13 / 72
CARDIOVASCULAR JOURNAL OF AFRICA • Volume 27, No 2, March/April 2016
AFRICA
67
the controls was estimated at 1.5
×
the total number of patients
(
±
five patients).
PPCM was defined according to the recommendations of
the Heart Failure Association of the European Society of
Cardiology working group on PPCM, and left ventricular (LV)
systolic dysfunction was defined as LV ejection fraction
<
50%.
7
At the study centres, physicians in the Internal Medicine,
and Obstetrics and Gynaecology Departments were approached
and requested to refer all patients with suspected PPCM to the
principal investigator (PI) for further evaluation. Patients were
then interviewed, clinically evaluated and recruited consecutively.
Hospital in-patients with PPCM were clinically assessed and
underwent investigations within the first 48 hours of admission.
Demographic data, relevant aspects of the history and
physical signs, results of investigations, co-morbid conditions,
and complications were included in a detailed questionnaire.
Baseline levels of serum urea, electrolytes and creatinine were
carried out in the laboratories of AKTH, while 12-lead ECGs
at rest, and transthoracic echocardiograms (for PPCM patients
only) were all carried out by the PI at the study sites, according to
standard recommendations.
8
The echocardiographic examination
was performed using a Sonoscape S8 Doppler ultrasound
(Shenzhen, China, 2010) and the ECG was recorded using a
Mindray DECG-03A digital electrocardiograph (Shenzhen,
China, 2008).
8,9
All ECG recordings were studied and interpreted by the
investigators in the standard fashion, and ECG intervals/
durations were measured using manual callipers.
10,11
The controls
were evaluated using the same protocol as the patients, including
the ECGs, but an echocardiogram was not performed.
Statistical analysis
Frequencies, mean, median and inter-quartile ranges were used
to describe patients’ characteristics. Chi-square, Fisher’s exact
probability, Student’s
t
- and Mann–Whitney
U
-tests were used
to compare categorical and continuous variables as appropriate.
Binary logistic regression models were used to determine
predictors of PPCM among the ECG variables, and values were
expressed as odds ratios (OR) and 95% confidence intervals (CI).
Pearson’s correlation coefficient and linear regression models
were used to further assess relationships between variables of
interest.
A simple score assigning 1 to each identified independent
ECG predictor was composed and its accuracy in predicting
PPCM was determined using the area under the receiver
operating characteristics (ROC) curve (AUC), and AUC
>
0.75
was considered satisfactory. A
p
-value
<
0.05 was considered
statistically significant. The statistical analysis was carried out
using SPSS version 16.0 software.
Results
A total of 54 PPCM and 77 controls satisfied all the inclusion
criteria and were consecutively recruited. PPCM patients were
recruited at the time of confirmation of diagnosis, when specific
heart failure treatment was also commenced.
The baseline characteristics of the subjects are presented
in Table 1. The mean age, body mass index (BMI), systolic
(SBP) and diastolic blood pressure (DBP), and prevalence of
pregnancy-induced hypertension were not significantly different
between the two groups (
p
>
0.05). However, mean serum level of
creatinine was higher (
p
=
0.045), and mean serum sodium and
potassium levels were significantly lower (
p
=
0.009 and
<
0.001
respectively) in patients compared to controls.
ECG findings are presented in Table 2. All subjects were
in sinus rhythm, and ectopic beats and PR interval were
not significantly different (
p
>
0.05) between the two groups.
However, patients had significantly faster heart rates, broader
QRS durations, prolonged QTc intervals, and more frequent
tachycardia and ST–T-wave abnormalities (T-wave inversion
with or without ST-segment depression in all leads except aVR,
V1 and V2) than the controls (
p
<
0.004 for all comparisons).
ECG predictors of PPCM
The results of the logistic regression models are presented
in Table 3. In the univariate analysis, heart rate, ST–T-wave
abnormalities, and QRS and QTc durations were all predictors
of PPCM (
p
≤ 0.003). In addition, heart rate
<
100 beats/min
reduced the risk of having PPCM by 89.7% (
p
<
0.001). The
presence of ST–T-wave abnormalities increased the odds of
PPCM almost 12-fold (
p
<
0.001), while QRS duration
>
110
ms and QTc duration
>
460 ms increased the odds 5.2-fold (
p
<
0.001) and 9.5-fold (
p
<
0.001), respectively.
Stepwise multivariate regression analyses were then carried
out to control for confounding factors. In the initial model,
including heart rate, ST–T-wave abnormalities, QRS duration
Table 1. Baseline characteristics of PPCM patients and controls
PPCM
patients
(
n
=
54)
Controls
(
n
=
77)
p
-value
Mean age (years)
26.6
±
6.7 25.7
±
5.7 0.450
Body mass index (kg/m
2
)
21.6
±
4.3 21.8
±
4.3 0.836
Systolic BP (mmHg)
119
±
24 123
±
16 0.293
Diastolic BP(mmHg)
86
±
18 82
±
12 0.099
Pregnancy-induced hypertension,
n
(%) 16 (41.0)
14 (28)
0.197
Serum creatinine (µmol/l)
93.2
±
67.1 74.7
±
19.3 0.045*
Serum sodium (mmol/l)
136.9
±
5.9 139.6
±
4.4 0.009*
Serum potassium (mmol/l)
3.9
±
0.8 4.6
±
0.7
<
0.001*
*
p
-value statistically significant; values are expressed as means
±
standard
deviations or as numbers with percentages in parentheses.
Table 2. ECG features of PPCM patients and controls
PPCM patients
(
n
=
54)
Controls
(
n
=
77)
p
-value
Sinus rhythm,
n
(%)
54 (100)
77 (100)
Premature ventricular or
atrial extrasystoles,
n
(%)
5 (9.3)
4 (5.2)
0.365
Heart rate, beats/min
111
±
16
90
±
16
<
0.001*
Tachycardia,
n
(%)
36 (66.7)
17 (22.1)
<
0.001*
QRS duration (ms)
109.9
±
23.6
98.6
±
12.8
0.004*
QRS duration
≥
110 ms
19 (35.2)
8 (10.4)
0.001*
QTc duration (ms)
445.0
±
34.2
421.2
±
18.9
<
0.001*
QTc duration
≥
460 ms
12 (22.2)
3 (3.9)
0.001*
PR interval (ms)
148.1
±
20.4
149.1
±
21.1
0.799
ST–T-wave abnormalities
37 (68.5)
13 (16.9)
<
0.001*
*
p
-value statistically significant; values are expressed as means
±
standard
deviations or as numbers with percentages in parentheses.