Length of intensive care unit (ICU) stay is a classical and important tool to evaluate quality of care and activity of the ICU. For clinicians, ensuring optimal care with a short length of stay is important because of the association between prolongation of ICU stay and complications, with an effect in terms of morbidity and mortality [5]. To our knowledge, no consensus exists that defines when a stay is considered to be prolonged; moreover, these settings might be different, depending on the characteristics of the ICU (surgical or not, and type of surgical ICU) [6]. However, an ICU stay of 14 days seems to be relatively consensual for a nonspecific ICU [7]. Over the years, probably related to patient evolution (with increasing comorbidities) and progress in intensive care, a new category of patients has appeared and is expected to increase: They experience a catastrophic illness and survive after a long ICU stay and aggressive therapy, but they remain dependent on a ventilator. Moreover, these patients suffer from consequences of multiple organ failure, including renal insufficiency, acquired weakness, malnutrition, and colonization with multidrug-resistant bacteria. We can designate them as chronic critically ill patients [8, 9].

We can find many definitions for prolonged weaning. Thus, in the United States, when a patient is ventilated more than 21 days, he or she can be systematically considered as a prolonged ventilated patient, regardless of the weaning process issue. This definition is directly related to economic considerations [10]. Indeed, some patients may not benefit from weaning attempts, because of the severity and persistence of the underlying disease that lead to prolonged invasive MV. Thereby, in 2007, an international consensus conference [1] proposed a classification of the weaning process according of the number of spontaneous breathing trials performed and the success of extubation. The third group in this new classification is referred to as “prolonged weaning,” defined as requiring more than three weaning attempts or 7 days to be separated from the ventilator. Although these prolonged weaning patients represent fewer than 10 % of the whole population of ICU patients, they paradoxically require a prolonged ICU stay, accounting for up to 40 % of ICU expenditures [11, 12].

The concept of a specialized weaning unit (SWU) was developed in the United States [13], based on epidemiologic (an increase of patients with difficult and prolonged ventilation) and economic (lower operating costs of these units and preservation of ICU bed availability) arguments. The notion of a SWU was also introduced by the international consensus conference on weaning, and discharge of these patients in specific units should now be encouraged [1]. Several modalities of organization have been described [4]. Schematically, a SWU could be supported by an ICU or located in a specific long-term care hospital.

The concept of chronic critically ill patients emerged in the last decade from cohort studies and can be applied to many ICU patients [14]. Recent studies confirm that prolonged weaning is associated with a higher mortality, longer ICU stay, and longer duration of MV [3]. This finding has led to the development of the SWU because of the high cost of conventional ICU wards and because ICUs may lack the necessary organization and specialized staff to manage prolonged weaning patients [15]. These specific units also facilitate more efficient use of critical care resources by increasing the availability of acute ICU beds. Thus, in their retrospective study, Lone et al. [16] showed that an SWU could decrease by 8–10 % the number of ICU day-beds occupied and permit a savings of nearly $500,000/year.

SWUs also offer expertise in prolonged weaning and specialized multidisciplinary teams composed of clinicians, physiotherapists, nurses, nutritionists, and psychotherapists. Moreover, these units are characterized by a suitable environment, with less noise and less activity, especially at night, both of which are sleep promoting. Thus, they can provide more specifically adapted care with reference to the patients’ needs, together with cost savings related to the reduced nurse-to-patient ratio, specific monitoring, and technical equipment.

To our knowledge, no study has compared different ICU criteria to define those used for discharging patients to an SWU. In relation to the purpose of such a unit, patients must be classified in the “prolonged weaning” category, provided that a well-conducted weaning from MV has been initiated but resulted in weaning failure. Consequently, a real weaning project should exist and be planned. Therefore, patients with persistent coma, spinal cord injury with diaphragmatic impairment, or terminal chronic respiratory failure do not appear to be good candidates for SWU admission. At the stage of prolonged weaning from MV, eligible patients for an SWU should have already been tracheostomized to facilitate their global and ventilatory management [15]. Recent studies have demonstrated that performing a tracheostomy does not alter outcome in terms of mortality or duration of MV [17]. It is generally acknowledged, however, that performing a tracheostomy in these ventilator-dependent patients permits reduction of the work of breathing and sedation level and more easily restores mobilization, nursing, swallowing, oral nutrition, and speech [18, 19]. In addition, no other vital organ failure than ventilatory dependence should persist, even though patients may present some consequences of multiple-organ failure (e.g., denutrition, renal failure). Thus, there are benefits from a transfer to an SWU for such patients: the patients benefit by the different conduct of their specific management, and the community benefits as it frees up some ICU beds.