Clinical Practice Guidelines and Consensus Statements About Pain Management in Critically Ill End-of-life Patients

A Systematic Review

Alejandro Durán-Crane, MD; Andrés Laserna, MD; María A. López-Olivo, MD, PhD, MSC; John A. Cuenca, MD; Diana Paola Díaz, MD; Yenny Rocío Cardenas, MD, MSC; Catherine Urso, BS; Keara O'Connell, BMBS; Kian Azimpoor, BS; Clara Fowler, MSLS; Kristen J. Price, MD; Charles L. Sprung, MD, JD, MCCM; Joseph L. Nates, MD, MBA, CMQ, MCCM


Crit Care Med. 2019;47(11):1619-1626. 

In This Article


Study Selection

We found 6,981 unique citations, of which 112 were retrieved in full text and 10 were finally included for data extraction. Figure 1 shows the flow diagram of document selection and the reasons for exclusion at each stage.

Figure 1.

Flow diagram of studies screened and reviewed.
BVS = Biblioteca Virtual en Salud.

Study Characteristics

The main study characteristics of the included publications are presented in Supplementary Table 1 (Supplemental Digital Content 1, The selected documents were published between 2002 and 2017. Four of them were developed in the United States,[15,14–16] two in Italy,[17,18] one in Spain,[19] one in Canada,[20] one in Belgium,[21] and one in India.[22]

Four of the 10 documents were updated versions of previously published guidelines.[14,16,17,22] None of the documents used a grading system for the quality of the evidence supporting the recommendations. One of the included documents was a position statement from the Italian National Committee for Bioethics,[23] originally written in Italian; therefore, we abstracted their recommendations from a summary published in English.[18] Two documents recommended quality indicators and measures for pain assessment and management in terminal care[5,14] (Supplementary Table 2, Supplemental Digital Content 1,

Risk of Bias Within Studies

The scores of the AGREE-II domains in each of the reviewed documents are shown in Supplementary Table 3 (Supplemental Digital Content 1, The fields with the lowest median scores were "applicability" (28.6%) and "rigor of development" (36.1%). The highest rated fields were "clarity of presentation" (66%) and "scope and purpose" (63.9%). For the overall assessment, five documents(50%) surpassed a threshold score of 60%.[5,15,16,20,22] However, only two were fully recommended by the reviewers,[5,20] whereas eight were recommended with modifications by more than half of the evaluators.[14–19,21]

Results of Individual Studies

Pain Assessment. Eight documents contain recommendations regarding pain assessment for terminally ill patients in the ICU[5,14–17,19,20,22] (Table 1). Two benchmarks for quality improvement suggest continuous pain assessment using a quantitative symptom scale as the quality indicator,[5,14] whereas three other publications mention the use of a nonspecified functional scale for pain assessment.[15,16,19] Only two articles recommend specific pain assessment scales such as the Pain Assessment Behavioral Scale, the Sedation Agitation Scale, the Visual Analogue Scale for conscious patients, the Campbell scale for unconscious patients, or monitoring of the bispectral index.[16,19] The Indian Society recommends a six-step approach (identify, assess, plan, provide, reassess, reflect). It emphasizes the importance of assessing physical symptoms and continuously reassess them, ensuring adequate pain control.[22] In one document, no recommendations regarding pain assessment are made, but the following barriers to adequate pain assessment in the ICU are listed as follows: 1) communication problems imposed by the ICU environment, 2) the severity of illness and the presence of multisystem organ failure, 3) the decreased level of consciousness of patients as a result of illness and drugs, 4) the critical care specialists' lack of knowledge of or difficulty in interpreting clinical signs, and 5) the unreliability of clinical signs.[20]

Medications and Interventions for Pain Management. Seven of the reviewed documents describe interventions for pain management.[14–17,19,20,22] Although all of them name pharmacologic interventions, only three (43%) refer to nonpharmacologic strategies for pain control, with two of them agreeing on the importance of the presence of family members, providing spiritual assistance, and ensuring a private and peaceful environment.[17,19,20] The suggested pharmacologic and nonpharmacologic strategies for pain management are presented in Table 2.

Dosing, Titration, and Route of Administration of Analgesics. Seven documents address the issue of analgesic dosing.[15–20,22] Of these, four suggest initial dosing for adult, opioid-naive patients (Table 3). Five agree that the starting dose of narcotics should depend on the previous amount of opioids received because of the rapid development of tolerance.[15,17,19,20,22] Three recommend adjusting the initial dosing based on the patient's age, previous alcohol or drug addiction, present clinical condition, current levels of consciousness and pain, and wishes concerning sedation.[17,19,20] Only one document recommends adjusting the initial dosing based on organ failure,[19] and only one recommends adjusting it on the basis of the level of available psychologic and spiritual support.[20]

Regarding the titration of medication, two papers agree that sedative and analgesic doses should be increased in response to signs or manifestations of inadequate pain control.[17,20] Two publications agree that reassessment of the drug's effect on the patient and titration of the opioid dose are the mainstays of successful treatment.[15,20] Four manuscripts state that no maximum dose exists and that the dose should be increased as needed to produce the desired effect or until intolerable side effects occur.[15,17,19,20] Two papers endorse the administration of preemptive dosing in anticipation of pain and suffering.[19,20] Concerning the legal standpoint, one paper states that exceeding the maximum opioid dosages necessary to relieve symptoms is morally suspect and illegal in most legislations and thus should be avoided since it may appear to be intended to hasten death rather than to offer symptom relief.[15]

Several routes of administration can be used for opioids and sedatives. One publication recommends the routine use of a bolus and infusion approach, specifying that IV bolus doses should be used for rapid effect when considering increasing the rate of infusion due to the reemergence of signs or symptoms of suffering.[16] Another document recommends the use of oral, combined oral and IV, or subcutaneous dosing, with transdermal opioids used only for chronic, stable pain, or dyspnea.[15] One document states that changing the route of analgesics administration, increase the doses, and liberally used as needed medications can improve pain management. In addition, the authors highlight that infusions are preferred in the ICU.[22] Finally, one document mentions that medications can be given subcutaneously; this has the advantage of reducing invasive treatment in patients without venous access, as long as the drugs are well tolerated.[17]

Management of Side Effects Related to Medication. Only two papers report on the management of side effects from analgesic medications.[15,16] They both recommend the initiation of a bowel stimulant and stool softener concomitantly with opioids, increasing the dose as opioid doses are escalated. For the relief of opioid-induced nausea, one paper recommends the use of antiemetics (e.g., prochlorperazine or metoclopramide) and opioid rotation in persistent cases.[15] Finally, one of the documents noted that morphine is associated with a greater risk of histamine release, which causes urticaria, pruritus, and flushing, recommending that these effects be relieved with antihistaminic therapy.[16]

Principle of Double Effect. Eight documents mention the principle of the double effect.[14–20,22] Seven of them use it to justify the use of analgesia and sedation to control pain and suffering, even though these medications may hasten death.[15,16,18–22] However, one paper addresses the principle of double effect as unnecessary based on a systematic review that found that patients in deep sedation did not have a lower survival rate in hospice settings compared with those who were not sedated.[18] The Indian Society of Critical Care states that the clear intention of symptom control versus shortening of the dying process is revealed by the process of titration of the administered medication, suggesting that protocols for palliative therapy should be in place and documentation should be meticulous.[22]

Use of Neuromuscular Blockade During Withdrawal of Life Support. Four documents agree that the effect of neuromuscular blockade must not be present when withdrawing life support to allow critical care specialists to accurately assess patients' pain and suffering.[15–17,20] One publication mentions that initiating neuromuscular blockade while withdrawing life support is unethical. The authors wrote that if the time required for neuromuscular blockers to wear off is too long, withdrawing life support should be initiated while ensuring that the patient comfortably proceeds through the dying process. During such a scenario, they noted that the detection of patient discomfort is difficult to assess.[16] Regarding this situation, another document states that intensivists who do not wait for the effects of neuromuscular blockade to wear off must document the need for their use appropriately.[20]

Palliative Care Consultation. Three documents agree that a palliative care consultation should be carried out when the limits of the critical care specialist's knowledge and skills have been reached[15,16,20] Such consultations can determine whether the patient has received all elements of traditional palliative care.[15] Other situations in which palliative care may intercede include inexperience with specific adjuncts for pain at end-of-life, assisting with severe psychologic issues with the family, or continuing care if the patient is discharged from the ICU or death is protracted.[20] Finally, one guideline suggests that routine palliative care consultations may improve end-oflife communication in the ICU.[16]