SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Case report
  4. Discussion
  5. Acknowledgements
  6. References

A 38-year-old woman experienced a massive postpartum haemorrhage 30 minutes after emergency caesarean delivery. The patient became severely haemodynamically compromised with an unrecordable blood pressure. Rapid fluid resuscitation was limited by the capacity of the intravenous cannula in place at the time and inability to establish additional vascular access using conventional routes in a timely manner. An intraosseous needle was inserted in the proximal humerus at the first attempt and administration of resuscitation fluid by this route subsequently enabled successful placement of further intravenous lines. Blood and blood products were deployed in conjunction with intra-operative cell salvage and transoesophageal Doppler cardiac output monitoring was used to assess adequacy of volume replacement. Haemorrhage control was finally achieved with the use of recombinant factor VIIa and hysterectomy.

The 2003–2005 UK Confidential Enquiries into Maternal Deaths recorded obstetric haemorrhage as the third highest cause of maternal death [1]. Severe obstetric haemorrhage is often associated with adverse postoperative outcomes including sepsis and acute renal failure requiring admission to the intensive care unit (ICU). Analysis of data collected from survivors of severe obstetric haemorrhage, so-called ‘near miss events’, has yielded important information to improve management of life-threatening conditions and may be able to serve as a marker of quality of delivery of healthcare [2–4].

Case report

  1. Top of page
  2. Summary
  3. Case report
  4. Discussion
  5. Acknowledgements
  6. References

A 38-year-old, 74-kg woman in her second pregnancy presented to our institution in established labour. She had had a caesarean section for a previous delivery but expressed a preference for vaginal delivery in this pregnancy and, on admission, requested epidural analgesia. Accordingly, a 16-G intravenous cannula was placed in the patient’s left arm and the epidural sited. Despite steady progress to full cervical dilation, delivery had not occurred after 2 h. In view of the presence of meconium and occasional fetal decelerations, she was transferred to theatre for delivery by caesarean section. The epidural was easily topped up using 10 ml of a 50:50 mix of 2% lidocaine and 0.5% bupivacaine, achieving a bilateral dermatomal block height to T4 (tested with ethyl chloride). Blood pressure was maintained without the need for vasopressor treatment during extension of the block. Pre-operative haemoglobin was 12.7 g.dl−1. A healthy 4.4-kg infant was delivered. Two bolus doses of oxytocin (5 IU) were given and an infusion of oxytocin 10 IU.h−1 commenced. Misoprostol (800 μg) was administered rectally at the conclusion of surgery, in line with departmental policy. Two and a half milligrams of diamorphine were given via the epidural catheter at the end of surgery, and the epidural catheter subsequently removed.

In view of the intra-operative blood loss of more than a litre, a blood transfusion was commenced via the existing 16-G cannula, shortly before the patient was transferred to the recovery area. Vital signs recorded on arrival in the recovery area were blood pressure 101/44 and pulse rate 100 beats.min−1. Observations were repeated every ten minutes and recorded in the notes as ‘satisfactory’. After approximately 30 minutes, the patient became progressively listless, coinciding with the observation of a brisk vaginal bleed by the attending midwife.

By the time medical staff arrived, the patient had become obtunded, diaphoretic and tachycardic, with a pulse rate of 120 beats.min−1, an unrecordable blood pressure and an impalpable radial pulse. High-flow oxygen was administered by facemask and the blood transfusion rate was increased by pressurising the infusion bag. However, this was inadequate to address the clinical situation. Several personnel tried to establish additional peripheral (i.e. all four limbs) and central, the patient (using right internal jugular and subclavian approaches) venous access simultaneously, but these attempts were unsuccessful. Unfortunately, ultrasound equipment was not immediately available. Concurrent measures were instituted to control the haemorrhage, including bimanual uterine compression, additional oxytocics including carboprost 250 mg intramuscularly, and repeated doses of intravenous oxytocin 10 IU, given as a slow bolus. A second dose of rectal misoprostol 800 μg was also administered. These events triggered simultaneous activation of the hospital’s major obstetric haemorrhage protocol, and initiated involvement of ICU personnel, who confirmed the requirement for urgent intravascular access to permit rapid fluid resuscitation. Accordingly, an intraosseous needle was obtained from our institution’s emergency department, and immediately placed at the first attempt in the patient’s right humeral head using an EZ-IO driver (Vidacare, San Antonio, TX, USA), within ten minutes of activation of the major haemorrhage protocol (Fig. 1). Rapid infusion using this 15-G device to supplement the existing peripheral access more than doubled the rate of fluid resuscitation. A volume of 500 ml Gelofusine® (B. Braun, Melsungen, Germany), was delivered using a syringe and three-way tap via the intraosseous needle within 5 minutes, filling the central volume sufficiently to facilitate the insertion of a right subclavian central line. Following the placement of a third reliable peripheral intravenous line (14-G) in the right antecubital fossa, the intraosseous needle was removed.

image

Figure 1.  Battery-operated EZ-IO driver equipped with intraosseous needle and accessories. From the top clockwise: extra long adult needle; standard adult needle; identification and warning bands; EZ-IO driver unit with adult needle attached; stabilising dressing; paediatric needle.

Download figure to PowerPoint

The patient was then taken back to the operating theatre. Rapid-sequence induction of anaesthesia was achieved using ketamine 80 mg, alfentanil 500 μg and suxamethonium 100 mg. Thereafter, general anaesthesia was maintained with sevoflurane 0.45–1.1%, nitrous oxide 50% and rocuronium. A radial arterial catheter was placed to monitor continuous arterial blood pressure, oxygenation, serial measurements of acid/base status and coagulation status. At the beginning of surgery, significant bleeding per vaginum was still occurring and uterine atony was evident. Further doses of carboprost 250 mg were administered both intramuscularly and intramyometrially at 15-min intervals. Surgical interventions included intra-uterine balloon tamponade and application of a uterine compression suture. Failure of these methods led to the joint decision to proceed to hysterectomy. Oedematous tissues compounded the problems of determining anatomical planes, and evidence of bladder trauma also contributed to a difficult surgical procedure.

Cell salvage (Electa; Sorin Group, Modena, Italy) was deployed and 9 l of shed blood collected producing 2.3 l of processed washed red cells. Unfortunately, concern that contamination of salvaged blood with urine and methylene blue (used to identify the site of bladder damage) may contraindicate re-infusion resulted in premature discontinuation of the transfusion, after only 360 ml of washed red cells had been returned. Thereafter, only donated blood was used.

An oesophageal Doppler probe (CardioQ, Chichester, West Sussex, UK) was placed to assist dynamic cardiac output monitoring and associated parameters, and to assess the adequacy of fluid resuscitation. Underfilling as determined by a corrected flow time (FTc) as low as 120 ms and stroke volume of 40 ml was used to guide bolus administration of crystalloid and colloid, including blood, to maximise cardiac output.

Blood component replacement was provided in collaboration with a consultant haematologist's support. Use of rFVIIa (Novo Nordisk, Crawley, West Sussex, UK) was sanctioned and 6 mg administered intravenously. Early fluid warming was initiated to reduce hypothermia and prevent further platelet dysfunction and loss. Haemostatic control was finally achieved after 2 h of surgery. Total blood loss was estimated as 13 l. The lowest recorded Hb was 6.4 g.dl−1, haematocrit 19% and platelets 56 × 109.1−1. Maximum serum lactate was recorded as 3.5 mmol.l−1. From the time of the initial delivery by caesarean section, the patient received 14 units of packed red blood cells, 2 units of cryoprecipitate, 1 pool of platelets, 4 units of fresh frozen plasma, 4.5 l of Gelofusine and 4 l of crystalloid solution.

The patient was transferred to the ICU for postoperative care. Her trachea was extubated the following morning and she was discharged to the ward 72 h later with no significant long-term sequelae. Thereafter, she and her husband received a full debrief to outline the extreme nature of the postpartum haemorrhage and the clinical steps instituted to treat her. Inspection of the intraosseous needle insertion site revealed no signs of superficial or deep tissue inflammation or infection.

Discussion

  1. Top of page
  2. Summary
  3. Case report
  4. Discussion
  5. Acknowledgements
  6. References

Our patient was older than 35 years and had previously delivered by emergency caesarean section. These are known to be significant risk factors for subsequent peripartum hysterectomy [5]. The Royal College of Obstetricians and Gynaecologists has produced recommendations to assist the prevention and management of postpartum haemorrhage (PPH) [6]. Uterine atony is the commonest cause of PPH, and first-line management incorporates uterine manual stimulation, manual compression and uterotonic drugs. Uterotonics were administered in this case in accordance with our departmental guidance. Second-line therapies in the management of PPH include haemostatic drugs, uterine packing, uterine balloon tamponade, artery ligation and uterine compression sutures. Arterial embolisation may be useful if interventional radiological expertise is available; however, this is not available in our institution out of hours. Peripartum hysterectomy for PPH has a mortality rate of 0.6% and is typically associated with transfusion of a median of 12 units of blood in the case of uterine atony [7].

We believe this case report to be noteworthy due to the use of an intraosseous needle to facilitate early fluid resuscitation. Josefson described an intraosseous infusion technique in humans in 1934 [8]. Latterly, the utility of intraosseous needles has become well established for paediatric emergencies [9]. Utilisation in adult emergency care has evolved more recently, and is generally associated with pre-hospital care, including treatment of battle zone casualties [10, 11]. The intraosseous space benefits from a rich blood supply, and is therefore a logical alternative to the intravenous route when secure reliable intravascular access is imperative. Training in the use of any new device is important. However, acquisition of this skill is rapid with a high degree of success, even in novice users [12]. The principal advantage of intraosseous devices is ease and speed of placement, which is typically achieved within 10–20 s. Unlike conventional intravenous access, the device does not need to be secured in place with adhesive dressings or sutures and is thus available for immediate use. A battery-operated driver is considered to ease passage of the intraosseous needle through thick cortical bone, making insertion quicker and possibly less painful. Flow rates in excess of 150 ml.min−1 can be achieved through intraosseous needles placed in the humerus with the aid of a pressure bag [13]. In our case, we used a syringe and a three-way tap to administer fluid rapidly. Resuscitation drugs can also be given via an intraosseous needle placed in the proximal humerus. Importantly, drugs administered via an intraosseous needle placed in the humeral head appear to reach the heart in a time comparable to those given via a central line [14]. Although no drugs were administered via the device in our patient, placement of an intraosseous needle was crucial. It provided rapid vascular access and permitted sufficient fluid resuscitation to expand the intravascular space in order that other lines could be inserted. Placement of the intraosseous needle coincided with application of bimanual uterine compression and attempts to secure intravenous access in the legs. Consequently, the humeral head was chosen because at the time of insertion, the shoulder was more accessible than the other recommended site, the medial aspect of the proximal tibia. The choice of the humeral approach may have been particularly advantageous because resuscitation fluid was administered rapidly into the central venous system, and may have preferentially filled these veins to enable easier insertion of a central venous line. The clinician using this device was experienced in use of the intraosseous needles for pre-hospital care. Recently, the Association of Anaesthetists of Great Britain and Ireland (AAGBI) has acknowledged the occasional requirement to use alternative methods to establish vascular access in the event of failure to secure intravenous access, when dealing with massive haemorrhage. Accordingly, the intraosseous route or surgical cut-down are recommended for this eventuality [15]. Given the significant contribution made by this device in this case report, it may be sensible to extend training in vascular access for all anaesthetists to include development of expertise with intraosseous needle insertion in both adult and paediatric patients.

In 2007, approximately 38% of UK maternity units used cell salvage [16]. Cell salvage for maternal haemorrhage is endorsed by the National Institute for Health and Clinical Excellence and other organisations [17, 18]. Specific concerns regarding the ability of cell salvage to remove amniotic fluid appear to have been largely addressed [19]. The exact significance of maternal alloimmunisation through re-infusion of fetal red cells is yet to be determined, although it may be that the final concentration in maternal circulation differs little from that found after spontaneous labour and delivery [20]. Whilst cell salvage may be particularly useful in obstetric emergencies [21], commentators still urge caution and comprehensive data collection to guide future best practice [22, 23].

Doppler cardiac output monitoring was reported in obstetric patients almost 20 years ago [24, 25]. At this time, it was suggested that oesophageal Doppler monitoring may have a useful role in resuscitation [26]. More recently, suprasternal Doppler cardiac output estimation has been used in a research capacity to study haemodynamic changes in parturients receiving regional anaesthesia for caesarean section and in response to vasopressor therapy [27–29]. However, we are not aware of the widespread use of non-invasive cardiac output measurement in the maternity setting, including in the management of haemorrhage. Work from our own department [30], supported by meta-analysis studies [31], confirms that optimising cardiac output and oxygen delivery reduces postoperative morbidity following abdominal surgery. Serum lactate measurement is a surrogate of tissue perfusion with a predictive value in determining outcome following traumatic injury. Oesophageal Doppler-guided fluid optimisation of multiple trauma victims can reduce serum lactate and consequently, the length of ICU stay [32]. We consider that there may be greater scope to develop such monitoring techniques to assist resuscitation in maternal haemorrhage.

In summary, success in this case relied upon the application of a number of techniques to facilitate resuscitation and optimise peri-operative care. A key element of the initial resuscitation process was the rapid utilisation of an intraosseous needle, a technique not previously reported in the management of massive maternal haemorrhage.

Acknowledgements

  1. Top of page
  2. Summary
  3. Case report
  4. Discussion
  5. Acknowledgements
  6. References

Published with the written consent of the patient. No external funding and competing interests declared.

References

  1. Top of page
  2. Summary
  3. Case report
  4. Discussion
  5. Acknowledgements
  6. References