CHAPTER 7

GENITAL TRACT SEPSIS

Summary

There were 16 deaths due to genital tract sepsis, a similar number to those which occurred in the previous triennium. Fourteen of the deaths are counted in this Chapter. Two others followed spontaneous abortions and are counted in Chapter 6, which discusses deaths occurring before 24 weeks' gestation. There was also one Late Direct death, which is counted and discussed in Chapter 14. Substandard care occurred in five of the cases discussed in this Chapter.

Genital tract sepsis: key recommendations Puerperal sepsis is not a disease of the past, and GPs and midwives must be aware of the signs and be prepared to institute immediate treatment and referral of any recently delivered woman with a fever and/or offensive vaginal discharge. There is clear evidence from controlled trials showing the benefit of prophylactic antibiotics for caesarean section. When infection develops and the patient is systemically ill, urgent and repeated bacteriological specimens, including blood cultures, must be obtained. The advice of a microbiologist must be sought at an early stage to assist with the use of appropriate antibiotic therapy. In serious cases doctors should be prepared to give parenteral antibiotics before the diagnosis can be confirmed.

The number of Direct deaths from sepsis compared with the previous two triennia is shown in Table 7.1

Table 7.1 - Maternal deaths from genital tract sepsis including abortion and ectopic pregnancy with rates per million maternities; United Kingdom 1985-96.

The 14 cases in this Chapter have been divided into sepsis before delivery, sepsis after vaginal delivery and sepsis after surgery. There were no cases of sepsis after ectopic pregnancy.

Sepsis before delivery

There was one case of sepsis before or during labour:

A grossly obese parous woman was admitted to hospital at 36 weeks' gestation with pyrexia and a 24-hour history of increasingly severe right-sided abdominal pain. On examination, the uterus was tender and there were no fetal heart sounds. A presumptive diagnosis of placental abruption or possibly a urinary tract infection was made. The initial coagulation screen was normal and the white cell count (WCC) was elevated. Intravenous antibiotics were administered and labour was induced with prostaglandin. During labour she developed peripheral cyanosis, hypotension, rapid respirations and coagulopathy. A stillborn infant was delivered vaginally. She then had a massive post-partum haemorrhage which responded to intensive treatment. Some hours later she had increasing respiratory difficulty and was transferred to an Intensive Care Unit (ICU) for elective ventilation. In spite of this she became bradycardic and hypotensive and had an asystolic arrest which failed to respond to resuscitation. It was later determined that blood cultures had grown Group A ß-haemolytic streptococcus. An autopsy was not performed.

Sepsis after vaginal delivery

Ten women died from sepsis of the genital tract after vaginal delivery:

A young woman who was 26 weeks pregnant developed severe shock and DIC some 12 hours after spontaneous labour and delivery of fresh stillborn twins. In spite of intensive treatment she died within six hours. A blood culture grew Group A ß-haemolytic streptococcus and autopsy findings supported the diagnosis of death from septic shock.

The pregnancy had been complicated by a twin-to-twin transfusion with polyhydramnios which was treated by intrauterine laser ablation of placental blood vessels. There had been no complications during or after this operation and it was considered to have been successful. It was most unlikely to have been associated with her death.

A parous woman, who had intermittent episodes of vaginal bleeding from 14 weeks' gestation, was admitted to hospital at 26 weeks in spontaneous labour with an intrauterine death. She had a breech delivery and the placenta was said to be offensive. Twenty-four hours after delivery she developed jaundice, shock and DIC, was transferred to ICU and in spite of resuscitative measures had a cardiac arrest and died 13 hours later. At autopsy the appearances in the uterus were typical of clostridrial septicaemia and necrosis and this diagnosis was confirmed by microbiology.

The medical notes in this case provided no information about the observations which were made in the 24 hours after delivery. Furthermore, antibiotics were not prescribed prior to her transfer to ICU in spite of clinical evidence of intrauterine infection at the time of delivery. There was therefore substandard care in this case.

A primigravida died from complications of necrotising fasciitis having had a normal pregnancy and delivery. She presented with a swollen left leg eight days after delivery and her GP arranged for immediate transfer to hospital. A provisional diagnosis of ileo-femoral thrombosis was made but her general condition rapidly deteriorated with extensive necrosis of the tissues extending from the vulva to the knee, confirming a diagnosis of necrotising fasciitis. Intensive antibiotic therapy and incision of the fascial planes of the leg were undertaken without maintained improvement. She subsequently had hyperbaric oxygen therapy, a hysterectomy to remove the focus of the infection and extensive debridement of the necrosed tissues. Her condition continued to deteriorate and she died 28 days post-partum. Histology of the excised uterus showed extensive haemorrhagic infarction and this was the presumed source of the infection. The organism identified from several tissues was Group A haemolytic streptococcus.

Necrotising fasciitis is extremely rare but similar cases in association with pregnancy have been reported in the last three triennial Reports.

A parous woman had an uneventful pregnancy until 34 weeks' gestation when she was admitted 24 hours after spontaneous rupture of membranes. Labour was induced with prostaglandin pessaries a further 24 hours after admission and she had an uncomplicated assisted breech delivery of a live infant. She was discharged home a day later. On the third day post-partum she complained of some abdominal pain at home and was seen by her GP and midwife who were satisfied with her condition. The following morning she was found dead at home. An autopsy examination confirmed inflammation of the uterine cavity with Group B ß- haemolytic streptococcus. It was concluded that her death was due to septic shock.

A possible risk factor in this case was the delay between rupture of the membranes and delivery, but this was no longer than 48 hours. She developed lower abdominal pain without obvious evidence of infection, followed by sudden collapse and death. There was no element of substandard care in this case.

A young parous woman had a planned home confinement with an uncomplicated labour and delivery at 39 weeks' gestation supervised by a midwife. Postnatal observations were satisfactory for four days. She then had an isolated episode of severe diarrhoea for which she received treatment after a phone contact with a locum GP. She was seen by the midwife the next day and on examination her temperature was 40°C and the pulse rate was 140/min. The lochia were normal and not offensive and there was no abdominal tenderness. The midwife contacted the GP, who attended the patient and after an examination he considered her condition satisfactory. However, he failed to record her temperature. When the midwife revisited the patient later the same day her temperature and pulse remained elevated. The next morning the patient's husband called an ambulance as her condition had suddenly deteriorated. She died soon afterwards. An autopsy confirmed that she had died from Group A haemolytic streptococcal septicaemia.

This case emphasises the serious nature of post-partum infection, which can be insidious and can rapidly progress to fatal septicaemia. Although the midwife alerted the GP to the pyrexia it was not fully investigated nor was appropriate action taken with antibiotic treatment and transfer to hospital. This case represents substandard care.

A primigravida had an uneventful pregnancy with spontaneous onset of labour at 39 weeks' gestation and delivery by ventouse extraction. She was discharged well, four days post-partum. Two weeks later she was admitted to the Accident & Emergency Department with a 24-hour history of diarrhoea and vomiting. She was extremely shocked and had a cardiac arrest shortly after arrival at the hospital. She was successfully resuscitated and transferred to the ICU. She then developed marked abdominal distension with rigidity. A laparotomy was performed to exclude perforation of an abdominal viscus. At operation there were two litres of peritoneal fluid and an extremely bulky uterus. A sub-total hysterectomy was performed. Despite intensive therapy her condition did not improve and she died two weeks post-partum.

An autopsy showed evidence of terminal multi-organ failure. On histological examination there were no residual products of conception or intrauterine infection. It was assumed that the diagnosis was toxic shock syndrome but no organism was isolated from various cultures.

A parous woman with a history of premature labour and recurrent abortions was admitted to hospital at 23 weeks' gestation because of suspected spontaneous rupture of the membranes. She had a history of chorioamnionitis in a previous pregnancy. A cervical suture had been inserted in the first trimester of the present pregnancy and she had repeated vaginal infections treated by courses of antibiotics. The day after admission there was obvious evidence of ruptured membranes. Some six hours later she was pyrexial with clinical signs of systemic infection. The cervical suture was removed, with the release of pus from the uterine cavity. At this stage antibiotics were commenced. Her condition rapidly deteriorated and she was transferred to the ICU, where she died three hours later after a cardiac arrest. A vaginal swab taken on admission but not reported until three days after her death showed a heavy growth of Escherichia coli. An autopsy was not performed.

This patient had an overwhelming infection which progressed rapidly after the first clinical signs of the condition. In view of her past history antibiotic treatment should have been initiated at an earlier stage. There was no attempt to obtain an urgent report from the microbiology laboratory.

A primigravida had an uneventful pregnancy until 37 weeks' gestation when she was admitted to hospital with vomiting, pyrexia and a pulse rate of 136/min. There was no obvious source of infection and she was said to have a flu-like illness. Some six hours later she complained of lower abdominal pain and was transferred to the labour ward in established labour. Within two hours she developed peripheral cyanosis and shock, and fetal death in utero was confirmed by ultrasound scan. She was started on antibiotics and transferred to the ICU where she rapidly developed DIC. She had a forceps delivery without excessive haemorrhage, but her condition deteriorated and she died less than 18 hours after her initial admission.

A blood sample taken before delivery grew Group A ß-haemolytic streptococci. An autopsy examination provided no evidence of any other cause of death apart from septicaemia. The origin of the infection was not determined. There was no significant inflammation within the lungs or the uterine cavity.

A young multiparous woman was booked for confinement in a GP unit. She had an uneventful pregnancy, spontaneous onset of labour at 42 weeks' gestation and a normal delivery of a healthy infant following a short labour. She was discharged home 24 hours post-partum. Some seven days later she complained of feeling unwell and lost her appetite. She was again seen by her GP on the 18th day post-partum with a flu-like illness and had developed a rash. The next day her condition had deteriorated and she was admitted to an ICU with pulmonary oedema. In spite of artificial ventilation and antibiotics her condition deteriorated and she died several days later. Vaginal swabs grew Group A haemolytic streptococcus and the autopsy confirmed that death was due to septic shock.

One further case of overwhelming sepsis following a normal delivery was reported but no further details could be obtained. It is included in the figures.

Sepsis after surgery

Three women died after surgery:

A primigravida had an uneventful pregnancy and was admitted in spontaneous labour at 38 weeks' gestation. She was delivered by ventouse extraction because of fetal distress in the late second stage of labour. The immediate postnatal period was uneventful but on the second day after delivery she complained of abdominal pain. She was apyrexial but her pulse was 100/min. On the following day she remained apyrexial but complained of increasing abdominal pain. The episiotomy scar was reddened and lochia were offensive. A vaginal swab revealed a strong growth of ß-haemolytic streptococci. She rapidly developed septic shock and a total abdominal hysterectomy and vulval debridement were performed. In spite of intensive therapy including hyperbaric oxygen her condition slowly deteriorated with multiple organ failure. She died six weeks after delivery.

This case again illustrates the rapidity and seriousness of postnatal infection. The complaint of abdominal pain was assumed to be after-pains and the rapid pulse rate in the absence of pyrexia was not recognised as significant. There was therefore some delay in commencing appropriate antibiotic therapy.

A primigravida was to have an elective caesarean section because of previous uterine surgery. She was admitted to hospital in spontaneous labour at 37 weeks' gestation and had an uncomplicated caesarean section without antibiotic cover. She was discharged on the fourth day post-partum but re-admitted three days later with a wound infection. This was treated with antibiotics and she was discharged after a few days. Four days later she was again re-admitted with abdominal pain and pyrexia and an ultrasound scan showed an intra-abdominal abscess. A laparotomy was performed and the abscess was drained. Her condition did not improve and she was found to have further abscess formation. Unsuccessful attempts were made to drain this by needle aspiration. A further laparotomy was performed, the abscess was drained and a sub-total hysterectomy was performed. She developed a post-operative haemorrhage and had a further laparotomy and splenectomy. In spite of intensive therapy her condition continued to deteriorate and she died several weeks later.

At autopsy there was massive liver necrosis. Both coliforms and streptococci were cultured from various sites. Prophylactic antibiotics would have been appropriate at the time of caesarean section. The attempt to drain the recurrent abscess by aspiration was inappropriate and as a result there was a delay in undertaking the repeat laparotomy.

A young primigravida concealed her pregnancy until 37 weeks' gestation when she booked for hospital confinement. The progress of her pregnancy at that stage was satisfactory. She was admitted to hospital the following day in an agitated state complaining of backache and uterine contractions; her pulse rate was 140/min and an ultrasound scan confirmed fetal death. Some five hours later she had a mottled appearance and her temperature was 38.1° C. A blood test taken on admission was then reported as profound thrombocytopenia. Medical induction of labour was commenced but her condition deteriorated with a clinical diagnosis of severe DIC. She received intensive supportive therapy and a caesarean section was performed with delivery of a fresh stillbirth. Her condition continued to deteriorate and she died 15 hours after her admission to hospital. At autopsy there was no evidence of amniotic fluid embolism or placental abruption but group B streptococcus was cultured from an intrauterine swab. It was concluded that death was due to fulminating septicaemia.

The serious nature of this patient's condition was not recognised on admission. There was a delay in analysing the blood to confirm the diagnosis of severe thrombocytopaenia. In the face of her rapidly deteriorating condition induction of labour in this primigravid woman was unlikely to produce an improvement in the time scale necessary and an earlier caesarean section would have been more appropriate. It was a difficult diagnostic problem but on balance, there was a degree of substandard care.

Deaths counted in other Chapters

Two further cases of genital sepsis associated with early pregnancy are counted and discussed in Chapter 6. Neither was associated with substandard care:

A young parous woman was seen by her GP at eight weeks' gestation with vaginal bleeding. On examination, the cervix was found to be 3 cm dilated and he assumed she had had a spontaneous abortion. A subsequent ultrasound scan, however, confirmed a normal intrauterine pregnancy. During the next 10 weeks she had intermittent episodes of vaginal bleeding but repeat ultrasound scans again confirmed normal progress of the pregnancy. At 19 weeks' gestation she was admitted to hospital with jaundice, further vaginal bleeding and uterine contractions. She delivered a macerated fetus and in view of a purulent vaginal discharge was started on intravenous antibiotics. Six hours later she underwent an evacuation of a retained placenta under general anaesthesia. Within 24 hours she developed DIC and severe shock and had a cardiac arrest which failed to respond to resuscitation. An autopsy confirmed that she had died from E. coli septicaemia of genital tract origin.

A parous woman was admitted with an incomplete miscarriage at 10 weeks' gestation. She was tachycardic and hypotensive but her condition improved after evacuation of the uterus. She was readmitted a few days later with signs of septic shock. Despite all attempts at treatment, including antibiotics and hysterectomy to remove the source of infection, she died. The organism was not identified.

There was one Direct Late death counted in Chapter 14 but described here:

A parous woman who had a completely normal pregnancy, delivery and puerperium was admitted to hospital six weeks post-partum with flu-like symptoms and a purpuric rash. Nothing abnormal was found on examination of the genital tract. Blood cultures grew Group A haemolytic streptococcus and appropriate antibiotic and supportive therapy was instituted. She fairly rapidly developed shock, DIC and multiple organ failure and died four days later despite intensive therapy. At autopsy a small amount of pus was expressed from the fallopian tubes and it was therefore assumed that the origin of the infection was the genital tract.

Comments

These cases demonstrate, once more, that infection must never be underestimated and that it continues to be an important cause of maternal mortality. The onset of infection, particularly after spontaneous rupture of the membranes, can be insidious and can rapidly progress to a fulminating septicaemia.

Puerperal sepsis is not a disease of the past, and GPs and midwives must be aware of the signs and be prepared to institute immediate treatment and referral of any recently delivered woman with a fever and/or offensive vaginal discharge.

As noted in the previous Report, there is clear evidence from controlled trials showing the benefit of prophylactic antibiotics for caesarean section. The Cochrane database ¹ states:

"Antibiotic prophylaxis markedly reduces the risk of serious postoperative infection, such as pelvic abscess, septic shock and septic pelvic vein thrombophlebitis. A protective effect of the same order of magnitude is seen for endometritis. The degree of reduction in risk of wound infection is somewhat less, but is still substantial. The evidence for these benefits is overwhelming."

When infection develops and the patient is systemically ill, urgent and repeated bacteriological specimens, including blood cultures, must be obtained. The advice of a microbiologist must be sought at an early stage to assist with the use of appropriate antibiotic therapy. In serious cases doctors should be prepared to give parenteral antibiotics before the diagnosis can be confirmed.

Clinicians must remain vigilant for early signs of invasive streptococcal disease. The signs of necrotising fasciitis are high fever plus swelling and marked tenderness localised to a muscle mass. Early treatment with antibiotics still seems to be the best way to prevent death.

There have been several recent reports² suggesting an increase in the number of cases of serious invasive streptococcal infections, with and without shock, suggesting spread of a more virulent clone. The course of these infections is dramatically rapid, as reflected in the cases reported here.

References

1. Enkin, M., Keirse, M.J.N.C., Renfrew, M. & Neilson, J. A Guide to Effective Care in Pregnancy and Childbirth 2nd Edn. Oxford: OUP, 1995; 322-7.

2. Holm, S.E. Invasive Group A streptococcal infections. New England Journal of Medicine 1996; 335: 590-1.

We welcome your comments on this site. Prepared 14 December 1998