Proximal tubular dysfunction in pregnant women receiving tenofovir disoproxil fumarate to prevent mother-to-child transmission of hepatitis B virus

Abstract

Background

Data evaluating the risk of proximal tubular dysfunction in women receiving tenofovir disoproxil fumarate for the prevention of mother-to-child transmission (PMTCT) of HBV are scarce.

Objectives

To assess the risk of proximal tubulopathy in pregnant women receiving tenofovir disoproxil fumarate for PMTCT of HBV.

Patients and methods

We used urine samples collected from HBV monoinfected pregnant women who participated in a Phase III, multicentre, randomized, double-blind, placebo-controlled clinical trial assessing a tenofovir disoproxil fumarate short course from 28 weeks gestational age (28-wk-GA) to 2 months post-partum (2-months-PP) for PMTCT of HBV in Thailand. Markers of tubular dysfunction, including retinol binding protein, kidney injury molecule-1, α1-microglobuin and β2-microglobulin, were assayed at 28- and 32-wk-GA and 2-months-PP visits. Proximal tubulopathy was defined as the presence of ≥2 of the following: tubular proteinuria, euglycaemic glycosuria and increased urinary phosphate.

Results

A total of 291 women participated in the study. No kidney-related adverse events were severe, and none led to tenofovir disoproxil fumarate discontinuation. At 2-months-PP, 3 of the 120 (3%) evaluated women in the tenofovir disoproxil fumarate group experienced proximal tubulopathy versus 3 of 125 (2%) in the placebo group (P = 1.00). None of the six women met the criteria for proximal tubulopathy at 12-months-PP but proteinuria persisted in three of them. No growth abnormalities were found at 1 year of age in infants born to mothers with proximal tubulopathy at 2-months-PP.

Conclusions

In these HBV-infected pregnant and breastfeeding women, tenofovir disoproxil fumarate administered from 28-wk-GA to 2-months-PP was not associated with a higher risk of proximal tubulopathy.

Introduction

In 2015, the WHO estimated that about 257 million people were living with chronic HBV infection in the world.¹ Most new HBV infections are due to mother-to-child transmission. In the absence of intervention, HBV-infected mothers with high HBV DNA loads, usually positive for hepatitis e antigen (HBeAg), have up to 85% risk of transmitting the virus to their infant and 80%–90% of infants infected during the first year of life develop chronic HBV infection. Immunoprophylaxis with hepatitis B specific immunoglobulin (HBIg) at birth and hepatitis B immunization starting at birth in infants born to HBV-infected mothers reduces the risk of mother-to-child transmission of HBV. However, it is considered that up to 10% of infants born to hepatitis B surface antigen (HBsAg)- and HBeAg-positive mothers are infected with HBV despite this immunoprophylaxis.² In addition to infant immunoprophylaxis, the 2020 WHO guidelines recommend that women with high HBV DNA load receive tenofovir disoproxil fumarate during the third trimester of pregnancy to prevent the risk of mother-to-child transmission.³

It has been estimated that 4.5 million women with chronic HBV infection give birth annually.⁴ Taking into account the worldwide distribution of HBV genotypes and hypothesizing that about 20% of pregnant women have high viral replication, about 1 million women may be eligible for tenofovir disoproxil fumarate prophylaxis each year. The occurrence of renal and bone side effects are the most severe complications related to tenofovir disoproxil fumarate use. Accumulation of tenofovir in renal proximal tubular cells may cause mitochondrial toxicity leading to renal impairment ranging from asymptomatic proximal tubular dysfunction to acute kidney injury with Fanconi syndrome.⁵ Fanconi syndrome has been rarely described in HBV-infected patients receiving tenofovir disoproxil fumarate but subclinical proximal tubular dysfunction is more frequent and has occurred in 20%–50% of HBV-infected patients on tenofovir disoproxil fumarate.^6–8 Tenofovir disoproxil fumarate tubulopathy is characterized by a chronic loss of phosphorus in urine and impaired vitamin D activation, which may lead to an altered equilibrium in calcium and phosphorus metabolism with progressive bone mineral density (BMD) decrease. This issue is of particular importance during pregnancy and the breastfeeding period, when mothers have to supply large amounts of calcium to ensure infant bone growth. This process can lead to a mean BMD reduction from 3% to 10% after 3–6 months of lactation.⁹ Tubular dysfunction related to tenofovir disoproxil fumarate may accelerate physiological bone turnover in mothers and modify the composition of breast milk, which may in turn impact the infant BMD. Investigating the risk of proximal tubular dysfunction during pregnancy and the post-partum period is therefore essential. In clinical trials for the prevention of mother-to-child transmission (PMTCT) of HBV,^10–13 no major kidney toxicities have been reported in women and their infants exposed to tenofovir disoproxil fumarate. However, in these studies, biological markers were not specific for the detection of tenofovir disoproxil fumarate-associated tubulopathy. Albuminuria, which reflects glomerular injury rather than proximal tubular dysfunction, is rarely reported in patients with kidney biopsy-proven tenofovir injury, making the use of urine dipsticks not useful to detect tenofovir disoproxil fumarate nephrotoxicity.¹⁴ Creatinine plasma levels may also remain within normal range despite substantial proximal tubular injury.¹⁵ Therefore, the routine investigation tools for renal safety monitoring fail to detect tubular dysfunction at an early stage. In recent years, several low-molecular-weight proteins in urine have been used as next-generation biomarkers for detecting kidney toxicity.¹⁶ These early kidney injury markers have been shown to be effective in detecting tubular dysfunction related to tenofovir disoproxil fumarate.¹⁷ Although tenofovir disoproxil fumarate has been extensively used to prevent mother-to-child transmission of HIV, and more recently for HIV pre-exposure prophylaxis (PrEP) and PMTCT of HBV, the risk of tenofovir disoproxil fumarate-associated tubular dysfunction in pregnant and breastfeeding women has not been assessed using these appropriate markers. The goals of this study were: (1) to evaluate whether a short course of tenofovir disoproxil fumarate given for PMTCT of HBV was associated with an increased risk of proximal tubulopathy during pregnancy and breastfeeding, using a sensitive and early kidney injury marker; and (2) to evaluate maternal BMD and infant bone development in women who experienced proximal tubulopathy.

Patients and methods

Study design and participants

We conducted a secondary analysis of the iTAP study, a Phase 3, multicentre, randomized, double-blind, placebo-controlled clinical trial assessing the efficacy and safety of tenofovir disoproxil fumarate 300 mg given daily from 28 weeks gestational age (28-wk-GA) to 2 months postpartum (2-months-PP) for HBV PMTCT in 17 public hospitals in Thailand (ClinicalTrials.gov identifier NCT01745822).¹³ Details of the protocol and study results have been published previously.^13,18 In short, it enrolled 331 pregnant women (168 on tenofovir disoproxil fumarate; 163 on placebo) 18 years of age or older, with positive HBsAg and HBeAg tests, an ALT level of ≤30 IU/L at first screening, no HIV and/or hepatitis C infection, no history of tenofovir disoproxil fumarate treatment at any time and no anti-HBV treatment during the current pregnancy, CL_CR > 50 mL/min (according to the Cockcroft–Gault formula), no proteinuria (>30 mg/dL), no normoglycaemic glycosuria, and no evidence of a fetal anomaly incompatible with life. Infants were followed until 12 months of age. Maternal serum creatinine and phosphorus were measured at 28-, 32- and 36-wk-GA, delivery, 1- and 2-months-PP and at birth in newborns. After ethical clearance of a first protocol amendment in April 2013, maternal urine samples were collected from consenting participants at 28-wk-GA, 32-wk-GA, 2-months-PP and 12-months-PP and kept stored frozen for the analysis of urine biomarkers of proximal tubulopathy. After a second protocol amendment in July 2015, mothers and infants who had not reached 12 months after delivery/birth were invited to participate in a BMD assessment at this timepoint (±1.5 months).¹⁹ Tenofovir plasma concentrations were measured at delivery in the tenofovir disoproxil fumarate group using a validated LC-triple quadrupole MS assay.²⁰

Definition of proximal tubulopathy

Proximal tubulopathy was defined as ≥ 2 of the following findings in urine collected at a given timepoint: (1) tubular proteinuria, defined as a protein/creatinine ratio in urine of >30 mg/mmol with a urine albumin/total protein ratio of <0.4 and/or a high concentration of at least one of the following urinary biomarkers: retinol binding protein (RBP)/urine creatinine ratio of >130 μg/g, kidney injury molecule-1 (KIM-1)/urine creatinine ratio of >1.58 μg/g, urine β2-microglobulin (β2M)/urine creatinine ratio of >300 μg/g or α1-microglobulin (α1M)/urine creatinine ratio of >15 mg/g; (2) euglycaemic glycosuria, defined as a positive urine glucose level (≥10 mg/dL); or (3) increased urinary excretion of phosphorus, defined as a fractional tubular resorption of phosphorus of <82% and/or a fractional tubular reabsorption of phosphate of <95% with serum phosphate of <2.6 mg/dL and/or a tubular maximum phosphate reabsorption (TmP) to glomerular filtration rate (GFR) ratio of <0.8 mmol/L.²¹ All these biomarkers have been previously used for early detection of proximal tubular dysfunction associated with tenofovir disoproxil fumarate.¹⁷ We used cut-off values based on previous studies.^22–25 We analysed all urine samples collected at 28-wk-GA and at 2-months-PP. In addition, we randomly selected 91 urine samples per group at 32-wk-GA to assess early tubulopathy after treatment initiation. Women with tubulopathy at 2-months-PP also had their urine samples at 32-wk-GA and 12-months-PP analysed to assess the earliness and reversibility of tubulopathy.

Laboratory analyses

Serum creatinine and phosphorus were measured in clinical site laboratories. Urine specimens were stored at −80°C until biomarker measurement. Creatininuria, glycosuria and phosphaturia were measured at the Division of Clinical Chemistry, Faculty of Associated Medical Science, Chiang Mai University, using the clinical chemistry and turbidimetry analyser BioSystems BA400 (Barcelona, Spain). The levels of RBP and α1M were measured at the study central laboratory using an α1M and RBP ELISA kit (Abcam, USA). The two panels of Bio-Plex RBM kidney toxicity assays (Bio-Rad, USA) were used to measure β2M and KIM–1. All laboratory staff performing the tests were blind to participants’ treatment group and clinical information.

Statistical analysis

The distribution of baseline characteristics in all patients participating in the iTAP study was already described in the original publication. The distribution of maternal characteristics included in the present study and of those excluded were compared using the Wilcoxon rank-sum test for continuous variables and Fisher’s exact test for categorical and discrete variables. At 2-months-PP, the primary analysis was performed in samples collected within a window of up to 14 days after treatment discontinuation and a sensitivity analysis was performed using a narrower window of up to 7 days. The windows for the other timepoints are specified in the Supplementary information, available as Supplementary data at JAC Online. Proportions of women with tubulopathy at 2-months-PP were compared between treatment groups using Fisher’s exact test. The distribution of characteristics of infants born to a mother with tubulopathy at 2-months-PP were compared with those of infants born to mothers without tubulopathy at 2-months-PP using the Wilcoxon rank-sum test. All P values were two-sided. All analyses were conducted using Stata software version 14.1.

Ethics

The study was approved by the Ethics Committee of the Faculty of Associated Medical Sciences, Chiang Mai University, Thailand (No. 367/2018). All women provided written informed consent.

Results

Participants’ baseline characteristics

Figure 1 summarizes the disposition of women and infants. One hundred and forty-eight consenting women in the tenofovir disoproxil fumarate group and 143 in the placebo group with a urine sample available at 28-wk-GA were included in the study. Characteristics of these women are described in Table 1. No characteristics were significantly different between women included and those excluded from this study, except for serum creatinine (median serum creatinine of 46.9 and 44.2 μmol/L, respectively). The median duration from enrolment to delivery was 10.7 weeks (IQR 10.0–11.6). One case of pre-eclampsia occurred in the tenofovir disoproxil fumarate arm and one in the placebo arm, but none was associated with proximal tubulopathy at any study timepoints. At 2-months-PP, 130 women in the tenofovir disoproxil fumarate group and 133 in the placebo group had a urine sample available. Proximal tubulopathy was observed in 28 of 128 (22%) women in the tenofovir disoproxil fumarate group and 24 of 129 (19%) in the placebo group (P = 0.54) at 28-wk-GA (Tables 2 and 3).

Figure 1.

Flow chart of women enrolled in the study. TDF, tenofovir disoproxil fumarate.

Table 1.

Women’s characteristics according to treatment group and participation in this substudy

Characteristics	Participated in the substudy (N = 291)			Did not participate in the substudy (N = 40)		P valuea (participated versus did not participate)
	TDF	Placebo	P valuea	TDF	Placebo
At 28-wk-GA	N = 148	N = 143		N = 20	N = 20
age (years), median (IQR)	25.8 (22.7–29.2)	26.7 (23.3–30.4)	0.14	23.5 (21.1–27.7)	29.6 (24.7–32.7)	0.95
gestational age (weeks), median (IQR)	28.3 (27.9–28.6)	28.1 (27.9–28.6)	0.29	28.1 (27.9–28.6)	28.6 (28.1–28.8)	0.05
weight (kg), median (IQR)	61.8 (56.0–70.4)	60.5 (54.0–68.0)	0.32	61.8 (55.3–73.0)	60.0 (55.6–74.0)	0.54
height (cm), median (IQR)	158 (154–162)	157 (152–160)	0.11	160 (153–163)	156 (151–158)	0.83
BMI (kg/m2), median (IQR)	24.8 (22.7–27.7)	24.6 (22.7–27.9)	0.79	24.8 (22.3–28.4)	25.6 (22.7–30.6)	0.49
systolic blood pressure (mmHg), median (IQR)	110 (100–118) (N = 99)	109 (100–117) (N = 94)	0.65	109 (100–120) (N = 18)	108 (102–110)	0.92
diastolic blood pressure (mmHg), median (IQR)	65 (60–71) (N = 99)	67 (60–70) (N = 94)	0.73	67 (60–70) (N = 18)	65 (60–70)	0.98
HBV DNA (log10 IU/mL), median (IQR)	8.1 (7.3–8.5)	8.0 (6.9–8.5)	0.38	8.1 (6.4–8.4)	7.5 (5.7–8.1)	0.06
HBV DNA > 200 000 IU/mL, n (%)	135 (91)	127 (89)	0.56	17 (85)	15 (75)	0.10
serum creatinine (μmol/L), median (IQR)	46.9 (42.4–52.6)	46.0 (40.7–53.0)	0.47	43.8 (37.6–44.2)	44.2 (38.9–46.4)	0.008
serum phosphorus (mg/dL), median (IQR)	3.4 (3.0–3.8) (N = 147)	3.5 (3.1 to 3.9) (N = 142)	0.26	3.4 (3.3–3.8) (N = 19)	3.5 (3.1–3.8)	0.81
Urine dipsticks, n (%)
glycosuria ≥2+	1 (1)	2/142 (1)	0.62	1 (5)	0	0.41
proteinuria ≥2+	0	1/142 (1)	0.49	0	0	1.00
At delivery	N = 142	N = 140		N = 20	N = 20
gestational age (weeks), median (IQR)	38.9 (38.3–39.7)	39.0 (38.1 to 40.0)	0.64	39.0 (38.3–39.5)	38.4 (37.9–39.4)	0.16
stillbirth antepartum, n (%)	0	0	1.00	1 (5)	0	0.12
prematurity (<37 weeks), n (%)	4 (3)	7 (5)	0.38	1 (5)	3 (15)	0.10
tenofovir plasma concentration (ng/mL), median (IQR)	57 (40–88)	—		65 (49–136)	—	0.19
tenofovir undetectable, n (%)	12 (8)	—		1 (5)	—	1.00
After delivery	N = 142	N = 140		N = 20	N = 20
breastfeeding, n (%)	136 (96)	134 (96)	1.00	19 (95)	20 (100)	1.00
duration of breastfeeding (months), median (IQR)	6.1 (3.0–12.0) (N = 136)	6.2 (4.0–12.0) (N = 134)	0.22	6.7 (4.1–11.3) (N = 19)	6.5 (3.2–12.1)	0.97

TDF, tenofovir disoproxil fumarate.

^aP values from the Wilcoxon rank-sum test or Fisher’s exact test.

Table 2.

Markers of tubular dysfunction by treatment group and visit

Markers of renal tubular function, median (IQR)	At 28-wk-GA			At 2-months-PP
	TDF	Placebo	P valuea	TDF	Placebo	P valuea
	N = 128	N = 129		N = 120	N = 125
Serum creatinine (μmol/L)	46.0 (42.4–52.6)	46.9 (40.7–53.0)	0.77	62.3 (56.6–72.5)	61.9 (54.8–67.2)	0.15
Serum phosphate (mg/dL)	3.4 (3.1–3.8)	3.5 (3.1–3.9)	0.32	3.9 (3.4–4.4)	3.9 (3.4–4.2)	0.83
Urine creatinine (mmol/L)	8.5 (5.1–12.5)	9.2 (4.8–13.9)	0.80	11.8 (8.2–16.5)	13.4 (7.3–19.8)	0.32
Urine phosphate (mg/dL)	29.9 (17.7–57.8)	27.9 (14.8–58.5)	0.75	37.9 (22.5–62.9)	47.6 (25.6–73.9)	0.08
Urine glucose (mg/dL)	5 (3–10)	5 (3–9)	0.47	4 (3–7)	5 (3–8)	0.06
Urine total protein to creatinine ratio (mg/mmol)	5.8 (4.2–8.2)	5.7 (4.2–8.8)	0.72	3.3 (2.2–5.2)	3.1 (2.1–5.0)	0.42
Urine albumin to creatinine ratio (mg/mmol)	0.63 (0.44–0.88)	0.60 (0.48–0.91)	0.59	0.67 (0.41–1.17)	0.69 (0.42–1.10)	0.93
Urine KIM-1 to creatinine ratio (ng/mmol)	20.4 (12.6–31.0)	21.4 (12.3–36.4)	0.38	25.8 (15.2–40.3)	28.2 (15.6–40.7)	0.56
Urine β2M to creatinine ratio (μg/mmol)	26.9 (13.6–59.3)	24.1 (14.0–52.6)	0.53	4.8 (3.7–7.9)	4.7 (3.5–6.7)	0.45
Urine RBP to creatinine ratio (μg/mmol)	25.4 (17.1–38.7)	22.0 (15.1–37.1)	0.22	9.6 (5.8–12.8)	8.0 (5.8–12.7)	0.30
Urine α1M to creatinine ratio (mg/mmol)	1.60 (1.03–2.35)	1.65 (0.90–2.63)	0.87	0.59 (0.37–1.13)	0.59 (0.31–1.10)	0.36

^aWilcoxon rank-sum test.

Table 3.

Frequencies and percentages of women with tubular dysfunction at each visit

Markers of renal tubular function, n (%)	28-wk-GA		32-wk-GA			2-months-PP
	TDF	Placebo	TDF	Placebo	P valuea	TDF	Placebo	P valuea
	N = 128	N = 129	N = 91	N = 93		N = 120	N = 125
Alteration of phosphate handling	3 (2)	4 (3)	6 (7)	3 (3)	0.33	1 (1)	0	0.49
fractional tubular reabsorption of phosphate <82%	1 (1)	0	1 (1)	1 (1)	1.00	0	0	1.00
tubular maximum phosphate reabsorption to eGFR ratio <0.8 mmol/L	2 (2)	2 (2)	4 (4)	2 (2)	0.44	1 (1)	0	0.49
fractional tubular reabsorption of phosphate <95% and serum phosphate <2.6 mg/dL	1 (1)	3 (2)	4 (4)	0	0.06	0	0	1.00
Alteration of glucose resorption	32 (25)	30 (23)	34 (37)	20 (22)	0.02	7 (6)	19 (15)	0.02
urine glucose ≥10 mg/dL	32 (25)	30 (23)	34 (37)	20 (22)	0.02	7 (6)	19 (15)	0.02
Tubular proteinuria	109 (85)	107 (83)	85 (93)	83 (89)	0.43	33 (28)	36 (29)	0.89
urine total protein to creatinine ratio >30 mg/mmol and urine albumin to total protein ratio <0.4 mg/mg	1 (1)	0	2 (2)	1 (1)	0.62	2 (2)	1 (1)	0.62
urine RBP/creatinine >130 μg/g	103 (80)	100 (78)	84 (92)	81 (87)	0.33	25 (21)	25 (20)	0.88
urine KIM-1/creatinine >1.58 μg/g	0	2 (2)	0	0	1.00	0	0	1.00
urine β2M/creatinine >300 μg/g	52 (41)	55 (43)	70 (77)	52 (56)	0.003	4 (3)	3 (2)	0.72
urine α1M/creatinine >15 mg/g	59 (46)	62 (48)	52 (57)	40 (43)	0.08	17 (14)	16 (13)	0.85
≥2 tubular dysfunctions	28 (22)	24 (19)	32 (35)	20 (22)	0.05	3 (3)	3 (2)	1.00
alteration of phosphate handling and glucose resorption	0	2 (2)	5 (5)	2 (2)	0.28	0	0	1.00
alteration of phosphate handling and tubular proteinuria	3 (2)	4 (3)	6 (7)	3 (3)	0.33	1 (1)	0	0.49
alteration of glucose resorption and tubular proteinuria	25 (20)	22 (17)	31 (34)	19 (20)	0.05	2 (2)	3 (2)	1.00
alteration of phosphate handling, glucose resorption and tubular proteinuria	0	2 (2)	5 (5)	2 (2)	0.28	0	0	1.00

TDF, tenofovir disoproxil fumarate; eGFR, estimated GFR.

^aFisher’s exact test.

Occurrence of proximal tubulopathy

At 32-wk-GA, i.e. 1 month after treatment initiation, a greater percentage of women on tenofovir disoproxil fumarate had high levels of β2M and glycosuria compared with women on placebo, resulting in a greater percentage of women with proximal tubulopathy on tenofovir disoproxil fumarate (32 of 91; 35%) compared with placebo (20 of 93; 22%) (P = 0.05). At 2-months-PP, proximal tubulopathy was observed in 3 of 120 women (3%) in the tenofovir disoproxil fumarate group and 3 of 125 (2%) in the placebo group (P = 1.00) (Table 3). The percentage of women with alteration of phosphate handling and tubular proteinuria was similar across groups but altered glucose handling was more frequent on placebo (19 of 125; 15%) than on tenofovir disoproxil fumarate (7 of 120; 6%) (P = 0.02). The sensitivity analysis using a narrower window after treatment discontinuation led to similar results (Table S1). At 2-months-PP, the concentrations of total proteinuria, glycosuria, phosphaturia and early markers of proximal tubular dysfunction (RBP, KIM-1, α1M and β2M) were not statistically different between treatment groups (Table 2). No severe or life-threatening kidney-related adverse events were reported, and no kidney-related adverse events led to tenofovir disoproxil fumarate discontinuation.

Earliness and reversibility of proximal tubulopathy

In the tenofovir disoproxil fumarate group, of the three women with proximal tubulopathy at 2-months-PP, all had tubulopathy at 28-wk-GA and two at 32-wk-GA (Table S2). At 12-months-PP, the alteration of phosphate and glucose handling disappeared but tubular proteinuria persisted in two of the three women. In the placebo group, one of the three women had proximal tubulopathy at 28-wk-GA and none of the three at 32-wk-GA. Tubular proteinuria persisted in one of the three women at 12-months-PP.

Infant consequences of tubulopathy

Characteristics of the six infants born to mothers with proximal tubulopathy at 2-months-PP are summarized in Table 4. Their growth measures at delivery, 6-months-PP and 12-months-PP were similar to those in infants born to women without tubulopathy at 2-months-PP. Of those six infants, three had BMD measurements at 12-months-PP and their median lumbar spine BMD was 0.312 g/cm², not statistically different from the 0.325 g/cm² observed in the other 99 infants assessed for BMD (P = 0.68).

Table 4.

Comparison of infant characteristics by maternal tubular dysfunction status at 2-months-PP

Children characteristics	Children of mothers with proximal tubulopathy at 2-months-PP		Children of mothers with no proximal tubulopathy at 2-months-PP		P valuea
	N = 6		N = 241
	N	median (range)	n	median (range)
At birth
weight (g)	6	3245 (2570–3794)	241	3070 (2140–4240)	0.57
length (cm)	6	52 (49–54)	240	50 (40–57)	0.21
head circumference (cm)	6	33 (31–34)	240	33 (29–37)	0.86
At 6 months (Z-score)
weight for age	6	0.6 (−0.9 to 1.8)	225	−0.4 (−4.1 to 2.2)	0.06
length for age	6	0.4 (−1.2 to 0.9)	225	−0.3 (−3.2 to 3.6)	0.47
head circumference for age	6	−0.2 (−1.6 to 2.1)	225	−0.6 (−3.5 to 3.7)	0.31
At 12 months
weight for age (Z-score)	6	0.2 (−1.1 to 0.5)	214	−0.4 (−3.1 to 2.8)	0.21
length for age (Z-score)	6	0.4 (−1.2 to 4.0)	214	−0.6 (−3.6 to 3.5)	0.09
head circumference for age (Z-score)	6	0.2 (−0.3 to 1.5)	213	−0.6 (−3.4 to 7.5)	0.03
lumbar spine bone mineral density (g/cm2)	3	0.312 (0.299–0.336)	99	0.325 (0.246–0.434)	0.68

^aWilcoxon rank-sum test.

Impact of tubulopathy on maternal BMD at 12-month-PP

Of the six women with proximal tubulopathy at 2-months-PP, four had BMD measurement at 12-months-PP, i.e. 10 months after treatment discontinuation. The median (range) maternal BMD was 0.816 (0.802–0.917) g/cm² for total hip and 0.881 (0.768–1.026) g/cm² for lumbar spine. The mean maternal BMD was not statistically different between those 6 women with proximal tubulopathy at 2-months-PP and the 118 without tubulopathy (Table S3).

Discussion

In these hepatitis B-monoinfected women participating in a multicentre, randomized, placebo-controlled trial in Asia, the use of early kidney injury markers did not show a higher risk of proximal tubulopathy after 5 months of tenofovir disoproxil fumarate prophylaxis compared with women on placebo. Proximal tubulopathy was observed only in three women in each group and was not found to be associated with infant growth and maternal BMD abnormalities.

Hyperfiltration leads to a 25% decline of serum creatinine concentration during the first months of pregnancy, followed by a gradual increase of creatinine in the third trimester and until 1-month-PP. These physiological changes, as well as the absence of a validated creatinine-based equation to estimate GFR during pregnancy, complicate the detection of kidney dysfunction during this period. It is known that tubular functions are also altered during pregnancy; pregnant women tend to have an increase in total urinary protein and albumin excretion as well as a loss of phosphorus and glucose in urine.²⁶ Proximal tubulopathy related to tenofovir disoproxil fumarate use may therefore be confused with pregnancy physiological changes. Therefore, a major strength of this study is to provide a comparison with a placebo group in the setting of a randomized controlled trial, allowing for the discrimination of tenofovir disoproxil fumarate renal toxicity from normal changes due to pregnancy. Indeed, the distribution of renal parameters was similar between groups after the randomization process, allowing for direct comparisons between the two groups.

About 20% of women had ≥ 2 markers of proximal tubular dysfunction before treatment initiation. These results must be interpreted with caution because pregnancy is associated with significant physiological tubular changes and the pathological threshold of each of these biomarkers was not determined in pregnant women. We used a panel of urine low-molecular-weight proteins (RBP, KIM-1, β2M, α1M) to detect proximal tubular injury at an early stage. These biomarkers have been shown to be highly sensitive and more reliable than other urine biomarkers to detect proximal tubulopathy related to tenofovir disoproxil fumarate.¹⁷ Another strength of our study was being able to describe the physiological evolution of these markers in pregnant and breastfeeding women in the placebo group. As expected, the decrease in proximal tubular reabsorption of low-molecular-weight protein led to high urine concentrations of RBP, β2M and α1M during the third trimester, which then normalized at 2-months-PP. Interestingly, urine KIM-1 concentrations during the third trimester were lower than values observed in healthy men and women, suggesting that this biomarker was less affected by pregnancy tubular changes. The proportion of women with proximal tubulopathy was higher in the tenofovir disoproxil fumarate group at 32-wk-GA, mainly due to higher β2M, α1M and glycosuria measurements. This strongly suggests that tenofovir disoproxil fumarate has an impact on renal physiology during the third trimester of pregnancy. However, this impact appeared to be limited in time and reversible during the post-partum period. Our findings in pregnant and breastfeeding women are in line with previous studies assessing the risk of proximal tubulopathy in healthy men and women receiving tenofovir disoproxil fumarate for HIV PrEP. In the Partners PrEP study, a randomized, placebo-controlled trial of daily oral tenofovir disoproxil fumarate alone or with emtricitabine conducted in HIV-uninfected African men and women, the frequency of proximal tubulopathy was 1.7% after a median 24 months on emtricitabine/tenofovir disoproxil fumarate versus 1.3% on placebo, but tubular proteinuria and uricosuria were more frequent in the tenofovir disoproxil fumarate group.²⁷ In a nested case–control analysis of the iPrEx trial, MSM receiving tenofovir disoproxil fumarate did not have higher levels of α1M, albuminuria or proteinuria than those on placebo.²⁸ However, in a longitudinal analysis including participants in the subsequent open-label phase of this trial, urine α1M and β2M concentrations increased by 22% and 14%, respectively, over 6 months of tenofovir disoproxil fumarate exposure.²⁹ In our study, all women on tenofovir disoproxil fumarate with proximal tubulopathy at 2-months-PP already had a tubulopathy before tenofovir disoproxil fumarate initiation at 28-wk-GA, suggesting that this was not attributable to tenofovir disoproxil fumarate. Most women with tubulopathy at 2-months-PP had no persistent tubular dysfunction or BMD impairment 10 months after study drug discontinuation. The consequences on infant bone growth of a proximal tubulopathy in mothers have not been extensively investigated, but in our study the presence of a proximal tubulopathy in mothers at 2-months-PP was not found to be associated with the development of abnormalities in infants. These safety data complement our previous study that showed no difference in BMD in mothers and infants evaluated at 12-months-PP for BMD.¹⁹ This is consistent with long-term safety data showing that tenofovir disoproxil fumarate administration for HBV PMTCT had no impact in infants on long-term growth, renal function and bone development up to 6–7 years after birth.³⁰

The 2020 WHO recommendations for the use of tenofovir disoproxil fumarate during pregnancy for HBV PMTCT or those receiving HIV PrEP in 2017 were based on cohort studies of HIV-infected women receiving or not receiving a tenofovir disoproxil fumarate-containing ART regimen. However, the validity of the comparisons in most of these studies was questionable due to the lack of a controlled comparison group (no placebo, no randomization). Our study significantly contributes to the evaluation of tenofovir disoproxil fumarate kidney toxicity during pregnancy and breastfeeding.

Our study has several limitations. First, we could not describe in detail the evolution of urine biomarkers after 32-weeks-GA and before 2-months-PP because we did not collect urine samples between these two timepoints. Second, women in the study were exposed to tenofovir disoproxil fumarate for a relatively short period of time and our results may not be generalizable to women who initiate tenofovir disoproxil fumarate earlier during pregnancy. Third, women enrolled in the study had a low risk of kidney dysfunction at baseline: most participants were less than 30 years old, none had hypertension or diabetes, and women with a CL_CR of <50 mL/min or who had confirmed proteinuria or glycosuria were excluded. Thus, caution should be exercised in pregnant women with altered renal function, pre-eclampsia, cardiovascular comorbidities or using nephrotoxic medications. Fourth, all participating women were Asians. Polymorphisms in tubular transporters have been associated with a higher risk of tenofovir disoproxil fumarate nephrotoxicity in some populations.³¹

In conclusion, the use of early kidney injury markers did not provide evidence of renal proximal tubular toxicity that would limit the use of tenofovir disoproxil fumarate prophylaxis of mother-to-child transmission of HBV in women with no contraindications.

Members of the iTAP Study Team

G. Jourdain, N. Ngo-Giang-Huong, L. Harrison, L. Decker, W. Khamduang, C. Tierney, N. Salvadori, T. R. Cressey, W. Sirirungsi, J. Achalapong, P. Yuthavisuthi, P. Kanjanavikai, O. P. Na Ayudhaya, T. Siriwachirachai, S. Prommas, P. Sabsanong, A. Limtrakul, S. Varadisai, C. Putiyanun, P. Suriyachai, P. Liampongsabuddhi, S. Sangsawang, W. Matanasarawut, S. Buranabanjasatean, P. Puernngooluerm, C. Bowonwatanuwong, T. Puthanakit, V. Klinbuayaem, S. Thongsawat, S. Thanprasertsuk, G. K. Siberry, D. H. Watts, N. Chakhtoura, T. V. Murphy, N. P. Nelson, R. T. Chung, S. Pol and N. Chotivanich.

Funding

The iTAP randomized controlled trial (the parent study) was supported by a grant from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) under a cooperative agreement (U01HD071889) among the NICHD, the Centers for Disease Control and Prevention, and the French National Research Institute for Sustainable Development. This substudy was supported by a grant from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (R03HD096131).

Transparency declarations

There were no conflicts of interest.

Author contributions

G.L., G.J., N.N. and J.Y.M. designed the study. G.L., G.J., N.S. and J.Y.M. developed the analysis. N.S. performed the analysis. N.N., P.B. and R.C. performed the laboratory analysis. J.A., P.K., O.P.N.A., S.P., T.S. and P.S. took part in the study at their sites. G.L. wrote the first draft of the report. All authors critically reviewed and approved the manuscript.

The preliminary results of the study were presented as a poster communication at the Twenty-Third International AIDS Conference, AIDS 2020 (virtual), 6–10 July 2021. Abstract PEB0192.

Supplementary data

Supplementary information and Tables S1 to S3 are available as Supplementary data at JAC Online.