Sickle Cell Disease and Kidney Injury: Circulating Uromodulin Allows Early Tissue Specific Diagnosis and Monitoring of Treatment

To the Editor,

1.

Sickle cell nephropathy (SCN) is a major determinant of morbidity and mortality in sickle cell disease (SCD), yet early injury remains difficult to recognize with conventional laboratory markers alone. Creatinine‐based assessment is problematic in SCD because low muscle mass, increased tubular secretion, and glomerular hyperfiltration may keep serum creatinine (SCr) deceptively low even in the presence of evolving nephropathy. Cystatin C (Cys‐C) is useful but can still remain within reference range in early disease, while albuminuria reflects predominantly glomerular injury and may miss an earlier phase of tubular damage. Serum uromodulin (sUmod), a circulating form of the kidney‐specific glycoprotein produced by the epithelial cells of the thick ascending limb of Henle's loop, may therefore provide a tissue‐specific marker of tubular integrity and an opportunity to detect SCN before conventional filtration markers become abnormal [1, 2, 3, 4, 5, 6, 7, 8].

We retrospectively evaluated the association of sUmod with kidney dysfunction in adults with SCD treated at the University Hospital Essen. Thirty‐two adults with confirmed SCD were included; all were assessed in steady state, without vaso‐occlusive crisis or intercurrent infection at the time of sampling. Mean age was 28.6 ± 8.8 years, 53.1% were male, 59.4% had HbSS, and 43.8% were receiving hydroxycarbamide (HC) (stable dosage) at baseline (Table 1). Further cohort and subgroup baseline details are provided in Tables S1 and S2. Kidney assessment included enzymatic serum creatinine (e‐SCr), CysC, beta‐2‐microglobulin (B2M), estimated glomerular filtration rate (eGFR), urine protein and albumin measurements, and sUmod measured by ELISA. Repeated measurements obtained during follow‐up were analyzed with mixed‐effect models including patient‐level random intercepts, allowing us to assess whether longitudinal changes in sUmod tracked changes in established kidney markers. The median follow‐up was 258 days (range, 69–473 days; Table S3).

TABLE 1.

Baseline characteristics of the study cohort.

Characteristic	Value
Age, years	28.6 ± 8.8
Male sex, n (%)	17 (53.1)
Genotype, n (%)
HbSS	19 (59.4)
HbS/beta‐thalassemia	6 (18.8)
HbSD‐Punjab	3 (9.4)
HbS/O‐Arab	1 (3.1)
HbSC	3 (9.4)
Hemoglobin, g/dL	9.4 ± 1.7
Absolute reticulocyte count, nL	240.1 ± 121.4
LDH, U/L	403.0 ± 232.9
Total bilirubin, mg/dL	2.9 ± 2.4
e‐SCr, mg/dL	0.6 ± 0.2
CysC, mg/L	0.9 ± 0.2
B2M, mg/L	1.6 ± 0.7
sUmod, ng/mL	162.4 ± 70.5
Proteinuria/albuminuria, n/N (%)	11/30 (36.7)
Hydroxycarbamide treatment, n (%)	14 (43.8)
Chronic exchange transfusion, n (%)	2 (6.3)

Note: Adult patients with sickle cell disease (N = 32).

Abbreviations: B2M, beta‐2‐microglobulin; CysC, cystatin C; e‐SCr, enzymatic serum creatinine; LDH, lactate dehydrogenase; sUmod, serum uromodulin.

At baseline, mean e‐SCr, CysC, and B2M were largely within reference ranges at the cohort level, whereas proteinuria or albuminuria was already present in 11 of 30 evaluable patients (36.7%). Mean sUmod was 162.4 ± 70.5 ng/mL (Table 1; Table S1). Two patients had clinically apparent kidney dysfunction at baseline, one with established focal segmental glomerulosclerosis and overt proteinuria and one with albuminuria and reduced eGFR despite e‐SCr within the reference range (Table S4). More broadly, proteinuria or albuminuria was observed in a larger subset of patients, consistent with earlier or less overt kidney involvement in many cases; albuminuria categories are summarized in Table S5. Because repeated urine measurements were available only for part of the cohort, persistence of albuminuria could not be confirmed uniformly in all affected patients (Table S3). The clinically important finding, however, was not the cohort mean but the pattern across serial measurements: lower sUmod consistently tracked with a less favorable kidney phenotype, even when conventional blood markers remained apparently unremarkable.

In mixed‐effects analyzes, lower sUmod was significantly associated with higher e‐SCr (beta = −0.0012, SE = 0.00033, p = 0.0005), higher CysC (beta = −0.00089, SE = 0.00031, p = 0.005), and higher B2M (beta = −0.0026, SE = 0.0012, p = 0.0242) (Figure 1 and Table 2, Table S6). Genotype terms were not significant in these models, suggesting that the relationship between sUmod and kidney dysfunction was not explained simply by genotype composition. Time terms were significant for the e‐SCr and CysC models, supporting a longitudinal relationship between declining tubular integrity and worsening kidney biomarker profile over follow‐up. Taken together, these findings indicate that sUmod is not merely a static cross‐sectional correlate but a dynamic marker that moves in parallel with kidney dysfunction over time. When kidney function was expressed as eGFR, higher sUmod remained associated with higher creatinine‐based eGFR (β = +0.64 per 10 ng/mL, 95% CI 0.17–1.11; p = 0.007; n = 103), higher cystatin C‐based eGFR (β = +0.81 per 10 ng/mL, 95% CI 0.14–1.47; p = 0.017; n = 93), and higher combined creatinine/cystatin C eGFR (β = +0.77 per 10 ng/mL, 95% CI 0.22–1.32; p = 0.006; n = 91), supporting the same interpretation in clinically familiar eGFR‐based terms (Table S7).

FIGURE 1.

Linear mixed models showing the dependency of serum uromodulin (sUmod) levels on (A) enzymatic creatinine (e‐SCr), (B) cystatin C (CysC), and (C) beta‐2‐microglobulin (B2M) in adult sickle cell disease (SCD) patients. Lower sUmod levels significantly associate with higher levels of all three biomarkers, indicating compromised tubular function. Abbreviations: SUmod, serum uromodulin; eSCr, enzymatic serum creatinine; CysC, cystatin C; B2M, beta‐2 microglobulin.

TABLE 2.

Key associations and paired diagnostic comparisons.

Analysis	Estimate	SE	p
Higher e‐SCr associated with lower sUmod	−0.0012	0.00033	0.0005
Higher CysC associated with lower sUmod	−0.00089	0.00031	0.0050
Higher B2M associated with lower sUmod	−0.0026	0.0012	0.0242
Abnormal sUmod with normal e‐SCr versus reverse	16 versus 0		0.0193
Abnormal sUmod with normal CysC versus reverse	16 versus 0		0.0193
Abnormal sUmod with normal B2M versus reverse	15 versus 5		0.1873

Note: Mixed‐effects models are based on repeated measurements; paired comparisons use repeated nonindependent observations in non‐HbSC patients.

Abbreviations: B2M, beta‐2‐microglobulin; CysC, cystatin C; e‐SCr, enzymatic serum creatinine; sUmod, serum uromodulin.

We next asked whether sUmod might signal abnormality earlier than conventional blood‐based kidney markers. In paired repeated observations from non‐HbSC patients, abnormal sUmod co‐occurred with normal e‐SCr in 16 of 104 measurements, whereas no observation showed abnormal e‐SCr with normal sUmod (p = 0.0193). The same pattern was observed for CysC: 16 observations had abnormal sUmod despite normal CysC, whereas none showed the reverse (p = 0.0193). For B2M, 15 observations showed abnormal sUmod with normal B2M and 5 showed abnormal B2M with normal sUmod; this difference did not reach statistical significance (p = 0.1873) (Table 2). These paired results suggest that sUmod may become abnormal before e‐SCr or CysC in a clinically relevant subset of patients with incipient nephropathy. Exploratory threshold‐based analyzes using sUmod < 100 ng/mL were directionally consistent showing moderate‐to‐good specificity for reduced eGFR but lower sensitivity for albuminuria (Table S8).

This interpretation is biologically plausible. The kidney medulla, and particularly the thick ascending limb, is highly susceptible to hypoxia and ischemia–reperfusion injury. Chronic medullary hypoxia is a central feature of SCD‐related kidney damage because recurrent sickling in the hyperosmolar, relatively hypoxic environment of the medulla promotes vaso‐occlusion, oxidative stress, and tubular injury [1, 4, 5, 9, 10, 11, 12]. Because uromodulin is produced exclusively by these tubular cells, reduced circulating concentrations may reflect early tubular loss or dysfunction before creatinine‐based filtration estimates become abnormal. In this sense, sUmod complements, rather than replaces, glomerular markers such as albuminuria: the two may capture distinct but related compartments of sickle cell nephropathy.

Our data also suggest that sUmod may provide clinically useful information even when albuminuria is absent or limited. Albuminuria remains important, but it is neither universal in early disease nor specific for tubular injury. In exploratory urine analyzes, pooled observations showed inverse correlation between sUmod and urine albumin‐creatinine ratio, but this relation was not significant in longitudinal mixed‐effects modeling of log‐transformed UACR (Table S7). Relationship among conventional kidney markers, including albuminuria and e‐SCr, is summarized in Table S9. We therefore view sUmod as complementary to albuminuria rather than a surrogate for it. The observation that sUmod was abnormal in repeated measurements despite normal e‐SCr and CysC supports a model in which tubular stress can be detected before overt changes in glomerular markers of filtration abnormalities are apparent. Such a marker could be particularly valuable in young adults with SCD, in whom hyperfiltration and low baseline creatinine often obscure early kidney decline. Baseline characteristics of the HbSS subgroup with sUmod < 100 ng/mL are provided in Table S10.

Exploratory analyzes further suggested that sUmod may be useful for longitudinal monitoring of treatment effects. HC‐treated patients showed numerically more stable sUmod trajectories than untreated patients, although the difference was not statistically significant in this modest cohort (Figure S1). We therefore interpret treatment‐related observations cautiously; they are hypothesis‐generating rather than confirmatory. Still, the direction of effect is clinically interesting because nephron stress in SCD is closely linked to hemolysis, vaso‐occlusion, and medullary hypoxia, processes that effective disease‐modifying therapy may attenuate [9, 10, 11, 12, 13, 14]. A biomarker that reflects tubular integrity could therefore become informative not only for early diagnosis but also for monitoring kidney benefit during therapy.

Our study has limitations. It was retrospective, single‐center, and modest in size. Not all urine measurements were available for every patient; healthy controls were not enrolled contemporaneously, and some subgroup analyzes were necessarily underpowered. Urine‐based analysis, particularly for UACR, should be interpreted cautiously. The HC signal should likewise be considered exploratory. Nevertheless, several features strengthen the main observation: repeated within‐patient measurements, consistent inverse associations across three kidney biomarkers, corroborating eGFR‐based analyzes, and paired analyzes showing that abnormal sUmod frequently preceded abnormalities in e‐SCr and CysC.

In summary, circulating uromodulin identified early kidney abnormalities in adults with SCD and tracked kidney dysfunction across time. The data support sUmod as a kidney‐specific biomarker that may complement established kidney markers, especially when creatinine and CysC remain within conventional ranges. Prospective multicenter studies should determine whether serial sUmod measurement can improve risk stratification, identify tubular injury earlier than current approaches, and support treatment monitoring in sickle cell nephropathy [15, 16].

Author Contributions

Ferras Alashkar and Jürgen E. Scherberich: conceptualization. Ferras Alashkar, Robert Hable, Denise Zwanziger, Raina Yamamoto, Yara Bestmann, and Jürgen E. Scherberich: methodology. Marco Kai, Ferras Alashkar, Alexander Röth, Friederike Poppenborg, Robert Hable, Denise Zwanziger, Raina Yamamoto, Yara Bestmann, and Jürgen E. Scherberich: formal analysis. Marco Kai, Ferras Alashkar, and Jürgen E. Scherberich: writing – original draft preparation. Ferras Alashkar, Robert Hable, Denise Zwanziger, Friederike Poppenborg, Raina Yamamoto, Hans Christian Reinhardt, Alexander Röth, Victor Herbst, Yara Bestmann, Marco Kai, and Jürgen E. Scherberich: writing – review and editing. All authors have read and agreed to the manuscript.

Funding

GBT/Pfizer funded in part the statistical analyzes provided by R.H., however GBT had no role in the collection, analysis, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.

Ethics Statement

All patients provided written informed consent prior to participation. Following consent, clinical and laboratory data were collected, and patients' disease course and kidney function were monitored in the context of a prospective, noninterventional observational study. The study was approved by the Ethics Committee of the University of Duisburg‐Essen (20‐9790‐BO) and was conducted in accordance with the Declaration of Helsinki.

Consent

All patients provided written informed consent prior to participation and for the use of their clinical and laboratory data in this study.

Conflicts of Interest

Ferras Alashkar received honorarium for participation in advisory boards from Avanzanite, Novartis, Global Blood Therapeutics/Pfizer, Vertex and Bristol Myers Squibb, received a speaker honorarium from Novartis, Global Blood Therapeutics/Pfizer, Vertex and Agios Pharmaceuticals and received an honorarium from Novartis, Bristol Myers Squibb, Vertex and Global Blood Therapeutics/Pfizer for consultancy. M.K., Y.B. and V.H. are employees of EUROIMMUN, a company that manufactures diagnostic tests and instruments. M.K., Y.B. and V.H. did not benefit from any potential or actual financial gain as a result of the work. J.E.S. is listed on US patent 16/488.601 EP 18706531.7, University Charite` Berlin. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflicts of interest.

Supporting information

Data S1: Supplementary Appendix. Online‐only supporting material for the correspondence, including Tables S1–S10, providing additional baseline and subgroup characteristics, follow‐up duration and repeated albuminuria measurements, details of two index cases with discordant kidney markers, albuminuria categories, mixed‐effect analyzes, additional eGFR‐ and UACR‐based analyzes, exploratory threshold‐based performance analyzes, relationships among conventional kidney markers, and baseline characteristics of HbSS patients with sUmod < 100 ng/mL.

Table S1: Baseline demographic and disease characteristics (N = 32).

Table S2: Baseline characteristics excluding HbSC patients.

Table S3: Follow‐up duration and repeated albuminuria measurements.

Table S4: Two HbSS index cases with discordant kidney markers.

Table S5: Albuminuria categories in patients without overt reduction in kidney function.

Table S6: Associations of e‐sCr, CysC, and B2M with sUmod.

Table S7: Additional eGFR‐ and UACR‐based analyzes of sUmod.

Table S8: Exploratory threshold‐based performance of sUmod < 100 ng/mL.

Table S9: Associations of CysC, B2M, and albuminuria with e‐SCr.

Table S10: Baseline characteristics of HbSS patients with sUmod < 100 ng/mL.

Figure S1: Association of hydroxycarbamide (HC) therapy with serum uromodulin (sUmod) levels in adult sickle cell disease (SCD) patients. While overall sUmod levels declined slightly in all patients, HC‐treated individuals maintained higher more stable levels, suggesting a potential protective effect on tubular function. Model: Linear mixed model; HC‐treated: β = 0.0028, p = 0.9148; non‐HC‐treated: β = −0.014, SE = 0.035, p = 0.7031.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.