Potential role for furosemide in the treatment of acute respiratory distress syndrome (ARDS) and an unusual presentation of pulmonary embolism in a complex patient

Abstract

An 81-year-old woman was admitted to our hospital after experiencing syncope. She was diagnosed with a large pulmonary embolism and was hemodynamically unstable therefore requiring endotracheal intubation and norepinephrine support. She presented with an upper gastrointestinal bleed which prevented her from receiving tissue plasminogen activator. She was treated with enoxaparin and ceftriaxone. Her blood, sputum and urine cultures were negative. When transferred to our ward, her antibiotic treatment was changed to piperacillin–tazobactam. A lumbar puncture was not suggestive of a central nervous system infection. Chest X-rays demonstrated rapid advancement of diffuse bilateral infiltrates which were not present at first and were interpreted by radiology consultation as suggestive of acute respiratory distress syndrome. An echocardiography showed right ventricle dilatation without left-sided heart failure. Diuretics were added and with this treatment, a quick respiratory improvement was noted as she regained consciousness and extubated shortly after.

Background

Although pulmonary embolism is a common problem, the incidence of an initial presentation of pulmonary embolism as loss of consciousness varies between different studies. A recent study conducted by Raynal et al showed that the prevalence of pulmonary embolism in patients with isolated syncope is only 2.2%,¹ while a study conducted by Secemsky et al found that the prevalence of syncope in patients experiencing massive pulmonary embolism is much higher and reaches 26.1%.² Syncope, in combination with other factors such as hypotension, right ventricle dysfunction and low systolic blood pressure, predicts a worse prognosis in patients presenting with pulmonary embolism and is suggestive of a ‘massive’ or ‘high-risk’ pulmonary embolism.³ Given this unusual presentation, we undertook extensive diagnostic measures to find an alternative aetiology for her comatose state. In addition, while mechanically ventilated, her chest X-ray and chest CT scans were suggestive of acute respiratory distress syndrome (ARDS). Furosemide was administered and a dramatic improvement of the respiratory and radiologic findings was noted. Subsequently, she regained consciousness and underwent extubation.

The pathophysiology of ARDS includes a protein-rich pulmonary oedema caused by extensive parenchymal and interstitial damage. Currently, restrictive fluid management including the use of furosemide to achieve diuresis, even in patients who do not show signs of clinical volume overload, is an accepted strategy in the treatment of patients with ARDS. Conservative fluid management has been shown to improve oxygenation and shorten mechanical ventilation, while positive fluid balance has been found to increase the risk for mortality.⁴ A study conducted by Martin et al found that furosemide with albumin compared with albumin improved fluid balance, oxygenation and hemodynamics in hypoproteinemic patients with acute lung injury.⁵ However, the use of furosemide was not found to directly reduce mortality and the studies examining its role in ARDS were limited by factors such as sample size and numerous exclusion criteria.⁶ To the best of our knowledge, there are no ‘head-to-head’ studies examining the effect of furosemide compared with placebo on mortality, length of hospital stay and ventilator free days. This case suggests that furosemide might have a therapeutic role in the management of patients with ARDS and especially in a subset of patients with elevated brain natriuretic peptide (BNP) levels.

This case also highlights the complexity involved in the diagnosis and management of patients hospitalised in the internal medicine ward and emphasises pulmonary embolism's reputation as ‘the great masquerader’.

Case presentation

We present the case of an 81-year-old woman with mild cognitive impairment yet independent in her basic activities of daily living. Her previous medical history consists of a provoked pulmonary embolism following an orthopaedic surgery 3 years prior, microcytic anaemia, osteoporosis and depression.

The patient's first pulmonary embolism was diagnosed after a reverse total shoulder replacement, 3 years prior to her current presentation. She received apixaban for an unknown duration. When currently presenting to the hospital she was not on anti-coagulation or antiplatelet therapy.

The patient's anaemia was present for at least 3 years prior to her current presentation with stable haemoglobin values ranging between 90 and 110 g/L and mean corpuscular volume (MCV) values ranging between 70 and 80 fL, ferritin levels ranging between 10 and 25 ng/mL (normal range, 10–120) and transferrin saturation ranging between 4%and 15% (normal range, 15%–50%). She did not have any endoscopy or additional workup for her anaemia.

On admission she was taking an unknown dose of quetiapine once daily, bisoprolol 2.5 mg once daily for an unknown indication, and an unknown dose of vitamin D and calcium supplements once daily.

Our patient was found to be suddenly unresponsive to sound or touch by her husband in their home. Emergency medical services arrived and were unable to get a blood pressure reading. She was administered a 200 mL bolus of normal saline and arrived at our hospital where her blood pressure was 100/60 mm Hg. A stroke was suspected in the emergency department and she underwent a brain and neck CT scan with contrast. This scan showed no intracranial abnormalities but demonstrated a pulmonary embolism in her right main pulmonary artery and in a lobar artery leading to the right upper lobe of the lung. Her initial venous blood pH was 7.28 with a pCO₂ of 37.4 mm Hg. Haemoglobin was 70 g/L and troponin was 155 ng/L (normal values, 0–50). Her blood pressure continued to drop, so she required norepinephrine support and underwent endotracheal intubation. She was treated with 5000 units of heparin and was promptly transferred to the cardiac intensive care unit (CCU), where she was supposed to receive tissue plasminogen activator. However, according to the CCU team, she developed hematemesis, thrombolysis was deferred, and she received weight adjusted therapeutic doses of low molecular weight heparin (100 mg two times per day).

During days 2–3 of her hospitalisation, she did not regain consciousness despite the fact that her blood gas tests normalised and there were no explaining metabolic disturbances. Arterial pH values at this time were 7.3–7.4, PaO₂ was 100–150 mm Hg and pCO₂ was between 30 and 40 mm Hg. FiO₂ ranged between 28% and 60%.

Through days 2–5 her mean arterial pressure (measured by non-invasive methods) remained low and she required norepinephrine support. Her physical examination did not show signs of volume overload. She did not have elevated jugular pressure and her extremities were warm without substantial peripheral oedema. Her C-reactive protein (CRP) levels rose steadily from 30 mg/L on day 1 to a peak of 316 mg/L (upper limit 5 mg/L) on day 5. She was presumed to have septic shock by the CCU team and ceftriaxone was administered after blood, urine and sputum cultures were taken. All cultures were found to be negative and there was no clear source of infection.

On days 4–5 she remained comatose and chest X-rays preformed on these days demonstrated rapid advancement of diffuse bilateral infiltrates which were not present at first and were interpreted by radiology consultation as suggestive of ARDS.

The patient was transferred to our ward on day 5 of her hospitalisation for further management and investigation. Our ward is an internal medicine ward with facilities to manage patients who require invasive mechanical ventilation but without the means for invasive haemodynamic monitoring. We initiated low tidal volume mechanical ventilation (4 mL/kg of ideal body weight) with a FiO₂–100%. Her blood tests on admission to our ward on day 5 were as follows: haemoglobin, 77 g/L; MCV, 74.1 fL; red cell distribution width, 27.0%; white blood cells, 10×10⁹/L, platelets, 199×10⁹/L. Creatinine was 0.71 mg/dL and blood urea nitrogen was 12 mg/dL. pH was 7.26, PaO₂ was 148 mm Hg and pCO₂ was 38.6 mm Hg. PaO₂/FiO₂ ratio was 148 and positive end expiratory pressure (PEEP) was ≥5 cm H₂O. Her thyroid functions were suggestive of sick euthyroid syndrome. Blood cortisol levels were normal. Cancer antigen 125 (CA125) was 125 U/mL (normal range 0–35) and cancer antigen 19-9 (CA19-9) was 42.6 U/mL (normal range 0–31). Cancer antigen 15-3 (CA15-3), alphafetoprotein (AFP) and carcinoembryonic antigen (CEA) were normal. Lupus anticoagulant, anticardiolipin IgM and IgG immunoglobulins, complement levels and rheumatic factor levels were normal. Serology tests for hepatitis B and C were normal.

Chest X-rays on days 5–9 and a chest and abdominal CT scan on day 9 showed bilateral diffuse infiltrates. Brain natriuretic peptide obtained on day 8 was 1267 pg/mL (normal range, 0–80). Since the patient was not regaining consciousness, we performed an additional brain CT on day 9 that showed no acute abnormalities. Repeated physical examinations during days 5–9 did not show signs of volume overload such as increased jugular venous pressure and extensive peripheral oedema. Net fluid balance on days 5–9 showed a total loss of 4500 mL.

A transthoracic echocardiography performed on day 5 showed the following: lack of inferior vena cava (IVC) collapsibility, moderate right ventricle dysfunction with a right atrial pressure of 20 mm Hg and a calculated pulmonary artery pressure of 68 mm Hg, which are considered higher than average normal values. The interventricular septum was compressed towards the left ventricle. Left ventricular ejection fraction was normal (60%) and left atrial filling pressure was 10 mm Hg.

Shortly after being transferred to our ward on day 5, and while treated with ceftriaxone, the patient developed a fever of 38.7°C. Throughout the course of her hospitalisation thus far she required norepinephrine support. We repeated blood, urine and sputum cultures and administered piperacillin–tazobactam. On day 7 we stopped the administration of norepinephrine because the patient's blood pressure had stabilised, and her urine output increased. We performed a lumbar puncture on day 10 and subsequently her cerebrospinal fluid cultures for bacteria, PCR for varicella zoster and West Nile virus IgM and IgG were negative. The aforementioned chest and abdominal contrast CT scan performed on day 9 did not demonstrate evidence of active infection or malignancy but supported the diagnosis of ARDS.

Given this clinical and radiologic presentation combined with the elevated BNP levels, we initiated furosemide on day 9. At this point her venous pH was 7.33, oxygen pressure was 44.5 mm Hg, venous pCO₂ was 35.5 mm Hg, and FiO₂ was 100%.

Shortly after commencing furosemide, the patient's respiratory and neurologic situation improved dramatically. She regained consciousness, her pulmonary infiltrates diminished, and she was successfully extubated on day 11.

After being extubated she had an uneventful recovery course except for elevated amylase levels of 405 U/L (normal range, 20–104 U/L) and lipase levels of 547 U/L (normal range, 9–51 U/L) that normalised without any intervention and were not accompanied by any clinical symptoms. Repeated abdominal CT scan showed no biliary stones, she was not consuming alcohol and was not receiving any medications that are known to cause pancreatitis. She was discharged on day 21 after awaiting placement in a rehabilitation facility.

Investigations

On admission we performed a head and neck CT which ruled out a stroke but demonstrated an unprovoked pulmonary embolism. Later during her hospitalisation, we repeated a head CT to further rule out any findings explaining her situation. We performed a lumbar puncture which was negative for West Nile virus serology and negative for herpes simplex virus and varicella zoster PCR.

Because of the unprovoked pulmonary embolism, we suspected an underlying malignancy as the cause of the pulmonary embolism and performed an abdominal and chest CT which confirmed the presence of a pulmonary embolism and was negative for findings suggestive of a tumour. The tumour markers AFP, CEA, CA19-9 were negative. Lupus anticoagulant and cardiolipin were negative. Partial thromboplastin time, prothrombin time and international normalised ratio were within the normal range.

Repeated chest X-rays on days 5–9 and the chest CT scan mentioned earlier showed bilateral diffuse infiltrates compatible with the diagnosis of ARDS.⁷ On day 5, her pH was 7.26, PaO₂ was 148 mm Hg, and pCO₂ was 38.6 mm Hg. PaO₂/FiO₂ ratio was 148 and PEEP was ≥5 cm H₂O. BNP level was 1267 pg/mL. Troponin was 155 ng/L (normal range, 0–50).

A transthoracic echocardiography preformed on day 5 showed the following: lack of IVC collapsibility, moderate right ventricle dysfunction with a right atrial pressure of 20 mm Hg and a calculated pulmonary artery pressure of 68 mm Hg. The interventricular septum was compressed towards the left ventricle. Left ventricular ejection fraction was normal (60%) and left atrial filling pressure was 10 mm Hg.

Differential diagnosis

The main differential diagnosis for ARDS we considered was left-sided heart failure.⁸ This was suspected in view of the patient's elevated BNP levels and quick improvement with furosemide. Elevated BNP levels are indicative of right and/or left heart failure.^{9 10} In this patient they were most likely caused by the pulmonary embolism, leading to the development of pulmonary hypertension and subsequently the development of right ventricular dysfunction. Pulmonary hypertension is defined by a pulmonary artery pressure ≥25 mm Hg. The transthoracic echocardiography performed on our patient showed a pulmonary pressure of 68 mm Hg, which was calculated by summing the right atrial pressure and the tricuspid regurgitation peak velocity.¹¹ The lack of IVC collapsibility demonstrated on echocardiography could be attributed to mechanical ventilation and pulmonary embolism as well as to volume overload due to left-sided heart failure.¹² However, physical examination suggested that the patient was euvolemic, and although clinically assessing volume status in ventilated patients can be challenging, net fluid balance during her hospitalisation in our ward was negative. High BNP levels are anecdotally described in patients suffering from septic shock, which was suspected in our patient, without overt left-sided heart failure.¹³ The imaging findings on chest X-ray and chest CT were more indicative of ARDS than of left-sided heart failure and included diffuse bilateral infiltrates, and the lack of a pleural effusion. The echocardiography we performed did not demonstrate left-sided heart failure which we would have expected to see had the patient's symptoms been caused by left-sided heart failure. To summarise, left-sided heart failure is indeed a plausible aetiology for the patient's clinical course, but we believe that the diagnosis of ARDS is more consistent with the findings mentioned above.

The main differential diagnosis for the patient's septic shock was obstructive shock due to her pulmonary embolism and right ventricular dysfunction. When diagnosed with septic shock in the CCU she did not have a high fever, an elevated leucocyte count and a clear source of infection. However, when transferred to our ward on day 5 she did develop a high fever. She did not have signs of clinical volume overload and her extremities were warm. Right systolic dysfunction on echocardiography was interpreted as ‘moderate’, which we would have expected to be more severe had her shock been caused by an obstructive aetiology. When switching her antibiotic treatment from ceftriaxone to piperacillin–tazobactam, normalisation of systemic temperature and blood pressure was noted and there was a sharp decrease in CRP levels. Since the patient was mechanically ventilated and developed diffuse infiltrates, it is possible that she had developed hospital-acquired pneumonia (HAP), which caused the septic shock. The HAP might have been masked clinically by diffuse crackles on auscultation and radiologically by the findings compatible with ARDS. It is possible that there was an obstructive component to the shock, but it is more likely that at that stage the infection was driving the clinical course.

Outcome and follow-up

We followed up with the patient's daughter 3 weeks after she was discharged. She was in a rehabilitation facility and doing well. There were no complaints of shortness of breath or abdominal pain.

Discussion

Acute respiratory distress syndrome is a form of acute lung injury caused by substantial alveolar damage leading to impairment of gas exchange, decreased lung compliance and increased pulmonary arterial pressure. Pulmonary hypertension occurs in up to 25% of patients with ARDS who undergo mechanical ventilation and can eventually lead to cor pulmonale.¹⁴ There are several pathologies which can cause ARDS such as sepsis, pneumonia, trauma, pancreatitis, burns, drowning and others.¹⁵ The Berlin criteria for diagnosis of ARDS require the following: (1) The onset of respiratory symptoms is new. (2) Bilateral opacities must be present on a chest radiograph or CT scan. These opacities must not be fully explained by pleural effusions, lobar collapse, lung collapse or pulmonary nodules. (3) The patient's respiratory failure must not be fully explained by cardiac failure or fluid overload, (4) The PaO₂/FiO₂ ratio is equal or lower than 300 mm Hg with a PEEP ≥5 cm H₂O.⁷ Our patient fulfilled these diagnostic criteria and improved dramatically after treatment with furosemide. The main differential diagnosis we considered was left- sided heart failure which was highly suspected because of the patient's elevated BNP levels. However, BNP levels can increase in right-sided heart failure which could have been caused by the patient's pulmonary embolism or by the ARDS.¹⁶ In addition, the patient's physical examination was not suggestive of left-sided heart failure because she did not have pedal oedema and elevated jugular venous pressure, and her net fluid balance was negative during the days when chest X-rays showed rapid progression of diffuse bilateral infiltrates. Left ventricle ejection fraction was 60%. It is also notable that basic BNP levels are higher in women and older individuals.¹⁷ Previous clinical trials demonstrated that conservative fluid-management strategy shortens the duration of assisted ventilation in ARDS. Small randomised trials of diuretics after shock reversal in patients with ARDS showed improved oxygenation and a trend toward a shorter duration of mechanical ventilation, but a larger trial did not suggest a reduction in mortality.¹⁸ To date no specific pharmacologic therapy has been shown to reduce either short-term or long-term mortality in ARDS. The experience in this case may suggest that furosemide has a therapeutic role in the treatment of ARDS, especially in patients with elevated BNP levels, and there is a need for further randomised controlled trials. To the best of our knowledge, previous studies which compare furosemide with placebo for the treatment of ARDS do not exist.

In 2016, Prandoni et al published a cross-sectional study showing that the prevalence of pulmonary embolism in patients presenting to the emergency department with syncope was 17.3%.¹⁹ This finding was challenged by other studies showing that the prevalence of pulmonary embolism in syncope ranged between 2.2% and 10%.^{1 20} In addition, the study conducted by Prandoni et al does not show a causative relationship between pulmonary embolism and syncope thus stressing even further the fact that syncope as a presentation of pulmonary embolism is a relatively uncommon finding. However, these studies have not distinguished between different types of pulmonary embolism. Our patient suffered from ‘massive’ or ‘high-risk’ pulmonary embolism in which the incidence of syncope is probably higher.^{2 3} Notably, although most current evidence does not support screening for malignancy in patients with venous thromboembolism,²¹ because of the patient's unusual presentation and previous pulmonary embolism, we took extensive diagnostic tests which all came back normal.

Learning points.

• Acute respiratory distress syndrome (ARDS) patients benefit from conservative fluid management and low tidal volume ventilation.

• Furosemide may have a therapeutic role in the treatment of ARDS and particularly in patients with high brain natriuretic peptide levels.

• Pulmonary embolism mimics many pathologies and can manifest as acute loss of consciousness.

• In cases which the patient's condition is not explained by a clear aetiology, a broad diagnostic approach is required.