Abstract
Acute respiratory failure is a rare occurrence during pregnancy, estimated to occur in 0.2–0.3% of pregnancies, but when this occurs it can have significant consequences for mother and child. Acute hypoxic respiratory failure associated with pregnancy requires attention to the inciting cause and distinction between cardiogenic and non-cardiogenic aetiologies. Specifically, acute respiratory distress syndrome (ARDS) is characterized by rapid onset of hypoxaemic respiratory failure associated with diffuse pulmonary opacities related to non-cardiogenic pulmonary oedema.
Acute respiratory failure is a rare occurrence during pregnancy, estimated to occur in 0.2–0.3% of pregnancies,1,2 but when this occurs it can have significant consequences for mother and child. Acute hypoxic respiratory failure associated with pregnancy requires attention to the inciting cause and distinction between cardiogenic and non-cardiogenic aetiologies. Specifically, acute respiratory distress syndrome (ARDS) is characterized by rapid onset of hypoxaemic respiratory failure associated with diffuse pulmonary opacities related to non-cardiogenic pulmonary oedema. Consequently, the alterations in physiology that characterize ARDS include hypoxaemia, decreased lung volumes and reduced lung compliance. Early recognition of this syndrome is important to prevent clinical deterioration and to initiate prompt treatment specific to the inciting cause. The following chapter will address the epidemiology, clinical features, risk factors and management of ARDS in pregnant patients.
Acute respiratory distress syndrome was initially described as a disorder occurring in adults that was sudden in onset associated with bilateral radiographic opacities and severe hypoxaemia refractory to supplemental oxygen.3 The initial description reported reduced lung compliance that contributed to the clinical and radiologic abnormalities, and that improved with mechanical ventilation and positive end-expiratory pressure (PEEP). Mortality was elevated in these patients and post-mortem examination of the lungs demonstrated pulmonary oedema, alveolar collapse and hyaline membrane involvement of the alveoli resembling neonatal hyaline membrane disease.4 To distinguish this from infant respiratory distress syndrome, this condition was termed adult respiratory distress syndrome, later changed to the present term, acute respiratory distress syndrome. However, ARDS was frequently confused with other conditions associated with cardiogenic pulmonary oedema. In order to distinguish ARDS (non-cardiogenic pulmonary oedema) from other causes of acute respiratory failure, a rigorous definition was developed by an American-European Consensus Conference (AECC), based on the following criteria:
1. PaO2/FiO2 <200
2. Presence of bilateral infiltrates on chest X-ray consistent with pulmonary oedema
3. Pulmonary artery wedge pressure ≤18 mmHg, or no clinical evidence of elevated left atrial pressure.5
In addition, the term acute lung injury (ALI) was introduced to describe patients with milder hypoxaemia (PaO2/FiO2 between 201 and 300) plus the previously mentioned radiographic pattern and haemodynamic parameters. Subsequently, concerns were raised about the large variation in physician interpretation of chest X-rays in patients with ARDS that resulted in delayed recognition or misdiagnosis,6,7,8 as well as the absence of information in the case definition about cause or timing, again potentially resulting in misdiagnosis. This was in part due to the initial description of ARDS as an acute process without a specific time limit. In addition, the AECC definition did not exclude subacute or chronic conditions such as hypersensitivity pneumonitis or interstitial lung disease associated with connective tissue disorder that could masquerade as ARDS.9 Furthermore, the AECC definition of hypoxaemia did not account for the effects of PEEP on PaO2/FiO2, and a proposal was made to revise the definition.10 Consequently, an international panel of critical care investigators in 2011 developed an updated definition of ARDS to address the issues regarding onset, chest X-ray features and origin of the radiographic infiltrates. This has been termed the Berlin definition for ARDS (Table 14.1).
Table 14.1 Berlin Criteria
| Timing | Symptoms of worsening respiratory or new respiratory symptoms that occur within one week of known clinical process |
| Chest imaging | Bilateral infiltrates that cannot be fully explained by atelectasis, lung nodules or effusions |
| Origin of oedema | Not completely explained by cardiac failure or fluid overload |
|
|
Notably, the Berlin definition takes into account that an early diagnosis of ARDS can be made in the setting of acute bilateral radiographic infiltrates and hypoxaemia in patients not mechanically ventilated, thereby leading to earlier recognition and treatment. It also allows use of either chest X-rays or computerized tomography to identify alveolar infiltrates. In addition, since pulmonary artery catheters are now used less often in the critical care setting, the consensus group removed wedge pressure as a criterion, substituting physician judgement as to the absence of a non-cardiac cause of pulmonary oedema.
Epidemiology
The epidemiology of ARDS has been assessed in the general population. In 1999, investigators reported the incidence of ARDS (using AECC criteria) in 150 Nordic intensive care units to be 13.5 per 100 000 per year.11 Similarly, a group of investigators described incidence of ARDS at a single urban hospital in Seattle to be 12.6 patients per 100 000 per year.12 In both of these populations, if Berlin criteria are utilized, the incidence would increase to 17.9 per 100 000 and would be categorized as mild ARDS. In 2002, an international, multicentre study of 5183 mechanically ventilated patients found that 9% were diagnosed with ARDS.13 More recently, a prospective study utilizing a standardized screening protocol examined 7944 patients using the Berlin criteria and reported the incidence of ARDS to be 27.6 per 100 000.14
The epidemiology of ARDS and pregnancy has been reported primarily from single centres that described an incidence of 15.9 to 130 per 100 000 deliveries.15,16,17 One group retrospectively reported the occurrence of ARDS over 14 years and identified 41 cases with an incidence of 70 patients per 100 000 deliveries.16 Another single-centre retrospective study from Argentina reported on the outcomes of critically ill obstetric patients from 1998 to 2005 and noted 30 out of 161 patients developed ARDS.17 However, there have been no contemporary studies on the epidemiology of ARDS in pregnant patients, indicating the need for additional information on this topic. More recently, investigators examined the incidence of ARDS in the United States from 2006 to 2012 using a large national database and ICD-9 codes to identify pregnant patients requiring mechanical ventilation for acute respiratory failure.18 These authors identified 2808 pregnant patients with ARDS and reported an incidence of 36.5 per 100 000 live births in 2006, increasing to 59.6 per 100 000 by 2012. During the hospital admission, 41% of patients underwent caesarean delivery while 14.6% underwent vaginal delivery; however, the authors were not clear on the pregnancy outcomes of the remaining 45% Notably, pneumonia was the most frequent cause for ARDS (25.9%), followed by pre-eclampsia–eclampsia (22.1%). The inpatient mortality for patients requiring >96 hours of mechanical ventilation was 14%. In earlier reports, investigators described ARDS-associated mortality that ranged from 24–39%.15,16,17 Review of the reported prevalence of ARDS in pregnant patients indicates regional and temporal variation, and also highlights concern about the diagnostic precision of ARDS. Future investigations regarding epidemiology are needed to clarify the prevalence and outcomes in pregnant and post-partum patients with ARDS.
Aetiology
The initial description of ARDS made it clear that pulmonary-specific conditions, such as pneumonia, and extrapulmonary conditions, such as sepsis, pancreatitis, trauma and multiple blood transfusions, can result in severe hypoxaemia and alveolar infiltrates3 The most frequent causes for ARDS are associated with sepsis, trauma, pneumonia, aspiration and blood transfusions.19,20 The aetiology of non-cardiogenic pulmonary oedema may be categorized as an insult that directly or indirectly causes lung injury. Direct lung injury would include aspiration pneumonitis, bacterial, viral or fungal pneumonia, inhalation injury, pulmonary contusion, fat emboli, near drowning and reperfusion injury. In contrast, indirect lung injury may be generated by sepsis, trauma, acute pancreatitis, disseminated intravascular coagulation, burns and blood transfusions.19,20 A recent multicentre study in 3022 non-obstetric patients with ARDS reported pneumonia (59%) as the most common cause of ARDS, followed by extrapulmonary sepsis (16%), aspiration (14%), trauma (4%) and blood transfusion (3.9 %). Notably, in this epidemiologic study, a specific risk factor leading to ARDS could not be identified in 8.3% of subjects.20
During pregnancy, non-obstetric causes for ARDS include pneumonia, sepsis, influenza, blood transfusion and trauma (Table 14.2).18 As has been noted for decades, pregnant patients are susceptible to aspiration of gastric contents leading to a chemical pneumonitis during labour.21 The observed physiologic changes occurring during pregnancy, specifically delay in gastric emptying and decreased lower oesophageal sphincter tone, can lead to aspiration of stomach contents as intra-abdominal pressure increases during labour and delivery. Sepsis due to pyelonephritis has been reported as a risk factor for ARDS.22,23,24 In 2014, a group of investigators performed a retrospective cohort study in Southern California and found a frequency of 0.5% suggesting that treatment of antepartum pyelonephritis may partially explain the decline, although this may be due to under-reporting.25 Another non-obstetric risk factor for ARDS is influenza A (H1N1), as reported in the 2009 pandemic, when various groups reported that pregnancy was associated with a higher risk for development of influenza-associated complications, including ARDS.18,26,27,28 ARDS can also be a result of obstetric causes, including pre-eclampsia–eclampsia, puerperal infection, amniotic fluid embolism, tocolytic therapy, septic abortion and retained products of conception.15,16,18 A recent publication identified pre-eclampsia–eclampsia as a risk factor for ARDS, seen in 22% of pregnant patients diagnosed with ARDS.18 Amniotic fluid embolism (AFE) is a rare event that tends to occurs late in labour or shortly after delivery.29 A proposed mechanism is that amniotic fluid enters the maternal circulation, which triggers a systemic inflammatory response via activation of proinflammatory mediators leading to cardiogenic shock, disseminated intravascular coagulation (DIC) and hypoxaemia. The underlying pathogenesis is not fully understood, and components of amniotic fluid are frequently found in the maternal circulation without ill effect. Non-cardiogenic pulmonary oedema has been described to develop in the majority of patients that survive the initial insult.30 Tocolytic-associated pulmonary oedema has also been associated with non-cardiogenic pulmonary oedema, and stems from use of certain drugs, usually β-adrenergic agents (such as terbutaline), but sometimes nifedipine or magnesium sulfate, to inhibit uterine contractions associated with pre-term labour.31 Administration of multiple doses of tocolytic agents and hypervolaemia (the first-line approach to pre-term labour has often been intravenous fluid loading) have been associated with development of this condition. The most frequent presenting symptoms include dyspnoea and cough, accompanied by hypoxaemia and bilateral radiographic infiltrates, during administration or up to 12 hours after discontinuation of the tocolytic.
Diagnosis/Recognition
ARDS is a form of acute hypoxic respiratory failure characterized by inflammation that leads to increased pulmonary vascular permeability with loss of aerated lung parenchyma resulting in hypoxaemia and radiographic pulmonary opacities. The most frequent symptoms associated with this syndrome include dyspnoea, cough, fever and fatigue. These non-specific symptoms can make this diagnosis especially challenging. Acute respiratory failure should raise suspicion of ARDS, especially in a patient with a predisposing condition such as sepsis, pneumonia, pancreatitis or blood transfusion. Physical findings include tachypnoea, accessory muscle use and bilateral crackles on chest auscultation. These signs and symptoms will frequently manifest within 6–72 hours after insult, and rapidly worsen. Chest radiography can demonstrate the characteristic features of diffuse, bilateral lung infiltrates; however, chest images may initially reveal asymmetric, dependent lobar opacities that evolve over time to bilateral involvement.32,33 Chest computerized tomography (CT) imaging has also been incorporated into the diagnosis of ARDS and may substitute for chest X-rays. Compared with chest X-ray, the resolution obtained with chest CT imaging is more sensitive in early detection of ARDS and superior in characterizing the nature of the lung opacities (Figure 14.1). While lung ultrasonography has not been established as an imaging criterion for ARDS, various authors have reported this as a useful modality in the diagnosis. In particular, patients with ARDS can demonstrate multiple B-lines that are caused by thickened interlobular septa, and/or ground glass opacities.34 In order to exclude cardiogenic pulmonary oedema, the clinician traditionally should assess for jugular venous distention (JVD) and perform cardiac auscultation for S3, S4 gallop or murmurs. Unfortunately, as pregnant women are in a hypervolaemic state with increased intra-abdominal pressure from the growing uterus, the presence of JVD as well as cardiac flow murmurs and S3 are not necessarily pathologic findings. In the absence of these findings, serum BNP has been shown to be useful in non-pregnant patients using a serum level <100 pg/ml to identify a patient with ARDS, with a sensitivity of 27%, specificity of 95%, positive predictive value of 90% and negative predictive value of 44%.35 However, serum levels >100 pg/ml did not exclude ARDS, and transthoracic echocardiography may be a useful diagnostic study to assess for mitral or aortic valvular disorders or severe LV systolic dysfunction. Lastly, right heart catheterization may be performed if cardiogenic oedema cannot be excluded using less invasive methods. As pregnancy acts as a physiological stress on the cardiovascular system, BNP levels may be higher than normal although a cutoff of 100 pg/ml may still be useful to exclude a cardiovascular event.36
(B) Chest X-ray the following day with interval development of bilateral air space opacities.
(C) CT image corresponding to (B) demonstrating bilateral opacities (as marked by an asterisk *) without pleural effusions.
Table 14.2 Aetiologies of ARDS
| Obstetric |
|
| Non-obstetric |
|
Supportive Care
Management of the pregnant or post-partum patient with ARDS involves identification and control of the inciting cause, antibiotic administration if indicated, fluid resuscitation, haemodynamic support, oxygen administration and mechanical ventilation. An important goal involves early identification and treatment of the underlying aetiology, which requires assessment for potential infectious complications. Diagnostic evaluation should be directed at infectious conditions such as pneumonia (bacterial, viral), chorioamnionitis, pyelonephritis, post-partum endometritis, necrotizing fasciitis or toxic shock syndrome.37,38,39 Notably, appropriate antibiotic selection and prompt antibiotic delivery are important, as administration of antibiotics within the first hour of septic shock is associated with improved survival.40,41, 42 Early identification of a source of infection amenable to source control, through inspection of a surgical site at least 24 hours after surgery to identify a soft tissue infection, abdominal imaging to drain an intra-abdominal abscess or removal of an intravenous catheter can help mitigate the systemic consequences of sepsis. Other conditions to consider include pancreatitis, amniotic fluid embolism or transfusion-related acute lung injury. Fluid resuscitation with serum lactate monitoring, especially in the setting of hypotension, is important in order to maintain adequate organ perfusion.43 Intravenous fluid resuscitation is vital to maintain adequate tissue perfusion, but an excess quantity may lead to worsening of lung compliance and progressive hypoxaemia. Consequently, supplemental oxygen administration should be provided by face mask or high flow nasal cannula to maintain a SpO2 of 88–95% with frequent assessment for the need for invasive mechanical ventilation.44
For the treatment of acute non-hypercapnic, hypoxaemic respiratory failure, high-flow nasal cannula oxygen administration provides reliable FiO2 and improved patient comfort, compared to conventional face mask administration.45 Observational case series have described successful outcomes with high-flow nasal cannula in the treatment of ARDS and influenza.46,47 However, a prospective, multicentre, randomized clinical trial conducted in France and Belgium, comparing high-flow nasal cannula, conventional face mask and non-invasive ventilation (NIV) in non-pregnant patients with acute hypoxaemic respiratory failure reported a similar intubation rate in all groups.48 The similar intubation outcome for high-flow nasal cannula and non-invasive ventilation in ARDS may be related to a delay in intubation and institution of mechanical ventilation.49 While NIV is frequently used in the management of acute exacerbations of chronic obstructive pulmonary disease (COPD) and cardiogenic pulmonary oedema, its use in ARDS remains controversial. A systematic review describing the use of NIV to avoid intubation in patients with ARDS concluded that there was little data to support the role of NIV in ARDS, and that this modality should be avoided as a first-line therapy.50 Furthermore, a post- hoc analysis indicated a greater mortality for patients with PaO2/FiO2 <200 treated with NIV compared to oxygen delivered via high-flow nasal cannula.51
With respect to pregnant patients with ARDS, case reports have described NIV as a means of ventilatory support.52,53 While there are no formal studies on the use of high-flow nasal cannula in pregnant patients with ARDS there is insufficient evidence to recommend NIV over high-flow nasal cannula as initial therapy in any patients with ARDS. If NIV is employed, it should be used early in the management of haemodynamically stable patients with acute hypoxic respiratory failure with PaO2/FiO2 of 200–300. Repeated assessments with arterial blood gas measurements 30 to 45 minutes after initiation of NIV should take place, and when signs of increasing tachypnoea, worsening mentation or declining oxygenation occur, the provider should promptly intubate and institute invasive mechanical ventilation. In addition to frequent maternal assessments, fetal monitoring may be indicated in some cases.
Mechanical Ventilation
Mechanical ventilation is an important component of critical care management of ARDS.
The goal of mechanical ventilation is to provide adequate gas exchange and rest the fatigued respiratory muscles, while not causing harm. In ARDS, epithelial and endothelial injury produces pulmonary oedema that results in hypoxaemia, resulting in decreased lung distensibility.54 Quantitative chest CT in severe ARDS demonstrates a significant reduction in the level of aerated lung measured at end-expiration, similar to that of a 5-year-old child (~200–500 g). As lung compliance correlates only with aerated lung, this reduced state popularized the concept of ‘the baby lung’.55 The ‘baby lung’ concept is a functional construct that represents the reduced amount of aerated lung available for gas exchange, and the need to deliver smaller tidal volumes to achieve adequate gas oxygenation, while avoiding lung overdistention and barotrauma.
Initiation of mechanical ventilation involves assessment of factors pertaining to tracheal intubation, with attention to the potential difficult airway. Prior to intubation, the clinician should be aware that upper airway anatomical changes often occur as a result of pregnancy, and these can lead to potential difficult airways. Also, the decrease in functional residual capacity associated with pregnancy can lead to more rapid oxygen desaturation during intubation attempts. Therefore, pre-oxygenation via facemask, high-flow oxygen or NIV is imperative to avoid oxygen desaturation during attempted intubation.56,57 In order to minimize gastric aspiration, rapid-sequence intubation with a sedative and neuromuscular agent has been recommended. For a difficult airway or failed intubation in the obstetric patient, algorithms and guidelines have been published.58
In the absence of randomized trials comparing ventilator strategies for pregnant patients with ARDS, mechanical ventilation of these patients is guided by a similar treatment strategy as in non-pregnant patients, namely a lung-protective ventilation strategy with low tidal volume and PEEP.59 Retrospective case series concerning mechanical ventilation during pregnancy for acute hypoxaemic respiratory failure have described the epidemiology, causes and maternal-fetal survival, although earlier studies did not follow a low tidal volume strategy and reported higher rates of barotrauma.15–18, 22 However, there is a paucity of data comparing the physiologic or clinical outcomes of different ventilation settings or strategies in the pregnant patient with respiratory failure. Mechanical ventilation in the pregnant patient is discussed in detail in Chapter 25 (Table 14.3).
