Kidney tubular abnormalities in the absence of

impaired glomerular function in HIV patients

treated with tenofovir
Pablo Labargaa, Pablo Barreiroa, Luz Martin-Carboneroa, Sonia
Rodriguez-Novoab, Carmen Soleraa, Jose Medranoa, Pablo Rivasa,
Marta Albalaterc, Francisco Blancoa, Victoria Morenoa,
Eugenia Vispoa and Vincent Sorianoa

Background: Tenofovir (TDF) is the most widely prescribed antiretroviral drug. Kidney
abnormalities are the main concern using the drug. As glomerular function is infre-
quently affected in patients treated with TDF, herein, we report the results of an
extensive examination of tubular function.
Methods: Cross-sectional study of plasma and 24 h urine markers of kidney tubulo-
pathy (glucosuria, hyperaminoaciduria, hyperphosphaturia, hyperuricosuria and
b2-microglobulinuria) could be allocated in three groups: patients under a TDF-
containing HAART; patients on HAART never exposed to TDF; and antiretroviral-
naive individuals. Significant tubular damage was defined when at least two of these
parameters were repeatedly present, being at least one part of the Fanconi syndrome
criteria (glucosuria, hyperaminoaciduria and hyperphosphaturia). Glomerular function
was assessed using creatinine clearance.
Results: A total of 284 consecutive HIV patients were examined, 154 on TDF, 49 on other
HAART regimens and 81 drug-naive. No significant differences in creatinine clearance
were observed when comparing distinct groups. The proportion of patients with tubular
damage in groups 1, 2 and 3 were 22, 6 and 12%, respectively. In a multivariate analysis
[odds ratio (OR) {95% confidence interval (CI)} P], the only independent predictors of
tubular dysfunction were TDF use (21.6, 4.1–113, <0.001) and older age (1.1 per year,
1.0–1.1, 0.01).
Conclusion: Exposure to TDF is associated with an increased risk over time of kidney
tubular abnormalities in the absence of significant impaired glomerular function.
Although long-term consequences of this tubulopathy are unknown, close monitoring
of accelerated bone mineral loss and renal insufficiency are warranted. Periodic screening
of tubular function parameters should be recommended to patients receiving TDF.
ß 2009 Wolters Kluwer Health | Lippincott Williams & Wilkins

AIDS 2009, 23:689–696

Keywords: bone, Fanconi syndrome, HIV, kidney, tenofovir

Introduction nucleotide reverse transcriptase inhibitor licensed for

the treatment of HIV infection [1]. The drug is also
Tenofovir disoproxil fumarate (TDF) (Viread; Gilead active against hepatitis B virus (HBV), and recently it has
Sciences, San Francisco, California, USA) is the first granted approval for this indication [2]. Given its

a
Infectious Diseases Department, bPharmacology Unit, Hospital Carlos III, Madrid, and cNephrology Department, Fundación
Jimenez Diaz, Madrid, Spain.
Correspondence to Dr Vincent Soriano, Department of Infectious Diseases, Hospital Carlos III, Calle Sinesio Delgado 10, Madrid
28029, Spain.
Tel: +34 91 4532500; fax: + 34 91 7336614; e-mail: vsoriano@dragonet.es
Received: 12 September 2008; revised: 15 November 2008; accepted: 11 December 2008.

DOI:10.1097/QAD.0b013e3283262a64

ISSN 0269-9370 Q 2009 Wolters Kluwer Health | Lippincott Williams & Wilkins 689
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
690 AIDS 2009, Vol 23 No 6
convenience (one pill once daily), coformulation with
other antiretrovirals (Truvada, Atripla), relatively good
safety profile and robust antiviral activity, TDF is
currently the most widely prescribed antiretroviral agent
in Western countries [3]. Up to the end of year 2007, the
cumulative patient TDF exposure in Europe and North
America was estimated to be 455 392 person-years [4].
The overall good safety profile of TDF has been
confirmed, with only sporadic cases of kidney disease,
which generally have been reported in patients with
predisposing renal illnesses or comorbidities such as
diabetes [5–11].
Given that another two monophosphate nucleoside
analogs, adefovir and cidofovir, licensed, respectively,
for the treatment of hepatitis B and cytomegalovirus
infections, are well known nephrotoxic agents in a dose-
dependent manner [12–14], concerns have recently
arisen regarding the long-term renal safety profile of TDF.
Tenofovir is uptaken at the kidney from the blood by the
organic anion transporter (OAT), located on the surface
of proximal tubular cells [15]. After two additional
phosphorylations, the drug is extruded by the apical
multidrug resistance protein (MRP) transporter into the
tubular space, in which urine is formed. In agreement
with this metabolism, most studies have emphasized that
although patients treated with TDF may rarely experi-
ence any impairment in their glomerular function, as
checked by measuring creatinine plasma concentrations
or creatinine clearance [4,16], tubular damage may be
more common and indeed case reports of kidney
abnormalities on TDF have mainly been reported as
Fanconi-like syndromes, a type of proximal tubular renal
acidosis [5–11]. Given that HIV itself [17] and other
antiretrovirals (e.g., indinavir) show nephrotoxic poten-
tial, there is a need to evaluate in detail renal tubular
parameters in patients treated with TDF, trying to split
out the influence of confounders.
Material and methods
Study population
A prospective observational study was started in June
2007 at the HIV outclinic of Hospital Carlos III, a
reference hospital for infectious diseases located in
Madrid, Spain. All consecutive HIV-infected patients
under regular follow-up were invited to participate in the
study, as long as they could be allocated in any of the
following three groups: patients under HAART currently
receiving TDF for at least 3 months; patients under
HAART never exposed to TDF; and HIV individuals
never exposed to antiretroviral therapy.
Main demographics (age, sex, risk group, weight, chronic
viral hepatitis and history of hypertension or diabetes
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
mellitus) and HIV status (plasma HIV-RNA and CD4þ
T-lymphocyte count) were recorded at the time of
recruitment in the study. Total prior exposure to
antiretroviral drugs was considered as the number of
months of uninterrupted anti-HIV treatment up to
inclusion, both in patients included in group 1 and 2;
periods of drug interruption for any reason, if occurring,
were discounted. This information was obtained by
checking clinical charts and pharmacy records, as all
patients recruited in the study had to obtain antiretroviral
drugs on a monthly basis from the hospital pharmacy.
Treatment was provided without any cost to patients.
The antiretroviral treatment modality, split out as based
on protease inhibitors, nonnucleoside analogs or three
nucleoside/nucleotide analogs, was recorded at the time
of inclusion in the study. It should be noted that TDF was
never given along with didanosine nor with stavudine in
this study, given the increased risk of side-effects [18] and
of selection of K65R [19], respectively using these
combinations. A specific questionnaire was provided to all
patients who accepted participation in the study asking
about the use of drugs with nephrotoxic potential,
particularly acetyl salicylic acid, nonsteroid anti-inflam-
matory drugs, valproic acid, cothrimoxazole, penthami-
dine, amphotericin B, gancyclovir or cidofovir. The study
was approved by the hospital ethical committee, and all
patients gave signed informed consent before inclusion in
the study.
Renal function tests
Within 1 month upon inclusion in the study, the main
kidney glomerular and tubular parameters were deter-
mined in all patients in blood, drawn in fasting conditions
and in 24 h urine. Venous blood biochemistry included
creatinine, sodium, potassium, phosphorus, total calcium,
uric acid, glucose, albumin, total proteins, pH and
bicarbonate. In parallel, creatinine, phosphorus, calcium,
uric acid, glucose, amino acids and b2-microglobulin
were measured in 24 h urine. Creatinine clearance was
calculated using the Dubois formula [(urine creatini-
ne
urine volume)/(plasma creatinine
time)].
Proximal tubular renal dysfunction was determined on
the basis of six criteria: glucosuria (urine glucose
>300 mg daily) with normal glycaemia (plasma glucose
<100 mg/dl); hyperaminoaciduria (any amino acid
in urine, with the exception of hystidine, glycine and
serine); fractional tubular resorption of phosphorus
[1  {(urine phosphorus
plasma creatinine)/(plasma
phosphorus
urine creatinine)}] lower than 0.82; total
excretion of phosphorus (urine phosphorus
urine
volume) greater than 1200 mg daily; fractional excretion
of uric acid [{(urine uric acid
plasma creatinine)/(urine
creatinine
plasma uric acid)}
100] greater than 15%;
and b2-microglobulinuria greater than 1 mg daily [20].
Clinically, meaningful proximal tubular damage was
defined when two or more of these parameters were
 Tenofovir-associated tubular damage Labarga et al. 691

present, being at least one of them any of the Fanconi
syndrome-defining alterations (glucosuria in nondiabetic
patients, hyperaminoaciduria or hyperphosphaturia) [21].
Arterial hypertension was considered in patients with
history of hypertensive disease, regardless of being treated
for this condition, or resting systolic blood pressure
greater than 140 mmHg or diastolic blood pressure
greater than 85 mmHg or both, at the time of recruitment
[22]. Diabetic patients were identified as those with
history of diabetes mellitus, receiving oral antidiabetic
drugs or insulin or having fasting glycaemia greater than
110 mg/dl at recruitment [23].
Determination of urine b2-microglobulin was per-
formed using inmunonephelometry (BN II DADE,
Behring, Barcelona, Spain) and amino acids using high
performance liquid chromatography (HPLC), following
the PicoTag method (empower chromatrography with
2487 detecting ultraviolet radiations), as reported else-
where [20].
Statistical analyses
Continuous values are given as median [interquartile
range (IQR)] and categorical data as percentages. The
three groups of patients analyzed were first compared in a
cross-sectional study using chi-squared for categorical
data and parametric tests for continuous variables.
Afterwards, a time-dependent study was performed
taking into account time under TDF-containing
HAART for group 1 versus time on antiretroviral
treatment that did not include TDF for group 2. A
Table 1. Main baseline characteristics of the study population.
Kaplan–Meier survival analysis was used to assess time
until appearance of proximal tubular damage and
comparisons between patients on HAART with and
without TDF were made using log-rank test. Factors
associated with tubular damage were examined by
univariate and multivariate Cox proportional hazard
regression models. A stepwise selection procedure was
used to assess the relative role of prognostic factors. The
level of significance was 0.05. All statistical analyses were
performed using the SPSS v11 software package (SPSS
Inc., Chicago, Illinois, USA).
Results
A total of 284 HIV patients were included in the study. At
recruitment, 154 were under TDF (group 1), 49 on other
HAART regimens and never exposed to TDF (group 2)
and 81 were antiretroviral-naive (group 3).
Characteristics of patients
Antiretroviral-naive patients were younger than patients
under any HAART regimen, and patients receiving
TDF had larger body weight than the rest (Table 1).
Total duration on antiretroviral drugs was comparable in
patients with and without TDF (59 versus 55 months,
respectively). In group 1, median exposure to TDF
was 36 months. The use of protease inhibitor tended to be
more common in patients receiving TDF than other
HAART regimens (57 versus 36%, respectively;
P ¼ 0.06). Plasma HIV-RNA levels and CD4 cell counts
were comparable between these two groups. The rate of

HAART with TDF HAART, never TDF Drug-naive Differences among

(Group 1) (Group 2) (Group 3) groups

No. of patients (%) 154 (54) 49 (17) 81 (29)

Median age (years) 44 (40–48) 46 (37–54) 37 (31–43) 1 versus 2, P ¼ 0.4
2 versus 3, P < 0.001
3 versus 1, P < 0.001
Male sex (%) 124 (80) 44 (92) 77 (95) P ¼ 0.004
Median body weight (kg) 68 (60–75) 74 (60–94) 71 (64–79) ANOVA, P ¼ 0.8
Median ART exposure (months) 59 (31–106) 55 (35–91) NA P ¼ 0.4
Median TDF exposure (months) 36 (14–48) NA NA NA
PI on HAART (%) 82 (57) 15 (36) NA P ¼ 0.06
Median plasma HIV-RNA (log copies/ml) <1.7 <1.7 4.5 (4.0–4.9) 1 versus 2, P ¼ 0.1
2 versus 3, P < 0.001
3 versus 1, P < 0.001
Median CD4 cell count (cells/ml) 487 (337–720) 572 (360–792) 273 (192–384) 1 versus 2, P ¼ 0.2
2 versus 3, P < 0.001
3 versus 1, P < 0.001
History of hypertension (%) 27 32 23 ANOVA, P ¼ 0.5
History of diabetes (%) 25 29 15 ANOVA, P ¼ 0.1
Use of other nephrotoxic drugs (%) 15 10 16 ANOVA, P ¼ 0.6
Serum HBsAgþ (%) 16 (11) 0 (0) 1 (1) 1 versus 2, P ¼ 0.01
2 versus 3, P ¼ 0.6
3 versus 1, P ¼ 0.01
Serum HCV-RNAþ (%) 52 (38) 3 (7) 13 (16) 1 versus 2, P < 0.001
2 versus 3, P ¼ 0.1
3 versus 1, P ¼ 0.007

Continuous variables are expressed as median (interquartile range). ANOVA, analysis of variance; ART, antiretrovirals; HCV, hepatitis C virus; NA,
not applicable; PIs, protease inhibitors; TDF, tenofovir.

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
692 AIDS 2009, Vol 23 No 6

arterial hypertension or diabetes mellitus was similar in
the three groups. Likewise, the history of exposure to
nephrotoxic drugs was comparable between all groups.
Finally, the prevalence of chronic hepatitis B and C was
greater in patients treated with TDF than in the other
two groups.
under TDF (250 mg/l) and drug-naive ones (235 mg/l)
than in patients on HAART never exposed to TDF
(210 mg/l). Finally, the rate of hypophosphoremia,
nondiabetic glucosuria or aminoaciduria or all was
greater in patients receiving TDF than in the rest,
although without reaching statistical significance.

Renal function Three cases of complete Fanconi syndrome, as defined by

Kidney glomerular filtration, as measured by plasma
creatinine levels or creatinine clearance or both, was
overall within normal limits and comparable among study
groups (Table 2). A total of 42 (14.8%) patients fulfilled
the criteria of significant tubular damage at recruitment,
using the definition previously described. Tubular
dysfunction was more frequent in patients receiving
TDF (22%) as compared with those never treated with
TDF (6%) or never exposed to antiretrovirals (12%). The
difference between drug-naive and patients on HAART
not exposed to TDF did not reach statistical significance.
abnormalities in all measured tubular parameters, were
identified. All were seen in patients receiving TDF. In all
cases, exposure to TDF had been for longer than 2 years.
All showed a low fractional tubular resorption of
phosphorus, ranging from 0.38 to 0.77, along with
severe aminoaciduria and glucosuria. Interestingly,
creatinine clearance was within normal values in all
three patients, ranging from 79 to 118 ml/min. In one
patient, urine abnormalities were accompanied by
acidosis (pH, 7.30), low uric acid (1.8 mg/dl) and low
phosphorus (1.5 mg/dl) in plasma.

Tubular impairment in patients under TDF as compared
with those on HAART never exposed to TDF and drug-
naive patients was mostly reflected by a lower fractional
tubular resorption of phosphorus (0.82, 0.85 and 0.87
units, respectively) and a greater fractional excretion of
uric acid (9, 7 and 7%, respectively). In agreement, plasma
levels of uric acid differed between TDF patients and the
other two groups (5.0, 5.6 and 5.6 mg/dl, respectively).
Interestingly, b2-microglobinuria was greater in patients
Considering the duration of exposure to antiretrovirals,
and comparing patients exposed and nonexposed to TDF,
Fig. 1 shows the incidence of significant tubular
dysfunction over time. Although the study began in
June 2007 recruiting patients prospectively, the Kaplan–
Meir curve shows the incidence of kidney tubular
abnormalities in this population took into consideration
prior to TDF exposure for each individual. After a
median follow-up of 462 and 225 patients-year, those

Table 2. Main baseline renal function parameters in the distinct patient populations.

HAART with TDF HAART, never TDF Drug-naive Differences among

(Group 1) (Group 2) (Group 3) groups

No. of patients (%) 153 (54) 49 (17) 81 (29)

Plasma creatinine (mg/dl) 0.9 (0.8–1.0) 0.9 (0.9–1.1) 0.9 (0.8–1.0) ANOVA, P ¼ 0.8
Creatinine clearance (ml/min) 109 (90–138) 119 (98–142) 123 (86–154) ANOVA, P ¼ 0.1
Plasma phosphorus (mg/dl) 3.4 (2.9–3.7) 3.3 (3.0–3.7) 3.4 (3.1–3.9) ANOVA, P ¼ 0.2
Plasma phosphorus <2.5 mg/dl (%) 13 (9.8) 3 (6.7) 2 (2.6) P ¼ 0.1
Fractional tubular resorption 0.82 (0.75–0.86) 0.85 (0.81–0.88) 0.87 (0.83–0.90) 1 versus 2, P ¼ 0.002
of phosphorus (units)
2 versus 3, P ¼ 0.11
3 versus 1, P < 0.001
Plasma uric acid (mg/dl) 5.0 (4.3–5.8) 5.6 (4.9–6.0) 5.6 (4.9–6.5) 1 versus 2, P ¼ 0.005
2 versus 3, P ¼ 0.5
3 versus 1, P < 0.001
Fractional excretion of uric acid (%) 9 (7–12) 7 (6–8) 7 (6–9) 1 versus 2, P < 0.001
2 versus 3, P ¼ 0.81
3 versus 1, P < 0.001
Nondiabetic glucosuria (%) 2 (2) 0 (0) 1 (2) ANOVA, P ¼ 0.7
b2-microglobulinuria (mg/l) 250 (210–664) 210 (204–219) 235 (208–434) 1 versus 2, P < 0.001
2 versus 3, P ¼ 0.01
3 versus 1, P ¼ 0.3
Aminoaciduria (%) 24 (19) 8 (17) 11(16) P ¼ 0.9
Alkaline phosphatase (IU/l) 242 (200–292) 224 (202–309) 192 (163–226) 1 versus 2, P ¼ 0.4
2 versus 3, P < 0.001
3 versus 1, P < 0.001
Tubular damageM (%) 30 (22) 3 (6) 9 (12) 1 versus 2, P ¼ 0.01
2 versus 3, P ¼ 0.3
3 versus 1, P ¼ 0.06

Continuous variables are expressed as median (interquartile range). MDefinition: At least two parameters of tubular function were altered, being one
of them nondiabetic glucosuria, reduced tubular resorption of phosphorus or pathologic aminoaciduria. ANOVA, analysis of variance; TDF,
tenofovir.

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

100
% of patients free of tubular dysfunction
80
60
40
20
No. of patients
Without tenofovir
With tenofovir
Fig. 1. Development of tubular damageM in patients on HAART, with and without tenofovir (Kaplan–Meier survival analysis).
Definition: At least two parameters of tubular function were altered, being one of them nondiabetic glucosuria, reduced tubular
resorption of phosphorus, or pathologic aminoaciduria.
receiving TDF had significantly greater risk for tubular
damage than patients never treated with TDF [odds ratio
(OR) 18.9; P < 0.001]. According to this analysis,
estimates for tubular dysfunction at 4 years were 25%
for patients on TDF and null for the rest.
Risk factors for kidney tubular damage
Of the 203 patients on HAART, complete data for the
examination of risk factors associated with tubular
damage could be obtained from 183 (90.1%) patients,
of whom 33 belonged to the group with tubular
dysfunction. Univariate analysis [OR, 95% confidence
interval (CI)] showed that treatment with TDF (10.6,
3.0–37.4) or protease inhibitor (2.3, 1.1–4.8) was
significantly associated with tubular damage (Table 3).
Patients with diabetes (1.8, 0.9–3.6), lower CD4 cell
counts (0.9, 0.9–1.1) and exposure to other nephrotoxic
drugs (2.2, 0.9–5.1) tended to have abnormal tubular
parameters. Multivariate analysis (OR, 95% CI, P)
identified TDF use (21.6, 4.1–113, <0.001) and older
age (1.1 per year, 1.0–1.1, 0.01) as the only independent
predictors of tubular damage in this population.
Discussion
Although gastrointestinal disturbances have been the most
frequent side-effects reported in clinical trials that have
assessed the efficacy and safety of TDF [4,24–26], kidney
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Tenofovir-associated tubular damage Labarga et al. 693
Log rank: 18.9 (P < 0.001)
With tenofovir
0 25 50 75 100 125 150 175
Months on follow-up
54 54 54 17 9 5 0 0
164 83 30 2 1 1 0 0
events are the most worrisome [27], especially consider-
ing the experience with other monophosphate nucleo-
side analogs such as adefovir and cidofovir [12–14]. Acute
renal failure and progressive renal insufficiency, however,
have only been reported sporadically in patients treated
with TDF and generally seen in patients with predis-
posing renal illnesses, comorbidities or use of potential
nephrotoxic agents or all [5–11]. It should be noted,
however, that a significant impairment in the renal
function mainly reflects a substantial compromise in the
glomerular filtrate. Primary tubular abnormalities, even
severe, may be missed until they affect the glomerular
function. Moreover, if tubular damage persists for long
periods in the absence of significant renal insufficiency,
clinical and laboratory manifestations other than those
typically associated with kidney failure may develop. This
is the case for premature osteoporosis due to bone mineral
loss. Our study focused on tubular dysfunction in a
relatively large number of HIV patients exposed to TDF
for long periods and controls. With this design, it was
permitted to collect relevant information on the unique
tubular-associated toxicity of TDF.
Our study assessed in parallel glomerular and tubular
parameters of kidney function in HIV individuals,
adjusting for exposure to TDF and other antiretroviral
agents and taking as controls HIV-untreated individuals.
The last group is important as HIV has been shown to be
directly involved in renal damage, particularly in blacks
[17]. The assessment of tubular damage was exhaustive,
694 AIDS 2009, Vol 23 No 6

Table 3. Predictors of renal tubular damage in HIV patients.

Tubular damage Univariate analysis Multivariate analysis

YesM No OR (95% CI) P OR (95% CI) P

No. of patients 33 150

Age (years) 47 (43–52) 44 (39–48) 1.0 (0.9–1.0) 0.26 1.06 (1.0–1.1) 0.01
Male sex (%) 75 85 0.5 (0.2–1.2) 0.18
Body weight (kg) 65 (57–73) 70 (62–79) 0.9 (0.9–1.0) 0.20
History of hypertension (%) 26 29 1.0 (0.4–2.3) 0.93
History of diabetes (%) 50 24 1.8 (0.9–3.6) 0.09
Plasma HIV-RNA (log copies/ml) 1.7 (1.7–1.7) 1.7 (1.7–1.7) 0.1 (0–11.4) 0.34
CD4 cell count (cells/ml) 494 (306–784) 506 (351–712) 0.9 (0.9–1.1) 0.07
Length of ART (months) 76 (39–118) 54 (30–98) 0.9 (0.9–1.1) 0.24
HAART with TDF (%) 91 71 10.6 (3.0–37.4) <0.001 21.6 (4.1–113) <0.001
HAART with PI (%) 64 50 2.3 (1.1–4.8) 0.02
Concomitant nephrotoxic drugs (%) 21 14 2.2 (0.9–5.1) 0.07
Serum HCV-RNA positive (%) 35 29 1.4 (0.6–2.9) 0.35
Serum HBsAg positive (%) 12 6 1.8 (0.6–5.1) 0.28

Continuous variables are expressed as median (interquartile range).

M
Definition: At least two parameters of tubular function were altered, one of them being nondiabetic glucosuria, reduced tubular resorption of
phosphorus or pathologic aminoaciduria.
ART, antiretroviral; HCV, hepatitis C virus; OR, odds ratio; PI, protease inhibitor; TDF, tenofovir.

measuring multiple variables in urine collected during
the last 24 h. In this regard, we are confident that our
estimates of tubular dysfunction are accurate. Although
no signs of renal insufficiency were noticed in the study
population, as reflected by creatinine plasma levels and
creatinine clearance in 24 h urine, up to 22% of HIV
patients treated with TDF showed two or more abnormal
parameters used to assess tubular dysfunction. In contrast,
this observation was found in only 6% of patients treated
with other HAARTregimens and in 12% of antiretroviral-
naive patients.
What may be the clinical relevance of chronic tubulo-
pathy in patients treated with TDF in the absence of a
significant impairment in the glomerular function? At
first glance, it is worth mentioning that a subset of patients
may progress to renal insufficiency as time passes by,
especially because older age reduces the threshold for
renal insufficiency. However, the most important caveat
regards the long-term consequences of chronic loss of
phosphorus in the urine. Premature osteoporosis due to
persistent hypophosphataemia may be a worrisome
complication.

The disconnection we found between the results of
glomerular filtrate and markers of tubular dysfunction
merits further discussion. Our results support an overall
lack of impact of TDF use on creatinine clearance, even
when measured by tools other than the most accurate we
used in 24 h urine (e.g., the Cockcroft–Gault method),
and are in agreement with data from others [28–30]. In
contrast with the numerous longitudinal studies that have
assessed the glomerular function in patients treated with
TDF, very few have been conducted checking prospec-
tively tubular dysfunction [31]. Indeed, most anecdotal
reports of acute or chronic renal failure in patients treated
with TDF have been preceded or accompanied by severe
tubular damage, including Fanconi syndrome [32,33]. In
some cases, TDF-associated tubular damage was sus-
pected following development of suggestive clinical
symptoms or laboratory signs such as metabolic acidosis
with low serum bicarbonate levels, hypophosphataemia
[32] or hypokalaemia [34]. The factor which so far was
still unknown is the true prevalence of TDF-associated
tubulopathy in the absence of any recognizable compro-
mise in the glomerular filtrate. It is noteworthy that our
22% rate being 6% in patients taking other HAART
regimens.
Given the cross-sectional nature of our study, we could
not assess adequately the time-frames for the develop-
ment of TDF-associated tubulopathy. From the literature,
it seems that some individuals are particularly susceptible,
and genetic predisposition may account for it [35–37]. In
our series, the only patient presenting with a complete
Fanconi syndrome with consequences recognizable in the
blood had been exposed to TDF for 37 months. In other
similar cases reported in the literature [5–11,32,33], the
lapsed time ranged between 6 and 18 months.
Comorbidities such as diabetes, concomitant adminis-
tration of nephrotoxic agents and use of protease
inhibitors have all been associated with an increased risk
of TDF-associated kidney disease. These conditions are
relatively common in HIV patients. However, in our
study, the multivariate analysis suggested that their role as
cause of tubulopathy was only marginal. A review from
the Food and Drug Administration noticed that some
cases of Fanconi syndrome in patients treated with TDF
occurred when using concomitantly either didanosine or
protease inhibitors or both, which could had enhanced
the nephrotoxicity of TDF [38]. In our series, no patient
was treated with didanosine because prior concerns about

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

CD4þ T lymphopenia [18]. Although protease inhibitors
were used by 57% of our patients treated with TDF and
their potential for increasing TDF-tubular damage has
already been claimed by others [39], we could not find a
significant association between their use and tubular
damage in the multivariate analysis. Therefore, their role
must be low, if any.
An interesting finding in our study was the recognition of
an increased rate of tubular dysfunction parameters in
antiretroviral-naive patients compared with those on
HAARTwho did not receive TDF. As b2 –microglobulin
concentrations in urine are a sensitive marker of tubular
damage, often preceding other signs of tubular injury
[40,41], we concentrated on this marker. Up to 9.4% of
drug-naive patients in our series showed a high urinary
excretion of b2 –microglobulin. This finding indirectly
supports that HIV itself might cause tubular damage to
some extent, as has been suggested by others [42–44].
The fact that our patients on HAART including TDF,
conversely, showed the greatest levels of b2 –microglo-
bulin in urine reinforces the unique potential for tubular
toxicity of TDF. Indeed, some authors have defended
monitoring urine b2 –microglobulin in patients on TDF
in an attempt to identify the subset of individuals at
increased risk for renal injury [45].
Conclusion
Our results confirm that exposure to TDF is associated
with signs of proximal tubular kidney damage, often in
the absence of any significant impairment of the
glomerular function, in a substantial proportion of HIV
patients. Although the clinical long-term consequences
of this observation are currently unknown, it seems
worth while to periodically monitor for signs of tubular
damage in HIV patients on TDF and to consider a
switch in therapy if signs of progressive deterioration are
manifest. Our findings in HIV patients should be
viewed as relevant also for chronic hepatitis B, as most
of these patients will need long periods (often years)
of treatment [2]. A periodic monitoring of tubular
parameters would be warranted in this population
as well.
Acknowledgements
This work was funded in part by grants from Fundación
Investigación y Educación en SIDA (IES), the European
NEAT consortium, Agencia Lain Entralgo and Red de
Investigación en SIDA (RIS, ISCIII-RETIC RD06). We
would like to thank Professor Juan Gonzalez-Lahoz for
comments and continuous support.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Tenofovir-associated tubular damage Labarga et al. 695
P.L., M.A. and V.S. designed the study. P.L., C.S., P.R.,
L.M.-C., P.B., F.B., V.M., E.V., S. R.-N. and V.S. enrolled
patients in the study and collected data. P.L., J.M., P.B.
and L.M.-C. did the statistical analyses and interpret-
ations. P.L. and V.S. wrote the first draft of the manuscript.
References
1. Hammer S, Eron J, Reiss P, Schooley R, Thompson M, Walmsley
S, et al. Antiretroviral treatment of adult HIV infection: 2008
recommendations of the International AIDS Society-USA panel.
JAMA 2008; 300:555–570.
2. Belongia E, Costa J, Gareen I, Grem J, Inadomi J, Kern E, et al.
NIH consensus development statement on management of
hepatitis B: draft. NIH Consens State Sci Statements (in press)
3. Jiménez-Nácher I, Garcı́a B, Barreiro P, Rodriguez-Novoa S,
Morello J, González-Lahoz J, et al. Trends in the prescription of
antiretroviral drugs and impact on plasma HIV-RNA measure-
ments. J Antimicrob Chemother 2008; 62:816–822.
4. Nelson M, Katlama C, Montaner J, Cooper D, Gazzard B, Clotet
B, et al. The safety of tenofovir disoproxil fumarate for the
treatment of HIV infection in adults: the first 4 years. AIDS
2007; 21:1273–1281.
5. Quimby D, Brito M. Fanconi syndrome associated with use of
tenofovir in HIV-infected patients: a case report and review of
the literature. AIDS Read 2005; 15:357–364.
6. Malik A, Abraham P, Malik N. Acute renal failure and
Fanconi syndrome in an AIDS patient on tenofovir treat-
ment: case report and review of literature. J Infect 2005; 51:E61–
E65.
7. Rifkin B, Perazella M. Tenofovir-associated nephrotoxicity:
Fanconi syndrome and renal failure. Am J Med 2004; 117:
282–284.
8. Gaspar G, Monereo A, Garcia-Reyne A, de Guzman M. Fanconi
syndrome and acute renal failure in a patient treated with
tenofovir: a call for caution. AIDS 2004; 18:351–352.
9. Karras A, Lafaurie M, Furco A, Bourgarit A, Droz D, Sereni D,
et al. Tenofovir-related nephrotoxicity in HIV-infected pa-
tients: three cases of renal failure, Fanconi syndrome, and
nephrogenic diabetes insipidus. Clin Infect Dis 2003; 36:1070–
1073.
10. Barrios A, Garcia-Benayas T, Gonzalez-Lahoz J, Soriano V.
Tenofovir-related nephrotoxicity in HIV-infected patients.
AIDS 2004; 18:960–963.
11. Peyriere H, Reynes J, Rouanet I, Daniel N, de Boever C,
Mauboussin J, et al. Renal tubular dysfunction associated with
tenofovir therapy: report of 7 cases. J Acquir Immune Defic
Syndr 2004; 35:269–273.
12. Tanji N, Tanji K, Kambham N, Markowitz G, Bell A, D’Agati V.
Adefovir nephrotoxicity: possible role of mitochondrial DNA
depletion. Hum Pathol 2001; 32:734–740.
13. Meier P, Dautheville-Guibal S, Ronco P, Rossert J. Cidofovir-
induced end-stage renal failure. Nephrol Dial Transplant 2002;
17:148–149.
14. Cihlar T, Lin D, Pritchard J, Fuller M, Mendel D, Sweet D. The
antiviral nucleotide analogs cidofovir and adefovir are novel
substrates for human and rat renal organic anion transporter 1.
Mol Pharmacol 1999; 56:570–580.
15. Ray A, Cihlar T, Robinson K, Tong L, Vela J, Fuller M, et al.
Mechanism of active renal tubular efflux of tenofovir. Anti-
microb Agents Chemother 2006; 50:3297–3304.
16. Javanbakht M, Khanlou H, Chien C, Denouden P, Voskanian A,
Kaur P, et al. Renal function changes at two years in
HIV-infected patients treated with tenofovir-DF [abstract
#MOPEB064]. In: 4th IAS Conference on HIV Pathogenesis,
Treatment and Prevention; 22–25 July 2007. Sydney, Australia.
17. Atta M, Lucas G, Fine D. HIV-associated nephropathy: epide-
miology, pathogenesis, diagnosis and management. Expert Rev
Anti Infect Ther 2008; 6:365–371.
18. Barreiro P, Soriano V. Suboptimal CD4 gains in HIV-infected
patients receiving didanosine plus tenofovir. J Antimicrob
Chemother 2006; 57:806–809.
696 AIDS 2009, Vol 23 No 6
19. Miller MD. K65R, TAMs and tenofovir. AIDS Rev 2004; 6:22–
33.
20. Kinai E, Hanabusa H. Renal tubular toxicity associated with
tenofovir assessed using urine-beta 2 microglobulin, percen-
tage of tubular reabsorption of phosphate and alkaline phos-
phatase levels. AIDS 2005; 19:2031–2033.
21. Izzedine H, Launay-Vacher V, Isnard-Bagnis C, Deray G. Drug-
induced Fanconi’s syndrome. Am J Kidney Dis 2003; 41:292–
309.
22. Khan N, Hemmelgarn B, Herman R, Rabkin S, McAlister F, Bell
C, et al. The 2008 Canadian Hypertension Education Program
recommendations for the management of hypertension: Part 2:
therapy. Can J Cardiol 2008; 24:465–475.
23. Bhattacharyya O, Shah B, Booth G. Management of cardiovas-
cular disease in patients with diabetes: the 2008 Canadian
Diabetes Association guidelines. CMAJ 2008; 179:920–926.
24. Gallant J, Parish M, Keruly J, Moore R. Changes in renal function
associated with tenofovir disoproxil fumarate treatment, com-
pared with nucleoside reverse transcriptase inhibitor treat-
ment. Clin Infect Dis 2005; 40:1194–1198.
25. Izzedine H, Hulot J, Vittecoq D, Gallant J, Staszewski S, Launay-
Vacher V, et al. Long-term renal safety of tenofovir disoproxil
fumarate in antiretroviral-naive HIV-1-infected patients. Data
from a double-blind randomized active-controlled multicentre
study. Nephrol Dial Transplant 2005; 20:743–746.
26. Arribas J, Pozniak A, Gallant J, DeJesus E, Gazzard B, Campo R,
et al. Tenofovir disoproxil fumarate, emtricitabine, and efavir-
enz compared with zidovudine/lamivudine and efavirenz in
treatment-naive patients: 144-week analysis. J Acquir Immune
Defic Syndr 2008; 47:74–78.
27. Izzedine H, Isnard-Bagnis C, Hulot J, Vittecoq D, Cheng A, Jais
C, et al. Renal safety of tenofovir in HIV treatment-experienced
patients. AIDS 2004; 18:1074–1076.
28. Vrouenraets S, Fernandez Garcia E, Wit F, Brinkman K, Hoek F,
Krediet R, et al. A comparison between different GFR-estima-
tions and [125I]-iothalamate, the gold standard for GFR-mea-
surement, in HIV-infected patients on HAART [abstract
#977b]. In: 15th Conference on Retroviruses and Opportunistic
Infections; 3–6 February 2008. Boston, MA.
29. Madruga J, Cassetti I, Koenig E, Etzel A, Zhou Y, Cheng A, et al.
Six year safety and efficacy of tenofovir DF in combination
with lamivudine and efavirenz in antiretroviral-naı̈ve patients
[abstract #WEPEB030]. In: 4th IAS Conference on HIV Patho-
genesis, Treatment and Prevention; 22–25 July 2007. Sydney,
Australia.
30. Gerard L, Chazallon C, Taburet A, Girard P, Aboulker JP, Piketty
C. Renal function in antiretroviral-experienced patients treated
with tenofovir disoproxil fumarate associated with atazanavir/
ritonavir. Antivir Ther 2007; 12:31–39.
31. Fux C, Christen A, Zgraggen S, Mohaupt M, Furrer H. Effect of
tenofovir on renal glomerular and tubular function. AIDS 2007;
21:1483–1485.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
32. Breton G, Alexandre M, Duval X, Prie D, Peytavin G, Leport C,
et al. Tubulopathy consecutive to tenofovir-containing antire-
troviral therapy in two patients infected with HIV-1. Scand J
Infect Dis 2004; 36:527–528.
33. Earle K, Seneviratne T, Shaker J, Shoback D. Fanconi’s syn-
drome in HIVR adults: report of three cases and literature
review. J Bone Miner Res 2004; 19:714–721.
34. Shepp D, Curtis S, Rooney J. Causes and consequences of
hypokalemia in patients on tenofovir disoproxil fumarate. AIDS
2007; 21:1479–1481.
35. Izzedine H, Hulot J, Villard E, Goyenvalle C, Dominguez S,
Ghosn J, et al. Association between ABCC2 gene haplotypes
and tenofovir-induced proximal tubulopathy. J Infect Dis 2006;
194:1481–1491.
36. Suzuki H, Sugiyama Y. Single nucleotide polymorphisms in
multidrug resistance associated protein 2 (MRP2/ABCC2): its
impact on drug disposition. Adv Drug Deliv Rev 2002;
54:1311–1331.
37. Bleasby K, Hall L, Perry J, Mohrenweiser H, Pritchard J. Func-
tional consequences of single nucleotide polymorphisms in the
human organic anion transporter hOAT1 (SLC22A6). J Phar-
macol Exp Ther 2005; 314:923–931.
38. Gupta S. Tenofovir-associated Fanconi syndrome: review of the
FDA adverse event reporting system. AIDS Patient Care STDS
2008; 22:99–103.
39. Goicoechea M, Liu S, Best B, Sun S, Jain S, Kemper C, et al.
Greater tenofovir-associated renal function decline with pro-
tease inhibitor-based versus nonnucleoside reverse-transcrip-
tase inhibitor-based therapy. J Infect Dis 2008; 197:102–108.
40. Rossi R, Pleyer J, Schafers P, Kuhn N, Kleta R, Deufel T, et al.
Development of ifosfamide-induced nephrotoxicity: prospec-
tive follow-up in 75 patients. Med Pediatr Oncol 1999; 32:177–
182.
41. Watanabe T, Yoshikawa H, Yamazaki S, Abe Y, Abe T. Sec-
ondary renal Fanconi syndrome caused by valproate therapy.
Pediatr Nephrol 2005; 20:814–817.
42. Kabanda A, Vandercam B, Bernard A, Lauwerys R, van Ypersele
S. Low molecular weight proteinuria in HIV-infected patients.
Am J Kidney Dis 1996; 27:803–808.
43. Kimmel P, Umana W, Bosch J. Abnormal urinary protein
excretion in HIV-infected patients. Clin Nephrol 1993; 39:
17–21.
44. Horberg M, Tang B, Towner W, Bersoff-Matcha S, Klein D,
Silverberg M, et al. Effect of tenofovir on renal function in
patients using HAART [abstract #975]. In: 15th Conference on
Retroviruses and Opportunistic Infections; 3–6 February 2008.
Boston, MA.
45. Gatanaga H, Tachikawa N, Kikuchi Y, Teruya K, Genka I,
Honda M, et al. Urinary beta2-microglobulin as a possible
sensitive marker for renal injury caused by tenofovir disoproxil
fumarate. AIDS Res Hum Retroviruses 2006; 22:744–748.