Initial side effects of roxadustat on thyroid function in patients undergoing hemodialysis

Abstract

Roxadustat, a hypoxia-inducible factor prolyl hydroxylase (HIF-PH) inhibitor, is effective for treating renal anemia, with efficacy noninferior to erythropoiesis-stimulating agents. However, data on its adverse effects remain scarce. In medical practice, roxadustat administration reportedly leads to central hypothyroidism; however, the frequency of its occurrence remains unclear. We investigated thyroid function in patients undergoing dialysis who received roxadustat at Saiseikai Yokohamashi Tobu Hospital, with particular focus on the timing of the decrease in serum thyroid-stimulating hormone (TSH) levels after roxadustat initiation and the subsequent course of thyroid function. HIF-PH inhibitors were administered to 43 chronic kidney disease patients including dialysis. After 2 weeks of roxadustat administration, we extracted the data of 16 hemodialysis patients who exhibited absence of residual urine and were able to undergo blood tests. All patients revealed decreased serum TSH levels. In contrast, changes in serum free triiodothyronine and free thyroxine levels were minor. None of the patients reported fatigue, bradycardia, or constipation, which are common symptoms of hypothyroidism. Our findings revealed that roxadustat may cause TSH reduction after 2 weeks of administration, though some patients showed subclinical hypothyroidism with high TSH levels. In a few cases, TSH levels decreased to one-tenth of the normal values. Despite these rapid changes, none of the patients exhibited clinical symptoms. Hence, regular thyroid hormone level monitoring is necessary with roxadustat administration, regardless of patients’ history of thyroid function.

LAY SUMMARY

Roxadustat is a relatively new medication used to treat anemia in patients with chronic kidney disease. While it is known to be as effective as conventional treatments, its side effects are not yet fully understood.

In this study, we examined how roxadustat affects thyroid function in patients undergoing hemodialysis. We found that levels of thyroid-stimulating hormone (TSH), which regulates thyroid function, decreased in all patients within 2 weeks after starting roxadustat. However, other free thyroid hormone levels showed only small changes, and none of the patients developed symptoms of hypothyroidism such as fatigue or constipation.

These results suggest that roxadustat can alter thyroid-related laboratory values without causing noticeable symptoms. Therefore, thyroid function should be monitored regularly from the early phase of roxadustat treatment, even in patients without a history of thyroid disease.

KEY POINTS

What is known

Roxadustat is an effective treatment for renal anemia, but its effects on thyroid function are not well established.

What this study adds

Roxadustat may cause an early reduction in TSH levels within 2 weeks of initiation in patients undergoing hemodialysis, without significant changes in free thyroid hormone level or clinical symptoms.

Potential impact

Routine monitoring of thyroid function should be considered from the early phase of roxadustat treatment, even in patients without a history of thyroid disease.

Introduction

Renal anemia is a well-known complication of advanced chronic kidney disease (CKD) [1]. Treatment of renal anemia is important because it not only worsens renal prognosis [1,2] but is also associated with an increased risk of morbidity, mortality, and hospitalization for cardiovascular diseases [3,4]. Traditionally, injection therapy with erythropoiesis-stimulating agents (ESAs) has been predominantly used to treat renal anemia [5]. One of the disadvantages of ESAs is injection-site pain in patients with advanced CKD [6].

Hypoxia-inducible factor prolyl hydroxylase (HIF-PH) inhibitors, such as roxadustat, have been beneficial in treating renal anemia in recent years [7–9]. An advantage associated with the administration of oral drugs is the lack of pain produced by injections [10]. However, regarding HIF-PH inhibitors, the accumulated data concerning associated side effects are currently insufficient [11]. In 2020, the Japanese Society of Nephrology issued recommendations for the proper use of HIF-PH inhibitors [12]. Roxadustat is an HIF-PH inhibitor that causes central hypothyroidism as a highlighted side effect, and a possible association between roxadustat and hypothyroidism was identified through an analysis of the pharmacovigilance database in Japan [13]. The frequency of occurrence and severity of roxadustat -induced central hypothyroidism remains unknown; however, an increasing number of studies have reported its occurrence [14–18]. In this context, we aimed to investigate thyroid function in patients with CKD who received roxadustat at our hospital to determine when central hypothyroidism develops after the administration of HIF-PH inhibitors and to clarify the frequency and subsequent course of thyroid function and clinical symptoms.

Materials and methods

Study design and participants

Roxadustat were administered to 43 patients (advanced CKD [n = 15], hemodialysis [HD] [n = 20], peritoneal dialysis [PD] [n = 6], and combined therapy with peritoneal dialysis and hemodialysis (PD+HD) [n = 2]) between January 2021 and July 2023 consecutively at Saiseikai Yokohamashi Tobu Hospital (Figure 1). Patients with absence of residual urine who were able to undergo blood tests after 2 weeks of roxadustat administration were included in the study. Patients who exhibited presence of residual urine, and who were unable to undergo blood tests after 2 weeks of drug administration were excluded from the study. Further, as reported in our previous study [19], a high possibility of excretion of thyroid hormone associated with proteinuria may lead to hypothyroidism. Therefore, to avoid confounding, we excluded patients with CKD and patients undergoing PD. Subsequently, we excluded 23 patients (advanced CKD [n = 15], PD [n = 6], combined therapy with PD+HD [n = 2] in whom the median proteinuria level was 4.80 g/gCr (1.73–4.81). Consequently, 16 patients who underwent blood tests 2 weeks after administration were included in the final analysis. Therefore, all participants included were patients who were hospitalized, undergoing HD, and had received roxadustat. Serum TSH, free triiodothyronine (FT3), and free thyroxine (FT4) levels were measured before and every 2 weeks after initiation of roxadustat administration. We evaluated the changes in thyroid function and checked for symptoms, such as fatigue, bradycardia, and constipation. In this study, the standard values for serum TSH, FT3, and FT4 levels were 0.61–4.23 μIU/mL, 2.30–4.40 pg/mL, and 0.90–1.70 ng/dL, respectively.

Figure 1.

Flowchart of study participants.

Abbreviations: CKD, chronic kidney disease; PD, peritoneal dialysis; HD, hemodialysis

Moreover, we retrospectively investigated patients’ age, sex, HD career, underlying disease, dose of roxadustat, dose of erythropoiesis-stimulating agents (ESAs) before roxadustat administration, blood hemoglobin level, serum C-reactive protein (CRP), ferritin, transferrin saturation, thyroid function, and body weight using an electronic chart.

Ethical considerations

This study was conducted in accordance with the principles of the Declaration of Helsinki and approved by the Ethical Review Board of the Saiseikai Yokohamashi Tobu Hospital (approval number 20230026). The requirement for written informed consent was waived by the Ethical Review Board of the Saiseikai Yokohamashi Tobu Hospital (approval number 20230026) due to the retrospective design of the study. Information about the study was posted on the institution’s website, and potential participants were free to opt out.

Statistical analysis

JMP software Pro 17.0 for Windows and Eazy R were used for statistical analyses. Within-group longitudinal changes were analyzed using repeated measures analysis of variance followed by Bonferroni’s multiple comparison tests. Given the treatment modifications during follow-up, including discontinuation and re-administration of roxadustat in some patients, these longitudinal analyses were considered exploratory. In addition, longitudinal changes in thyroid function were analyzed separately according to whether roxadustat was continued or discontinued, and individual patient courses are presented in the Supplementary Figures (Supplementary Figures S1 and S2). Paired comparisons were also performed to evaluate individual-level changes from baseline. In addition, the relationship between thyroid hormone levels and body weight was assessed using Spearman’s rank correlation analysis.

Results

During this study, 43 patients were administered roxadustat, regardless of CKD stage. Of these, after excluding 27 patients based on our exclusion criteria referring to Study design and participants, 16 patients were included in the final analysis (Figure 1).

Patient backgrounds are summarized in Table 1. Regarding the patients’ background before roxadustat administration, the median age was 72.0 years, 15 patients were male, and dialysis was performed for approximately 7.0 years. The most common cause of CKD was diabetic nephropathy (14/16 patients), and a history of chronic limb-threatening ischemia was present in 13 patients (Table 1).

Table 1.

Patient characteristics before roxadustat administration.

Parameter	Median (IQR or %)
Age (year)	72.0 (61.75–76.0)
Sex male, n (%)	15 (93.8%)
Hemodialysis duration (year)	7.00 (4.00–11.25)
Underlying disease, n (%)	Diabetic nephropathy 14 (87.5%)
	Chronic glomerulonephritis 1 (6.25%)
	Nephrosclerosis 0
	Membranoproliferative 0 glomerulonephritis 1 (6.25%)
History of CLTI, n (%)	13 (81.25%)
Hemoglobin (g/dL) (13.7–16.8)	8.15 (7.48–8.68)
CRP (g/dL) (0.0–0.2)	5.75 (1.49–10.33)
Ferritin (ng/mL) (39.0–465.0)	382.5 (228.5–575.0)
TSAT (%)	18.5 (14.3–24.6)
TSH (μIU/mL)	2.20 (1.36–4.03)
FT4 (ng/dL)	1.35 (1.15–1.50)
FT3 (pg/mL) Urination Body weight	1.70 (1.48–2.23) Undetected 59.4 (51.0–70.6)

Abbreviations: IQR, interquartile range; CLTI, chronic limb-threatening ischemia; CRP, C-reactive protein; TSAT, transferrin saturation; TSH, thyroid-stimulating hormone; FT4, free thyroxine; FT3, free triiodothyronine.

Table 2 shows the changes in thyroid hormone levels in the study cohort. Thyroidal dysfunction was categorized into four conditions: overt hypothyroidism, defined as low FT4 levels with elevated TSH levels; subclinical hypothyroidism (SH), defined as elevated TSH levels with normal FT4 levels; low T3 syndrome, defined as low FT3 levels with normal FT4 and TSH levels; and low T4 syndrome, defined as decreased FT3 and FT4 levels with normal TSH levels. Findings other than these four thyroid conditions were considered indicative of normal thyroid function. Nine (cases 3, 6, 7, 10, 12, 13, 14, 15, and 16) (56.3%) patients had low T3 syndrome, four (cases 4, 8, 9, and 11) (25.0%) had SH, two (cases 1 and 2) (12.5%) had normal thyroid function, and one (Case 5) had hyperthyroidism before roxadustat administration.

Table 2.

Changes in thyroid hormone before and after every 2 weeks of roxadustat administration.

Table 2-A. (Patient background)

Case	Age/Sex	ESA before Rox	First dose of Rox (mg)
1.	74/M	EPO 9000 IU/week	100
2.	74/M	DA 60 μg/week	100
3.	61/M	DA 20 μg/week	100
4.	82/M	DA 80 μg/week	100
5.	62/M	DA 120 μg/week	100
6.	73/F	DA 60 μg/week	100
7.	71/M	DA 120 μg/week	100
8.	82/M	DA 120 μg/week	100
9.	60/M	CERA 75 μg/2 week	100
10.	83/M	DA 120 μg/week	100
11.	65/M	DA 120 μg/week	100
12.	51/M	DA 120 μg/week	100
13.	50/M	DA 120 μg/week	100
14.	74/M	DA 60 μg/week	100
15.	85/M	DA 60 μg/week	100
16.	65/M	DA 180 μg/week	100

Table 2-B. (TSH)
Case	TSH (μIU/mL) (standard value: 0.61–4.23)
	Pre	2 weeks	4 weeks	6 weeks	8 weeks
1.	1.28	0.43	0.28	1.37
2.	1.46	0.05	0.35	0.20	1.53
3.	2.97	0.44	0.38	0.41	2.19
4.	4.71	0.29	1.53	0.20
5.	3.12	0.16
6.	2.83	1.79
7.	1.54	0.33	0.04	0.03
8.	5.23	0.82
9.	6.34	1.59	0.07	0.69	0.35
10.	1.39	0.94	0.31	4.57
11.	8.00	0.09	0.15
12.	1.56	0.11	0.43	0.89
13.	3.80	3.36	0.74	1.71
14.	0.84	0.05	0.61	0.08
15.	1.14	0.12	0.22
16.	0.72	0.13	0.44
Median (IQR)	2.20 (1.36–4.03)	0.31 (0.12–0.85)	0.35 (0.22–0.44)	0.55 (0.2–1.25)	1.53 (0.94–1.86)

Table 2-C. (FT3)
Case	FT3 (pg/mL) (standard value: 2.30–4.40)
	Pre	2 weeks	4 weeks	6 weeks	8 weeks
1.	2.4	1.9	2.0	2.1
2.	2.3	1.9	1.3	2.1	2.3
3.	2.0	1.8	1.5	1.0	1.5
4.	1.1	1.0	1.1	1.0
5.	1.1	1.6
6.	1.6	1.5
7.	1.5	1.8	1.9	1.9
8.	1.9	1.8
9.	3.6	4.4	1.7	1.6	1.2
10.	2.2	1.8	1.8	2.0
11.	2.8	2.4	2.7
12.	1.7	1.6	1.3	1.6
13.	1.7	1.6	1.8	1.5
14.	1.6	1.3	1.2	1.5
15.	1.4	1.1	1.5
16.	1.3	1.2	1.6
Median (IQR)	1.7 (1.48–2.23)	1.7 (1.45–1.83)	1.6 (1.30–1.8)	1.6 (1.5–1.98)	1.5 (1.35–1.9)

Table 2-D. (FT4)
Case	FT4 (ng/mL) (standard value: 0.90–1.70)
	Pre	2 weeks	4 weeks	6 weeks	8 weeks
1.	1.3	0.9	0.8	0.7
2.	1.4	1.5	1.3	0.8	0.7
3.	1.4	1.4	1.1	1.0	1.0
4.	1.0	0.5	0.7	0.5
5.	1.9	2.0
6.	1.5	1.0
7.	1.3	1.3	1.5	1.6
8.	1.0	0.9
9.	1.4	1.5	0.7	1.2	1.0
10.	1.3	1.0	0.8	0.8
11.	0.9	0.8	0.5
12.	1.7	1.5	1.3	2.2
13.	1.5	1.1	0.8	0.8
14.	1.7	1.3	0.9	0.8
15.	1.2	1.1	1.2
16.	1.0	0.8	1.0
Median (IQR)	1.35 (1.15–1.5)	1.1 (0.9–1.43)	0.9 (0.8–1.2)	0.8 (0.8–1.15)	1 (0.85-1)

Abbreviations: TSH, thyroid-stimulating hormone; FT4, free thyroxine; FT3, free triiodothyronine; EPO, erythropoietin; DA, darbepoetin alfa; CERA, continuous erythropoietin receptor activator; IQR, interquartile range.

According to the Japanese package insert for roxadustat, the recommended initial dose when switching from erythropoiesis-stimulating agents (ESAs) is based on the previous ESA dose. For patients receiving epoetin, the recommended starting dose is 70 mg for those receiving less than 4500 IU/week and 100 mg for those receiving 4500 IU/week or higher. For patients receiving darbepoetin alfa, the recommended starting dose is 70 mg for those receiving less than 20 μg/week and 100 mg for those receiving 20 μg/week or higher. In our study, ESAs were prescribed to patients based on once a week before switching to roxadustat. Therefore, the initiation of roxadustat was performed according to the recommended dosing criteria for patients switching from ESA therapy. In addition, at the time of this study, only the 100 mg formulation of roxadustat was available at our institution. As a result, all patients were initiated on 100 mg of roxadustat when switching from ESAs. After 2 weeks of using roxadustat, serum TSH levels decreased in all patients (Supplementary Figures S1 and S2). Compared with the change in median TSH levels, the changes in median serum FT3 and FT4 levels remained minor (first quartile point, third quartile point) (before roxadustat administration [TSH: 2.20 {1.36–4.03}, FT3: 1.70 {1.48–2.23}, FT4: 1.35 {1.15–1.50}]; 2 weeks after roxadustat administration [TSH: 0.31 {0.12–0.80}, FT3: 1.70 {1.45–1.83}, FT4: 1.10 {0.90–1.43}]). In two patients (Cases 2, 14), TSH levels decreased to one-tenth of the usual value (0.0611 μgIU/mL).

Figure 2 shows the changes in thyroid hormone levels before and after every 2 weeks of roxadustat administration, analyzed using repeated measures analysis of variance (ANOVA). Significant differences were observed between TSH and FT4 levels. In addition, paired comparisons demonstrated significant differences in these parameters at 2 weeks (p < 0.05), whereas no significant difference was observed in FT3 levels (p = 0.176). We evaluated the relationship between body weight and the degree of TSH reduction using Spearman’s rank correlation analysis; no significant correlation was observed (Spearman’s rank correlation coefficient=-0.433, 95%CI −0.765 to 0.0795, p = 0.0936). None of the patients experienced fatigue, bradycardia, or constipation, which are symptoms of hypothyroidism.

Figure 2.

Changes in thyroid hormone before and at 2-week intervals during roxadustat administration. Among 16 cases, three cases discontinued roxadustat, and 13 cases (81.25%) continued roxadustat.

(a) The change of TSH, (b) FT3, and (c) FT4 levels

TSH, thyroid-stimulating hormone; FT3, free triiodothyronine; FT4, free thyroxine

Table 2, Figure 3, and Supplementary Table show the changes in TSH levels after roxadustat administration and the resulting outcomes. In 11 out of the 16 patients (cases 3, 4, 5, 7, 9, 10, 11, 12, 14, 15, and 16), roxadustat was stopped when TSH levels decreased after roxadustat administration. In six out of 16 patients (cases 3, 4, 9, 10, 12, and 14) (37.5%), the TSH levels improved after discontinuation of roxadustat. After discontinuation of roxadustat, TSH levels improved to standard values naturally within 2 weeks in five patients (cases 3, 4, 9, 10, and 14), whereas they normalized after 4 weeks in one patient (Case 12). Three patients (cases 1, 2, and 13) (18.8%) continued roxadustat treatment, but their TSH levels improved without any additional intervention. In three patients (cases 4, 9, and 14), roxadustat was readministered when TSH levels returned to normal after stopping roxadustat. In these patients, TSH levels decreased again 2 weeks after the re-administration of roxadustat, indicating reproducibility. Three patients (cases 11, 15, and 16) showed improvement in TSH levels compared with those before discontinuation, but the values did not reach the normal range. One patient (Cases 7) (6.25%) showed no improvement in TSH levels after discontinuing roxadustat; however, no clinical symptoms of hypothyroidism were noted.

Figure 3.

Outcomes of thyroid dysfunction.

Discussion

Our study showed that roxadustat reduced TSH levels after 2 weeks of roxadustat administration, though some patients showed subclinical hypothyroidism with high TSH levels. Furthermore, TSH levels reversibly improved in most cases, when the drug was discontinued or continued, although a few cases showed persistent abnormalities. Regular thyroid hormone level monitoring is necessary for patients who are administered roxadustat, regardless of patients’ history of thyroid function.

Thyroid hormone receptors (THRs) are nuclear receptors that exist as THRα and THRβ isoforms [20]. According to Yao et al., roxadustat has a structure similar to that of triiodothyronine (T3), and it is considered to be a selective agonist for THRβ [21] (Figure 4). Moreover, roxadustat has stronger affinity for THRβ than does T3 [21]. Furthermore, THRβ inhibits thyrotropin-releasing hormone (TRH) secretion from the hypothalamus through negative feedback, and secretion of TSH from the anterior pituitary gland may be more effective than that of THRα [22,23]. In the human body, when blood T3 levels are low, the hypothalamus secretes TRH, which stimulates the production of TSH by TSH-producing cells in the pituitary glands [24]. Subsequently, TSH acts on the thyroid gland and stimulates the production and secretion of T3 and T4, which are the thyroid hormones [24]. The secretion of T3 and T4 may provide feedback to the pituitary gland [24]. T3 binds to the thyroid hormone receptors of TSH-producing cells and may decrease TSH production [23]. Based on these physiological mechanisms, roxadustat, as a potential THRβ agonist, may activate negative feedback pathways at the pituitary level despite the absence of excessive circulating thyroid hormone levels, which could contribute to the observed reduction in TSH levels [21]. However, the precise mechanism remains unclear because TRH levels and dynamic testing of the hypothalamic-pituitary-thyroid axis were not evaluated in this study.

Figure 4.

Chemical structures of triiodothyronine (T3) and roxadustat. Chemical structures and molecules formulas were obtained from PubChem (CID 11256664 for Roxadustat and CID 5920 for T3) [41,42].

(a) T3, (b) Roxadustat

In our study, TSH levels decreased in all patients after 2 weeks of roxadustat administration. Although previous studies have reported a positive correlation between TSH and body weight [25], no significant association between body weight and TSH was observed in this study. Significantly decreased TSH levels below the standard values (0.61–4.23 μIU/mL) were observed in 12 of 16 patients (75.0%). The serum TSH values had decreased to less than one-tenth of the standard value after 2 weeks of roxadustat administration in two of these patients (cases 2 and 14). According to previous reports, some of these cases were treated with levothyroxine as hormone replacement therapy (HRT) [15,17]. In our study, roxadustat administration decreased TSH levels, but we only observed serum TSH levels and did not initiate HRT because of the absence of any symptoms. In six out of 16 patients, TSH levels naturally improved to standard values after roxadustat discontinuation; in five of these patients (cases 3, 4, 9, 10, and 14), the levels improved after 2 weeks, whereas in one patient (Case 12) they normalized after 4 weeks. Three of these patients (cases 11, 15, and 16) revealed improved values compared with the values before discontinuation; however, the values did not reach the normal range. Thyroid hormone levels generally require several weeks (approximately 4–6 weeks) to reach a steady state, because of the long half-life of thyroxine [26]. Therefore, the changes observed at 2 weeks may reflect early dynamic alterations rather than steady-state conditions. Our study suggests that the TSH decline associated with roxadustat may not necessarily depend on thyroid hormone replacement, as the TSH levels improved reversibly after drug discontinuation. Several case reports have described similar clinical courses [14,27].

In our study, in 12 out of 16 patients (75.0%), TSH levels improved or showed improvement tendency without the use of thyroid hormone preparations. Our findings revealed that although the TSH level returned to the normal range with continued administration of roxadustat, the FT4 level decreased to below 0.9 ng/dL (cases 1, 2, and 13). This suggests that, in addition to decreasing TSH levels, roxadustat may attenuate the responsiveness of the thyroid to TSH during treatment. From another perspective, the decrease in TSH levels due to roxadustat may diminish overtime in some cases despite its continued administration, leading to spontaneous improvement [28].

Our findings raise the possibility that the effects of roxadustat on the pituitary-thyroid axis may attenuate over time, although the underlying mechanisms remain unclear. Previous case reports have documented that TSH levels decrease after roxadustat administration [15,29]. Among these cases, a history of hypothyroidism due to chronic thyroid disease was frequently noted [15,29]. In our study, four patients (cases 4, 8, 9, and 11) (25.0%) had a medical history of SH and nine patients (cases 3, 6, 7, 10, 12, 13, 14, 15, and 16) (56.3%) had low T3 syndrome before roxadustat administration. Notably, none of the patients were administered drugs such as steroids and amiodarone that cause thyroid dysfunction. In our previous study, we reported that patients with CKD tend to have hypothyroidism [19,30]. Careful monitoring is strongly required when using roxadustat in patients with a history of thyroid dysfunction and CKD because thyroid function may fluctuate after roxadustat administration.

As a treatment modality for renal anemia in patients on HD, roxadustat is not inferior compared with darbepoetin alfa, which is an erythropoiesis-stimulating agent (ESA) and is also expected to have an effect, such as a state of inflammation, on ESA-resistant renal anemia [31]. In our patients, the CRP levels were higher than normal. As we used roxadustat, which is resistant to ESAs, and because some patients on dialysis had chronic limb-threatening ischemia (CLTI), many of them had an elevated inflammatory response due to infection. Additionally, low T3 syndrome has been reported to be associated with malnutrition and inflammation states [32]. It is expected that HIF-PH inhibitors will continue frequently in patients with low T3 syndrome accompanied by inflammation. In our study, 13 out of 16 patients (81.25%) had elevated inflammatory responses associated with CLTI, which may also explain the high prevalence of low T3 syndrome (62.5%) before roxadustat administration. Infections are the second leading cause of death during dialysis [33]. Roxadustat is expected to be used more frequently for renal anemia in the inflammation state. However, data accumulation on the efficacy and safety of HIF-PH inhibitors, including that of roxadustat, in such situations remains insufficient.

According to this study, TSH levels improve after the discontinuation of roxadustat, and there were still few cases in which thyroid function did not recover after discontinuation. Patients treated with roxadustat have significantly lower levels of TSH, FT3, and FT4 than those treated with recombinant human erythropoietin (p < 0.05) [34]. Another meta-analysis showed that the incidence of suppression of thyroid function was significantly higher in patients treated with roxadustat than in those receiving ESA therapy (odds ratio, 6.45; 95% confidence interval, 3.39–12.27) [35]. Therefore, monitoring thyroid function during roxadustat administration is clinically important.

Regarding the onset of hypothyroidism associated with roxadustat, Tanaka et al. reported a mean onset time of 92 days, with >50% of cases developing it within 100 days of treatment initiation [13]. Similarly, Kouki et al. found that approximately 50% of the cases develop hypothyroidism within 50 days of treatment initiation [36]. However, our study demonstrated that in all patients, TSH levels decreased within 2 weeks, which suggests that thyroid dysfunction can occur earlier than that reported previously. Although previous reports have suggested that roxadustat-induced thyroid dysfunction may be reversible [15,37], our study indicates that this is not always the case.

Notably, in our study, despite the presence of thyroid dysfunction, no clinical symptoms were observed. Similarly, Haraguchi et al. reported that despite decreases in FT4 and TSH levels following roxadustat administration, no symptoms of hypothyroidism were observed [38]. However, there have also been case reports of roxadustat-induced hypothyroidism presenting with symptoms such as psychomotor slowing, apathy, and somnolence [16]. In patients on dialysis, symptoms of hypothyroidism, such as muscle cramps, shoulder stiffness, constipation, cold intolerance, dry skin, and decreased motivation, may be overlooked, as they can be attributed to dialysis-related conditions [39]. Therefore, careful monitoring of a patient’s clinical course is essential. Based on our findings, thyroid function should be assessed at baseline and within the first few weeks after initiation of roxadustat. Subsequent monitoring should be guided by clinical findings. However, given the limited follow-up period, the optimal monitoring intervals and duration remain to be established. We observed changes in thyroid function in patients treated with roxadustat for up to 8 weeks; however, the long-term effects of continued roxadustat administration on thyroid function remain unclear. Further studies are required to evaluate the long-term efficacy and safety of these drugs.

This study has some limitations. This was a retrospective, single-center study with a small sample size. In addition, the observation period was limited to 8 weeks, and thyroid function was not assessed within the first week after roxadustat initiation; therefore, the precise timing of the earliest TSH changes could not be determined. Furthermore, the long-term effects, reversibility and dose dependency could not be fully evaluated, and these issues remain subjects for future investigation. In addition, the decision to continue or discontinue roxadustat was left to the clinical judgment of the treating physicians, taking into account changes in TSH and FT4 levels. Therefore, factors associated with treatment discontinuation could not be fully evaluated. Future studies with larger sample sizes and predefined discontinuation criteria are required to clarify the clinical significance of thyroid function changes during roxadustat treatment. Given the small sample size and the limited duration of follow-up, large-scale, multicenter prospective studies with extended follow-up are warranted to validate the preliminary findings of the present study. In our study, 16 hemodialysis patients were included, the majority of whom were male and had a high prevalence of diabetic nephropathy and CLTI. All 16 patients were hospitalized cases, and many had severe comorbidities, particularly CLTI. Previous reports have suggested that hospitalized patients with CLTI are more frequently male [40], which may partly explain the predominance of male patients in our study. In addition, according to the Japanese Society for Dialysis Therapy guidelines, the target hemoglobin level in hemodialysis patients is uniformly set at 10–12 g/dL regardless of gender. As a result, treatment with ESAs or HIF-PH inhibitors is initiated when hemoglobin levels fall below this threshold regardless of sex, which may have contributed to the higher proportion of male patients by chance. Furthermore, more than half of patients had low T3 syndrome at baseline and had elevated inflammatory markers associated with CLTI complications. These baseline conditions may have acted as confounding factors, independently affect TSH and free thyroid hormone levels and influencing the interpretation of thyroid function changes after roxadustat initiation. Nevertheless, this highly selected study population may introduce selection bias and limited generalizability of our findings. Therefore, caution is warranted when extrapolating these findings to broader hemodialysis populations. This study also lacked a control group consisting of hemodialysis patients maintained on conventional ESAs. Therefore, changes in thyroid function associated with uremic status, intercurrent illness, nutritional alterations, or intra-individual variability could not be fully excluded, and the causal relationship between roxadustat and changes in TSH levels could not be definitively established. Despite these limitations, previous studies in Japanese dialysis and pre-dialysis patients have described various thyroid abnormalities, including low T3 syndrome, and overt hypothyroidism with elevated TSH levels. However, primary TSH suppression has rarely been reported in these populations. Although our findings should be interpreted with caution because of the absence of a control group, they suggest that roxadustat may contribute to this characteristic pattern of TSH reduction. Furthermore, TRH levels were not measured, and TRH stimulation tests were not performed. Therefore, we could not definitively distinguish between primary thyroid suppression and pituitary-mediated TSH reduction or confirm the involvement of the hypothalamic-pituitary-thyroid axis. Future studies incorporating TRH assessment are warranted to clarify the mechanisms underlying roxadustat-associated changes in thyroid function.

According to American thyroid association guidelines, thyroid function is generally reassessed 4–6 weeks after initiation or dose adjustment of thyroid hormone therapy [26]. However, based on our findings and observed early changes at 2 weeks, we suggest that thyroid function should be assessed at baseline and within the first few weeks after initiation of roxadustat, with subsequent monitoring performed as clinically indicated. Further studies are needed to determine the optimal monitoring intervals and duration. Furthermore, as HIF-PH inhibitors are expected to be used in inflammatory states, physicians should be cautious when using roxadustat, especially in patients with a history of thyroid dysfunction and inflammatory CKD.

Funding Statement

This work was not supported by any funding agency.

Disclosure statement

The authors report there are no competing interests to declare.

Data availability statement

Raw data were generated at Saiseikai Yokohamashi Tobu Hospital. Derived data supporting the findings of this study are available from the corresponding author H.Y. on request.