Acute kidney injury after ruxolitinib: Common complication, uncommon cause

To the Editor:

Ruxolitinib is a selective inhibitor of janus associated kinase 1 and 2 that is approved for treatment of intermediate or high-risk myelofibrosis, polycythemia vera, and acute steroid-refractory graft vs host disease (aGVHD).^1-3 Early trials leading to its approval did not report nephrotoxicity; however, a recent trial in patients with steroid-refractory aGVHD reported 18% of participants experienced elevated serum creatintine, and 4% of patients experienced a severe adverse event due to acute kidney injury (AKI).⁴ A query of the food and drug administration adverse events reporting system (FAERS) Dashboard as of October 2019 showed 426 “Renal Injury” reports for ruxolitinib.⁵

To address whether or not ruxolitinib causes unexplained AKI, we identified all patients prescribed ruxolitinib in the Partners healthcare network, which includes patients seen at Massaschusetts General Hospital Cancer Center and Dana Farber Cancer Institute in Boston, MA. Patients were identified by querying the Partners Research Patient Data Registry. Our primary outcome was moderate-to-severe AKI (at least 2-fold rise in creatinine from baseline) occurring within 90 days of starting ruxolitinib. Two nephrologists reviewed charts of all cases to determine the primary etiology of AKI. We categorized AKI into the following etiologies: (a) prerenal azotemia occurring from volume depletion and rapidly reversing with hydration, (b) hemodynamic intrinsic AKI due to sepsis or shock lasting >48 hours despite supportive measures, (c) nephrotoxic intrinsic AKI related to a known nephrotoxic agent that improves upon discontinuation, and (d) unexplained AKI occurring in patients without hemodynamic disturbances or exposure to alternative nephrotoxins. We excluded patients who did not start ruxolitinib, did not have at least one baseline and follow-up creatinine, were less than 18 years old, or were on dialysis. Baseline characteristics, comorbidities, and concomitant medications were determined by chart review. The Partners IRB approved this study and waived the need for informed consent.

We identified and reviewed all patients with a prescription for ruxolitinib between 2012-2019 (N = 414) to determine the start date and baseline creatinine. After applying the exclusion criteria, 311 patients were included in the final analysis. Baseline characteristics are shown in Table 1. Mean age was 64 years, 56% were male, and 89% were white. The most common indication for ruxolitinib was aGVHD (47%); the breakdown of other indications is shown in Table 1. Among patients with aGVHD, 66 (46%) were treated with calcinurin inhibitors and 24 (17%) received sirolimus concurrently. Moderate-to-severe AKI occurred within 90 days of starting ruxolitinib in 30 patients (10%), 25 of whom (83%) were being treated for aGVHD. A univariable comparison of the baseline characteristics of patients with and without moderate-to-severe AKI is shown in Table 1. On average, patients with AKI were taking lower doses of ruxolitinib, typically because lower doses are used to treat aGVHD, due to baseline cytopenias and drug-drug interactions in the population. Review of the medical record uncovered clear alternative causes for all but one case of AKI. Of the 30 patients with moderate-to-severe AKI, 13 (43%) had hemodynamic intrinsic AKI from sepsis or shock, 12 (40%) had prerenal azotemia that resolved with intravenous fluid administration, and four (13%) had AKI due to other nephrotoxins (trimethoprim-sulfamethoxazole, deferasirox, or supratherapeutic tacrolimus in two patients). Thirteen of the 30 patients experienced severe AKI (stage 3 AKI defined by ≥3-fold rise in creatinine) and seven required renal replacement therapy. We found only one case of unexplained AKI (3% of AKI, 0.3% incidence in entire cohort). The patient was an elderly, white female with stage 4 chronic kidney disease attributed to hypertension and a baseline creatinine of 2.5 mg/dL (estimated glomerular filtration rate [eGFR] of 16 mL/min/1.73m²). She received ruxolitinib 5 mg daily, dose reduced for eGFR, to treat myelofibrosis. Six weeks after ruxolitinib initiation, her serum creatinine increased to 3.8 mg/dL. There were no concurrent illnesses, hemodynamic disturbances, or nephrotoxin use noted. Ruxolitinib was continued and her serum creatinine continually increased over the next 3 months to 5.7 mg/dL; she was offered dialysis but instead entered hospice and passed away 5.5 months after starting ruxolitinib. The patient had been evaluated by a nephrologist, but a kidney biopsy was not performed.

TABLE 1.

Baseline characteristics of patients receiving ruxolitinib

	Overall cohort	Moderate- to-severe AKI	No AKI within 90 d
Characteristics	N = 311	n = 30	n = 281
Age, n (%)
22-65	150 (48)	18 (63)	131 (46)
≥66	161 (52)	12 (36)	150 (53)
Gender, n (%)
Male	174 (56)	16 (53)	158 (56)
Race, n (%)*
White	278 (89)	23 (77)	255 (91)
Asian	7 (2)	2 (7)	5 (2)
Black	5 (2)	1 (3)	4 (1)
Hispanic	5 (2)	2 (7)	3 (1)
Other	16 (5)	2 (7)	14 (5)
Cancer Type, n (%)*
aGVHD	145 (47)	25 (83)	120 (43)
HLHa	9 (3)	2 (7)	7 (3)
Polycythemia vera	51 (16)	2 (7)	49 (17)
Myelofibrosisb	93 (30)	1 (3)	92 (33)
Other	13 (4)	0 (0)	13 (4)
Baseline kidney function, n (%)*
eGFR >90 mL/min/1.73m2	102 (33)	18 (60)	84 (30)
eGFR 60-89 mL/min/1.73m2	128 (41)	6 (20)	122 (43)
eGFR 30-59 mL/min/1.73m2	70 (23)	5 (17)	65 (23)
eGFR <30 mL/min/1.73m2	11 (4)	1 (3)	10 (4)
Medical comorbidities
CAD, n (%)	36 (12)	1 (3)	35 (12)
Diabetes (%)	73 (23)	9 (30)	64 (23)
HTN, n (%)	209 (67)	17 (57)	192 (68)
Baseline medication use
ACEi/ARB, n (%)	45 (14)	5 (17)	40 (14)
NSAIDS, n (%)	18 (6)	0 (0)	18 (6)
TMP-SMX, n (%)	73 (23)	9 (30)	64 (23)
PPI, n (%)	117 (38)	12 (40)	105 (37)
CNI, n (%)*	76 (24)	15 (50)	61 (23)
Sirolimus, n (%)*	26 (8)	8 (27)	18 (6)
Diuretic, n (%)	53 (17)	4 (13)	49 (17)
Ruxolitinib dose/d (mg), mean (SD)*	20 (10)	16 (6)	21 (10)

^(*)Note: Baseline characteristics of patients receiving ruxolitinib broken down by the overall cohort and by AKI status. AKI defined by at least a 2-fold rise in serum creatinine within 90 d of starting ruxolitinib. Any variable that was statistically significant in univariable comparison is shown with an asterisk .

Abbreviations: ACEi/ARB, angiotensin converting enzyme inhibitor or angiotensin receptor blockade; aGVHD, acute graft versus host disease; CAD, coronary artery disease; CNI, calcineurin inhibitor; eGFR, estimated glomerular filtration rate; HLH, hemophagocytic lymphohistiocytosis; HTN, hypertension; NSAIDS, nonsteroidal anti-inflammatory drugs; PPI, proton pump inhibitors; SD, standard deviation; TMP-SMX, trimethoprim-sulfamethoxazole.

^aThe causes of HLH were genetic in three, associated with cancer in two (leukemia and lymphoma), infection in two, due to autoimmune disease in one, and unknown in one.

^bMyelofibrosis cases included both primary myelofibrosis and post-essential thrombocythemia myelofibrosis. The majority of patients prescribed calcineurin inhibitors were taking tacrolimus (N = 74); two were prescribed cyclosporine.

Our study is the first to systematically evaluate kidney function in patients receiving ruxolitinib and included all patients seen at two major cancer centers over an eight-year period. We detected a very low rate of unexplained moderate-to-severe AKI (0.3%) among consecutive recipients of ruxolitinib. Our study has several limitations, particularly the potential bias of retrospective series. We only evaluated moderate-to-severe cases of AKI and it is possible we may have missed milder cases of unexplained AKI due to ruxolitinib. Because we only reviewed AKI events that occurred in the first 90 days after starting ruxolitinib, we may have missed events that occurred more than 90 days after starting ruxolitinib if there is a cumulative dose effect that leads to AKI. Additionally, our study included only small numbers of patients with advanced CKD, because the kidney eliminates both active and inactive metabolites of ruxolitinib, rates of adverse events could be higher in this group. Finally, “ruxolitinib withdrawal syndrome”, which may cause a septic shock-like hemodynamic compromise, could potentially lead to AKI; however, only two of our patients had discontinued ruxolitinib prior to developing AKI, and they had undergone an appropriately slow taper.

In conclusion, clinicians should be reassured that, despite a large number of reports in the FAERS system of worsening renal function in patients receiving ruxolitinib, our study detected a very low rate of unexplained moderate-to-severe AKI (0.3%) among consecutive recipients of ruxolitinib. In the complex mileu of aGVHD after hematopoetic stem cell transplant, ruxolitinib did not appear to be a culprit of AKI in our series. AKI was rare in patients receiving ruxolitinib for myelofibrosis and polycythemia vera; indeed, a recent study showed that ruxolitinib improved kidney function in a substantial number of patients with primary myelofibrosis.⁶ Given that rare cases of unexplainted AKI may still occur, any case of unexplained AKI after ruxolitinib warrants further investigation including a nephrology referral and consideration of kidney biopsy to determine the underlying pathology.