Endgames Case Report

Postpartum fever and shortness of breath

BMJ 2013; 346 doi: http://dx.doi.org/10.1136/bmj.f391 (Published 24 January 2013) Cite this as: BMJ 2013;346:f391
  1. Adam Morton, staff specialist, obstetric medicine and endocrinology
  1. 1QLD Diabetes Centre, Mater Hospital, Raymond Tce, South Brisbane, Australia 4101
  1. Correspondence to: A Morton adam.morton{at}mater.org.au

A 26 year old gravida 3 para 2 woman developed fever and malaise after an uncomplicated spontaneous vaginal delivery at 38 weeks’ gestation. Postpartum vaginal loss was unremarkable. Her baby, a boy with a birth weight of 3420 g, was well. Her symptoms were initially attributed to a viral illness. However, she gradually became more unwell over the next two days with lightheadedness, abdominal pain, fever, and shortness of breath, and she was noted to have tachycardia and hypotension. On the third day postpartum, while still an inpatient, she was extremely unwell. Pulse rate was 140 beats/min, blood pressure was 80/20 mm Hg, and oxygen saturations were 91% while breathing oxygen at 6 L/min. Heart sounds were dual, chest examination showed bibasal crackles, and she had hypogastric tenderness. Initial investigations were haemoglobin 96 g/L (pregnancy specific reference range 115-165), white blood cell count 11.2×109/L (4-15), platelets 155×109/L (150-400), prothrombin time 21 s (11-16), activated partial thromboplastin time 52.7 s (23-38), fibrinogen 13.2 µmol/L (5.1-11.8; 1µmol/L=3401 mg/dL), creatinine 136 µmol/L (1 µmol/L=0.01 mg/dL;30-70), estimated glomerular filtration rate 41 mL/min (80-120), serum bicarbonate 10 mmol/L (1 mmol/L=1 mEq/L; 22-33), and venous lactate 6.6 mmol/L (1 mmol/L=9.01 mg/dL; 0.5-2.2).

She was given intravenous antibiotics and hydrocortisone and transferred to the intensive care unit (ICU). Chest radiography showed diffuse bilateral lower zone infiltrates. A non-contrast computed tomogram of the abdomen and pelvis was unremarkable. Echocardiography showed bilateral reduction in left and right ventricular function, with a left ventricular ejection fraction (LVEF) of 40% and no evidence of endocarditis. Cardiac troponin I was raised at 0.34 µg/L (<0.05), with a subsequent peak value of 22 µg/L.


  • 1 What is the most likely cause of the patient’s initial deterioration?

  • 2 What were the possible factors in her cardiac dysfunction?

  • 3 What are the risks for a future pregnancy?

  • 4 How would you manage a future pregnancy?


1 What is the most likely cause of the patient’s initial deterioration?

Short answer

The most likely cause is shock owing to puerperal sepsis. Other causes of severe illness postpartum are severe pre-eclampsia, postpartum haemorrhage, acute fatty liver of pregnancy, thrombotic thrombocytopenic purpura, amniotic fluid embolism, massive pulmonary embolism, disseminated herpes simplex infection, viral or bacterial pneumonia, catastrophic antiphospholipid syndrome, phaeochromocytoma, severe thyrotoxicosis, adrenal insufficiency, and maternal pulmonary hypertension.

Long answer

The presentation with multiorgan failure, hypotension, and coagulopathy is consistent with septic shock. Puerperal sepsis is thought to cause at least 75 000 maternal deaths each year, mainly in low income countries. Studies from high income countries report maternal morbidity as a result of sepsis in 0.1-0.6 per 1000 deliveries.1 Infection is often caused by more than one organism. Endometritis, wound infection, mastitis, urinary tract infection, and septic thrombophlebitis (including ovarian vein thrombosis) are the chief causes of puerperal infection.

Several other conditions could have caused some features of this presentation. Severe pre-eclampsia can manifest with acute left ventricular dysfunction, pulmonary oedema, and renal failure. Coagulopathy can occur with disseminated intravascular coagulation, but this patient’s normal platelet count and fibrinogen concentration are not consistent with this condition. Shock may occur in pre-eclampsia owing to hepatic or splenic subcapsular bleeding or rupture. The absence of previous hypertension and proteinuria make pre-eclampsia unlikely in this patient, and there were no features of HELLP (haemolysis, elevated liver enzymes, or low platelets) syndrome.2

Postpartum haemorrhage may result in cardiovascular collapse, tachycardia, disseminated intravascular coagulation, and multiorgan failure. Concealed blood loss may rarely occur postpartum with perivaginal pelvic haematoma, adrenal haemorrhage, or splenic artery rupture.3 The gradual onset of the patient’s symptoms and signs without a serious fall in haemoglobin concentration are against haemorrhage as a cause of her deterioration.

Acute fatty liver of pregnancy may be associated with coagulopathy without thrombocytopenia, pulmonary oedema, and renal failure (as a result of acute tubular necrosis). Abnormal liver function test results are always seen with this condition, together with greatly reduced concentrations of antithrombin III, and patients often have hypoglycaemia and hyperammonaemia. Cardiomyopathy is not usually seen with this condition.2

Thrombotic thrombocytopenic purpura or haemolytic uraemic syndrome may present in the postpartum period as fever, acute renal failure, and cardiogenic shock. Thrombocytopenia and microangiopathic haemolytic anaemia are a feature in this condition.2

Amniotic fluid embolism can present with severe hypotension and hypoxia as a result of cardiogenic shock and disseminated intravascular coagulation. This condition would usually present during delivery or immediately postpartum.4

Massive pulmonary embolism may present with hypotension and hypoxia, although it would usually be associated with pulmonary hypertension and not left ventricular dysfunction on echocardiography.5

Disseminated infection with herpes simplex virus may present with fever, hypotension, coagulopathy without thrombocytopenia, pulmonary infiltrates, renal failure, and left ventricular dysfunction. The absence of hepatitis or encephalopathy make this diagnosis unlikely.6

Other viral infections, especially influenza A H1N1, may be complicated by pneumonia (both primary viral and secondary bacterial), myocarditis, coagulopathy, and renal failure.7

Catastrophic antiphospholipid syndrome in pregnancy is commonly associated with HELLP syndrome, acute renal failure, cerebrovascular ischaemia, and less commonly cardiogenic shock as a result of dilated cardiomyopathy.8

A predominantly adrenaline secreting phaeochromocytoma may manifest with hypotension, tachycardia, dilated cardiomyopathy, pulmonary oedema, and acute renal failure.9

Severe thyrotoxicosis may be associated with fever, tachycardia, left ventricular dysfunction and pulmonary oedema, pulmonary hypertension, multiorgan failure, and disseminated intravascular coagulation.10

Adrenal insufficiency may manifest as fever, hypotension, and tachycardia. This may occur postpartum owing to bilateral adrenal haemorrhage (especially in antiphospholipid syndrome), Sheehan’s syndrome as a result of peripartum hypotension, or lymphocytic hypophysitis.11

Maternal pulmonary hypertension is associated with a 30-50% risk of death as a result of pregnancy and may present as sudden cardiovascular collapse or death during delivery or in the first month after delivery.12

A systematic review of 40 studies of ICU admissions in postpartum women and those under obstetric care found that rates varied from 0.7 to 13.5 per 1000 deliveries.13 The median rate was 2.7 ICU admissions per 1000 deliveries, or 1 ICU admission per 370 deliveries. Pregnant or postpartum women accounted for 0.4-16.0% of ICU admissions in study centres. Hypertensive disorders of pregnancy were the most prevalent indication for ICU admission (median 0.9 cases per 1000 deliveries; range 0.2-6.7), followed by obstetric haemorrhage (0.7/1000; 0.1-2.3), with sepsis (0.2/1000; 0.0-2.3) and other direct obstetric complications (0.1/1000; 0.0-1.6) occurring less often. The incidence and profile of mothers admitted to the ICU did not differ between developing and developed countries.

2 What were the possible factors in her cardiac dysfunction?

Short answer

Left ventricular dysfunction might have been caused by sepsis related cardiomyopathy or peripartum cardiomyopathy (PPCM).

Long answer

Left ventricular systolic function is severely reduced on echocardiography (left ventricular ejection fraction (LVEF) <40%) in 20-60% of patients with septic shock. In one study patients with septic shock and normal troponin all had normal systolic function, whereas half of those with raised troponin had left ventricular systolic dysfunction.14 Isolated diastolic dysfunction was seen in 20-40% of patients with septic shock, with a further 20% having both systolic and diastolic dysfunction. Raised troponin concentrations and cardiac dysfunction on echocardiography typically resolve within 7-10 days of presentation with septic shock.15 16 A recent study of 45 critically ill patients with left ventricular dysfunction found that echocardiographic changes resolved within one month of initial presentation in all but one patient, who had restrictive diastolic dysfunction only, which persisted at follow-up two years later.17

Our patient’s cardiac dysfunction persisted for more than six months postpartum. Although her initial left ventricular dysfunction was probably caused by sepsis, the persistence of this problem suggests an additional cause of cardiomyopathy, with peripartum cardiomyopathy being most likely in the absence of alternative causes. Peripartum cardiomyopathy is defined as the development of heart failure, with left ventricular systolic dysfunction, in the last month of pregnancy or within five months of delivery in the absence of another identifiable cause. There are wide geographical variations in incidence, with risk factors including African ethnicity, age greater than 30 years, multiparity, pregnancy with multiple fetuses, and hypertensive disorders of pregnancy. Prognosis varies widely, depending on the population studied, with maternal mortality rates of 6-23%, and cardiac transplantation rates of 4-7%. About half of patients regain LVEF above 50%, the degree of recovery being greatest in those with baseline LVEF >30%. It was previously thought that almost all of the recovery of left ventricular function occurred within six months of diagnosis, but several studies have shown that many people are “slow responders,” with more than half taking 12-18 months from diagnosis to recover fully. One patient did not fully recover until 48 months after presentation.18

3 What are the risks for a future pregnancy?

Short answer

The risk of deterioration in left ventricular function and death in any subsequent pregnancy is high. The degree of risk depends on the degree of recovery of left ventricular function.

Long answer

Subsequent pregnancies were associated with maternal death in 7-33% and worsening heart failure in 44-53% of women with previous peripartum cardiomyopathy with incomplete recovery of left ventricular function.19 20 21 22 23 Doctors should advise women with persistent left ventricular dysfunction after peripartum cardiomyopathy or LVEF less than 25% at the time of diagnosis against subsequent pregnancy and give advice on contraception or sterilisation. The levonorgestrel releasing intrauterine device is the safest and most effective method of contraception for women with heart disease. Surrogacy should be discussed as a possible option for couples who want more children. Termination should be discussed in the event of an unplanned pregnancy.

In a series of women who regained LVEF greater than 50% after peripartum cardiomyopathy, subsequent pregnancy was associated with a reduction in mean LVEF from 56% to 49%, and LVEF fell by more than 20% in 21% of women.21 Twenty one per cent of women developed symptoms of heart failure. These women are therefore also at risk of serious morbidity and appropriate contraception should be discussed. Exercise stress echocardiography may help stratify the risk of deterioration in left ventricular function in subsequent pregnancies in women who recover normal cardiac function.24

Currently, no data are available on pregnancy outcomes in women who have had cardiac transplantation because of previous peripartum cardiomyopathy. Rates of graft rejection and death are known to be higher in these women compared with other age matched heart transplant recipients.25 Sixty three pregnancies have been reported in 39 women who have undergone cardiac transplantation for reasons other than peripartum cardiomyopathy.26 27 In these women 47% of the pregnancies were complicated by hypertension and 10% by pre-eclampsia. Although graft rejection occurred in 22% of pregnancies, this was mild and no episodes of graft loss were reported within two years of pregnancy. The live birth rate was 70-80%, with a mean gestation of 37-38 weeks.

4 How would you manage a future pregnancy?

Short answer

Management of peripartum cardiomyopathy requires a multidisciplinary team approach involving midwives and an obstetrician, cardiologist, anaesthetist, paediatrician, and intensive care doctor. Diuretics, digoxin, β blockers, hydralazine, and nitrates can all be used safely during pregnancy. Anticoagulation with low molecular weight heparin should be used if LVEF is less than 35%. Dopamine agonists may have a role in treatment and possibly prevention of peripartum cardiomyopathy.

Long answer

Management of peripartum cardiomyopathy requires a multidisciplinary team approach involving midwives and an obstetrician, cardiologist, anaesthetist, paediatrician, and intensive care physician. Angiotensin converting enzyme (ACE) inhibitors and angiotensin II receptor blockers are fetotoxic and cannot be used in the second or third trimester. A combination of hydralazine and nitrates may be used as an alternative. With the exception of atenolol, β blockers are safe during pregnancy. Spironolactone is said to be contraindicated because large doses caused demasculinisation of exposed male rat fetuses, although there is no evidence of this in humans. The use of eplerenone has been reported in only two pregnancies. Vaginal delivery with early epidural and assisted second stage with forceps or vacuum is preferred over caesarean section. Postpartum non-steroidal anti-inflammatory drugs should be avoided, and ACE inhibitors started. Captopril, enalapril, benzapril, and quinapril are safe for breastfeeding mothers. ACE inhibitors and β blockers should be continued for at least one year after recovery of left ventricular function.

It has recently been recognised that a 16 kDa prolactin fragment impairs cardiomyocyte function, promotes endothelial cell apoptosis, disrupts capillary structures, and could play a pathophysiological role in peripartum cardiomyopathy. A proof of concept pilot study showed that the addition of bromocriptine to standard treatment seemed to improve LVEF in women with acute severe peripartum cardiomyopathy.28 Six case reports describe the rapid improvement in cardiac function in peripartum cardiomyopathy with dopamine agonists (five with bromocriptine, one with cabergoline). Thus, breastfeeding should probably be discouraged in these mothers, and the use of dopamine agonists considered in addition to standard heart failure management. Randomised controlled trials testing the efficacy of dopamine agents in the treatment of peripartum cardiomyopathy are needed.

Patient outcome

Blood cultures from the mother grew Streptococcus pyogenes (group A, biotype 3, BA exotoxin, emm49). Histological analysis of samples obtained by endometrial curettage showed inflamed decidua and grew the same organism on culture, indicating that endometritis was the source of puerperal sepsis. Nasopharyngeal aspirate was negative for respiratory viruses including influenza A and influenza B.

She developed progressive multiorgan failure, requiring 27 days in the ICU. Repeat echocardiography on the sixth day postpartum showed LVEF of 25%, despite high dose inotropes. She was subsequently treated with bisoprolol and enalapril. She elected to breast feed. Two months postpartum her LVEF was unchanged at 25% but improved to 40% by six months. One year postpartum cardiac function was normal, with LVEF 60%. Three years later she asked for her drugs to be stopped, with a view to considering another pregnancy. She would not consider altruistic surrogacy.

Enalapril then bisoprolol were slowly weaned and stopped, with her cardiac function remaining stable on echocardiography. Exercise stress echocardiography showed normal resting LVEF of 60%, with an increase in contractility with exercise to 65%. She is contemplating another pregnancy having discussed the risks.


Cite this as: BMJ 2013;346:f391


  • Competing interests: The author has completed the ICMJE uniform disclosure form at www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declares: no support from any organisation for the submitted work; no financial relationships with any organisations that might have an interest in the submitted work in the previous three years; no other relationships or activities that could appear to have influenced the submitted work.

  • Provenance and peer review: Not commissioned; externally peer reviewed.

  • Patient consent obtained.