Cardiovascular Journal of Africa: Vol 29 No 5 (September/October 2018)

52 / 74

CARDIOVASCULAR JOURNAL OF AFRICA • Volume 29, No 5, September/October 2018 314 AFRICA
a sensitivity of 89% and specificity of 68% (AUC on the ROC
curve of 0.9).
In univariate logistic regression, MPI z-score, AFI, EFW, UA
Doppler, CPR category, DV Doppler and MCA Doppler were
assessed separately as potential predictors of adverse outcome.
The only significant predictor of adverse outcome was MPI
z-score. Treating this as a continuous variable, the odds ratio was
7.8 (95% CI: 2.3–26.1), which can be interpreted as follows: for
a one unit higher MPI z-score, there is an approximately eight
times higher risk of adverse outcomes. This study suggests that
even in stable placental-mediated disease, an elevated MPI can be
a predictor of adverse outcome later in the pregnancy.
Our previous study and other studies5,19,20 have shown that the
MPI becomes abnormal much earlier than arterial redistribution,
and DV anomalies and MPIs deteriorated with worsening grades
of growth restriction. This study expands the notion that cardiac
dysfunction is probably the initial quantitative parameter to
become abnormal in placental-mediated disease and tracks the
severity of it. It is the first study to have shown that a single
elevated MPI in the context of a perceived mild/stable placental-
mediated disease scenario can be predictive of deterioration later
in the pregnancy.
In the study group (across all groups), adverse outcomes were
reported in 49% of cases, including three intra-uterine deaths,
11 cases eventually having decelerative tococardiography, four
who subsequently developed imminent eclampsia, two with
HELLP syndrome, five who developed severe oligohydramnios
later in the pregnancy, and two abruptios, with the highest
number recorded in the PE + IUGR group (67%). This would be
consistent with a more advanced placental maladaptive process.
The question is why would this be the case? Intrinsic cardiac
function plays a pivotal role in the compensatory mechanisms
of the growth-restricted foetus. Cardiac flow and cardiac
contractility may be directly impaired by early hypoxaemia
before Doppler changes in MCA can occur, while polycythaemia
resulting from blood viscosity changes may alter preload.21,22
Elevated MPIs beyond ‘buffer coping zones’ may be reflecting
early hypoxaemia and therefore could predict adverse obstetric
outcome.5 Pre-eclampsia on the other hand affects foetal
cardiac function by causing an increase in afterload. This
is due to the abnormal placental remodelling process and
reduced placental perfusion, causing placental vessel injury
and placental vasoconstriction, leading to increased placental
vascular resistance and thus increased foetal cardiac afterload.
This process can certainly impact on cardiac function.
Our initial study investigating MPI in severe pre-eclampsia
demonstrated that MPIs deteriorate, with worsening placental
vascular resistance in severe pre-eclampsia and this was linked
to adverse neonatal outcomes.6 This is probably on the basis
of angiogenic disparity, tipped in favour of anti-angiogenic
substances such as soluble fms-like tyrosine kinase (sFlt-1),
which are able to block the effects of vascular endothelial
growth factor and placental growth factor (PLGF) by
inhibiting interactions with its receptors, leading to widespread
vasoconstriction.23-26 Therefore elevation of MPI would reflect
these pathophysiological mechanisms, which would directly
impact on the foetal heart.
An abnormal CPR (< p5) for gestational age has been
reported to be an indicator of foetal hypoxaemia and impaired
long-term neurological outcome.27 Foetuses were therefore also
Table 3. MPI z-score cut-off points for predicting adverse outcomes
Cut-off
point
Sensitivity
(%)
Specificity
(%)
Correctly
classified (%) LR+ LR–
(≥ 2.98 ) 100.00 0.00 49.09 1.0000
(≥ 3.23 ) 100.00 3.57 50.91 1.0370 0.0000
(≥ 3.27 ) 100.00 7.14 52.73 1.0769 0.0000
(≥ 3.53 ) 100.00 10.71 54.55 1.1200 0.0000
(≥ 3.54 ) 100.00 14.29 56.36 1.1667 0.0000
(≥ 4.07 ) 96.30 25.00 60.00 1.2840 0.1481
(≥ 4.15 ) 96.30 32.14 63.64 1.4191 0.1152
(≥ 4.16 ) 96.30 39.29 67.27 1.5861 0.0943
(≥ 4.29 ) 96.30 42.86 69.09 1.6852 0.0864
(≥ 4.31 ) 96.30 53.57 74.55 2.0741 0.0691
(≥ 4.45 ) 96.30 57.14 76.36 2.2469 0.0648
(≥ 4.49 ) 92.59 60.71 76.36 2.3569 0.1220
(≥ 4.5 ) 88.89 67.86 78.18 2.7654 0.1637
(≥ 4.58 ) 85.19 67.86 76.36 2.6502 0.2183
(≥ 4.64 ) 85.19 71.43 78.18 2.9815 0.2074
(≥ 4.67 ) 85.19 75.00 80.00 3.4074 0.1975
(≥ 4.83 ) 85.19 78.57 81.82 3.9753 0.1886
(≥ 4.89 ) 81.48 78.57 80.00 3.8025 0.2357
(≥ 4.97 ) 81.48 82.14 81.82 4.5630 0.2254
(≥ 5.11 ) 77.78 82.14 80.00 4.3556 0.2705
(≥ 5.23 ) 74.07 85.71 80.00 5.1852 0.3025
(≥ 5.24 ) 74.07 89.29 81.82 6.9136 0.2904
(≥ 5.29 ) 70.37 89.29 80.00 6.5679 0.3319
(≥ 5.35 ) 66.67 92.86 80.00 9.3333 0.3590
(≥ 5.48 ) 62.96 96.43 80.00 17.6296 0.3841
(≥ 5.54 ) 59.26 96.43 78.18 16.5926 0.4225
(≥ 5.62 ) 55.56 100.00 78.18 0.4444
(≥ 5.69 ) 51.85 100.00 76.36 0.4815
(≥ 5.76 ) 48.15 100.00 74.55 0.5185
(≥ 5.77 ) 44.44 100.00 72.73 0.5556
(≥ 6.32 ) 40.74 100.00 70.91 0.5926
(≥ 6.96 ) 37.04 100.00 69.09 0.6296
(≥ 7.08 ) 33.33 100.00 67.27 0.6667
(≥ 7.69 ) 29.63 100.00 65.45 0.7037
(≥ 7.95 ) 25.93 100.00 63.64 0.7407
(≥ 8.15 ) 22.22 100.00 61.82 0.7778
(≥ 8.98 ) 18.52 100.00 60.00 0.8148
(≥ 9.84 ) 14.81 100.00 58.18 0.8519
(≥ 10.19 ) 11.11 100.00 56.36 0.8889
(≥ 10.33 ) 7.41 100.00 54.55 0.9259
(≥ 11.36 ) 3.70 100.00 52.73 0.9630
(> 11.36 ) 0.00 100.00 50.91 1.0000
LR+ = likelihood ratio positive, LR– = likelihood ratio negative.
Table 4.The utility of CPR in predicting adverse outcomes
Adverse outcome No adverse outcome
CPR < p5 18 16
CPR > p5 9 12
Sensitivity (%) 66.7
Specificity (%) 42.9
PPV (%) 52.9
NPV (%) 57.1
CPR = cerebro-placental ratio, PPV = positive predictive value, NPV = negative
predictive value.