חומר רקע
CLINICAL PHARMACOLOGY & THERAPEUTICS | VOLUME 0 NUMBER 0 | Month 2024
1
Medications and Acute Hemolysis in
G6PD-Deficient Patients – A Real-World Study
Naomi Gronich1,2,*
, Bar Rosh2
, Nili Stein1
and Walid Saliba1,2,3
Many drug labels contain precautions of use in G6PD-deficient patients due to hemolytic concerns, but much of this
is based on scarce clinical, epidemiological, or structural data. In this real-world study, we aimed to examine if the
administration of presumably risky medications for G6PD-deficient patients was followed by hemolysis. The study is
based on data from Clalit Health Services database that provides inclusive health care for more than half of the Israeli
population (~ 4.7 million). Within the database, we identified all G6PD-deficient patients by G6PD <6 U/g Hb. Within the
G6PD-deficient cohort, we identified all hospitalizations with a discharge diagnosis of hemolysis (January 1, 2010 to
December 31, 2022), validated the cases, and identified the culprit event. For the rest of the G6PD-deficient patients
with no-hemolysis, we recorded filled prescriptions of medications listed as presumably risky. We identified 31,962
G6PD-deficient patients. Within the cohort, there were 71 cases of major hemolysis requiring hospitalization (0.2%
of the cohort), of whom 51 (71.8%) had been caused by ingestion of fava beans, six (8.5%) were associated with an
infection, and three (4.2%) suggested to be associated with medications (nitrofurantoin, phenazopyridine, and a “pain
killer”). Within the 31,875 patients with no major hemolysis, nitrofurantoin has been prescribed safely to 1,366 G6PD-
deficient males and females; hundreds/thousands of G6PD-deficient patients had been prescribed safely ciprofloxacin,
glibenclamide, ofloxacin, phenazopyridine, sulfamethoxazole/cotrimoxazole, sulfasalazine, hydroxychloroquine,
glimepiride, mesalazine, and sulfacetamide. In this real-world study, we are showing that a list of medications,
suspected previously as carrying risks for hemolysis in G6PD-deficient patients, have been prescribed safely to G6PD-
deficient patients, providing reassurance to patients, prescribers, and regulators.
Glucose-6-phosphate dehydrogenase (G6PD) deficiency is a com-
mon X-linked inherited enzyme deficiency, with over 150 reported
alleles. The mutation is estimated to be present in almost 6% of the
world’s population,1,2 with higher prevalence in populations orig-
inating in the tropical and sub-tropical areas of the Eastern hemi-
sphere, and as high as 60–70% carriage among Kurdish Jews.3,4
Received February 25, 2024; accepted May 20, 2024. doi:10.1002/cpt.3333
1Department of Community Medicine and Epidemiology, Lady Davis Carmel Medical Center, Haifa, Israel; 2Ruth and Bruce Rappaport Faculty
of Medicine, Technion-Israel Institute of Technology, Haifa, Israel; 3Research Authority, Lady Davis Carmel Medical Center, Haifa, Israel.
*Correspondence: Naomi Gronich ([email protected])
Study Highlights
WHAT IS THE CURRENT KNOWLEDGE ON THE
TOPIC?
; Glucose-6-phosphate dehydrogenase (G6PD) deficiency is a
common X-linked inherited enzyme deficiency. Oxidative stress
events put carriers at risk, as they might serve as potential cata-
lysts for acute hemolytic anemia events. Food and medications
may precipitate the acute hemolytic event in G6PD-deficient
patients, hence, precaution and awareness are needed to avoid
these precipitating factors.
; Many drug labels contain warnings or precautions of use in
G6PD-deficient patients due to hemolytic risk, but, much of
this is based on scarce clinical and epidemiological studies or
even on the resemblance of a medication with a group of medi-
cations known to precipitate hemolysis.
WHAT QUESTION DID THIS STUDY ADDRESS?
; Was the administration of presumably risky medications to
G6PD-deficient patients followed by a hemolytic event, in the
real world?
WHAT
DOES
THIS
STUDY
ADD
TO
OUR
KNOWLEDGE?
; Hundreds or thousands of G6PD-deficient males and fe-
males (G6PD activity level of 1–6 U/g Hb; some with severe
deficiency G6PD < 1 U/g Hb) had been prescribed safely ni-
trofurantoin, ciprofloxacin, glibenclamide, ofloxacin, phena-
zopyridine, sulfamethoxazole/cotrimoxazole, sulfasalazine,
hydroxychloroquine, glimepiride, mesalazine, and sulfaceta-
mide that appeared on various lists as carrying risk for hemoly-
sis in G6PD-deficient patients, or that their risk was suspected
but was unknown due to lack of data.
HOW MIGHT THIS CHANGE CLINICAL PHARMA-
COLOGY OR TRANSLATIONAL SCIENCE?
; From real-world data, it seems that nitrofurantoin, cip-
rofloxacin,
glibenclamide,
ofloxacin,
phenazopyridine,
sulfamethoxazole/cotrimoxazole, sulfasalazine, hydroxychloro-
quine, glimepiride, mesalazine, and sulfacetamide might safely
be administered to male and female G6PD-deficient patients.
ARTICLE
VOLUME 0 NUMBER 0 | Month 2024 | www.cpt-journal.com
2
G6PD plays a key role in the production of ribose
5-phosphate, and the generation of nicotinamide adenine di-
nucleotide phosphate (NADPH) in the hexose monophos-
phate pathway, which is the main NADPH-generation process
in mature red cells (lacking the citric acid cycle).4,5 Oxidative
stress events put carriers at risk, as they might serve as potential
catalysts for an acute hemolytic anemia event. Acute hemolytic
anemia may manifest with a wide range of severity depending
also on the mutation variant.4 In heterozygous females the ex-
pression, and therefore the potential severity of hemolysis is
highly variable, subject to the phenomenon of X-chromosome
inactivation.4 However, in general, males are many times more
severely affected than women. A universal screening newborn
biochemical test to identify G6PD deficiency had been started
in Israel in mid-2021.
Foods and medications may precipitate an acute hemolytic
event in G6PD-deficient patients, hence, awareness and precaution
are needed to avoid these precipitating factors.4 Several antibiotics
have been associated with hemolysis in G6PD-deficient patients,
although it is difficult to assess the role of infection distinctively
as the cause for the oxidative stress and hemolysis, in antibiotic-
treated patients.4,6,7
More than fifty medications have been listed as carrying a
risk for hemolysis in G6PD-deficient patients. Thus, many drug
labels contain warnings or precautions to be taken in G6PD-
deficient patients due to hemolytic risk, but, many of these are
based on scarce clinical and epidemiological studies or even on
the resemblance of a medication with a group of medications
known to precipitate hemolysis (for example: the label of glime-
piride8). This and more, the “risky” medications- lists from dif-
ferent sources are not identical4,6 (Table S1). Out of this long
list, only seven drugs (dapsone, methylthioninium chloride, ni-
trofurantoin, phenazopyridine, primaquine, rasburicase, and to-
lonium chloride) were listed as carrying hemolytic risk following
a broad literature review in 2002.7 Clinical Pharmacogenetics
Implementation Consortium (CPIC) guideline9 lists seven
high-risk medications, two medium-risk medications, 26 low-
to-no risk medications, and 16 drugs for which recommendation
could not be made due to lack of even one related case report in
the searchable literature. In the paucity of valid data on which
to establish recommendations whether to allow treatment with
presumably risky medications, patients might be subjected to a
great amount of uncertainty and distress, and prescribers might
be reluctant to prescribe these medications.
In this retrospective cohort study, we aimed to examine, in real-
world data, if administration of presumably risky medications for
G6PD-deficient patients was followed by a hemolytic event. In
particular, we aimed to “clear” the risk in medications with low or
unknown risk, and to describe the drugs that caused hemolysis in
G6PD-deficient patients.
METHODS
The study was approved by the institutional review board of Lady Davis
Medical Center (CMC-0137-22) and the Data Utilization Committee
of Clalit Health Services (CHS). Owing to the retrospective nature of
the study, a waiver of informed consent was granted by the institutional
review board.
Data source
The study is based on data from the CHS database. CHS provides in-
clusive health care for more than half of the Israeli population (~ 4.7
million). Health care coverage in Israel is mandatory according to the
National Health Insurance Law (1995) and is provided by four groups
akin to not-for-profit health maintenance organizations (HMOs),
which are charged with providing a broad package of benefits stip-
ulated by the government. The four HMOs are both healthcare in-
surers and providers, thus financing and supplying health services.
Membership in a specific HMO is voluntary and members can freely
switch to another HMO. All members of the different HMOs have
similar health insurance plans and similar access to health services,
including low medication copayment. CHS maintains a database that
receives data from multiple sources including records of primary care
physicians, community specialty clinics, hospitalizations, laboratories,
and pharmacies. A registry of chronic disease diagnoses is compiled
from these data sources. Diagnoses are captured in the registry by
diagnosis-specific algorithms, employing International Classification
of Diseases Ninth Revision (ICD-9) code reading, laboratory test
results, and disease-specific drug usage. A record is kept of the data
sources and dates used to establish the diagnosis, with the earliest
recorded date from any source considered to be the defining date of
diagnosis. Designed for purposes of administrative and clinical man-
agement, the database is available for clinical studies.
The cohort
Within CHS database, we identified all G6PD-deficient patients by a re-
cord of at least one laboratory spectrophotometric quantitative test, mea-
suring G6PD activity normalized to hemoglobin (Hb) concentration,
throughout the entire insured follow-up, of < 6 U/g Hb. The biochemical
tests were performed in hospital laboratories or CHS central laboratories.
Although in the literature a threshold of 7 U/g Hb is sometimes used,4
we have used a stricter definition of < 6 U/g Hb. All G6PD-deficient pa-
tients insured in CHS between January 1, 2010 and December 31, 2022,
were included in the cohort.
Outcome
Within the G6PD-deficient cohort, we identified all hospitalizations
with a discharge diagnosis of hemolysis between January 1, 2010 and
December 31, 2022, using ICD-9 codes (Table S2). Cases with available
discharge letters of the index hospitalization were manually reviewed by
two researchers (NG, BR) for adjudication of the hemolytic event and
classifying the culprit exposure.
Exposure
We generated a list of medications that appear on either Luzzatto et al.,4
or CPIC lists of medications, previously suggested as carrying the risk
of hemolysis to G6PD-deficient patients. All available discharge letters
that included diagnoses of hemolysis were manually read to identify the
culprit medication that was associated with the hemolysis. For the rest
of G6PD-deficient patients, with no record of hemolysis, we recorded all
prescriptions of medications from the generated presumably risky list of
medications that were dispensed between 1.1.2010 and 31.12.2022. If
at least one prescription was dispensed to a patient, the medication was
counted as one, and further dispensing of the medication type was not
added to the sum.
All medication types dispensed to fewer than hundred patients were
manually validated by reading the medical files in the outpatient records
or in hospitalization discharge letters surrounding the filled prescription
date, to search for the description of the actual intake of the medication.
Statistical analysis
Statistical analysis was performed using IBM Statistics (SPSS) version
28 software. Continuous variables are presented by means ± standard
ARTICLE
15326535, 0, Downloaded from https://ascpt.onlinelibrary.wiley.com/doi/10.1002/cpt.3333 by Cochrane Israel, Wiley Online Library on [28/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
CLINICAL PHARMACOLOGY & THERAPEUTICS | VOLUME 0 NUMBER 0 | Month 2024
3
deviation and medians with interquartile range (IQR) as appropri-
ate. Categorical variables are presented as numbers and percentages.
Differences in baseline characteristics between patients with G6PD
< 1 U/g Hb and patients with G6PD 1–6 U/g Hb, separated into hemo-
lysis and no-hemolysis cohorts, were compared using Chi-square for the
categorical variables and Mann–Whitney for the continuous variables.
P < 0.05 was considered statistically significant.
RESULTS
Within all CHS enrollees between January 1, 2010 to December
2022, we identified 31,962 people with G6PD deficiency (by a
laboratory test result of G6PD < 6 U/g Hb) (~ 0.7% of the pop-
ulation). Patients with the lowest G6PD activity levels (< 1 U/ g
Hb) were mostly male, corresponding to the X-linked inheritance
of the gene.
Cases of acute hemolysis
Within the G6PD-deficient patients’ cohort, we identified 71 cases
of true major hemolytic events requiring hospitalization (0.2% of
the cohort). The mean G6PD activity level of these patients was
2.02 (SD 1.8) IQR [0.4; 3.2]. Twenty nine of them (40.8%) had
G6PD < 1 U/g Hb (not at the time of hemolysis). Cohort demo-
graphic characteristics are presented in Table 1. Additional 16
suspected cases did not have available discharge letters and were
excluded from the cohort.
Hemolysis cases by year of event are presented in Figure S1.
Median age at hemolysis was 12.4 years IQR [3.3; 53.6]; 39
(54.9%) were below age 18 at hemolysis. The most severe case was
in an 18 month-old baby boy with G6PD activity level of 0.2 U/g
Hb, who presented with signs of hemolysis, and Hb level of 3.4 g/
dL. The causes of the 71 hemolysis cases are described in Table 2.
Fifty one cases (71.8%) were of favism (hemolytic disease due to
the ingestion of fava beans (Vicia faba)), presenting usually within
2 days following the ingestion of fava beans, with jaundice, fever,
abdominal pain, vomiting, and laboratory results compatible with
hemolytic anemia in various severities. Six cases (8.5% of the he-
molytic cases) had been associated with an infection, and only
three cases (4.2% of the hemolytic cases) were suspected as being
caused by medications: one case by nitrofurantoin, one case by
phenazopyridine, and an additional case had been recorded receiv-
ing a “pain killer.” The culprit exposure remained unknown in 11
patients (15.5% of the hemolytic cases). Importantly, all patients
were discharged from the acute hemolysis hospitalizations in good
health condition.
Medications in G6PD-deficient patients with no hemolysis
hospitalization
Within the cohort of 31,875 G6PD-deficient patients with no major
hemolysis events, we recorded all prescription fills of medications,
from the list of presumably risky medications, between January
1, 2010 and December 31, 2022. Only 11 patients had been pre-
scribed medications appearing as carrying high risk for hemolysis
(Table 3 and Table S1; Figure 1). Nitrofurantoin, associated
previously with definitive4 or medium9 risk has been prescribed
safely to 1,366 G6PD-deficient males and females, almost 500 of
whom with G6PD < 1 U/g Hb. Importantly, also, hundreds or
thousands of G6PD-deficient patients had been prescribed safely
ciprofloxacin, glibenclamide, ofloxacin, phenazopyridine, sulfa-
methoxazole/cotrimoxazole, sulfasalazine, hydroxychloroquine,
glimepiride, mesalazine, and sulfacetamide that appear on various
G6PD-risky medications lists, either under definitive risk category,
low-risk category, or under a section of medications associated in
the past with risk, but that their risk has been recently regarded
“unknown” due to lack of data9 (Tables 3 and S1; Figure 1).
For medications administered to only a few patients, we manu-
ally explored the medical file for records of actual intake. In some
cases, we have found notes from outpatient clinics verifying expo-
sure to the medication. These are detailed in the right column in
Table 3.
Table 1 Demographic characteristics of a cohort of G6PD-deficient patients, CHS, 2010–2022
G6PD Activity, U/g Hb
Hemolysis cases (N = 71)
G6PD cohort with no hemolysis (n = 31,875)
< 1 (N = 29)
1–6 (N = 42)
P-value
< 1 (N = 20,349)
1–6 (N = 11,526)
P-value
Male N %
24 (82.8)
26 (61.9)
0.058
16,647 (81.8)
2,512 (21.8)
< 0.001
Age at diagnosis, median (IQR)
9 (3; 39)
29 (3; 57)
0.006
16 (0.12; 41.7)
16 (0.84; 8)
0.490
Jews N (%)
16 (55.2)
21 (50.0)
0.668
18,448 (90.7)
9,460 (82.1)
< 0.001
Socioeconomic status
0.066
< 0.001
Low
13 (44.8)
29 (75.2)
4,668 (22.9)
3,336 (28.9)
Middle
12 (41.4)
9 (22.5)
9,184 (45.1)
4,738 (41.1)
High
4 (13.8)
2 (5.0)
4,767 (23.4)
2,572 (22.3)
Missing
0
2
1730 (8.5)
880 (7.6)
Table 2 Causes of hemolysis in G6PD-deficient patients,
CHS, 2010–2022
Exposure medication/
food
Number of validated
hemolysis cases
Percentage
Medication-related
3
4.2
Nitrofurnatoin
1
Phenazopyridine
1
Pain killer (no details)
1
Fava beans
51
71.8
Infection
6
8.5
Unknown
11
15.5
Total
71
100
ARTICLE
15326535, 0, Downloaded from https://ascpt.onlinelibrary.wiley.com/doi/10.1002/cpt.3333 by Cochrane Israel, Wiley Online Library on [28/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
VOLUME 0 NUMBER 0 | Month 2024 | www.cpt-journal.com
4
Table 3 Medications prescribed to G6PD-deficient patients without record of hospitalization with hemolysis, CHS
(2010–2022)
Medicationsa with
suggested risk for
G6PD deficient
patients
ATC level 5
Number
of treated,
% within
total cohort
(N = 31,875)
Number of treated, %
within G6PD-deficient
male cohort (total male
cohort N = 19,159)
Number of treated, % within
G6PD-deficient female
cohort (total female cohort
N = 12,716)
Manual validation notesb
G6PD actvity level
(U/g Hb)
< 1
N = 16,647
1–6
N = 2,512
< 1
N = 3,702
1–6
N = 9,014
By Luzzato et al.4
Dapsone-containing
combinations
J04BA02
9 (0.03)
5 (0.03)
1 (0.04)
0
3 (0.03)
Recorded mild anemia,
abdominal pain in two
women. In addition, a baby
boy treated. Treatment was
withheld in all three. Three
additional patients were
prescribed following allogeneic
bone marrow transplantation.c
In additional three, true
exposure could not be
validated
Primaquine
P01BA03
2
2
Exposure could not be
validated
Ciprofloxacin
J01MA02
5,424 (17.0)
2,293 (13.8)
255 (10.0)
1,042 (28.1)
1834 (20.3)
Glibenclamide
(Glyburide)
A10BB01
290 (0.9)
124 (0.7)
21 (0.8)
50 (1.3)
95 (1.0)
Moxifloxacin
J01MA14
1
1
A woman (G6PD 5.1 U/g
Hb) was hospitalized with
pneumonia. Received
moxifloxacin in the hospital
and for three additional days
at discharge
Nalidixic acid
J01MB02
0
Nitrofurantoin
J01XE01
1,366 (4.3)
108 (0.6)
23 (0.9)
379 (10.2)
856 (9.5)
Norfloxacin
J01MA06
0
Ofloxacin
J01MA01
2,635 (8.3)
869 (5.2)
102 (4.1)
558 (14.8)
1,116 (12.4)
Phenazopyridine
(“sedural”)
G04BX06
1707 (5.4)
309 (1.9)
46 (1.8)
401 (10.8)
951 (10.6)
Sulfamethoxazole/
cotrimoxazole
J01EE01
912 (2.9)
327 (2.0)
86 (3.4)
152 (4.1)
347 (3.8)
Chloroquine
P01BA01
0
Quinidine
C01BA01
2
2
0
0
0
A 69-year-old man (G6PD
0.2 U/g Hb) with atrial
fibrillation was discharged
from cardiology department
(2014) with hydroxyquinidine
prescription; reported as still
receiving when hospitalized
6 years later; recommended to
stop because of amiodarone
co-treatment
Quinine
P01BC01
21 (0.1)
7 (0.04)
0
5 (0.1)
9 (0.1)
A woman (G6PD 0 U/g Hb)
reported being treated for 2
months for leg pain. Stopped
due to rash, no hemolysis. For
additional ten men and women
(G6PD levels 0–5.6 U/g Hb)
prescription filled, but we
could not find any outpatient
notes describing exposure
(Continued)
ARTICLE
15326535, 0, Downloaded from https://ascpt.onlinelibrary.wiley.com/doi/10.1002/cpt.3333 by Cochrane Israel, Wiley Online Library on [28/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
CLINICAL PHARMACOLOGY & THERAPEUTICS | VOLUME 0 NUMBER 0 | Month 2024
5
DISCUSSION
In this study, we are showing that a list of medications, suspected
previously to carry various risks for hemolysis in G6PD-deficient
patients, have been prescribed safely to male and female G6PD-
deficient patients. Prescriptions of nitrofurantoin, ciprofloxacin,
glibenclamide, ofloxacin, phenazopyridine, sulfamethoxazole/cotri-
moxazole, sulfasalazine, hydroxychloroquine, glimepiride, mesala-
zine, and sulfacetamide have been filled by hundreds or thousands
of patients with laboratory-confirmed G6PD-deficiency, suggesting
that these medications might be safe for G6PD-deficient patients.
Within the same G6PD-deficient cohort, we identified all hos-
pitalizations with a discharge diagnosis of hemolysis, validated
the cases manually for true acute hemolysis, and recorded the
culprit exposure. Surprisingly, most of the 71 hemolysis cases had
been caused by ingestion of fava beans. A few by an infection, and
only three were suspected as being related to medication: one by
Medicationsa with
suggested risk for
G6PD deficient
patients
ATC level 5
Number
of treated,
% within
total cohort
(N = 31,875)
Number of treated, %
within G6PD-deficient
male cohort (total male
cohort N = 19,159)
Number of treated, % within
G6PD-deficient female
cohort (total female cohort
N = 12,716)
Manual validation notesb
Aspirin
B01AC06
63 (0.2)
33 (0.2)
9 (0.3)
11 (0.3)
10 (0.1)
325 mg/day
Sulfadiazine
J01EC02 (PO for
toxoplasmosis)
0
D06BA01
Silver
Sulfadiazine
(topical
“silverol”)
1,441 (4.5)
525 (3.2)
95 (3.8)
256 (6.9)
565 (6.3)
Sulfasalazine
A07EC01
137 (0.4)
42 (0.3)
5 (0.2)
29 (0.8)
61 (0.7)
Chloramphenicol
D06AX02
topical (skin)
3,607 (11.3)
1,636 (9.9)
201 (8.0)
586 (15.8)
1,184
(13.2)
S01AA01 (eyes)
11,488
(36.0)
5,475 (32.9)
701 (27.9)
1,585 (42.8)
3,727 (41.3)
Vitamin K analogs
B02BA
29 (0.1)
13 (0.1)
1 (0.04)
2 (0.05)
13 (0.14)
Most patients were on chronic
warfarin, but no validation of
exposure
Additional - only by CPIC list9
4-Aminosalicylic
acid
J04AA01
0
Hydroxychloroquine
P01BA02
369 (1.2)
80 (0.5)
9 (0.4)
92 (2.5)
188 (2.1)
Tolbutamide
A10BB03
0
Vitamin C
A11GA
1,493 (4.7)
625 (3.8)
66 (2.6)
287 (7.8)
515 (5.7)
Chlorpropamide
A10BB02
0
Dabrafenib
L01EC02
1
0
0
0
1
A 34-year-old woman with
lupus and melanoma
(G6PD 3.4 g/U Hb) received
dabrafenib with trametinib
Glimepiride
A10BB12
333 (1.0)
150 (0.9)
25 (1.0)
78 (2.1)
80 (0.9)
Glipizide
A10BB07
77 (0.2)
28 (0.2)
8 (0.3)
17 (0.5)
24 (0.3)
Mesalazine
A07EC02
218 (0.7)
96 (0.6)
11 (0.4)
40 (1.1)
71 (0.8)
Nitrofural
D08AF02
0
Probenecid
M04AB01
1
1
0
0
0
A man with gout (G6PD 0.3
g/U Hb). Stopped febuxostat
due to rash. No notes
describing probenecid intake
Sulfacetamide
S01AB04
290 (0.9)
117 (0.7)
20 (0.8)
46 (1.2)
107 (1.2)
Trametinib
L01EE01
1
0
0
0
1
A 34-year-old woman
(described above) with lupus
and melanoma (G6PD 3.4 g/U
Hb) received dabrafenib with
trametinib
aMedications ordered by suggested risk and the origin of the list. bPerformed for medications prescribed to < 50 G6PD-deficient patients. cFollowing successful
bone marrow transplantation, G6PD activity of the donor is expected.
Table 3 (Continued)
ARTICLE
15326535, 0, Downloaded from https://ascpt.onlinelibrary.wiley.com/doi/10.1002/cpt.3333 by Cochrane Israel, Wiley Online Library on [28/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
VOLUME 0 NUMBER 0 | Month 2024 | www.cpt-journal.com
6
nitrofurantoin, one by phenazopyridine, and an additional case re-
corded receiving a “pain killer.”
Recently published Clinical Pharmacogenetics Implementation
Consortium (CPIC) guideline9 authors performed an extensive
literature review and noted that there was a large number of drugs
that carried regulatory or literature warnings for use in patients with
G6PD deficiency, but that the vast majority were based on limited
and old case reports, and sometimes with inadequate measures to
rule out other causes of anemia. Additional medications have inher-
ited such labeling warnings simply by virtue of being in the same
chemical or pharmacologic class. Our data on the safety of these
medications in the real world might establish the ground for re-
thinking of these label warnings.
Specifically, the recommendation to use nitrofurantoin with
caution (medium-risk) in G6PD-deficient patients had been pre-
sented in the CPIC guideline with an optional strength, given that
evidence was moderate at best, and that the paucity of case reports
was surprising, considering how commonly the drug had been used
worldwide.9 Our data might serve to further lower the caution
level for this drug. Our results were in line with CPIC’s low-to-no
risk level of sulfamethoxazole, stating that the evidence linking it to
acute hemolytic anemia was weak.9
In the case of favism, the triggers are two components present in
fava beans: vicine and convicine.10 In the intestine, the two beta-
glucosides are converted into their respective aglycones, divicine, and
isouramil. Their action is to increase the production of free radicals
that eventually lead to the oxidation of glutathione. The presence of
these glucosides is the discriminating factor between broad beans
and other legumes. The quantity of vicine and convicine determines
the onset or not of the hemolytic crisis and, in the case of crisis, of its
severity. Thus, several factors influence whether or not a hemolytic
crisis occurs: the amount of fava beans ingested; the quality of the
fava beans (i.e., raw, boiled, or canned); and the degree of ripeness,
which corresponds to the amount of glucosides present in the le-
gume. Finally, age of onset can influence the clinical presentation.10
Our study has several strengths. First, CHS is the largest health-
care provider in Israel, insuring more than half of the Israeli popu-
lation who are considered to be a large representative sample of the
general Israeli population. Second, the CHS comprehensive data-
base allowed us to include laboratory-confirmed G6PD-deficient
patients which is more accurate than relying on ICD-9 codes.
Based on the level of G6PD, we were able to look at the safety of
risk drugs both in patients with severe deficiency and patients with
higher residual function.11 In addition, the database enabled adju-
dication of hemolytic cases and the culprit exposure.
The study had also several limitations. First, we did not have data on
the specific mutation, which is known to be associated with the sever-
ity of hemolysis.12 However, we used laboratory data on G6PD activ-
ity level. It is currently widely accepted to establish diagnoses of G6PD
deficiency on tests of enzyme activity rather than genotype, due to
the unpredictable phenotype for heterozygous persons.2,13 Moreover,
due to X-linked mosaicism, it is impossible to predict women’s G6PD
activity based on genotype alone. Second, our study was performed
in Israel where the most frequent mutation is the Mediterranean
(NM_001360016.2(G6PD): c.563C>T (p.Ser188Phe))14 type class
II mutation which is also the most frequent mutation in Caucasians15
and is characterized by only a few hours half-life of the enzyme, severe
hemolysis upon oxidative stress, and no chronic hemolysis. Due to
higher residual function, the class III African mutation is less severe.
It might be assumed, although not proven,9 that medications safe in
patients harboring more severe mutation will be safe also among pa-
tients with the less severe form of the disease. Nonetheless, patients
harboring the rare, type I, most severe mutation, which causes chronic
hemolysis, might suffer from increased hemolysis when prescribed
medications that are safe in type II and III mutation carriers.9 Hence,
our results may not be generalized to patients with class I mutation,
and as a whole, lack of validation in other data sets and other popu-
lations limits generalization. Third, values defining G6PD deficiency
are not universally determined. However, we took a conservative ap-
proach to define deficiency by activity of less than 6 U/g Hb. Fourth,
it should be stressed that G6PD-deficient patients might have been
warned not to take certain medications and had not been prescribed
the presumably risky medication. Thus, it should not be assumed
that medications that are not on our “safe list” are not safe. Fifth, we
are not sure that patients actually took the medicine. However, we
manually read the medical files for medications with only a few pa-
tients prescribed, while for medications with hundreds or thousands
of patients with evidence of filled prescriptions, an actual exposure
might be assumed for at least some. Sixth, a few of the medications
listed on the presumably risky medications are not registered in Israel
and could not be assessed, and for the top two riskiest drugs, there
may not have been a true exposure, limiting significance. Seventh, the
number of G6PD-deficient patients might be an underestimation of
true disease prevalence in Israel, as additional individuals suspected
to have G6PD deficiency by family history and clinical suspicion, did
not undergo laboratory tests, or might have been tested years before
the database started. For 11 validated hemolysis cases a culprit fac-
tor was not identified by the treating physician, thus misclassification
might have happened if the patients were exposed to certain medi-
cations not identified. For additional 16 suspected hemolysis cases,
discharge letters were not available for the study, and these patients
have been excluded from the cohort, which might have caused an un-
derestimation of the true percentage of hemolytic cases. However, the
maximum additional cases would have caused an additional 0.05%
Figure 1 Number of patients with at least one prescription between
the years 2010–2022, CHS. Additional medications prescribed to 21
patients or less, with no record of hemolysis (the number of patients
in parenthesis: Quinine (21), dapsone (9: a few of them following
allogeneic bone marrow transplantation), Primaquine (2), Quinidine
(2), Moxifloxacin (1), Dabrafenib (1), Probenecid (1), Trametinib (1)).
ARTICLE
15326535, 0, Downloaded from https://ascpt.onlinelibrary.wiley.com/doi/10.1002/cpt.3333 by Cochrane Israel, Wiley Online Library on [28/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
CLINICAL PHARMACOLOGY & THERAPEUTICS | VOLUME 0 NUMBER 0 | Month 2024
7
added. Nonetheless, additional medications might have been associ-
ated with these cases. Eighth limitation is the fact that the extent of
G6PD deficiency and clinical symptoms varies between and within
individuals and is dependent not only on the type of G6PD allele,
the number of X chromosomes a person has, and other inherited
factors affecting erythrocyte physiology9 but also on the triggering
agent and its dosage, and the presence of concurrent infection. Thus,
inter-individual variation is expected. However, the robustness of our
results was established by the large number of administrations of the
studied medications.
CONCLUSION
In this real-world study, we are showing that a list of medications,
previously suspected as carrying various risks for hemolysis in
G6PD-deficient patients, have been prescribed safely to males and
females who are G6PD-deficient, providing reassurance to pa-
tients, prescribers, and regulators.
SUPPORTING INFORMATION
Supplementary information accompanies this paper on the Clinical
Pharmacology & Therapeutics website (www.cpt-journal.com).
FUNDING
No funding was received for this work.
CONFLICTS OF INTEREST
The authors declared no competing interests for this work.
AUTHOR CONTRIBUTIONS
N.G. and B.R. wrote the manuscript. N.G., N.S., and W.S. designed the
research. N.G., B.R., and N.S. performed the research. N.G., N.S., and
W.S. analyzed the data.
© 2024 The Author(s). Clinical Pharmacology & Therapeutics published
by Wiley Periodicals LLC on behalf of American Society for Clinical
Pharmacology and Therapeutics.
This is an open access article under the terms of the Creative Commons
Attribution-NonCommercial-NoDerivs License, which permits use and dis-
tribution in any medium, provided the original work is properly cited, the
use is non-commercial and no modifications or adaptations are made.
1. Malaria Policy Advisory Group Meeting Technical consultation to
review the classification of glucose-6-phosphate dehydrogenase
(G6PD). <https://cdn.who.int/media/docs/default-source/malar
ia/mpac-documentation/mpag-mar2022-session2-technical-
consultation-g6pd-classification.pdf?sfvrsn=1f36be5e_12>
2. Nkhoma, E.T., Poole, C., Vannappagari, V., Hall, S.A. & Beutler,
E. The global prevalence of glucose-6-phosphate dehydrogenase
deficiency: a systematic review and meta-analysis. Blood Cells
Mol. Dis. 42, 267–278 (2009).
3. Oppenheim, A., Corine, J., Rund, D., TomJ, V. & Luzzatto, L.
G6PD Mediterranean accounts for the high prevalence of G6PD
deficiency in Kurdish Jews. Hum. Genet. 91, 91–294 (1993).
4. Luzzatto, L., Ally, M. & Notaro, R. Glucose-6-phosphate
dehydrogenase deficiency. Blood 136, 1225–1240 (2020).
5. Entry – *305900 – GLUCOSE-6-PHOSPHATE DEHYDROGENASE;
G6PD – OMIM. <https://www.omim.org/entry/305900>
Accessed February 4, 2024.
6. FDA. Table of pharmacogenomic biomarkers in drug labeling
<https://www.fda.gov/drugs/science-and-research-drugs/table
-pharmacogenomic-biomarkers-drug-labeling> Accessed May 6,
2024.
7. Youngster, I. et al. Medications and glucose-6-phosphate
dehydrogenase deficiency: an evidence-based review. Drug Saf.
33, 713–726 (2010).
8. <https://www.accessdata.fda.gov/drugsatfda_docs/label/2009/
020496s021lbl.pdf> Accessed May 6, 2024.
9. Gammal, R.S. et al. Expanded clinical pharmacogenetics
implementation consortium guideline for medication use in the
context of G6PD genotype. Clin. Pharmacol. Ther. 113, 973–985
(2023).
10. Beretta, A., Manuelli, M. & Cena, H. Favism: clinical features at
different ages. Nutrients 15, 343 (2023).
11. WHO Working Group. Glucose-6-phosphate dehydrogenase
deficiency. Bull. World Health Organ. 67, 601–611 (1989).
12. Migeon, B.R. X-linked diseases: susceptible females. Genet. Med.
22, 1156–1174 (2020).
13. Nannelli, C., Bosman, A., Cunningham, J., Dugué, P.-A. & Luzzatto,
L. Genetic variants causing G6PD deficiency: clinical and
biochemical data support new WHO classification. Br. J. Haematol.
202, 1024–1032 (2023).
14. Riskin, A. et al. The Genetics of Glucose-6-Phosphate-
Dehydrogenase (G6PD) and uridine diphosphate
glucuronosyl transferase 1A1 (UGT1A1) promoter gene
polymorphism in relation to quantitative biochemical G6PD
activity measurement and neonatal hyperbilirubinemia. Children
(Basel) 10, 1–14. (2023). https://doi.org/10.3390/children10
071172.
15. CPIC® Guideline for G6PD – CPIC. <https://cpicpgx.org/guide
lines/cpic-guideline-for-g6pd/> Accessed February 4, 2024.
ARTICLE
15326535, 0, Downloaded from https://ascpt.onlinelibrary.wiley.com/doi/10.1002/cpt.3333 by Cochrane Israel, Wiley Online Library on [28/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License