This study was a randomised, open-label, multicentre, parallel-group study comparing the safety and efficacy of an analgesia-based regime using remifentanil with a conventional hypnotic-based regime in critically ill patients requiring mechanical ventilation for 3 to 10 days. The study was conducted in accordance with good clinical practice and within the guidelines set out in the Declaration of Helsinki. Informed consent/assent was obtained from all patients or their representatives. After approval from local and national ethics committees, 105 patients from 15 centres in 10 countries were recruited. Patients were randomised in a 1:1 ratio to receive either a remifentanil-based regime or a comparator hypnotic-based regime using midazolam with either morphine or fentanyl for analgesia.
Inclusion and exclusion criteria
The target population were those patients requiring long-term mechanical ventilation for medical reasons. Post-surgical patients requiring extended mechanical ventilation as a result of post-surgical complications were also included. Patients were eligible if they were more than 18 years old, had been admitted to the intensive care unit (ICU) within the previous 30 hours, were expected to require mechanical ventilation for longer than 96 hours and required analgesia and sedation. Females were eligible to enter the study if they were of non-childbearing potential or had a negative pregnancy test at screening and agreed not to fall pregnant for 12 days after stopping the study drug.
Patients were excluded from the study if their medical condition prevented assessment of depth of sedation, if it required the frequent down-titration of analgesics/sedatives for assessment, if it was likely to require surgery or tracheostomy during the treatment period, if it required neuromuscular blocking drugs by infusion, if it required epidural blockade, if it required sedatives or anaesthetic agents other than study drugs specified in the treatment period, or if there was a contraindication to the use of remifentanil, morphine, fentanyl or midazolam. Other exclusions were sensitivity to the drugs or class of drugs specified in the study, a history of alcohol or drug abuse, a concurrent or previous entry into this or other investigational drug studies within 30 days, or pregnancy or lactation. Protocol-specified treatment regimes had to be appropriate for the management of the patients. After 30 patients had entered the study, a protocol amendment allowed the inclusion of patients who had been receiving mechanical ventilation for up to 30 hours irrespective of the time in the ICU, allowed the inclusion of patients requiring surgery of less than 6 hours' duration during the treatment period and reduced the required duration of mechanical ventilation from 96 to 72 hours.
The study was divided into four periods: screening, treatment, post-treatment and follow-up.
The screening period was from ICU admission to the start of the study drug and included the time for considering eligibility, obtaining consent/assent, randomisation and assessment of the patient's SAPS II score. Baseline demographics, physiological variables, Sedation–Agitation Scale (SAS) score and pain intensity (PI) score were also assessed during this time. The SAS is a seven-point scoring system, and a SAS score of 3 or 4 was defined as optimal sedation in this study (see Additional file 1
]). The PI score is a six-point score where 1 or 2 represents no pain or mild pain (see Additional file 2
). Baseline liver function tests and creatinine clearance were also measured.
The treatment period was from the start of the study drug to permanent discontinuation of the study drug, or after 10 days of administration, or death, whichever was the sooner. SAS, PI, heart rate (HR) and mean arterial pressure (MAP) were continuously monitored throughout the treatment period and were recorded at the time of each bolus dose and/or change in infusion rate of any of the study drugs. These parameters were recorded again when optimal sedation and pain control had been established or re-established. These variables were also recorded at least every 8 hours in the event of no change in study drug. Study drugs and amount administered were also recorded during the treatment period. Liver function tests and creatinine clearance were assessed daily. Before the start of administration of the study drugs the existing sedative/analgesic regime was discontinued. Patients were then sedated to an optimal SAS and PI score by a remifentanil-based regime or a hypnotic-based regime.
The remifentanil infusion was started at 6 to 9 μg kg-1
(0.1 to 0.15 μg kg-1
). The remifentanil infusion was titrated in 1.5 μg kg-1
(0.025 μg kg-1
) increments at 5 to 10 min intervals to achieve an optimum level of sedation/analgesia based on clinical judgement. Bolus doses of remifentanil were not permitted. Once the remifentanil infusion reached a rate of 12 μg kg-1
(0.2 μg kg-1
), boluses of midazolam (not more than 2 mg) could be used if required after clinical assessment. Remifentanil was not used as the sole agent for sedation at infusion rates greater than 18 μg kg-1
(0.3 μg kg-1
). Above this rate, midazolam boluses were used. Further increases in the remifentanil rate were allowed for the treatment of pain and in anticipation of short stimulating procedures, up to a maximum rate of 45 μg kg-1
(0.75 μg kg-1
). The remifentanil dosing regime is depicted in Fig. [5
Midazolam was used by infusion and/or boluses as the sedative agent, and was titrated to an optimum level of sedation based on clinical judgement and in accordance with standard clinical protocols. Either morphine or fentanyl was used as the analgesic agent, titrated to obtain adequate pain control. The initial dose and subsequent adjustments of sedative and analgesic agents were at the investigators' discretion and in accordance with routine clinical practice to obtain an optimum SAS and PI score.
Weaning and extubation
The decision to begin the weaning process was based on clinical judgement and was defined as the time point at which the investigator first adjusted the study drug infusion rate or decided not to give any more boluses of the study drugs so as to encourage spontaneous respiration with the result of extubating the patient. For those patients extubated within 10 days, the study drugs were down-titrated in accordance with clinical judgement until a decision was made to extubate the patient. For patients in the comparator hypnotic-based treatment group this was performed in accordance with routine clinical practice at the investigational site. As a guide, for patients in the remifentanil group who were eligible for extubation, the remifentanil infusion was decreased to 6 μg kg-1 h-1 (0.1 μg kg-1 min-1) either immediately or in increments at the investigator's discretion, and no further midazolam boluses were given. Remifentanil was discontinued after extubation and, if necessary, suitable alternative methods of pain relief were instituted.
For patients not extubated within 10 days, the study drugs were discontinued in both groups and the time to offset of pharmacodynamic effects was recorded. As soon as there was a demonstrable change in haemodynamic variables, SAS score or PI score, alternative sedation and analgesic regimes were instituted as soon as clinically indicated.
The post-treatment period was from the end of the treatment period until 24 hours later. MAP, HR, SAS and PI were recorded at 15 min intervals for the first 2 hours, hourly for the next 4 hours then 6-hourly until the end of the post-treatment period.
The follow-up period was from the end of the post-treatment period until 6 days later.
The primary endpoint was the time from the start of study drug to extubation. Secondary endpoints were the time from start of study drug until start of weaning, the time from start of weaning until extubation, the time from start of study drug to ICU discharge, descriptive PI and SAS during the treatment and post-treatment periods, total exposure to study drugs and concomitant sedative requirements.
The safety endpoints were the offset of pharmacodynamic effects of study drugs after permanent discontinuation, haemodynamic effects, clinical adverse events and the requirement for re-intubation. Haemodynamic variables were monitored continuously throughout the study and recorded at the times stated above. Adverse events were recorded from the start of the study drug until the end of the post-treatment period. Serious adverse events were defined as adverse events that resulted in any of the following outcomes: death, life-threatening event, prolongation of hospitalisation, or a disability or incapacity. Important medical events that did not result in death or were not life-threatening were considered serious adverse events when, on the basis of appropriate medical judgement, they jeopardised the patient and required medical or surgical intervention to prevent one of the outcomes listed above. In addition, serious adverse events possibly attributable to study medication were recorded throughout the 6-day follow-up period.
The time to event endpoints were analysed with the generalised Wilcoxon test with a two-sided α level of 5% judged to indicate a statistically significant difference between the treatment groups. The data for patients who did not experience the event were censored in accordance with predetermined rules. The results of these analyses were summarised by using 75th centiles, difference between 75th centiles and its 95% confidence interval because too few patients achieved each event to allow estimates based on median times to be determined with any precision. The confidence intervals were calculated with methods described by Collett [19
The percentage time of optimal analgesia/sedation was calculated and summarised by the median in each treatment group and the median of all possible differences between the groups and the 95% confidence interval around that difference to give the best estimate of median difference. Treatments were compared by using the Wilcoxon rank sum test.
With the exception of the incidence of re-intubation, no formal statistical analyses were performed on the demographic, baseline or safety data. These data were summarised either by means and standard deviations (SD) or by frequency tables as appropriate to the data.