Rare complication after pericardial window: symptomatic diaphragmatic hernia containing bowel and liver

  1. Rebecca Harsten 1,
  2. Mark Kelly 2,
  3. Madeleine Garner 2 and
  4. Peter Roberts 1
  1. 1 Critical Care, Queen Elizabeth Woolwich, London, UK
  2. 2 General Surgery, Guy's and St Thomas' Hospitals NHS Trust, London, UK
  1. Correspondence to Dr Rebecca Harsten; rebecca.harsten@nhs.net

Publication history

Accepted:22 Oct 2020
First published:30 Nov 2020
Online issue publication:30 Nov 2020

Case reports

Case reports are not necessarily evidence-based in the same way that the other content on BMJ Best Practice is. They should not be relied on to guide clinical practice. Please check the date of publication.

Abstract

A 37-year-old woman presented to her local district general hospital with a cough, pleuritic chest pain and intermittent cyanosis. Eight months prior, she underwent a successful pericardial window for recurrent, symptomatic pericardial effusions. On presentation she was hypoxic but haemodynamically stable. Her chest radiograph raised the suspicion of a diaphragmatic hernia, confirmed by CT imaging. This identified herniation through the diaphragm of the transverse colon and left lobe of the liver resulting in cardiac compression and right ventricular dysfunction. She continued to deteriorate and required emergency intubation to allow safe transfer to a tertiary upper gastrointestinal unit. She underwent a laparotomy and repair of the diaphragmatic hernia with an uneventful inpatient recovery. In the literature, diaphragmatic liver herniation is a recognised complication secondary to trauma or congenital defects, however, to our knowledge, there are currently no cases described following pericardial windowing.

Background

In healthy individuals, the pericardial cavity (the potential space between the visceral and parietal layers of the serous pericardium) contains between 15 and 50 mL of serous fluid.1 A pericardial effusion is defined as an abnormal excess of fluid in this compartment, most frequently attributed to neoplastic (36%), idiopathic (32%) or uremic (20%) aetiologies.2 There are limited data regarding their frequency, but an Italian study revealed the mean annual incidence and prevalence in an urban general hospital in Northern Italy was 3% and 9%, respectively.3

While small effusions without haemodynamic compromise can be managed conservatively with serial surveillance echocardiograms, larger effusions are treated with pericardiocentesis which is both diagnostic and therapeutic.

Treatment choice depends on aetiology and clinical status of the patient. There are different methods for drainage, including needle pericardiocentesis via subxiphoid or an anterior thoracic approach, percutaneous balloon pericardiotomy, emergent thoracotomy and pericardiotomy, and surgical pericardial window via subxiphoid, anterior mini-thoracotomy or a video-assisted thoracoscopic surgery approach.4

Case presentation

A 37-year-old woman, with a background of Trisomy 21, asthma, hypothyroidism, spina bifida, vitamin D deficiency and previous pericardial window procedure, presented to her local emergency department with 3 weeks of worsening dyspnoea, dry cough and cyanotic spells. She had no gastrointestinal (GI) symptoms. At baseline, she mobilised with a wheelchair and her mother assisted with activities of daily living.

The pericardial window was performed electively 8 months prior this presentation due to recurrent, symptomatic pericardial effusions, likely related to hypothyroidism. This procedure was attempted thoracoscopically, but due to difficulty identifying the central diaphragmatic tendon, was converted to a subxiphoid approach. A 600 mL of pericardial fluid was evacuated and the pericardial biopsy taken showed reactive changes. At her 2-month follow-up, the patient was well, with healed surgical wounds and no evidence of further reaccumulation of the effusion.

On initial examination in the emergency department, although she was not clinically shocked (blood pressure 126/98 mm Hg, heart rate 85 beats/min, peripheral warm with a capillary refill time less than 2 s), she was hypoxic requiring high-flow oxygen. Her chest was clear and her abdomen was soft, non-tender and not distended.

Investigations

Initial investigations included a normal arterial blood gas, unremarkable blood tests (white cell count 10.3×109/L, C reactive protein 4 mg/L), a chest radiograph (figure 1) suggestive of bowel loops in the thorax and a suboptimal beside echocardiogram showing minimal effusion, no outflow obstruction and no significantly raised cardiac pressures. Due to her symptoms and chest radiograph findings, both CT of the abdomen and pelvis and CT pulmonary angiogram (CTPA) were performed (figure 2). These scans showed a right anterior paramedian diaphragmatic hernia with bowel and liver herniation. There was no evidence of bowel obstruction, but the liver and bowel were compressing and displacing the heart with secondary decompensation of the right cardiac chambers. No pulmonary emboli were identified.

Figure 1

An erect chest radiograph on admission to the district general hospital showing signs of bowel loops in the chest.

Figure 2

A chest CT coronal view, soft tissue window, taken on admission to the district general hospital demonstrating liver herniation into the chest.

Based on the clinical picture, the patient was admitted to the intensive care unit for invasive monitoring while awaiting the decision for transfer to the tertiary centre for definitive surgical treatment. Hypotensive episodes recorded during this period (lowest of 83/44) were managed successfully with fluid boluses. The cause of hypotension was likely due to the combined effect of right-sided cardiac compression and hypovolemia secondary to starving restrictions initiated in anticipation for surgery. After initial improvement in oxygenation and weaning of supplementary oxygen, she again became hypoxic and a decision to intubate prior to transfer was deemed safest. There was no specific event that precipitated this drop in oxygen saturations; it was likely a result of the compressive atelectasis and reduced pulmonary compliance caused by the large diaphragmatic hernia. Under ultrasound guidance, a right femoral central venous catheter was sited (favourable anatomy compared with the standard internal jugular vein site), and propofol and fentanyl infusions commenced. During transfer, a norepinephrine infusion (0.2 µg/kg/min) was started and on arrival at the tertiary centre, her blood gas showed a pH of 7.46, partial pressure of carbon dioxide of 3.55, partial pressure of oxygen of 30.76, bicarbonate of 18.9, BE of −3 mmol/L and lactate of 1.8.

Preoperatively, a second echocardiogram showed good left ventricular function with no regional wall motion abnormality, mild tricuspid regurgitation and a posterior pericardial effusion not large enough to cause circulatory compromise.

Treatment

Prior to surgical intervention, a multidisciplinary plan was formulated, led by the upper GI surgical team in collaboration with both a cardiac surgeon and cardiac anaesthetist. The patient was consented for both a sternotomy and upper midline incision with the operation primarily performed by the upper GI surgeon, supported by the cardiac team.

Access was acquired via an upper midline incision. The transverse colon and the left lobe of the liver were partially herniated and there were omental adhesions surrounding the diaphragmatic defect. The adhesions were divided circumferentially allowing the herniated bowel to be reduced and the left lobe of the liver mobilised to fully oppose the 8×7 cm diaphragmatic defect. On inspection, the entire bowel was deemed viable. The defect was repaired with Ethibond sutures followed by mass closure of the abdominal wall with polydioxanone 1–0 and skin staples. No mesh was used for the diaphragmatic repair as the defect was satisfactorily closed without one.

Outcome and follow-up

Immediate postoperative recovery was uneventful. She was started on both chemical (low molecular weight heparin) and mechanical (compression stockings) venous thromboembolic prophylaxis, and a chest radiograph showed clear lung fields bilaterally. Norepinephrine was weaned cautiously, and the patient was extubated day 1 postoperatively. She was stepped down from intensive care on day 3. On the ward, new right arm swelling was noted and a duplex ultrasound confirmed thrombophlebitis which was treated with a course of oral flucloxacillin.

A month after discharge the patient developed a cough and chest pain and was readmitted. A CTPA diagnosed pulmonary emboli. This provoked event was attributed to a combination of postoperative stress response and immobility on a background of underlying poor mobility secondary to spina bifida. She was started on rivaroxaban 15 mg two times per day which was then converted to 20 mg once daily. She has been followed up in thrombosis clinic with no new concerns.

The patient has also been reviewed by cardiology and has undergone serial echocardiograms demonstrating recurrence of the pericardial effusion. However, the latest scan showed this was minimal in size and the plan is to closely monitor for progression. Follow-up with critical care and upper GI was converted to virtual clinics due to the COVID-19 pandemic.

Discussion

Down’s syndrome (Trisomy 21) is the most common chromosomal disorder with an incidence of 1 in 750.5 The incidence of hypothyroidism in Down’s syndrome is high, but diagnosis is often missed or delayed for lack of specific clinical features.6 It is well understood that pericardial effusion is common in hypothyroidism with an incidence of 3%–37%.7

According to a retrospective single-centre observational study, pericardial window via subxiphoid approach usually leads to less postoperative pain and earlier postoperative extubation8 compared with a transthoracic approach. In another study by Allen et al, which compared percutaneous drainage to subxiphoid pericardiostomy, the latter had no operative deaths and a complication rate of 1.1%. In contrast, percutaneous drainage had significantly higher mortality (4%) and complication rates (17%).9 These findings echoed that of Susini et al who stated that the subxiphoid method provided both superior diagnostic access and less complications compared with the percutaneous approach.10 The subxiphoid approach however is associated with higher rates of recurrence and need for further surgery than other techniques.8 11

There have been rare cases described in the literature where patients have developed iatrogenic diaphragmatic hernias after subxiphoid pericardial window, for example in Docekal and Fabian12 when a patient had bowel loops incarcerated into the pericardial cavity. In that case, the patient presented much earlier than ours and was grossly symptomatic. It is also worth noting that the patient in the Docekal and Fabian’s case study was almost two times as old as our patient and immunosuppressed—this could explain the earlier symptoms. The patient underwent similar treatment with laparotomy and evacuation of the herniated bowel. Again, the bowel did not require resection, and the pericardial defect was closed with interrupted figure-of-eight sutures.

Another case study by Murari et al described a large intrapericardial herniation of stomach following pericardial window performed via a left thoracotomy. This was managed with laparotomy, reduction of the stomach and repair of the phrenopericardial defect with a Goretex patch.13 This proves that this type of complication is not isolated to the subxiphoid approach.

It is difficult to determine whether the patient’s asthma contributed to the clinical picture, but as her CT scan showed mosaic attenuation of the aerated lung parenchyma, obstructive airway disease may well have contributed to her breathlessness.14

In summary, we have presented an unusual case of diaphragmatic bowel and liver herniation following elective pericardial windowing. We feel that although it is rare, clinicians should have a low diagnostic threshold to consider this complication in patients presenting with breathlessness and cyanosis even at a relatively late stage postoperatively.

Patient’s perspective

The whole experience of our daughter being admitted to hospital and intensive care was initially distressing and we were surprised and shocked to learn about her diagnosis. We all wondered for how long it had been going on as she had been well at home and not reported anything. We are however very happy and thankful with the care our daughter received at both the hospitals she was admitted to.

Learning points

  • Subxiphoid pericardial windowing is a safe, well-established treatment option for recurrent, symptomatic pericardial effusion.

  • Iatrogenic diaphragmatic hernias are a recognised complication of pericardial windowing.

  • Symptoms can be non-specific and present late. A high index of suspicion by the treating clinician is required with low threshold to investigate further.

Footnotes

  • Contributors RH cared for the patient during admission, gained consent and has written the case report draft. MK was the surgical consultant who operated on the patient and has provided valuable intraoperative information, and edited the revised manuscript. MG has helped gather imaging, documentation and investigations from the tertiary centre, as well as helped with technical editing of the manuscript and helped get better resolution of the images. PR was the intensive care consultant who was medically responsible for the patient in the DGH. PR has additionally proofread, corrected and advised on appropriate academic terminology of the case report.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Patient consent for publication Parental/guardian consent obtained.

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

References

Use of this content is subject to our disclaimer