To the best of our knowledge, our study was the first randomized controlled trial aiming to compare two sedation strategies in a developing country with a low nursing staff level. Although it was underpowered to show any difference in duration of mechanical ventilation between intermittent sedation and daily interruption of sedation, it suggests that both strategies seem to be safe in terms of complications, and are not different regarding nurse workload - two very important issues in units with profiles similar to ours. Both strategies are also secure in terms of psychological distress six months after hospital discharge. These results indicate that lighter sedation approaches may be feasible and safe even in lower nursing staff level ICUs.
Deep sedation is a huge problem in critically ill patients on mechanical ventilation. Continuous sedation has been associated with prolonged times on mechanical ventilation for more than a decade . Since then, many studies have shown that protocols of sedation which allow patients to be awake or easily arousable are associated with fewer days on mechanical ventilation and in the ICU [8, 18]. However, in ICUs with very high staff levels, protocols of sedation may not be necessary .
Daily interruption of sedative infusions was another strategy that has shown to be effective in reducing timing on mechanical ventilation [6, 7]. It is not associated with excessive complications rate as compared to the usual practice , and it is safe in terms of psychological distress after hospital discharge [5, 21]. A systematic review suggested that daily interruption of sedatives had a higher level of evidence than sedation protocols , but comparisons between the two are scarce.
There have been five studies before ours comparing daily interruption of sedatives with a specified sedation protocol. A Boston study included 74 medical patients and compared daily interruption with a sedation protocol. In that study, daily interruption of sedatives was associated with a prolonged duration of mechanical ventilation (almost three days), ICU length of stay (seven days), and length of hospital stay (eleven days) . A pilot Canadian trial compared a sedation protocol with the same sedation protocol associated with daily interruption of sedative infusions in 65 patients. They found no difference between the two approaches in terms of duration of mechanical ventilation or ICU length of stay . A subsequent multicenter trial based on this pilot was conducted in Canada and the United States with 423 patients, and also evidenced no difference in time to successful extubation, ICU and hospital length of stay . A Greek trial included 97 patients, and also showed no difference in duration of mechanical ventilation, ICU and hospital stay between daily interruption of sedative infusions and a practice of early awakening of patients . A recent Danish study with 113 patients suggested that a protocol of no-sedation, that kept patients with scheduled infusions of morphine and instituted continuous infusion of propofol for short periods of time, if necessary, was associated with more days without mechanical ventilation and shorter ICU and hospital stays .
Except for the Canadian trials, all of them were single-center trials. Therefore, they expressed local practices, and their results can only be compared and externally applied with some caution. The Boston study  targeted a deeper level of sedation than all other studies and the original description of daily interruption of sedative infusions . Also, this is the only study which included only medical patients. Consequently, their results may be very different from the others. The Canadian trials [24, 25] had the exact same target as ours, but they studied continuous infusion of sedatives on both arms. The Greek trial  had a target of sedation similar to ours, although their protocol used continuous infusion of sedatives and remifentanyl as the analgesic drug. It seems that the study that is most similar to ours is the Danish one . Like ours, that study allowed patients to be maintained without any sedative, used propofol or midazolam in the daily interruption arm (although in a different fashion), but kept patients with scheduled morphine infusions, while we used fentanyl as analgesic. The Danish study ICU also had a higher nurse-to-patient ratio. However, their patients had fewer days free of mechanical ventilation. We may postulate that the impact of reducing sedative agents appears only when patients need to be on mechanical ventilation for longer periods of time. Nevertheless, we have shown that a strategy of intermittent sedation is feasible even in a lower staff level ICU.
Recently, it has been shown that not only are deeper levels of sedation associated with worse outcomes, but also early deep sedation (that is at the first 48 hours after admission) is associated with delayed extubation and higher mortality [3, 27]. Most clinical trials aiming to reduce the burden of sedation only enrolled patients after 48 hours of mechanical ventilation [7, 18, 28]. In our trial, both interventions started very early, since all patients were randomized in less than 48 hours, and we used a very small amount of sedatives in both arms compared to previous [6, 7, 23–25], but not all studies . Like ours, a recent pilot clinical trial has shown the feasibility of achieving an early lighter sedation .
It has been thought that deeper sedation levels were associated with higher levels of comfort. However, recent findings from clinical trials were not able to demonstrate those ideas. The greater cardiovascular safety associated with deeper sedation is also speculative. The daily sedative interruption is not associated with higher cardiovascular events in high risk patients . In this trial, the physiological variables were similar in both groups, a fact that points out the safety of light sedation levels in mixed critically ill patients.
In our study, tidal volumes were higher in the intermittent sedation group. Although we excluded patients with severe hypoxemic respiratory failure, it is worth noting that this strategy may be harmful for these patients, since tidal volumes would not be strictly controlled .
Nurse workload was similar in both groups and decreased during the first five days of mechanical ventilation (Table 3). By contrast, in the Danish no-sedation trial, the awake patients needed an extra person to take care of them . Although our data does not show significant increase in workload, it is advisable to exercise caution before widely adapting this sedation strategy in a lower staff ICU. As we mentioned before, we began the intermittent-sedation strategy in 2006. Perhaps, at that time, this approach may have been associated with a higher nurse workload, but then, an adaptive process took place. Nevertheless, in this study, lighter sedation was not associated with a number of adverse events that may be commonly related to lower staffing levels, such as self-extubation or accidental removal of catheters. More importantly, these adverse events were not different from those presented in other studies [9, 25].
Lighter levels of sedation have been associated with safety [21, 32] or even better psychological outcomes when compared to deeper sedation levels [5, 18]. Our study, like the Danish study , suggests that intermittent sedation is not associated with worse psychological outcomes than daily interruption of sedative infusion. However, our conclusions are derived from a very small sample of patients since we were not able to reach a large number of patients who were discharged alive from the hospital.
We believe our trial has some strengths. First, it has shown that lighter sedation levels are feasible and safe even in a lower nursing staff ICU. Second, we studied two approaches which prioritize pain assessment and management. This is in accordance with what is recommended in international guidelines . Therefore, it seems that it is possible to adopt light sedation strategies even in low nursing staff ICUs in order to reduce the duration of mechanical ventilation.
On the other hand, our study also has several limitations. First, it was conducted in a single ICU in a teaching hospital. Therefore, our results may not be generalizable to services with different profiles. Second, and most importantly, we were not able to reach the target sample. Nevertheless, as commented before, it would be necessary to analyze a larger sample than it was initially calculated in order to find a statistically significant result that could still not be clinically meaningful. Third, we hypothesized that intermittent sedation would increase the number of mechanical ventilation-free days. However, since both strategies reached very close levels of sedation, a non-inferiority design would probably have been a better approach to address this question. Fourth, we did not use dexmedetomidine as a sedative agent, which has been associated with positive findings when compared to benzodiazepines [28, 34], but not with propofol .