV-seronegative MSM were found to have a small but significant decrease in creatinine clearance after TDF exposure,21 but novel urinary biomarkers among this population have not been evaluated. Our findings underscored the importance of subclinical renal function monitoring in HIV-infected MSM treated with TDF, and the findings might also give clues for early detection of TDF-related toxicity in non-infected MSM.

Here we chose four urinary candidate markers to indicate early renal dysfunction: albuminuria is a classical indicator of early glomerular injury;22 α1M, β2M, and NAG are tubular biomarkers, but their role in diagnosing and monitoring early tubular damage is still undetermined. Of all of the selected markers, urinary α1M seems to be the most reliable for determining TDF-associated renal dysfunction in several ways. First, urinary α1M was the only biomarker that could distinguish the TDF group from the other two control groups in the cross-sectional study. Second, α1M was also the only marker with notable elevation after ART initiation assessed in the longitudinal study. In a review paper, TDF-related renal damage usually appears within 5–12 months after initiation of therapy.23 Our observation confirmed the sensitivity of urinary α1M, which increased significantly already after three months of TDF-containing therapy. Third, we observed no significant differences in albumin or NAG in cross-sectional or longitudinal studies. Urinary β2M might also be a helpful biomarker, which has been recommended along with urinary α1M as a useful screening tool for TDF-induced kidney tubulopathy in a diagnostic accuracy study,10 and showed a higher rate of proteinuria in the TDF group than the non-TDF group in our study. However, urinary β2M is highly dependent on serum β2M and the pH value of the urine.24,25 Unlike β2M, which is unstable in acidic urine, α1M has been reported to be very stable in urine at different pHs.26

It is worth mentioning that the inflammatory status of HIV infection may disturb the interpretation of urinary biomarkers in tubular dysfunction because α1M and β2M are both inflammatory biomarkers.26,27 In our observation, urinary α1MCR and β2MCR were both correlated to serum β2M, which is a non-specific marker of chronic inflammation and could be increased even after ART.27,28 However, serum β2M seemed to have a stronger effect on urinary β2MCR than urinary α1MCR. In multivariate analyses, TDF exposure was an independent risk factor for UQ α1MCR before and after serum β2M adjustment. By contrast, after adjusting for serum β2M, all traditional risk factors of tubular proteinuria29 lost their significance in association with UQ β2MCR. In the longitudinal study, we also observed elevated α1MCR levels after three and six months on TDF exposure, but no obvious change was found in β2MCR after ART, probably due to the decreased level of serum β2M. Nevertheless, in another study with a longer observation time after TDF than in our study, urinary β2M was highly elevated.19 It reminds us that urinary β2M may still be a useful marker after long-term TDF exposure and encourages further study.

Our study provided information about the presence of early renal dysfunction in Chinese HIV-infected patients. Compared with another HIV-infected cohort of African Americans and Caucasians in America, our patients had a lower rate of albuminuria (12.1% versus 28%). A recent study in Germany reported a higher rate of albuminuria (20%), but a lower rate of tubular proteinuria (41%) than that seen in our study (54.4%), as measured by their definition (α1MCR >14mg/g).30 More studies using similar biomarkers and criteria of early renal dysfunction are needed. Additionally, because TDF only became a first-line drug to treat naive patients in China in 2012, future studies are required to determine its long-term toxicity in Chinese patients.

Urinary α1M seems to be a promising biomarker for early detection of TDF-associated tubular dysfunction among HIV-infected MSM. Urinary β2M may help indicate tubular dysfunction, but its clinical interpretation should also take inflammatory conditions into account. More studies will be necessary to address the clinical role of these biomarkers in diagnosing and monitoring TDF-associated tubular injury.

This study was supported by the National Natural Science Foundation of China (81201345).

The authors declare no conflicts of interest.