28 / 74
CARDIOVASCULAR JOURNAL OF AFRICA • Volume 29, No 5, September/October 2018
290
AFRICA
(WHO risk group class II–IV) presenting at the dedicated
cardiac-obstetric clinic between 1 July 2010 and 31 December
2015, were screened for warfarin use. The research was approved
by the University of Cape Town Human Research Ethics
Committee (261/2015).
Of the 33 pregnancies identified, 29 (in 23 patients) were
included, with four excluded due to incomplete data. Additional
information, particularly maternal post-partum bleeding,
ischaemic and thrombotic events, and time hospitalised were
collected retrospectively from maternity folders stored at Groote
Schuur Hospital.
Preconception counselling in patients attending tertiary care
included advice to conceive on warfarin and to attend a
maternity clinic as soon as a period was missed or pregnancy
was suspected. Echocardiography was not routinely performed
pre-conception.
Antenatal anticoagulation comprised warfarin up to six
weeks’ gestation, unfractionated, intravenous heparin (heparin
sodium–fresenius) from six to 12 weeks’ gestation, warfarin
from 12 to 36 weeks’ gestation, and unfractionated heparin
from 36 weeks’ gestation. During periods of heparin infusion,
the activated partial thromboplastin time (aPTT) target was
2.5 to 3.5 times the control value for all patients, achieved
through regular monitoring and adjustment of the infusion
rate. International normalised ratio (INR) levels were used to
monitor and adjust the warfarin dose to achieve an INR value
between 2.5 and 3.5. Patients found to be sub-therapeutic while
on warfarin were admitted for heparin infusion concurrent with
warfarin until INR levels were again therapeutic.
Peripartum anticoagulation involved stopping warfarin at
36 weeks’ gestation and admitting for heparinisation. Heparin
was omitted from the onset of labour and reinitiated six hours
post-partum in the absence of clinical concern of haemorrhage.
Warfarin was restarted at the consultant’s discretion, usually the
day after delivery, with concurrent heparin until warfarin was
therapeutic. Induction of labour and caesarean sections were
performed only for obstetric indications.
Parameters collected included baseline characteristics such
as age, gravidity, HIV status and warfarin dosage, as well as
occurrence of bleeding, thrombotic and ischaemic complications.
Gestational age was calculated by the obstetrician and measured
either from the last normal menstrual period if the date was
certain, or by early foetal ultrasound dating. Failing either of
these methods, a late ultrasound was used. In patients who had
more than one pregnancy, all pregnancies occurring during the
study period were included.
Major haemorrhage was defined as bleeding necessitating
return to theatre or bleeding associated with both transfusion as
well as a drop in haemoglobin of
≥
2.0 g/dl. Minor bleeding was
defined as all other bleeding including gum bleeding, epistaxis
or troublesome bleeding from drip sites. Table 1 outlines the
parameters recorded.
Statistical analysis
Data were analysed using GraphPad Prism version 5.00 for
Windows (GraphPad Software, La Jolla California, USA).
Results are expressed as mean
±
SD and percentages. Unpaired
t
-tests with Welch correction were used to establish whether
differences in maternal outcome according to the HIV status
were statistically significant.
Results
Demographic data for the 23 patients are shown in Table 2. The
majority of patients were black (78%) and spoke isiXhosa (61%).
Most (78%) had reached high school but none had tertiary
education. Sixty-five per cent of patients declared a monthly
income of
<
ZAR 300 and none had a monthly income
>
ZAR
10 000.
Table 3 shows baseline maternal characteristics. All patients
had baseline effort tolerance of NYHA I or II. The mean heart
rate was 90 beats per min, mean systolic blood pressure (SBP)
was 111 mmHg and mean diastolic blood pressure (DBP) was
72 mmHg. The left ventricular mean ejection fraction was 54.4%.
Seven patients had abnormal cardiac rhythms. Twenty-eight of
29 patients had rheumatic valve disease, while just one patient
had a valve replacement for Takayasu’s disease. Nine patients (10
pregnancies) were HIV infected. Other co-morbidities included
syphilis, psoriasis and hearing impairment. Just two patients
presented prior to six weeks’ gestation and five presented after
24 weeks’ gestation. No patient was known to have had a
thrombotic event (deep-vein thrombosis, pulmonary embolus,
stroke) prior to her pregnancy.
In this cohort there were two deaths, both occurring post-
partum. The first was an 18-year-old, HIV-negative patient with
a double valve replacement (mitral and aortic). In the third
Table 1. Maternal and foetal outcomes recorded
Maternal,
bleeding
Antepartum haemorrhage, postpartum haemorrhage, haemor-
rhagic stroke, need for transfusion, other bleeding complications
Maternal,
thrombotic
Valve thrombosis, ischaemic or embolic stroke, other thrombotic
complications
Maternal,
other
Maternal mortality rates, maternal days in hospital, rates of
caesarian section and indication (obstetric or medical), new-onset
atrial fibrillation, infective endocarditis, new-onset or worsening
heart failure
Foetal/
infant
Warfarin embryopathy, analysed according to warfarin dosages;
other congenital anomalies; rates of miscarriage; rates of stillbirth
Table 2. Demographic data of 23 patients
Demographics
Number (%)
Ethnicity
African or black
18 (78)
Mixed
4 (18)
White
0 (0)
Other (Arab, Indian, other)
1 (4)
Language
isiXhosa
14 (61)
Afrikaans
4 (17)
English
3 (13)
Other
2 (9)
Education level
Year 1–5
5 (22)
Year 6–10
18 (78)
Year
>
10
0 (0)
Income per month (ZAR)
<
300
15 (65)
300–999
3 (13)
1 000–9 999
5 (22)
>
10 000
0 (0)