TO THE EDITOR:

Erythrocytosis referrals to hematology are frequently triggered by hemoglobin and hematocrit (Hgb/Hct) levels that exceed conventional criteria thresholds for polycythemia vera (PV), (Hgb/Hct levels above >16.5 g/dL [49%] in males and >16 g/dL [48%] in females).1 Foremost, PV should always be excluded by the absence of a JAK2 mutation (V617F exons 14 and 12) and/or a subnormal serum erythropoietin (EPO) level,2 after which a systematic workup for secondary erythrocytosis is recommended.3 The latter includes a review of medications associated with erythrocytosis: testosterone, androgen, erythropoiesis-stimulating agents, diuretics, and contemporary antidiabetic agents (sodium-glucose cotransporter-2 [SGLT-2] inhibitors: canagliflozin, empagliflozin, dapagliflozin, and ertugliflozin). In addition to management of diabetes mellitus, the US Food and Drug Administration’s approval of SGLT-2 inhibitors has recently expanded to include patients with heart failure and chronic renal disease.4-6 Importantly, studies have implied shared molecular mechanisms for SGLT-2 inhibitor–associated erythrocytosis and cardioprotection, whereby a drug-induced hypoxic and nutrient-deprived state activates sirtuin 1 and hypoxia-inducible factor (HIF-2α and -1α).7,8 The current literature on the subject is limited to 5 cases of marked erythrocytosis (Hct >53%) induced by SGLT-2 inhibitor therapy, 2 of which occurred in the context of coadministration of testosterone9; 1 case resulted in the unmasking of PV,10 whereas the remaining 2 were without another predisposing factor.11 Therefore, we sought to investigate the presenting features, degree, and duration of erythrocytosis and the management strategies and outcomes of 30 consecutive patients with SGLT-2 inhibitor–induced erythrocytosis who underwent hematological evaluation at our institution from December 2015 to August 2021.

Patients with concomitant erythrocytosis defined by World Health Organization 2016 thresholds of Hgb/Hct levels >16.5 g/dL (49%) in males and >16 g/dL (48%) in females1 and by receiving SGLT-2 inhibitor therapy for management of diabetes mellitus were retrospectively recruited after obtaining institutional review board approval. Foremost, PV was excluded by negative JAK2 mutation analysis in all cases, along with negative CALR and MPL mutation testing in 11 informative cases, after which an evaluation for secondary erythrocytosis was performed that, at the minimum, included abdominal imaging and overnight oximetry. In addition, bone marrow examination and hereditary erythrocytosis testing was obtained in 3 and 2 patients, respectively.

Baseline Hgb/Hct levels before initiation of the SGLT-2 inhibitor and peak Hgb/Hct levels during therapy were carefully recorded. The duration of therapy was noted, with particular attention to Hgb/Hct levels after discontinuation of therapy. Thrombotic events included major arterial thrombosis, myocardial infarction, angina, cerebrovascular accidents, peripheral artery occlusion or major venous thromboses, deep vein thrombosis, and pulmonary embolism. Therapeutic interventions were dependent on physician discretion and included phlebotomy, blood donation, antiplatelet therapy, and systemic anticoagulation.

A total of 30 patients with SGLT-2 inhibitor–associated erythrocytosis (median age, 64 years; range, 29-81; 67% men) were studied. Table 1 provides details of clinical and laboratory features, including drug dosage and duration of therapy for all patients. At the time of erythrocytosis evaluation, the majority of the patients were receiving empagliflozin (n = 17; 57%), followed by canagliflozin (n = 9; 30%) and dapagliflozin (n = 4; 13%). Erythrocytosis was identified in 8 (27%) patients who had an additional predisposition which included obstructive sleep apnea, alone (n = 5) or in association with smoking (n = 2), and smoking alone (n = 1). All patients with concomitant obstructive sleep apnea underwent sleep evaluation to ensure the absence of hypoxia.

Table 1.

Clinical and laboratory characteristics of 30 patients with SGLT2 inhibitor associated JAK2 wild-type erythrocytosis

Variables at time of erythrocytosis evaluationData
(N = 30)
Age in years, median (range) 64 (29-81) 
Men, n (%) 20 (67) 
SGLT2 inhibitor and dose, n (%)
 Canagliflozin 300 mg
 Canagliflozin 100 mg
 Canagliflozin 100 mg/Empagliflozin* 10 mg
 Canagliflozin 100 mg/Dapagliflozin* 10 mg
 Canagliflozin 100 mg/Empagliflozin 10mg/Dapagliflozin* 10 mg
 Empagliflozin 10 mg
 Empagliflozin 12.5 mg
 Empagliflozin 25 mg
 Dapagliflozin 10 mg
 Empagliflozin/linagliptin 25/5 mg
Duration of therapy in years, median (range) 

6 (20)
3 (10)
2 (7)
1 (3)
1 (3)
6 (20)
1 (3)
7 (23)
2 (7)
1 (3)
1.7 (0.2-5.7) 
Additional predisposition for erythrocytosis, n (%)
 OSA
 OSA+smoking
 Smoking 
8 (27)
5 (17)
2 (7)
1 (3) 
Prior thrombosis, n(%)
 Major arterial thrombosis
 Major venous thrombosis 
4 (13)
1 (3)
3 (10) 
Baseline hemoglobin g/dL, median (range)
 Male
 Female 
15.8 (13.4-16.9)
16 (13.4-16.9)
14.5 (13.4-16.5) 
Baseline hematocrit %, median (range)§
 Male
 Female 
45.7 (39-48.2)
46.1 (40.7-48.2)
42.6 (39-46.3) 
Peak hemoglobin g/dL, median (range)
 Male
 Female 
17.9 (15.3-19.8)
18.3 (16.5-19.8)
16.9 (15.3-18) 
Peak hematocrit %, median (range)
 Male
 Female
 Increase in hematocrit %, median (range)
 Time to peak hematocrit in months, median (range) 
52.9 (48-60.1)
54.4 (50.7-60.1)
51.2 (48-52.8)
7.4 (2-14.1)
11.5 (1-57) 
Serum erythropoietin mIU/mL, median (rangeǁ 10.3 (6.1-37.3) 
Absolute reticulocyte count, median (range) 83.1 (83.1-135) 
Serum ferritin μg/L, median (range)# 54.5 (9-840) 
Treatment, n (%)
 Phlebotomy/blood donation
 Antiplatelet therapy (aspirin or clopidogrel)
 Anticoagulation 

7 (23)
14 (47)
5 (17) 
Outcomes  
 Follow-up in years, median (range) 2.2 (0.2-5.7) 
 Thrombosis during therapy, n (%)
  Major arterial thrombosis, n (%)
  Major venous thrombosis, n (%) 
2 (7)
2 (7)
0 (0) 
 Treatment discontinuation, n (%)
Reason for discontinuation, n (%)
  Erythrocytosis
  Poorly controlled diabetes mellitus
  Yeast infection 
7 (23)
3 (43)
3 (43)
1 (14) 
 Resolution of erythrocytosis, n (%)**
  Median hemoglobin g/dL, median (range)
  Median hematocrit %, median (range)
  Time to resolution in months, median (range) 
6 (100)
15.3 (14.5-16.1)
45.4 (42.5-47)
2 (1-12) 
Variables at time of erythrocytosis evaluationData
(N = 30)
Age in years, median (range) 64 (29-81) 
Men, n (%) 20 (67) 
SGLT2 inhibitor and dose, n (%)
 Canagliflozin 300 mg
 Canagliflozin 100 mg
 Canagliflozin 100 mg/Empagliflozin* 10 mg
 Canagliflozin 100 mg/Dapagliflozin* 10 mg
 Canagliflozin 100 mg/Empagliflozin 10mg/Dapagliflozin* 10 mg
 Empagliflozin 10 mg
 Empagliflozin 12.5 mg
 Empagliflozin 25 mg
 Dapagliflozin 10 mg
 Empagliflozin/linagliptin 25/5 mg
Duration of therapy in years, median (range) 

6 (20)
3 (10)
2 (7)
1 (3)
1 (3)
6 (20)
1 (3)
7 (23)
2 (7)
1 (3)
1.7 (0.2-5.7) 
Additional predisposition for erythrocytosis, n (%)
 OSA
 OSA+smoking
 Smoking 
8 (27)
5 (17)
2 (7)
1 (3) 
Prior thrombosis, n(%)
 Major arterial thrombosis
 Major venous thrombosis 
4 (13)
1 (3)
3 (10) 
Baseline hemoglobin g/dL, median (range)
 Male
 Female 
15.8 (13.4-16.9)
16 (13.4-16.9)
14.5 (13.4-16.5) 
Baseline hematocrit %, median (range)§
 Male
 Female 
45.7 (39-48.2)
46.1 (40.7-48.2)
42.6 (39-46.3) 
Peak hemoglobin g/dL, median (range)
 Male
 Female 
17.9 (15.3-19.8)
18.3 (16.5-19.8)
16.9 (15.3-18) 
Peak hematocrit %, median (range)
 Male
 Female
 Increase in hematocrit %, median (range)
 Time to peak hematocrit in months, median (range) 
52.9 (48-60.1)
54.4 (50.7-60.1)
51.2 (48-52.8)
7.4 (2-14.1)
11.5 (1-57) 
Serum erythropoietin mIU/mL, median (rangeǁ 10.3 (6.1-37.3) 
Absolute reticulocyte count, median (range) 83.1 (83.1-135) 
Serum ferritin μg/L, median (range)# 54.5 (9-840) 
Treatment, n (%)
 Phlebotomy/blood donation
 Antiplatelet therapy (aspirin or clopidogrel)
 Anticoagulation 

7 (23)
14 (47)
5 (17) 
Outcomes  
 Follow-up in years, median (range) 2.2 (0.2-5.7) 
 Thrombosis during therapy, n (%)
  Major arterial thrombosis, n (%)
  Major venous thrombosis, n (%) 
2 (7)
2 (7)
0 (0) 
 Treatment discontinuation, n (%)
Reason for discontinuation, n (%)
  Erythrocytosis
  Poorly controlled diabetes mellitus
  Yeast infection 
7 (23)
3 (43)
3 (43)
1 (14) 
 Resolution of erythrocytosis, n (%)**
  Median hemoglobin g/dL, median (range)
  Median hematocrit %, median (range)
  Time to resolution in months, median (range) 
6 (100)
15.3 (14.5-16.1)
45.4 (42.5-47)
2 (1-12) 

OSA, obstructive sleep apnea.

*

SGLLT2 inhibitor at the time of erythrocytosis evaluation.

Major arterial thrombosis included cerebrovascular accident, transient ischemic attack, myocardial infarction, angina, peripheral arterial thrombosis. Major venous thrombosis included deep vein thrombosis, pulmonary embolism, splanchnic vein thrombosis.

Reference range: women (11.6-15 g/dL), men (13.2-16.6 g/dL).

§

Reference range: women (35.5% to 44.9%), men (38.3% to 48.6%).

ǁ

n = 27. Reference range: 2.6-18.5 mIU/mL.

n = 3. Reference range: 30.4-110.9 × 109/L.

#

n = 9. Reference range: 24-336 μg/L.

**

n = 6.

A history of thrombosis was documented in 4 (13%) patients, 3 of whom had had provoked venous events in the form of deep venous thrombosis/pulmonary embolism, whereas 1 had experienced a prior cerebrovascular accident. Median and range for baseline Hgb/Hct levels before initiation of therapy were 15.8 g/dL (13.4-16.9 g/dL) and 45.7% (39% to 48.2%), respectively. Patients received SGLT-2 inhibitor therapy for a median of 1.7 years (range, 0.2-5.7 years); the corresponding levels of peak Hgb/Hct during therapy were 17.9 g/dL (15.3-19.8 g/dL) and 52.9% (48% to 60.1%). Notably, a substantial increase in Hct level (median, 7.4%; range, 2% to 14.1%) was noted; with time to peak Hct of 11.5 months (range, 1-57 months) while receiving therapy.

A dose-dependent correlation with peak Hct level was demonstrated in patients receiving canagliflozin (100 mg vs 300 mg; P = .04), but not in patients receiving empagliflozin (10 mg vs 25 mg; P = .72).

Serum erythropoietin measurement was obtained in 27 patients and was found to be inappropriately normal or increased in all cases, with a median of 10.3 mIU/mL (range, 6.1-37.3 mIU/mL). In contrast, the serum ferritin level showed variability (median, 54.5; range, 9-840 μg/L), although the level was available in only a subset of patients (n = 9), of which 3 underwent genetic testing to rule out hemochromatosis.

Management consisted of regular blood donation in 3 patients, whereas monthly phlebotomy was instituted in 4 cases, with a target Hct <52%. Half of the patients (n = 14) received antiplatelet therapy that included aspirin (n = 12) and clopidogrel (n = 2), whereas 5 received systemic anticoagulation (3 for prior venous thromboses), and 2 received primary prophylaxis for atrial fibrillation. At a median follow-up of 2.2 years (range, 0.2-5.7), thrombotic complications in the form of unstable angina occurred in 2 male patients (7%) aged 56 and 60 years, despite the use of phlebotomy in both cases. However, one of them was not receiving aspirin therapy at the time of the event.

Therapy with the SGLT-2 inhibitor was discontinued in a minority of patients (n = 7; 23%). Reasons for discontinuation included erythrocytosis (n = 3), poorly controlled diabetes mellitus (n = 3), and yeast infection (n = 1). Resolution of erythrocytosis was documented in all 6 patients in whom Hgb/Hct levels were available; median and range Hgb/Hct levels were 15.3 g/dL (14.5-16.1 g/dL) and 45.4% (42.5% to 47%), respectively, with median time to resolution of 2 months (range, 1-12 months) after discontinuation of therapy.

Our observations highlight the self-limiting characteristics of SGLT-2 inhibitor-induced erythrocytosis. Despite a substantial increase in Hct level, discontinuation of therapy led to universal resolution of the disease erythrocytosis.11,12 Treatment with empagliflozin in the context of a large, placebo-controlled study (n = 7020) resulted in an increase in Hgb/Hct levels, with a baseline increase in Hct levels of 4.8% ± 5.5% in patients treated with empagliflozin 10 mg and 5.0% ± 5.3% in those receiving empagliflozin 25 mg, vs 0.9% ± 4.7% in the placebo arm.13 Similarly, Hgb changes were 0.8 ± 1.3 g/dL with empagliflozin vs −0.1 ± 1.2 g/dL with placebo.13 All 30 patients in our study had baseline Hgb/Hct levels within the normal range that predisposes to development of erythrocytosis during SGLT-2 inhibitor therapy.

The putative mechanisms underlying erythrocytosis include increased EPO production via hypoxia-induced activation of HIF2α, modulation of iron metabolism through hepcidin, and hemoconcentration.7,14 Our findings underscore the lack of uniformity in the diagnostic workup and management of secondary erythrocytosis related to unsubstantiated concerns regarding heightened thrombosis risk, in view of which 10% of patients underwent bone marrow examination, despite negative JAK2 testing and normal/elevated EPO levels. Moreover 23% of the patients underwent regular phlebotomy, consistent with a recent retrospective report in which ∼33% of patients with secondary erythrocytosis were subjected to bone marrow examination and 42% to periodic phlebotomy.15 

Notwithstanding, the inherent limitations of a retrospective study, specifically because Hb/Hct levels were not obtained at predefined intervals, this is the first series to focus on SGLT-2 inhibitor–induced erythrocytosis. Given the foreseeable increase in erythrocytosis referrals within a hematology practice, our study prompts formal recognition of SGLT-2 inhibitor–induced erythrocytosis within the diagnostic algorithm. Nonetheless, implementation of phlebotomy in such situations is unlikely to be beneficial, given that the development of erythrocytosis correlates with the cardioprotection conferred by the drug.16 

Contribution: N.G. and A.T. designed the study, compiled the data, performed the analyses, and wrote the paper; N.S. developed the study concept; H.A., A.A., and A.P. recruited and enrolled the patients; and all authors reviewed the final draft of the paper.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Naseema Gangat, Division of Hematology, Department of Medicine, Mayo Clinic, 200 First St SW, Rochester, MN 55905; e-mail: gangat.naseema@mayo.edu; and Ayalew Tefferi, Division of Hematology, Department of Medicine, Mayo Clinic, 200 First St SW, Rochester, MN 55905; e-mail: tefferi.ayalew@mayo.edu.

1.
Arber
DA
,
Orazi
A
,
Hasserjian
R
, et al
.
The 2016 revision to the World Health Organization classification of myeloid neoplasms and acute leukemia
.
Blood.
2016
;
127
(
20
):
2391
-
2405
.
2.
Tefferi
A
,
Vannucchi
AM
,
Barbui
T
.
Polycythemia vera: historical oversights, diagnostic details, and therapeutic views [published online ahead of print 3 September 2021]
.
Leukemia.
doi:10.1038/s41375-021-01401-3.
3.
Gangat
N
,
Szuber
N
,
Pardanani
A
,
Tefferi
A
.
JAK2 unmutated erythrocytosis: current diagnostic approach and therapeutic views
.
Leukemia.
2021
;
35
(
8
):
2166
-
2181
.
4.
McMurray
JJV
,
Solomon
SD
,
Inzucchi
SE
, et al;
DAPA-HF Trial Committees and Investigators
.
Dapagliflozin in patients with heart failure and reduced ejection fraction
.
N Engl J Med.
2019
;
381
(
21
):
1995
-
2008
.
5.
Packer
M
,
Anker
SD
,
Butler
J
, et al;
EMPEROR-Reduced Trial Investigators
.
Cardiovascular and renal outcomes with empagliflozin in heart failure
.
N Engl J Med.
2020
;
383
(
15
):
1413
-
1424
.
6.
Heerspink
HJL
,
Stefánsson
BV
,
Correa-Rotter
R
, et al;
DAPA-CKD Trial Committees and Investigators
.
Dapagliflozin in patients with chronic kidney disease
.
N Engl J Med.
2020
;
383
(
15
):
1436
-
1446
.
7.
Mazer
CD
,
Hare
GMT
,
Connelly
PW
, et al
.
Effect of empagliflozin on erythropoietin levels, iron stores, and red blood cell morphology in patients with type 2 diabetes mellitus and coronary artery disease
.
Circulation.
2020
;
141
(
8
):
704
-
707
.
8.
Sano
M
,
Goto
S
.
Possible mechanism of hematocrit elevation by sodium glucose cotransporter 2 inhibitors and associated beneficial renal and cardiovascular effects
.
Circulation.
2019
;
139
(
17
):
1985
-
1987
.
9.
Motta
G
,
Zavattaro
M
,
Romeo
F
,
Lanfranco
F
,
Broglio
F
.
Risk of erythrocytosis during concomitant testosterone and SGLT2-inhibitor treatment: a warning from two clinical cases
.
J Clin Endocrinol Metab.
2019
;
104
(
3
):
819
-
822
.
10.
Das
L
,
Bhansali
A
,
Walia
R
.
Unmasking and aggravation of polycythemia vera by canagliflozin
.
Diabet Med.
2018
;
35
(
11
):
1613
-
1616
.
11.
Gupta
R
,
Gupta
A
,
Shrikhande
M
,
Tyagi
K
,
Ghosh
A
,
Misra
A
.
Marked erythrocytosis during treatment with sodium glucose cotransporter-2 inhibitors-report of two cases
.
Diabetes Res Clin Pract.
2020
;
162
:
108127
.
12.
Chin-Yee
B
,
Solh
Z
,
Hsia
C
.
Erythrocytosis induced by sodium-glucose cotransporter-2 inhibitors
.
CMAJ.
2020
;
192
(
42
):
E1271
.
13.
Zinman
B
,
Wanner
C
,
Lachin
JM
, et al;
EMPA-REG OUTCOME Investigators
.
Empagliflozin, cardiovascular outcomes, and mortality in type 2 diabetes
.
N Engl J Med.
2015
;
373
(
22
):
2117
-
2128
.
14.
Ghanim
H
,
Abuaysheh
S
,
Hejna
J
, et al
.
Dapagliflozin suppresses hepcidin and increases erythropoiesis
.
J Clin Endocrinol Metab.
2020
;
105
(
4
):
dgaa057
.
15.
Nguyen
E
,
Harnois
M
,
Busque
L
, et al
.
Phenotypical differences and thrombosis rates in secondary erythrocytosis versus polycythemia vera
.
Blood Cancer J.
2021
;
11
(
4
):
75
.
16.
Packer
M
.
SGLT2 inhibitors produce cardiorenal benefits by promoting adaptive cellular reprogramming to induce a state of fasting mimicry: a paradigm shift in understanding their mechanism of action
.
Diabetes Care.
2020
;
43
(
3
):
508
-
511
.