Management of suspected acute heart failure dyspnea in the emergency department: results from the French prospective multicenter DeFSSICA survey

This large French observational survey of acute HF coordinated by emergency medicine investigators has successfully collected epidemiologic, diagnostic, and therapeutic data, from emergency admission to hospital discharge. As such, DeFSSICA provides information on HF presentation and the French pathway of care.

Not surprisingly, patients in DeFSSICA were elderly, with a median age of 83 years. However, DeFSSICA patients were generally older than those in other HF registries from France (Observatoire Français de l’Insuffisance Cardiaque Aiguë [OFICA]) [4], Switzerland (Acute Heart Failure Global Survey of Standard Treatment [ALARM-HF]) [5], Spain (Epidemiology of Acute Heart Failure in Emergency Departments [EAHFE]) [6], and the US (Acute Decompensated Heart Failure National Registry Emergency Module [ADHERE-EM] [7], Organized Program to Initiate Lifesaving Treatment in Hospitalized Patients with Heart Failure [OPTIMIZE-HF]) [8]. Mean ages in these studies ranged from 73 to 79 years (details of these registries can be found in the Additional file 1). Compared with these registries [4‐8], patients in DeFSSICA were less likely to have chronic respiratory failure or chronic HF, but more likely to have chronic renal failure (Web Appendix). Compared with the French OFICA study [4], patients in DeFSSICA were more likely to have hypertension (71 % vs 62 %) and atrial fibrillation (45 % vs 38 %). As atrial fibrillation and a rapid heart rate have been closely linked to mortality [9], detection of atrial fibrillation should be considered systematically.

Furosemide, oxygen, and nitrates were the most commonly used treatments in DeFSSICA patients with HF (75, 68, and 19 %, respectively). The limited use of nitrates may be related to the median (IQR) SBP of 140 (121–160) mmHg. However, our use of nitrates was similar to those in the EAHFE [6] (20.7 %) and OPTIMIZE-HF [8] (14.3 %) registries. Only 10 % of our HF patients received non-invasive ventilation or Boussignac continuous positive airway pressure (CPAP), similar to in the OFICA study (12.4 %) [4]. This ventilation assistance improves outcome in acute HF patients, especially those with pulmonary edema [10, 11]. However, it requires patient compliance and cooperation, which is not necessarily obvious in elderly [12, 13].

Current recommendations for diagnostic investigations in patients with suspected HF include first the assessment of HF probability related to the clinical history (history of coronary artery diseases, or hypertension, or use of diuretics…), the physical examination (bilateral ankle oedema, jugular venous dilatation…) and the ECG (Class I, Level C) [14].

If one of these symptoms or abnormal ECG signs are present, the measurement of BNP or NT-proBNP is the second step in the algorithm of the diagnosis (Class IIa, Level C) with also blood chemistry, and complete blood count (Class I, Level C). Finally, comprehensive assessment of patients with HF comprises, chest X-Ray, transthoracic echocardiography (Class I, Level C), and lung ultrasound (Classe IIb, Level C) [14]. However, in DeFSSICA, only 15 % of all patients underwent echography, due to a lack of equipment and training of emergency physicians. When echography was performed by a cardiologist (60 % of cases), most considered the results to be satisfactory (66 %), but when it was performed by emergency physicians (40 % of cases), only 21 % considered the results to be satisfactory. Therefore, we have identified a need for additional effort for training emergency physicians and providing equipment, as prompt use of echography in the ED has been shown to improve the diagnosis of HF [15].

In line with guidelines, the proportions of patients who underwent ECG, biological analysis, or chest X-ray were all >90 % in DeFSSICA. Similarly, BNP or pro-BNP was measured in 93 % of patients, compared with 82 % of patients in the OFICA study [4]. Not surprisingly, BNP and pro-BNP values were significantly higher among those with versus without HF in DeFSSICA. Several studies have shown the diagnostic and prognostic values of BNP (and troponin) in HF [16‐18]; however, they have limited specificity and may be elevated secondary to a variety of causes [9]. Two recent papers have also suggested that neprilysin [19] and soluble CD146 [20] may have a role in the diagnosis of acute HF. Although BNP may be helpful when the diagnosis of HF is in doubt [21], ultrasound remains the gold standard. [22] The use of ultrasound in the ED has been reported [23] to accelerate the diagnosis of HF and the initiation of treatment, and shorten the length of stay. For example, in this study [24] the use of ultrasound could help early diagnosis and therapy for the patient who may need to transfer to the intensive care unit and reduce the length of stay for in-hospital patients. Several others studies have shown the feasibility of ultrasound in the ED and prehospital emergency management of HF patients [24‐27]. As the diagnosis of acute HF remains difficult, ultrasound enables emergency physicians to improve their diagnostic performance, optimize their therapeutic strategy, and improve prognosis [28].. In the DeFSSICA survey, only 19 patients received lung ultrasound (LUS) evaluation in the ED. This result clearly shows that, for now, lung ultrasound is scarcely used by French emergency physicians. This fact is really different from the American or Italian practices where LUS is being used frequently. Lung ultrasound is an excellent differential diagnostic tool for distinguishing cardiac causes from pulmonary causes of dyspnea [29].

Pivetta et al. [30] definitively confirmed the excellent diagnostic performance of lung ultrasound to detect acute HF in an ED setting. In their multicenter study that included 1005 patients admitted with acute dyspnea, LUS sensitivity and specificity were respectively 97 % [95 % CI, 95–98.3 %] and 97.4 % [95 % CI, 95.7–98.6 %], and were superior to the one observed with clinical evaluation, chest X-ray and natriuretic peptides. Importantly, the net reclassification index of the LUS-implemented approach compared with standard workup was 19.1 % [30], which suggest the high diagnostic value of LUS in this setting. This performance of LUS has been confirmed in a prehospital setting by Prosen et al. (100 % sensitivity, 95 % specificity) [31]. So it’s probably a key to extend the training facilities for LUS in Europe, especially in France.

To date, the HF pathway of care has been poorly examined in the literature, in particular, patient destinations after ED discharge. It has been suggested that up to half of patients with HF may not require hospitalization [32]. However, in DeFSSICA, 95 % of HF patients were hospitalized, most commonly (28 %) in cardiology units. Approximately 14 % of HF patients were moved to ED hospitalization units, which provide close observation in the absence of a bed in the appropriate unit, or for patients due to be discharged the following day. Over one third of HF patients were hospitalized in non-cardiology units, particularly geriatric units (12 %). If the demonstration was made that the management of HF patients is more optimal in cardiology units [3], the choice of another destination may have been governed by a lack of beds in cardiology and/or the advanced age of some patients requiring a geriatric unit. Considering that majority of patients with AHF should be hospitalized in cardiology or ICU, we asked emergency physicians to assess whether the final destination was appropriate. In more than 75 % of cases they considered it adapted based on the diagnosis and the patient’s clinical status in ED. In 24 % of cases, the ED physician considered that the destination unit was not appropriate, mainly due to lack of beds availability in the unit perceived as appropriate.

In DeFSSICA, 18 % of HF patients had creatinine clearance <30 mL/min. Close cooperation with nephrologists is necessary in these patients, to determine whether dialysis is indicated [28, 33].

Although the incidence of HF decreased between 2000 and 2010 in relation to improvements in the management of myocardial infarction. Regarding to mortality, we have different data between Europe and the United States. While the mortality rate appears stable in the United States between the years 2000-2010 [34], there is a decrease between the years 1987–2008 in Europe [35]. In-hospital mortality of HF patients in DeFSSICA was 6.4 %, slightly lower than in the OFICA study [4] (8.2 %). At 30 days, 6.8 % of DeFSSICA HF patients were still hospitalized, while the median (IQR) length of stay was 7 (3–12) days. In the same idea, the length of stay in DeFSSICA is 7 days in median, like in EAHFE registry [3] and ALARM-HF [5] study which include the patients from the Emergency Department. However in OFICA [4] registry the length of stay is twice. This is probably explained by the inclusion of the patients admitted in intensive cardiac unit who are probably more serious.

Improved interdisciplinary cooperation has been highlighted as a key factor for the improvement of HF patient care [36]. An acute HF syndromes clinical trials network has been set up to improve clinical trials in this patient population [37]. Telemedicine [38] and wireless monitoring could also offer opportunities for improved HF management.