10 / 78
CARDIOVASCULAR JOURNAL OF AFRICA • Volume 28, No 6, November/December 2017
348
AFRICA
of warfarin were atrial fibrillation (AF) (57 patients), deep-vein
thrombosis (DVT) (24 patients) and heart valve replacements (29
patients) (Table 1).
The median (IQR) admission INR was 8.49 (6.38–10) with
the median (IQR) discharge INR 1.98 (1.28–2.82). The cause of
warfarin toxicity was identified and addressed in 14.3% (18/126)
of patients. Where the cause was identified, 55.6% (10/18) were
due to dosing errors, 16.7% (3/18) DDIs, 11.1% (2/18) acute
illnesses and 11.1% (2/18) due to inability to control INR despite
best effort.
In cases of dosing errors, seven were due to physician error, two
were due to patient error, and one was due to both physician and
patient error. Physician error was due to a too-aggressive increase
in warfarin dosage in response to previously sub-therapeutic
episodes, and patient error was ascribed to incorrect and/
or inconsistent usage of warfarin. In 85.7% of patients with
warfarin toxicity, the cause was not identified (Table 2).
Eighty-five per cent (107/126) of patients were using
concomitant medication on admission with 77% (97/126) of
patients using one or more medicines with a known DDI with
warfarin. The median (IQR) number of possible DDIs was
two (one to three) per patient. The potential number of DDIs
with warfarin per patient were: one DDI 18% (23/126), two
DDIs 25% (31/126), three DDIs 18% (23/126), four DDIs 10%
(13/126), five DDIs 4% (5/126), six DDIs 1% (1/126), and seven
DDIs 1% (1/126). The most frequent drugs used found to have
a DDI with warfarin were simvastatin (57 patients), aspirin (33
patients) and atenolol (29 patients). Table 3 reports on all major
DDIs with warfarin.
Twenty-eight per cent (35/126) of patients presented with
major bleeding, 18% (23/126) with non-major bleeding and 54%
(68/126) without bleeding. The most frequent sites of bleeding
were upper gastrointestinal tract (31%, 18/58), haemoptysis
(19%, 11/58) and epistaxis (17%, 10/58). Seven cases (12%, 7/58)
of intracranial haemorrhage were reported. The median INRs
for the major bleeding, non-major bleeding and non-bleeding
groups were not significantly different (
p
=
0.05) at 10, 7.59 and
7.65, respectively.
We found no statistically significant relationship between the
presence of DDIs and the occurrence of bleeding. Furthermore,
although 36 patients were using concomitant antiplatelet
medicines, no statistically significant relationships were found
between the concomitant usage of antiplatelet medicines together
with warfarin and the occurrence of bleeding (see Table 4).
The median number of treatment interventions was two, with
33.3% (42/126) of patients not receiving any interventions and
35.7% (45/126) and 23.8% (30/126) of patients receiving one and
two treatment interventions, respectively. Nine (7.14%) patients
received three or more interventions. Five per cent (6/126) of
patients received three, 2% (2/126) received four and 1% (1/126)
received five interventions, respectively.
The most frequently used interventions were vitamin K (45
patients), FFP (43 patients) and packed red blood cells (RBC)
(34 patients). Factor PCC (Haemosolvex
®
) was administered in
eight patients. Other interventions used were cryoprecipitate (one
patient), tranexamic acid (two patients) and platelet products
(three patients). See Table 5 for median (IQR) total dose given
for the most frequently used interventions.
Table 2. Causes of warfarin toxicity
Causes
No of patients Percentage
Cause identified
18
14.3
Dosing error
10
7.9
Physician
7
5.6
Patient
2
1.6
Both
1
0.8
Drug–drug interaction
3
2.4
Acute illness
2
1.6
Inability to control INR despite best effort
2
1.6
Other (liver injury)
1
0.8
Cause not identified
108
85.7
Total
126
Table 3. Major DDIs with warfarin
Drug
Number of patients
using drug
Quality of evidence
of interaction
Cardiovascular medicines
Simvastatin
57
Excellent
Aspirin
33
Fair
Clopidogrel
4
Fair
Amiodarone
3
Excellent
Antimicrobial, including antiretroviral medicines
Efavirenz
6
Fair
Amoxicillin
1
Good
Amoxicillin/clavulanic acid
1
Good
Ciprofloxacin
1
Good
Cotrimoxazole
1
Excellent
Moxifloxacin
1
Excellent
Metronidazole
1
Good
Central nervous system medicines
Fluoxetine
4
Good
Citalopram
1
Good
Mirtazapine
1
Excellent
Valproic acid
1
Good
Table 4. Bleeding versus antiplatelet medicines
Bleeding
Aspirin Clopidogrel
Aspirin and clopidogrel
Major bleeding (
n
)
8
1
0
Non-major bleeding (
n
)
4
0
0
No bleeding (
n
)
20
2
1
Table 5. Most frequent interventions given
Intervention
Median total dose given (IQR)
Vitamin K (oral/IV) (mg)
10 (5–20)
FFP (IU)
3 (2–4)
Packed RBC (IU)
2.5 (2–5)
4-factor PCC (IU)
1 250 (1 000–2 000)
IQR
=
interquartile range, IV
=
intravenous, FFP
=
fresh frozen plasma, RBC
=
red blood cells, IU
=
international units, PCC
=
prothrombin complex concen-
trate.
Table 6. Calculated treatment cost
Component
Cost average
Total range
Hospital stay
R 7 464
(R 627 – R70 224)
Vitamin K
R 21
(R 1 – R 81)
FFP
R 3 948
(R 1 193 – R 10 737)
Packed RBC
R 4 617
(R 2 434 –R 15 821)
4-factor PCC
R 4 312
(R 1 568 – R 6 273)
Total cost to treat
R 10 578
(R 627 – R 79 762)
FFP
=
fresh frozen plasma, RBC
=
red blood cells, PCC
=
prothrombin
complex concentrate.