Postpartum to volume
J. C. F. E.
cardiac overload?
failure-heart
failure
due
E. Sanderson, M.R.C.P. 0. Adesanya, M.D. I. Anjorin, M.B. H. 0. Parry, F.R.C.P.
Zaria, Nigeria
The development of heart failure without any obvious cause after a pregnancy in a previously healthy woman is now a well-recognized although rare event, and it is generally accepted that there is a specific heart disease of the puerperium.1-3 In Zaria, this condition accounts for almost 10 per cent of female admissions in our hospital at certain times of the year, and although like others we have called this peripartum cardiac failure (PPCF), nearly all our patients develop symptoms postpartum.“-6 The clinical picture differs from the typical cases previously described’-” because edema (including facial) is gross, treatment is rapidly effective in nearly all patients, and there are very few re1apses.j.6 The purpose of the present study was to try to assessmyocardial function in these patients to determine if there was severe heart muscle failure or not. Patients
and methods
Patients. Forty-three patients who presented to the hospital with the typical features of PPCF within 6 months of delivery were studied. All patients had carried out the usual Hausa tradition of taking food rich in added salt and lying on heated beds during the postpartum period for at least 40 days. Patients with severe anemia (PCV From Zaria,
the Department Nigeria.
of Medicine,
Ahmadu
Be110 University
This work was supported by a grant from the Institute for Medical Research, and a Research Wellcome Trust (Dr. Sanderson). Received
for publication
May
12, 1978.
Accepted
for publication
Sept.
27, 1978.
Hospital,
Nigerian Fellowship
Reprint requests: J. E. Sandenon, Dept. of Clinical Cardiology, Postgraduate Medical School, Hammersmith Hospital, London OHS England.
0002-8703/79/050613
+ 09$00.90/O
0 1979 The
National of the
Royal W12
C. V. Mosby
Co.
Table
I. Clinical data on admission (43 patients)
Age: Mean 24 years (range 15 to 36 years) Number of previous History of previous Time from delivery months
pregnancies: mean 3; 6 primigravida postpartum swelling: 1 patient to presentation: mean 2.25 ? I.3
Symptoms and signs: Shortness of breath and edema: all patients Jugular venous pressure: raised in all patients 9 f 2.4 cm.) Heart rate: 116 * 15 beats per min.-’ Blood pressure: systolic, 129 2 20 mm. Hg diastolic, 98 5 16 mm. Hg Systolic murmur at apex: 16 patients Gallop rhythm: 32 patients Pleural effusion: 16 patients Ascites: 32 patients
(mean
Investigations: Weight: 55.5 I 8 Kg. PCV: 37 f 7% Cardiothoracic ratio: 0.62 * 0.06 Electrocardiogram: Sinus rhythm-all patients T wave inversion widespread: 19 patients T wave inversion lateral chest leads: 12 patients Others: 5 (3 bundle branch block; 2 small complexes) Normal: 1 patient
< 25 per cent), organic valvular heart disease, long-standing cardiomegaly, toxemia, chronic renal disease, or fever were excluded. On admission all patients were examined and weighed. A chest radiograph and electrocardiogram were done and the PCV was checked. The clinical features of the patients are shown in Table I. All patients were treated with diuretics and digoxin. Twenty-seven patients were restudied during the convalescent period (mean interval of 15 days
American
Heart
Journal
6 13
Sanderson
et al.
after admission; normal Nigerian
range 5 to 51 days). Twelve subjects were also studied.
Methods.
Echocardiography. All patients had an echocardiogram done on admission using the Electronics for Medicine V3270 ultrasonoscope (frequency 2.23 MHz, repetition frequency 1,000 sec. I). The output was displayed on an Electronics for Medicine multichannel stripchart recorder at a paper speed of 50 mm. sec.-‘, with a simultaneous electrocardiogram and phonocardiogram. The technique used has been described before.7 Measurements of the left ventricular dimension were only made on those recordings showing clear continuous echoes from the septum and posterior wall (41 patients). Analysis of the echocardiograms. The recordings were analyzed in the usual fashion.’ Left ventricular volumes were estimated by the method of cubing the dimension.” It was recognized that this overestimates the volume of a large cavity and that individual values are only an estimate, but it is sufficiently accurate for studying groups of patients.” The usual indices were used to assessleft ventricular function and were given by the following formulae”‘: EDD - ESD EDD of circumferential fiber short-
fractional shortening = mean velocity ening
EDD - ESD (Mean
Vcf)
=
EDD
ejection fraction =
x
ET
EDV - ESV EDV
where EDV = end-diastolic volume, ESV = endsystolic volume, EDD = end-diastolic dimension, ESD = end-systolic dimension, and ET = ejection time derived from the externally recorded carotid pulse and the aortic valve echogram, if available.’ A diagnosis of pericardial effusion was only made if there was a clear posterior echo-free space between the left ventricular epicardium and pericardium with flattening of the normal movement of the pericardium, and disappearance of the space behind the left atrium when a scan was made from the left ventricle to the aortic root.:. I; Systolic time intervals. These were recorded as has been previously described,” at a paper speed of 100 mm./sec. The carotid pulse was recorded by a solid state externally applied transducer, and
614
left ventricular ejection time (LVET) was measured from the onset of the carotid upstroke to the dicrotic notch. The pre-ejection period (PEP) was obtained by subtraction of the LVET from the Q-S, interval. As all the intervals are affected by heart rate, they were corrected to the appropriate systolic time index using the formulae derived by Weissler and associates.” Right heart catheterization and cardiac output determinations. Ten patients had right heart catheterization. In eight patients miniature heart catheters were used and cardiac output was measured by the thermodilution technique as described by Branthwaite and Bradley.” A cardiac output computer was used to integrate the total area under the thermodilution curve and instantaneously calculate the cardiac output. All thermodilution curves were recorded and those with an uneven outline were discarded. Although the cardiac output computer assumesa hematocrit of 42 per cent, no correction was made for lower values because even at a hemocrit of 30 per cent the calculated error is less than 1 per cent.” Pressure measurements were made through the miniature catheters connected to a pressure transducer in the normal way.’ i In two patients Swan-Ganz balloon catheters were used and in one patient a satisfactory pulmonary capillary wedge pressure was obtained. In another two patients the miniature catheters were left in situ in the pulmonary artery and repeat cardiac output measurements were made 24 hours later. As the patients were studied sitting at 45 degrees, the choice of a zero reference point was difficult. The sternal angle was chosen and a +5 mm. Hg constant was added to the pressures recorded. Since the sternal angle does not have a constant relation in all positions to the mid point of the right atrium, it is probable that the pressures are underestimated. Results A. On admission.
Clinical data (Table I). The mean systolic (129 + 20 mm. Hg) and diastolic (98 + mm. Hg) blood pressures were raised. The diastolic blood pressure was less than 95 mm. Hg in only 11 patients. The cardiothoracic ratio on the chest radiograph was greater than 50 per cent in all the patients. The electrocardiograms were abnormal in all but one of the patients, with the majority showing non-specific T wave changes.
May, 1979, Vol. 97, No. ii
Postpartum
Table
failure
II. Echocardiographic data on admission (Mean t S.D.) Index
measured
1. Basic: Left ventricular end-diastolic dimension (cm.) Left ventricular end-systolic dimension (cm.) Left atrium (cm.) Aortic root (cm.) Mitral valve mean diastolic closure rate (ems. sec.- ‘) 2. LVfunction: Mean Vcf (circ. sec. ‘) Ejection fraction (%) Fractional shortening 3. Estimated volumes: Stroke volume (ml.) Cardiac output (liters min.-‘)
*= Normal t = Normal
range from Feigenbaum.: data from The Heart, ed. by J. Willis Hurst,
Patients (n = 41-43)
Normals (n = 12)
5.7 4.5 4.2 2.6 13.5
4.6 3.1 2.7 2.8 13
Heart
Journal
k * * 2 *
0.6 0.7 0.7 0.3 5
1.2 + 0.4 50 t 13 0.2 t 0.08 100 * 46 11.4 f 4.4
” * t k -c
0.4 0.3 0.4 0.4 5
Normal range*
3.5-5.7 2.5-3.7 1.9-4.0 2.0-3.7 5-20
1.2 + 0.18 70 2 7 0.32 i 0.04
1.0-1.9 50-80 0.24-0.4
68 j, 20 4.8 i 1.8
50-loot 3.5-6.0t
1974, p. 83.
Echocardiograms (Table II). The results from the 12 normal Nigerian subjects are also shown. In our studies we have found that our normal values do not differ significantly from those of Feigenbaum.’ Twenty-four (55 per cent) of the patients had pericardial effusion and these accounted for the very large hearts seen on the chest radiographs of some patients. The mean left ventricular end-diastolic dimension was at the upper limit of normal. Thirteen patients had a LVEDD greater than 6 cm., but only two patients had a dimension greater than 6.5 cm., (6.8 cm. each) (Fig. 1). Except for these two patients, the left ventricles were dilated, but not grossly so as in congestive cardiomyopathy.“. ‘I, The left atrium was larger than normal (4 cm.) in 28 patients (Fig. 1). The mean right ventricular end-diastolic dimension was normal (1.7 + 0.6 cm.; normal range 0.8 to 2.4 cm.). However, the measurement of the size of the right ventricle by echocardiography is recognized to be inaccurate.” The indices of left ventricular function showed considerable scatter (Fig. 2). The mean values for the ejection fraction and the mean Vcf were just within the normal range, Table II. Only nine patients had a mean Vcf below 0.9 circs. sec.-’ (Fig. 2). The two patients with the lowest mean Vcf (0.51 and 0.53, respectively), approaching the range expected if there is a severe heart muscle
American
cardiac
failure,” were the only two patients who did not respond quickly to treatment and who had signs of left ventricular failure when restudied. Two other patients had low mean Vcf values (0.6 each) but they appeared to respond well to hospital treatment although they did not return for repeat measurements. There was a good correlation within individuals between the ejection fraction and the mean Vcf, (r = 0.86) (Fig. 3). Fractional shortening was depressed, possibly artificially so because of the increase in cavity size. The estimated stroke volume was greater than 50 ml. in all patients. As the heart rates were high, the estimated cardiac outputs were also high, although there was considerable scatter. The two patients who responded poorly to treatment had lower cardiac outputs (6.6 litres min.-’ each). The three lowest estimated outputs were in three patients who had normal LVEDDs (4.0, 4.3, and 5.0 cm.). Systolic time intervals. (Table III and Fig. 4). There was prolongation of the PEP1 (156 t 27 msec.; normal range 100 to 155 msec.), and shortening of the LVETI (373 + 22 msec.; normal range 395 to 435 msec.). The mean of the ratio PEPULVETI, a better index of left ventricular function, was at the upper limit of normal, 0.41 + 0.08; (normal range 0.27 to 0.41).18 Although the ratio increases with heart failure, half the patients had normal values and the results do not suggest there is severe heart muscle failure in all the patients.
615
Sanderson
et al.
7.0
.
.. .
l .+. .H .. 2:. :. . .. .
6.0.
4.0,
:. ‘2’ .E
l :* .:: ”
:*
ems.
3.0, 2.0. 1.0. 0. 1. Scattergram
1.5
circs.
III. Systolic time intervals on admission (mean t S.D.)
I
. . c
. :.*
t
sec. -1 1.0
....
of individual values of the basic echocardiographic measurements. LVEDD = left ventricular end diastolic dimension; LVESD = left ventricular end systolic dimension; LA = left atria1 dimension. The vertical bars represent the normal ranges. Fig.
20
loo-
” .
:. ..
I :.:I .0;.. Liz
5.0,
.
2.0
.. ..:...
0.5
0
; ” .
50 i %-
i .. :
m Mean Vcf
.a. 2 . 2;.*I ‘$::
-
: : . :.
o&J” fraction
r
15
‘! 1 . .
litres
. > i . E I
min.
-1
10
. In Estimated cardiac output
Fig. 2. Scattergram of individual values of indices of left ventricular function and estimated cardiac output derived from the echocardiograms. The vertical bars represent the normal range.
diogram was done first prior to catheterization. In two patients the echocardiograms were destroyed due to a technical error. In the remaining four patients who also had thermodilution measurements done the mean difference was 1.0 liters min.-’ (Table V).
Table
PEP1 (msec.) LVETI (msec.) PEP/ LVET t = normal
range
156 k 27 373 -C 22
127 f 20 395 i 18
0.41
0.32
from
‘-e 0.08 W&&r
-t 0.04
100-155 395-435 0.27-0.41
et al. 1968.”
Cardiac catheterization data. (Table IV). All the patients had raised atria1 pressures on admission. The right ventricular end-diastolic pressure was raised in all patients and all the patients had pulmonary hypertension. The pulmonary artery end-diastolic pressure (an index of the left ventricular filling pressure) was also raised in all the nine patients in whom it was measured. The mean cardiac output measured by the thermodilution technique for the whole group was 6.3 rt 1.5 liters min.-’ In all patients an echocar-
616
B. Convalescent.
Clinical data. At restudy all the patients had improved. The mean weight loss was 11.5 & 6.2 kilograms (P < 0.001). Two patients had shortness of breath, but none had edema. The blood pressure of these 27 patients had fallen (systolic 116 ? 15 mm. Hg, P < 0.01; diastolic 79 I 15 mm. Hg, P < 0.01). Only three patients had a diastolic blood pressure above 95 mm. Hg. The heart rate had decreased to 94 + 16 beats min. ’ (P < 0.001). The jugular venous pressure was slightly raised in only one patient. Seven patients had a gallop rhythm still and six patients still had an apical systolic murmur. No patient had a pleural effusion or ascites. The cardiothoracic ratio had decreased to 0.57 & 0.08 and only three patients had a ratio above 0.6. The electrocardiograms had changed little; in three widespread T wave inversion had
May, 1979, Vol. 97, No. 5
Postpartum
cardiac
failure
. . .
:*
.:
‘. . l*. .
. .
.l.O
.*
-m--‘-r-, C. l l
r= 0.86
‘:
. . .*
: I :
0
50
Ejection
fraction
.
200,
..$ ...
(%I
been replaced by T wave inversion in the lateral chest leads only. The PCV had risen to 40 t 6 per cent. Echocardiograms. (Table VI). The major changes were the reduction in the size of the chambers and the estimated cardiac output. There was no change in the indices of left ventricular function. The fall of the mitral valve diastolic closure rate from 13 + 3 cm. sec.-’ probably reflects the reduction of blood flow across the valve.“’ Systolic time intervals. (Table VI). There was little change: PEP1 = 135 + 19 msec., LVETI = 359 & 21 msec., PEPI/LVETI = 0.4 2 0.07. Discussion
The accurate assessmentof the performance of heart muscle is still difficult. As a pump the heart’s product is cardiac output, but this is determined by the interaction of many factors such as myocardial contractility, aortic pressure (or afterload), degree of ventricular hypertrophy, and the magnitude of the venous return (or preload) which itself is affected by the peripheral vascular resistance, blood volume, and over-all vascular tone. To isolate the function of the muscle of the heart out of all these variables is almost impossible. However, certain pragmatic approaches are now recognized as being valid for defining groups. The ejection phase indices of left
100
0
. . i. .
. . . fI I I I..
z
100
Fig. 3. Relationship of mean Vcf to ejection fraction, both derived from the echocardiogram. Dotted lines are the lower limits of the normal ranges.
Heart Journal
.
msecs.
1,,, ,i ,, ,,,
American
300.
I
l ..
r-l
LVETI
: l t:{..
0.5
..i;..
0
I PEP I
PEPI LVET
Fig. 4. Scattergram of the systolic time intervals for individual patients. The vertical bars represent the normal ranges.
veritricular function (mean velocity of circumferential fiber velocity (Vcf), ejection fraction, and fractional shortening) have been shown to distinguish between groups of patients with normal and abnormal left ventricular function.“, II1An important advantage is that they can be calculated from echocardiograms. The calculation of volumes from the left ventricular dimensions measured by echocardiography is more debatable.‘” It was clearly recognized in this study that the individual values for any volume were only an approximatation. However, for studying groups of patients this method is sufficiently accurate to decide whether the stroke volume is very low, normal, or very high. Support for this comes from the results using the thermodilution technique for the measurement of cardiac output. The differences were not great in those individuals whose cardiac outputs were measured by both techniques.
617
Sanderson
et al.
NECHANISM
OF HEART FAILURE
IN PFCF
MEAN ARTERIAL STROKE VOLUM3 TOWARDS NOFJ4AL RENAL
HIGH NA INTAKE FUID
ARTERIAL
(=
RETENTION
EDEMA ______*
HEART FAILURE WITH NORMAL C.O.)
PERIPHERAL (A~RI~AD)
PRDLAC’W’
HUMIDITY-+~FLUID
Loss THROUGH SWEATING Fig.
5. Outline
IV. Cardiac catheterization
Table
RA f&d
mechanism
PPCF
14 13 11
53119 55117 53/10
53/35 55/35 53/30
4.8 + 0.49 7.0 k 0.59 8.8 k 0.69
4 (9
20
50/22
(ii)
13
45/15
50/37 45/30
7
13 10 8 20
35/5-20 45/10 25/9 60/17
35/10-20 25/15 60/32
8 9
13 15
52/10 60/18
52133 60/35
5
6 6) (ii)
10
13
range*
-l-+8
in Zaria.
data PA (mm. W
1 2 3
25113 15-28/O-8
RA = right atria1 pressure; RV = right ventricular Figures in brackets are the number of determinations * = from “The Heart” ed. by J. Wilis Hurst, 1974,
25/19 15-28/5-16
Cardiac (liters
output min. y
(beats
HR min.~ I/ 110
(3) (9) (3)
78 130
44 90 61
4.7 t 0.3 (5) 5.2 + 0.4 (8)
126 105
37 49
6.3 k 0.46 (4) 6.2 +- 0.9 (9) 5.2 -' 0.5 (6) -
140 100 102
45 62 51 -
-
-
-
-
-
3.5-6.0
pressure; HR = heart rate;SV = stroke volume; PCW for which the thermodilution cuwes were satisfactory. p, 83.
The surprising fact that emerged from this study was that, despite the clinical evidence of advanced circulatory congestion and edema which was present in all the patients studied, most had maintained relatively good left ventricular function and the cardiac output was high. By contrast, the two patients who had very low values for the mean Vcf and ejection fraction were the only two patients who had continuing signs of heart failure, similar to the pattern described in patients with a congestive cardio-
618
to explain
RV (mm. Hg)
Patient
Normal
Mean (mm.
of the proposed
50-110 = pulmonary
capillary
wedge
pressure.
myopathy.“. l6 In the remainder, the clinical course and echocardiographic findings were not like congestive cardiomyopathy. Instead, the hemodynamic findings were very similar to those seen in patients with acute glumerulonephritis,‘” although none of our patients had any evidence of renal disease.PPCF in Zaria, therefore, may be an example of congestion of the circulatory system which simulates congestive heart failure, but which is not primarily due to it.” But why should apparently normal postpar-
May,
1979,
Vol.
97, No.
5
Postpartum
V. Combined echocardiographic and thermodilu tion measurements Table
Table
cardiac
failure
VI. Follow-up data (27 patients) (mean t
SD.) Measurement
On admission
Convalescent
1. Echocardiographic
~ 1. 2. 5. 6. LVEDD fraction;
6.2 6.0 4.0 5.2
26 42 65 43
63 91 42 61
6.7 8.7 6.1 6.5
= left ventricular end diastolic dimension; SV = stroke volume; CO. = cardiac output.
4.8 7.0 6.3 6.2
k * & f
0.49 0.59 0.46 0.9
EF = ejection
turn woman accumulate such large amounts of fluid? (Fig. 5). It has been suggested previously that the Hausa postpartum practices of eating food rich in salt and lying on heated mud beds may be responsible for overriding the normal control mechanisms of the extracellular volume.” The heating produces a low total peripheral resistance so that the mean arterial pressure cannot be raised except by increasing cardiac output. As the mean arterial pressure is the major determinant of urinary output,” this inability to raise the arterial pressure as efficiently as usual, in women already taking large amounts of sodium, inevitably leads to their retaining excess sodium and water. In such postpartum women, prolactin, known to have sodium retaining propert,ies >I. “? may also play a part. Fluid loss through sweating will also decrease as the humidity increases at the onset of the rains’” (and this probably accounts for the peak of admissions at this time’). The cardiac output will continue to rise in response to the increasing blood volume and venous return in an attempt to excrete the excessive fluid, until the reserves available from left ventricular dilatation and the Frank-Starling mechanism are fully used. This limit will vary for each individual and’those with poor myocardial reserves, from whatever cause, will be unable to sustain the required cardiac output and will develop a form of high output failure. The combination of high venous pressures and a large blood volume rapidly produces gross edema (Fig. 5). A further complication arises if the peripheral resistance should begin to increase. This is likely to happen once edema has formed because of a direct effect of edema on the vessels” and because the increased sodium content of a vessel wall
American
Heart
Journal
LVEDD (cm.) LVESD (cm.) LA (cm.) Aortic root (cm.) MV mean DCR (cm. sec. ‘) Mean Vcf (circs. sec.-‘) Ejection fraction (%) Fractional shortening Estimated S.V. (ml.) Estimated C.O. (liters min. ~‘) 2. Systolic time intervals LVETI (msec.) PEP1 (msec.) LVET/PEP
5.8 4.6 4.3 2.5 14
2 f + + k
0.6 0.7 0.6 0.2 5
5.4 4.2 3.2 2.5 9
1.2 49 0.21 99 11.8
t k f i i-
0.4 12 0.07 39 4.8
1.0 f 0.3 50 If- 11 0.2 f 0.07 82 t 35* 7.5 i 3.3**
373 f 22 156 + 22 0.42 k 0.07
Significance of change from admission **p = i 0.001
to convalescent:
k xk k + +
0.7* 0.7 0.9** 0.2 3**
359 f 21 135 + 19 0.4 + 0.07 *p = < 0.1;
limits the ability of an arteriole to dilate.‘” Cooling will have the same effect. When the filling pressures of the left ventricle are high and the sarcomeres are fully extended, the stroke volume becomes very sensitive to change in afterload, dropping sharply with any further increase.‘!’ This will decrease the cardiac output back towards normal, thus exacerbating the situation even further. The combination.of a high inflow volume with an elevated peripheral resistance produces an elevation of the left ventricular work load far above that present in comparable high output states in which the mean arterial pressure is low. It is possible that these combined loads are sufficient to produce myocardial damage, and this may account for the slightly poor left ventricular function seen in some of the patients. (Exactly the same event will occur on the right side of the heart if the pulmonary artery pressure rises). The fact that the women were able to cope with the hemodynamic load of pregnancy without any sign of heart failure suggests that they did not have pre-existing heart muscle disease. The provoking factors of salt and heat are clearly important in explaining why PPCF is so common around Zaria. But the syndrome we see is similar to some of the previously reported cases. Thus, Demakis and colleagues”” found that the chest radiograph became normal in 14 of 27 patients with PPCF, and in these women the
619
Sanderson
et al.
prognosis was excellent with no further heart failure in 21 subsequent pregnancies. Johnson and associates” reported a patient who was catheterized and in whom the cardiac output was normal and the hemodynamic data was interpreted as showing “biventricular failure, probably of the high output type.” Another study reported cardiac output results from ten patients. In seven patients the values were normal (3.6 to 5.1 liters min. I), although the authors conclude that the “haemodynamic findings were characterised by low output biventricular failure.” The other reports in the literature which contain cardiac output data are six case reports”-“ (two were normal and four were low), and that of Pearce and co-workers,,‘!’ who found that the cardiac outputs were low in three patients. It is possible therefore, that PPCF in temperate climates also may not be initially a primary myocardial problem. A more radical concept is that during the postpartum period the extracellular volume may expand more than is normally possible while the control systems are resetting after the pregnancy. Sodium retention and volume expansion may occur, helped by the high prolactin levels. An expanding blood volume will, as above, produce an increasing venous return, and although the mother is able to cope with this during the pregnancy, she may only be able to do so if the peripheral resistance is low. However, it may be that during the postpartum period in certain women the peripheral vessels respond to volume loads and excess sodium more vigorously. This rise in peripheral resistance may produce a small rise in the blood pressure. (It is possible that this is the mechanism of the well recognized postpartum rise in blood pressure seen in normal women.)“’ But, more importantly, this rise in the peripheral resistance will markedly increase the work load of an already dilated ventricle. Again, as in our own patients, certain individuals may be incapable of coping with the combined demands of a high inflow volume and an increased afterload, so that myocardial damage and a true cardiomyopathy may follow. This theory reconciles two apparently conflicting views of the etiology of PPCF. Most authors have considered it to be a form of cardiomyopathy or intrinsic heart muscle disease. But as was pointed out by Brockington,” there are several objections to this concept. The major difficulty is that it does not explain the hypertension found in
620
most cases, and it seems highly unlikely that any left ventricle that is already in failure due to a myocardial fault is able to maintain a higher t,han normal blood pressure. However, although Brockington concludes that PPCF is hypertensive heart failure, he provides no plausible explanation for the fact that the presenting blood pressures are only moderately raised, they fall to normal quickly after treatment with diuretics, and the women do not need further treatment for hypertension. Therefore, it seems reasonable to suggest that the primary event is volume overloading, which by dilating the ventricles makes them very sensitive to any increase in arterial pressure (which is also due to the same initial volume expansion). The consequent reduction of stroke volume in the face of a high venous return will accelerate the development of the symptoms and signs of congestive heart failure and possibly in a few result in permanent myocardial damage. Therefore, although there are special provoking factors in this area, the Hausa postpartum practices may be exposing a mechanism that is common to postpartum women throughout the world. Summary
Ventricular function has been studied in 43 patients with the peripartum cardiac failure (PPCF) syndrome which occurs around Zaria. All patients had an echocardiogram on admission and 10 patients had right heart catheterization. Despite the gross edema, left ventricular function assessedby echocardiography and systolic time intervals was relatively good and the estimated cardiac outputs were high. At catheterization, although the pressures were high, the cardiac outputs were greater than normal in four out of six patients. No patient had a low cardiac output. These findings are not compatible with a severe heart muscle disorder, or cardiomyopathy. We suggest that the primary event in PPCF of Zaria is fluid retention which leads to a form of high output cardiac failure. The postpartum practices in this area (taking high sodium diets and lying on heated beds) almost certainly cause the fluid to accumulate initially, but the heart may be unable to meet the demands either because of preexisting heart muscle disease or, more likely, because of a rise of the peripheral resistance due to the volume expansion, overburdens such dilated hearts and leads to myocardial damage.
May.
1979,
Vol.
97, No.
5
Postpartum
Since there and PPCF that there traditional
are similarities between this condition in temperate climates, it is possible is a common mechanism which the practices of this area have unveiled.
22. 23.
REFERENCES 1. 2.
J. C. F. E.
cardiac overload?
failure-heart
failure
due
E. Sanderson, M.R.C.P. 0. Adesanya, M.D. I. Anjorin, M.B. H. 0. Parry, F.R.C.P.
Zaria, Nigeria
The development of heart failure without any obvious cause after a pregnancy in a previously healthy woman is now a well-recognized although rare event, and it is generally accepted that there is a specific heart disease of the puerperium.1-3 In Zaria, this condition accounts for almost 10 per cent of female admissions in our hospital at certain times of the year, and although like others we have called this peripartum cardiac failure (PPCF), nearly all our patients develop symptoms postpartum.“-6 The clinical picture differs from the typical cases previously described’-” because edema (including facial) is gross, treatment is rapidly effective in nearly all patients, and there are very few re1apses.j.6 The purpose of the present study was to try to assessmyocardial function in these patients to determine if there was severe heart muscle failure or not. Patients
and methods
Patients. Forty-three patients who presented to the hospital with the typical features of PPCF within 6 months of delivery were studied. All patients had carried out the usual Hausa tradition of taking food rich in added salt and lying on heated beds during the postpartum period for at least 40 days. Patients with severe anemia (PCV From Zaria,
the Department Nigeria.
of Medicine,
Ahmadu
Be110 University
This work was supported by a grant from the Institute for Medical Research, and a Research Wellcome Trust (Dr. Sanderson). Received
for publication
May
12, 1978.
Accepted
for publication
Sept.
27, 1978.
Hospital,
Nigerian Fellowship
Reprint requests: J. E. Sandenon, Dept. of Clinical Cardiology, Postgraduate Medical School, Hammersmith Hospital, London OHS England.
0002-8703/79/050613
+ 09$00.90/O
0 1979 The
National of the
Royal W12
C. V. Mosby
Co.
Table
I. Clinical data on admission (43 patients)
Age: Mean 24 years (range 15 to 36 years) Number of previous History of previous Time from delivery months
pregnancies: mean 3; 6 primigravida postpartum swelling: 1 patient to presentation: mean 2.25 ? I.3
Symptoms and signs: Shortness of breath and edema: all patients Jugular venous pressure: raised in all patients 9 f 2.4 cm.) Heart rate: 116 * 15 beats per min.-’ Blood pressure: systolic, 129 2 20 mm. Hg diastolic, 98 5 16 mm. Hg Systolic murmur at apex: 16 patients Gallop rhythm: 32 patients Pleural effusion: 16 patients Ascites: 32 patients
(mean
Investigations: Weight: 55.5 I 8 Kg. PCV: 37 f 7% Cardiothoracic ratio: 0.62 * 0.06 Electrocardiogram: Sinus rhythm-all patients T wave inversion widespread: 19 patients T wave inversion lateral chest leads: 12 patients Others: 5 (3 bundle branch block; 2 small complexes) Normal: 1 patient
< 25 per cent), organic valvular heart disease, long-standing cardiomegaly, toxemia, chronic renal disease, or fever were excluded. On admission all patients were examined and weighed. A chest radiograph and electrocardiogram were done and the PCV was checked. The clinical features of the patients are shown in Table I. All patients were treated with diuretics and digoxin. Twenty-seven patients were restudied during the convalescent period (mean interval of 15 days
American
Heart
Journal
6 13
Sanderson
et al.
after admission; normal Nigerian
range 5 to 51 days). Twelve subjects were also studied.
Methods.
Echocardiography. All patients had an echocardiogram done on admission using the Electronics for Medicine V3270 ultrasonoscope (frequency 2.23 MHz, repetition frequency 1,000 sec. I). The output was displayed on an Electronics for Medicine multichannel stripchart recorder at a paper speed of 50 mm. sec.-‘, with a simultaneous electrocardiogram and phonocardiogram. The technique used has been described before.7 Measurements of the left ventricular dimension were only made on those recordings showing clear continuous echoes from the septum and posterior wall (41 patients). Analysis of the echocardiograms. The recordings were analyzed in the usual fashion.’ Left ventricular volumes were estimated by the method of cubing the dimension.” It was recognized that this overestimates the volume of a large cavity and that individual values are only an estimate, but it is sufficiently accurate for studying groups of patients.” The usual indices were used to assessleft ventricular function and were given by the following formulae”‘: EDD - ESD EDD of circumferential fiber short-
fractional shortening = mean velocity ening
EDD - ESD (Mean
Vcf)
=
EDD
ejection fraction =
x
ET
EDV - ESV EDV
where EDV = end-diastolic volume, ESV = endsystolic volume, EDD = end-diastolic dimension, ESD = end-systolic dimension, and ET = ejection time derived from the externally recorded carotid pulse and the aortic valve echogram, if available.’ A diagnosis of pericardial effusion was only made if there was a clear posterior echo-free space between the left ventricular epicardium and pericardium with flattening of the normal movement of the pericardium, and disappearance of the space behind the left atrium when a scan was made from the left ventricle to the aortic root.:. I; Systolic time intervals. These were recorded as has been previously described,” at a paper speed of 100 mm./sec. The carotid pulse was recorded by a solid state externally applied transducer, and
614
left ventricular ejection time (LVET) was measured from the onset of the carotid upstroke to the dicrotic notch. The pre-ejection period (PEP) was obtained by subtraction of the LVET from the Q-S, interval. As all the intervals are affected by heart rate, they were corrected to the appropriate systolic time index using the formulae derived by Weissler and associates.” Right heart catheterization and cardiac output determinations. Ten patients had right heart catheterization. In eight patients miniature heart catheters were used and cardiac output was measured by the thermodilution technique as described by Branthwaite and Bradley.” A cardiac output computer was used to integrate the total area under the thermodilution curve and instantaneously calculate the cardiac output. All thermodilution curves were recorded and those with an uneven outline were discarded. Although the cardiac output computer assumesa hematocrit of 42 per cent, no correction was made for lower values because even at a hemocrit of 30 per cent the calculated error is less than 1 per cent.” Pressure measurements were made through the miniature catheters connected to a pressure transducer in the normal way.’ i In two patients Swan-Ganz balloon catheters were used and in one patient a satisfactory pulmonary capillary wedge pressure was obtained. In another two patients the miniature catheters were left in situ in the pulmonary artery and repeat cardiac output measurements were made 24 hours later. As the patients were studied sitting at 45 degrees, the choice of a zero reference point was difficult. The sternal angle was chosen and a +5 mm. Hg constant was added to the pressures recorded. Since the sternal angle does not have a constant relation in all positions to the mid point of the right atrium, it is probable that the pressures are underestimated. Results A. On admission.
Clinical data (Table I). The mean systolic (129 + 20 mm. Hg) and diastolic (98 + mm. Hg) blood pressures were raised. The diastolic blood pressure was less than 95 mm. Hg in only 11 patients. The cardiothoracic ratio on the chest radiograph was greater than 50 per cent in all the patients. The electrocardiograms were abnormal in all but one of the patients, with the majority showing non-specific T wave changes.
May, 1979, Vol. 97, No. ii
Postpartum
Table
failure
II. Echocardiographic data on admission (Mean t S.D.) Index
measured
1. Basic: Left ventricular end-diastolic dimension (cm.) Left ventricular end-systolic dimension (cm.) Left atrium (cm.) Aortic root (cm.) Mitral valve mean diastolic closure rate (ems. sec.- ‘) 2. LVfunction: Mean Vcf (circ. sec. ‘) Ejection fraction (%) Fractional shortening 3. Estimated volumes: Stroke volume (ml.) Cardiac output (liters min.-‘)
*= Normal t = Normal
range from Feigenbaum.: data from The Heart, ed. by J. Willis Hurst,
Patients (n = 41-43)
Normals (n = 12)
5.7 4.5 4.2 2.6 13.5
4.6 3.1 2.7 2.8 13
Heart
Journal
k * * 2 *
0.6 0.7 0.7 0.3 5
1.2 + 0.4 50 t 13 0.2 t 0.08 100 * 46 11.4 f 4.4
” * t k -c
0.4 0.3 0.4 0.4 5
Normal range*
3.5-5.7 2.5-3.7 1.9-4.0 2.0-3.7 5-20
1.2 + 0.18 70 2 7 0.32 i 0.04
1.0-1.9 50-80 0.24-0.4
68 j, 20 4.8 i 1.8
50-loot 3.5-6.0t
1974, p. 83.
Echocardiograms (Table II). The results from the 12 normal Nigerian subjects are also shown. In our studies we have found that our normal values do not differ significantly from those of Feigenbaum.’ Twenty-four (55 per cent) of the patients had pericardial effusion and these accounted for the very large hearts seen on the chest radiographs of some patients. The mean left ventricular end-diastolic dimension was at the upper limit of normal. Thirteen patients had a LVEDD greater than 6 cm., but only two patients had a dimension greater than 6.5 cm., (6.8 cm. each) (Fig. 1). Except for these two patients, the left ventricles were dilated, but not grossly so as in congestive cardiomyopathy.“. ‘I, The left atrium was larger than normal (4 cm.) in 28 patients (Fig. 1). The mean right ventricular end-diastolic dimension was normal (1.7 + 0.6 cm.; normal range 0.8 to 2.4 cm.). However, the measurement of the size of the right ventricle by echocardiography is recognized to be inaccurate.” The indices of left ventricular function showed considerable scatter (Fig. 2). The mean values for the ejection fraction and the mean Vcf were just within the normal range, Table II. Only nine patients had a mean Vcf below 0.9 circs. sec.-’ (Fig. 2). The two patients with the lowest mean Vcf (0.51 and 0.53, respectively), approaching the range expected if there is a severe heart muscle
American
cardiac
failure,” were the only two patients who did not respond quickly to treatment and who had signs of left ventricular failure when restudied. Two other patients had low mean Vcf values (0.6 each) but they appeared to respond well to hospital treatment although they did not return for repeat measurements. There was a good correlation within individuals between the ejection fraction and the mean Vcf, (r = 0.86) (Fig. 3). Fractional shortening was depressed, possibly artificially so because of the increase in cavity size. The estimated stroke volume was greater than 50 ml. in all patients. As the heart rates were high, the estimated cardiac outputs were also high, although there was considerable scatter. The two patients who responded poorly to treatment had lower cardiac outputs (6.6 litres min.-’ each). The three lowest estimated outputs were in three patients who had normal LVEDDs (4.0, 4.3, and 5.0 cm.). Systolic time intervals. (Table III and Fig. 4). There was prolongation of the PEP1 (156 t 27 msec.; normal range 100 to 155 msec.), and shortening of the LVETI (373 + 22 msec.; normal range 395 to 435 msec.). The mean of the ratio PEPULVETI, a better index of left ventricular function, was at the upper limit of normal, 0.41 + 0.08; (normal range 0.27 to 0.41).18 Although the ratio increases with heart failure, half the patients had normal values and the results do not suggest there is severe heart muscle failure in all the patients.
615
Sanderson
et al.
7.0
.
.. .
l .+. .H .. 2:. :. . .. .
6.0.
4.0,
:. ‘2’ .E
l :* .:: ”
:*
ems.
3.0, 2.0. 1.0. 0. 1. Scattergram
1.5
circs.
III. Systolic time intervals on admission (mean t S.D.)
I
. . c
. :.*
t
sec. -1 1.0
....
of individual values of the basic echocardiographic measurements. LVEDD = left ventricular end diastolic dimension; LVESD = left ventricular end systolic dimension; LA = left atria1 dimension. The vertical bars represent the normal ranges. Fig.
20
loo-
” .
:. ..
I :.:I .0;.. Liz
5.0,
.
2.0
.. ..:...
0.5
0
; ” .
50 i %-
i .. :
m Mean Vcf
.a. 2 . 2;.*I ‘$::
-
: : . :.
o&J” fraction
r
15
‘! 1 . .
litres
. > i . E I
min.
-1
10
. In Estimated cardiac output
Fig. 2. Scattergram of individual values of indices of left ventricular function and estimated cardiac output derived from the echocardiograms. The vertical bars represent the normal range.
diogram was done first prior to catheterization. In two patients the echocardiograms were destroyed due to a technical error. In the remaining four patients who also had thermodilution measurements done the mean difference was 1.0 liters min.-’ (Table V).
Table
PEP1 (msec.) LVETI (msec.) PEP/ LVET t = normal
range
156 k 27 373 -C 22
127 f 20 395 i 18
0.41
0.32
from
‘-e 0.08 W&&r
-t 0.04
100-155 395-435 0.27-0.41
et al. 1968.”
Cardiac catheterization data. (Table IV). All the patients had raised atria1 pressures on admission. The right ventricular end-diastolic pressure was raised in all patients and all the patients had pulmonary hypertension. The pulmonary artery end-diastolic pressure (an index of the left ventricular filling pressure) was also raised in all the nine patients in whom it was measured. The mean cardiac output measured by the thermodilution technique for the whole group was 6.3 rt 1.5 liters min.-’ In all patients an echocar-
616
B. Convalescent.
Clinical data. At restudy all the patients had improved. The mean weight loss was 11.5 & 6.2 kilograms (P < 0.001). Two patients had shortness of breath, but none had edema. The blood pressure of these 27 patients had fallen (systolic 116 ? 15 mm. Hg, P < 0.01; diastolic 79 I 15 mm. Hg, P < 0.01). Only three patients had a diastolic blood pressure above 95 mm. Hg. The heart rate had decreased to 94 + 16 beats min. ’ (P < 0.001). The jugular venous pressure was slightly raised in only one patient. Seven patients had a gallop rhythm still and six patients still had an apical systolic murmur. No patient had a pleural effusion or ascites. The cardiothoracic ratio had decreased to 0.57 & 0.08 and only three patients had a ratio above 0.6. The electrocardiograms had changed little; in three widespread T wave inversion had
May, 1979, Vol. 97, No. 5
Postpartum
cardiac
failure
. . .
:*
.:
‘. . l*. .
. .
.l.O
.*
-m--‘-r-, C. l l
r= 0.86
‘:
. . .*
: I :
0
50
Ejection
fraction
.
200,
..$ ...
(%I
been replaced by T wave inversion in the lateral chest leads only. The PCV had risen to 40 t 6 per cent. Echocardiograms. (Table VI). The major changes were the reduction in the size of the chambers and the estimated cardiac output. There was no change in the indices of left ventricular function. The fall of the mitral valve diastolic closure rate from 13 + 3 cm. sec.-’ probably reflects the reduction of blood flow across the valve.“’ Systolic time intervals. (Table VI). There was little change: PEP1 = 135 + 19 msec., LVETI = 359 & 21 msec., PEPI/LVETI = 0.4 2 0.07. Discussion
The accurate assessmentof the performance of heart muscle is still difficult. As a pump the heart’s product is cardiac output, but this is determined by the interaction of many factors such as myocardial contractility, aortic pressure (or afterload), degree of ventricular hypertrophy, and the magnitude of the venous return (or preload) which itself is affected by the peripheral vascular resistance, blood volume, and over-all vascular tone. To isolate the function of the muscle of the heart out of all these variables is almost impossible. However, certain pragmatic approaches are now recognized as being valid for defining groups. The ejection phase indices of left
100
0
. . i. .
. . . fI I I I..
z
100
Fig. 3. Relationship of mean Vcf to ejection fraction, both derived from the echocardiogram. Dotted lines are the lower limits of the normal ranges.
Heart Journal
.
msecs.
1,,, ,i ,, ,,,
American
300.
I
l ..
r-l
LVETI
: l t:{..
0.5
..i;..
0
I PEP I
PEPI LVET
Fig. 4. Scattergram of the systolic time intervals for individual patients. The vertical bars represent the normal ranges.
veritricular function (mean velocity of circumferential fiber velocity (Vcf), ejection fraction, and fractional shortening) have been shown to distinguish between groups of patients with normal and abnormal left ventricular function.“, II1An important advantage is that they can be calculated from echocardiograms. The calculation of volumes from the left ventricular dimensions measured by echocardiography is more debatable.‘” It was clearly recognized in this study that the individual values for any volume were only an approximatation. However, for studying groups of patients this method is sufficiently accurate to decide whether the stroke volume is very low, normal, or very high. Support for this comes from the results using the thermodilution technique for the measurement of cardiac output. The differences were not great in those individuals whose cardiac outputs were measured by both techniques.
617
Sanderson
et al.
NECHANISM
OF HEART FAILURE
IN PFCF
MEAN ARTERIAL STROKE VOLUM3 TOWARDS NOFJ4AL RENAL
HIGH NA INTAKE FUID
ARTERIAL
(=
RETENTION
EDEMA ______*
HEART FAILURE WITH NORMAL C.O.)
PERIPHERAL (A~RI~AD)
PRDLAC’W’
HUMIDITY-+~FLUID
Loss THROUGH SWEATING Fig.
5. Outline
IV. Cardiac catheterization
Table
RA f&d
mechanism
PPCF
14 13 11
53119 55117 53/10
53/35 55/35 53/30
4.8 + 0.49 7.0 k 0.59 8.8 k 0.69
4 (9
20
50/22
(ii)
13
45/15
50/37 45/30
7
13 10 8 20
35/5-20 45/10 25/9 60/17
35/10-20 25/15 60/32
8 9
13 15
52/10 60/18
52133 60/35
5
6 6) (ii)
10
13
range*
-l-+8
in Zaria.
data PA (mm. W
1 2 3
25113 15-28/O-8
RA = right atria1 pressure; RV = right ventricular Figures in brackets are the number of determinations * = from “The Heart” ed. by J. Wilis Hurst, 1974,
25/19 15-28/5-16
Cardiac (liters
output min. y
(beats
HR min.~ I/ 110
(3) (9) (3)
78 130
44 90 61
4.7 t 0.3 (5) 5.2 + 0.4 (8)
126 105
37 49
6.3 k 0.46 (4) 6.2 +- 0.9 (9) 5.2 -' 0.5 (6) -
140 100 102
45 62 51 -
-
-
-
-
-
3.5-6.0
pressure; HR = heart rate;SV = stroke volume; PCW for which the thermodilution cuwes were satisfactory. p, 83.
The surprising fact that emerged from this study was that, despite the clinical evidence of advanced circulatory congestion and edema which was present in all the patients studied, most had maintained relatively good left ventricular function and the cardiac output was high. By contrast, the two patients who had very low values for the mean Vcf and ejection fraction were the only two patients who had continuing signs of heart failure, similar to the pattern described in patients with a congestive cardio-
618
to explain
RV (mm. Hg)
Patient
Normal
Mean (mm.
of the proposed
50-110 = pulmonary
capillary
wedge
pressure.
myopathy.“. l6 In the remainder, the clinical course and echocardiographic findings were not like congestive cardiomyopathy. Instead, the hemodynamic findings were very similar to those seen in patients with acute glumerulonephritis,‘” although none of our patients had any evidence of renal disease.PPCF in Zaria, therefore, may be an example of congestion of the circulatory system which simulates congestive heart failure, but which is not primarily due to it.” But why should apparently normal postpar-
May,
1979,
Vol.
97, No.
5
Postpartum
V. Combined echocardiographic and thermodilu tion measurements Table
Table
cardiac
failure
VI. Follow-up data (27 patients) (mean t
SD.) Measurement
On admission
Convalescent
1. Echocardiographic
~ 1. 2. 5. 6. LVEDD fraction;
6.2 6.0 4.0 5.2
26 42 65 43
63 91 42 61
6.7 8.7 6.1 6.5
= left ventricular end diastolic dimension; SV = stroke volume; CO. = cardiac output.
4.8 7.0 6.3 6.2
k * & f
0.49 0.59 0.46 0.9
EF = ejection
turn woman accumulate such large amounts of fluid? (Fig. 5). It has been suggested previously that the Hausa postpartum practices of eating food rich in salt and lying on heated mud beds may be responsible for overriding the normal control mechanisms of the extracellular volume.” The heating produces a low total peripheral resistance so that the mean arterial pressure cannot be raised except by increasing cardiac output. As the mean arterial pressure is the major determinant of urinary output,” this inability to raise the arterial pressure as efficiently as usual, in women already taking large amounts of sodium, inevitably leads to their retaining excess sodium and water. In such postpartum women, prolactin, known to have sodium retaining propert,ies >I. “? may also play a part. Fluid loss through sweating will also decrease as the humidity increases at the onset of the rains’” (and this probably accounts for the peak of admissions at this time’). The cardiac output will continue to rise in response to the increasing blood volume and venous return in an attempt to excrete the excessive fluid, until the reserves available from left ventricular dilatation and the Frank-Starling mechanism are fully used. This limit will vary for each individual and’those with poor myocardial reserves, from whatever cause, will be unable to sustain the required cardiac output and will develop a form of high output failure. The combination of high venous pressures and a large blood volume rapidly produces gross edema (Fig. 5). A further complication arises if the peripheral resistance should begin to increase. This is likely to happen once edema has formed because of a direct effect of edema on the vessels” and because the increased sodium content of a vessel wall
American
Heart
Journal
LVEDD (cm.) LVESD (cm.) LA (cm.) Aortic root (cm.) MV mean DCR (cm. sec. ‘) Mean Vcf (circs. sec.-‘) Ejection fraction (%) Fractional shortening Estimated S.V. (ml.) Estimated C.O. (liters min. ~‘) 2. Systolic time intervals LVETI (msec.) PEP1 (msec.) LVET/PEP
5.8 4.6 4.3 2.5 14
2 f + + k
0.6 0.7 0.6 0.2 5
5.4 4.2 3.2 2.5 9
1.2 49 0.21 99 11.8
t k f i i-
0.4 12 0.07 39 4.8
1.0 f 0.3 50 If- 11 0.2 f 0.07 82 t 35* 7.5 i 3.3**
373 f 22 156 + 22 0.42 k 0.07
Significance of change from admission **p = i 0.001
to convalescent:
k xk k + +
0.7* 0.7 0.9** 0.2 3**
359 f 21 135 + 19 0.4 + 0.07 *p = < 0.1;
limits the ability of an arteriole to dilate.‘” Cooling will have the same effect. When the filling pressures of the left ventricle are high and the sarcomeres are fully extended, the stroke volume becomes very sensitive to change in afterload, dropping sharply with any further increase.‘!’ This will decrease the cardiac output back towards normal, thus exacerbating the situation even further. The combination.of a high inflow volume with an elevated peripheral resistance produces an elevation of the left ventricular work load far above that present in comparable high output states in which the mean arterial pressure is low. It is possible that these combined loads are sufficient to produce myocardial damage, and this may account for the slightly poor left ventricular function seen in some of the patients. (Exactly the same event will occur on the right side of the heart if the pulmonary artery pressure rises). The fact that the women were able to cope with the hemodynamic load of pregnancy without any sign of heart failure suggests that they did not have pre-existing heart muscle disease. The provoking factors of salt and heat are clearly important in explaining why PPCF is so common around Zaria. But the syndrome we see is similar to some of the previously reported cases. Thus, Demakis and colleagues”” found that the chest radiograph became normal in 14 of 27 patients with PPCF, and in these women the
619
Sanderson
et al.
prognosis was excellent with no further heart failure in 21 subsequent pregnancies. Johnson and associates” reported a patient who was catheterized and in whom the cardiac output was normal and the hemodynamic data was interpreted as showing “biventricular failure, probably of the high output type.” Another study reported cardiac output results from ten patients. In seven patients the values were normal (3.6 to 5.1 liters min. I), although the authors conclude that the “haemodynamic findings were characterised by low output biventricular failure.” The other reports in the literature which contain cardiac output data are six case reports”-“ (two were normal and four were low), and that of Pearce and co-workers,,‘!’ who found that the cardiac outputs were low in three patients. It is possible therefore, that PPCF in temperate climates also may not be initially a primary myocardial problem. A more radical concept is that during the postpartum period the extracellular volume may expand more than is normally possible while the control systems are resetting after the pregnancy. Sodium retention and volume expansion may occur, helped by the high prolactin levels. An expanding blood volume will, as above, produce an increasing venous return, and although the mother is able to cope with this during the pregnancy, she may only be able to do so if the peripheral resistance is low. However, it may be that during the postpartum period in certain women the peripheral vessels respond to volume loads and excess sodium more vigorously. This rise in peripheral resistance may produce a small rise in the blood pressure. (It is possible that this is the mechanism of the well recognized postpartum rise in blood pressure seen in normal women.)“’ But, more importantly, this rise in the peripheral resistance will markedly increase the work load of an already dilated ventricle. Again, as in our own patients, certain individuals may be incapable of coping with the combined demands of a high inflow volume and an increased afterload, so that myocardial damage and a true cardiomyopathy may follow. This theory reconciles two apparently conflicting views of the etiology of PPCF. Most authors have considered it to be a form of cardiomyopathy or intrinsic heart muscle disease. But as was pointed out by Brockington,” there are several objections to this concept. The major difficulty is that it does not explain the hypertension found in
620
most cases, and it seems highly unlikely that any left ventricle that is already in failure due to a myocardial fault is able to maintain a higher t,han normal blood pressure. However, although Brockington concludes that PPCF is hypertensive heart failure, he provides no plausible explanation for the fact that the presenting blood pressures are only moderately raised, they fall to normal quickly after treatment with diuretics, and the women do not need further treatment for hypertension. Therefore, it seems reasonable to suggest that the primary event is volume overloading, which by dilating the ventricles makes them very sensitive to any increase in arterial pressure (which is also due to the same initial volume expansion). The consequent reduction of stroke volume in the face of a high venous return will accelerate the development of the symptoms and signs of congestive heart failure and possibly in a few result in permanent myocardial damage. Therefore, although there are special provoking factors in this area, the Hausa postpartum practices may be exposing a mechanism that is common to postpartum women throughout the world. Summary
Ventricular function has been studied in 43 patients with the peripartum cardiac failure (PPCF) syndrome which occurs around Zaria. All patients had an echocardiogram on admission and 10 patients had right heart catheterization. Despite the gross edema, left ventricular function assessedby echocardiography and systolic time intervals was relatively good and the estimated cardiac outputs were high. At catheterization, although the pressures were high, the cardiac outputs were greater than normal in four out of six patients. No patient had a low cardiac output. These findings are not compatible with a severe heart muscle disorder, or cardiomyopathy. We suggest that the primary event in PPCF of Zaria is fluid retention which leads to a form of high output cardiac failure. The postpartum practices in this area (taking high sodium diets and lying on heated beds) almost certainly cause the fluid to accumulate initially, but the heart may be unable to meet the demands either because of preexisting heart muscle disease or, more likely, because of a rise of the peripheral resistance due to the volume expansion, overburdens such dilated hearts and leads to myocardial damage.
May.
1979,
Vol.
97, No.
5
Postpartum
Since there and PPCF that there traditional
are similarities between this condition in temperate climates, it is possible is a common mechanism which the practices of this area have unveiled.
22. 23.
REFERENCES 1. 2.
