National survey

We conducted a MEDLINE search using the keywords “analgesia”, “sedation”, “delirium” and “ICU” to identify the most important aspects in the literature regarding this area. The survey had five parts: 1) professional profile and ICU characterization (8 questions); 2) generic information about guideline application and follow-up (4 questions); 3) sedation practices (9 questions); 4) analgesia practices (4 questions); and 5) delirium, rehabilitation/mobilization and sleep improvement (17 questions) (Appendix 1S - Supplementary material). The questions referred to participants’ perceptions of their ICU daily practice. The format of the questionnaire was tick boxes and blank spaces for completion, with an average duration of 15 minutes, and questions were not mandatory. To ensure that the survey was robust, sensitive, and reliable, it was presented to the GASD (Analgesia, Sedation and Delirium Study Group). This group is a critical care research group that includes intensivists and ICU trainees who commented on, discussed and approved the final version.

Between 1 September 2016 and 30 April 2017, the survey was distributed through a link sent by e-mail to all ICU physicians registered in the Portuguese Intensive Care Society, irrespective of their working place being the National Health Service or a private setting. Neonatal and pediatric ICUs were excluded. The survey was designed using a web-based provider (MedQuest) accompanied by a cover letter, which informed the respondent of the details required for its completion.

The survey did not contain any data that could identify the respondents. Participation in the survey was anonymous, voluntary, and noninterventional.

To reinforce participation, the survey link was sent two times during this period. There was a limitation in the survey response; only the professionals who had not responded to the first link could open and fill the second survey link that was sent (to avoid duplicate answers).

Point prevalence study

A point prevalence study (PPS) was performed on the 26^th of January 2018 in adult ICUs. We excluded high dependency units and cardiac and cardiothoracic ICUs. Invitations were sent to the directors of 20 ICUs in

Portugal from north to south, and 17 agreed to participate. The study was approved by the Research and Ethics Committee of Centro Hospitalar de Lisboa Ocidental. Each hospital received ethics approval from its local Institutional Review Board and was granted a waiver of informed consent for this observational, minimal risk study. No incentives were offered to participants. No financial rewards were granted to participating centers.

The participating ICU team was asked to complete a short data collection form (Appendix 2S - Supplementary material) for each patient (≥ 18 years) in the ICU between 00:00 and 24:00h on the 26^th of January 2018. The exclusion criteria were death during the study period and withholding or withdrawing life-sustaining treatment decisions.

In each ICU, there was one physician dedicated to conducting the study, checking the eligibility, and collecting all the data during the study day (24 hours). Most of the data were collected directly from the patients’ records; there was no intervention in clinical practice, no additional assessments, and no changes in the normal routine in the ICU. The collection form was returned on paper. All data were entered into a dedicated electronic database exclusively created for this study and managed by the authors who ensured its confidentiality. Patients were given a code number to secure their identification.

Data and statistical analysis

Standard descriptive statistics were used as appropriate, and variables were reported as numbers (%). Continuous variables are expressed as the mean ± standard deviation, and categorical variables are expressed as n (%). As the number of respondents varied across the questions, with some missing answers, the proportions displayed in the results section and tables were not constant. Statistical analysis was performed using IBM SPSS Statistics v26.0 (IBM, Somers, NY, USA).

National survey

The overall response rate was 28% (117/418 physicians). The majority of respondents (91.2%; 103/113) worked exclusively in the ICU and specialized in intensive care medicine (68%; 78/114), with a mean age of 47 years and a mean of 12 years of ICU practice. The ICUs had a mean of 13 beds (5 beds level 2 and 10.7 beds level 3) and were predominantly distributed in the north and center of the country (84%), and most of them had a mixed medicalsurgical patient case mix (89.8%) (Table 1).

Table 1 - Survey: characteristics of respondents

The existence of protocols for ASD was considered useful by 95% (104/110) of the physicians, but less than 50% referred to having such protocols in their ICUs (Table 2).

Table 2 - Survey analysis of protocols regarding analgesia, sedation and delirium

Sedation assessment

Concerning sedation, 92% (102/111) of the respondents reported daily sedation assessment. The most commonly used scales were Richmond Agitated Sedation Scale (RASS)⁽¹⁹⁾ 41.7% (98/235) and Ramsay Sedation Scale (RSS)⁽⁶⁾ 17% (40/235). Some physicians (15.7%, 37/235) reported the use of clinical evaluation without specific sedation assessment tools (Table 2).

Most of the respondents described doing such an assessment three times per day (40.2%; 47/117). Spontaneous awakening trials with daily sedation interruption were reported by 54% (60/111) of respondents.

The most frequently used drugs were propofol (91.4% of physicians 107/117), followed by opioids (79%; 93/117), midazolam (68%; 80/117) and dexmedetomidine (50%; 59/117) (Figure 1).

Figure 1 - Survey: first-line therapy for pain management (n = 282), sedation (n = 393) and hyperactive delirium treatment (n = 316). NSAIDs - nonsteroidal anti-inflammatory drugs.

Download

Figure 1 - Survey: first-line therapy for pain management (n = 282), sedation (n = 393) and hyperactive delirium treatment (n = 316). NSAIDs - nonsteroidal anti-inflammatory drugs.

Concerning the adequacy of sedation, 46% (52/112) considered that patients were oversedated most of the time, 2% (2/112) reported insufficient sedation, and 52% (58/112) considered sedation to be appropriate.

Pain assessment

Pain was described as a frequent problem in the ICU (99%; 111/112), and 93% of the respondents (101/109) reported monitoring pain daily, but only 71% (75/106) adopted standardized scales. The Behavioral Pain Scale (BPS) and Visual Analog Scales (VAS) were the most commonly used scales (Table 2), and opioids were the most commonly used analgesic (94%; 110/117), followed by acetaminophen (77%; 90/117) (Figure 1).

Analgesia by regional blockade was reported by 67.6% of the physicians (75/111) as part of multimodal analgesia^(20,21) (use was reported in 94% (100/106)).

Delirium assessment

Most physicians (70%; 77/110) reported performing daily delirium monitoring. The Confusion Assessment Method for Intensive Care Unit (CAMI CU)⁽²²⁾ was the most reported delirium assessment tool (54.5%; 55/101), and the Intensive Care Delirium Screening Checklist (ICDSC)⁽²³⁾ was reported by only 4% (4/101). Some physicians described delirium assessments based on clinical evaluation without the use of scales (41.6%; 42/101). Almost half of the respondents considered delirium scales easy to apply (37.4%; 34/91).

Delirium prevention and treatment therapies included both pharmacological and nonpharmacological approaches. The most common approaches were sleep promotion (19.5%; 106/543), noise reduction (18.2%; 99/543) and pain treatment (17.7%; 96/543). The family was involved in delirium treatment in only 14.2% (77/543) (Figure 2). Measures for sleep promotion included reducing hearing with ear plugs (28.7%; 115/401), minimizing light (22.9%; 92/401) and noise (20.2%; 81/401), cognitive stimulation (14.8%; 59/401) and pharmacological therapies (13.5%; 54/401).

Figure 2 - Nonpharmacological treatment for intensive care unit delirium (n = 543).

Download

Figure 2 - Nonpharmacological treatment for intensive care unit delirium (n = 543).

When analyzing the pharmacological approach to hyperactive delirium, physicians reported the use of haloperidol (91.5%; 107/117), quetiapine (71.8%; 84/117) and alpha 2 agonists (43.6%; 51/117) (Figure 1). Benzodiazepines were reported to be used by 23.9% (28/117). Subsyndromal delirium^(24,25) was reported by 76% of physicians (68/117) and delirium treatment was reported by 46.2% (54/117) of physicians.

Point prevalence study

Twenty ICUs were invited to participate in the PPS. A total of 17 ICUs were included, from the North, Center and South of Portugal, with a total of 189 beds. We included 192 patients (37.5% females and 62.5% males). The number of patients was higher than the number of beds available since, in some cases, during the study duration, a patient was discharged from the ICU and another was admitted to the same bed (Table 3).

Table 3 - Point prevalence study: demographic and clinical characteristics of patients

Most of the ICUs were mixed, just one was specific for neurocritically ill patients, and the main type of admission was medical (61.5%). The mean Acute Physiology and Chronic Health Evaluation (APACHE) II and Simplified Acute Physiology Score (SAPS) II scores at admission were 19 and 42, respectively. The Sequential Organ Failure Assessment (SOFA) score at admission was 7.3 and on the study day was 5.9. Fifty-six percent of the patients were under invasive mechanical ventilation. There were some missing values in the case report form that are mentioned in the results presentation (Table 3).

Sedation

Sedation was used in 87 patients (45.3%), with propofol being the most used drug (58.6%; 51/87), followed by midazolam (27.5%; 24/87), remifentanil (8.0%; 7/87) and dexmedetomidine (4.6%; 4/87). The mean propofol daily dose was 2400mg/day, and midazolam was 157mg/day. Apart from these 24 patients with an intravenous perfusion of midazolam, 20 other patients were also receiving the enteric benzodiazepines lorazepam, alprazolam, oxazepam and bromazepam regularly (Table 4).

Table 4 - Point prevalence study: sedative and analgesic intravenous medication

A sedation assessment was performed in 88.5% of patients (170/192), at a minimum of 1 time per day and a maximum of 17 times/per day. The most commonly used scale was the RASS (58.2%; 99/170). The Glasgow coma scale (GCS),⁽²⁶⁾ was used alone in 18.2% (31/170) and in association with the RASS scale in 15.9% (27/170) (Table 5).

Table 5 - Point prevalence study: sedation, analgesia and delirium assessment tools (individual use or in combination)

To assess the adequacy of sedation (oversedation and undersedation), we compared the sedation target for a specific day with the sedation presented by patients. Figure 3 shows the variation in sedation level, in RASS points, between the target and the measured RASS.

Figure 3 - Sedation target compared with real sedation; sedation level assessed with the Richmond Agitated Sedation Scale. RASS - Richmond Agitation-Sedation Scale.

Download

Figure 3 - Sedation target compared with real sedation; sedation level assessed with the Richmond Agitated Sedation Scale. RASS - Richmond Agitation-Sedation Scale.

In 147 patients, 57% (n = 84) achieved the sedation target desired for that day, 32% (47 patients) were oversedated compared with the targets, and 11% (16 patients) were undersedated.

Analgesia

Analgesia was used in 162 patients. Acetaminophen was the most frequently used analgesic (34.6%; 83/240) at a mean dose of 3000mg/day (SD ± 404), followed by opioids: fentanyl (21.3%; 51/240), morphine (9.2%; 22/240), tramadol (7.9%; 19/240) and remifentanil (4.2%; 10/240). The mean fentanyl dose was 2.16mg/day (SD ± 1.6), and the mean morphine dose was 24 mg/day (SD ± 31.4) (Table 4).

Acetaminophen was the most commonly used agent in monotherapy (n = 61), followed by metamizole (n = 34) and fentanyl (n = 34). The most frequent analgesic association was fentanyl with acetaminophen (n = 14), followed by morphine with acetaminophen (n = 12).

All ICUs used pain scales, and furthermore, some applied more than one scale in the same patient (Table 5). The BPS⁽¹⁹⁾ was the most used scale (45.7%; 80/175), followed by the Numeric Rating Scale (NRS)⁽²⁷⁾ (20.6%; 36/175), VAS⁽²¹⁾ (9.7%; 17/175) and Escala de Conductas Indicadoras de Dolor (ESCID)⁽²⁸⁾ (5.7%; 10/175).

Delirium

Delirium assessment was reported in 8.3% (16/192) of the patients. In 17.7% (34/192), it was not possible to access delirium due to deep sedation (RASS -4 and -5), and 14% (27/192) of the patients had missing information. The majority of the patients (59.9%, 115/192) did not have any delirium assessment.

Concerning the 8.3% that were assessed for delirium, the method used had a significant percentage of missing values (91.7%). Assessment with CAM-ICU was reported only in 5 patients. Subsyndromal delirium was not assessed in the ICUs. Physical restraints were used in 18 patients (9.4%).