Pain at the End of Life

Paul L. DeSandre, DO
Chief, Palliative and Supportive Care, Grady Memorial Hospital
Fellowship Director, Hospice and Palliative Medicine
Assistant Professor, Emergency Medicine
Emory University School of Medicine
Atlanta, GA

Eashwar B. Chandrasekaran, MD MSc
Assistant Professor of Clinical Emergency Medicine
Palliative Care Services
Department of Emergency Medicine
Indiana University School of Medicine
Indianapolis, IN



For patients with serious illness nearing the end of life, uncontrolled suffering is among their greatest fears. The cause of suffering can be complex and is often due not just to pain and other physiologic changes that reduce function and capacity, but also psychosocial distress and spiritual distress. As symptoms escalate, patients and their caregivers may feel they have little choice other than to seek help in a hospital setting; even patients receiving home hospice services may come or be brought to the hospital in search of relief. While acute care clinicians recognize the need for effective and efficient interventions for these patients, they may feel ill-equipped to provide appropriately aggressive treatment to alleviate pain at the end of life. (Smith 2009, Grudzen 2012) A widely held fear limiting appropriate titration of medications–particularly opioids–near the end of life is that they may hasten death. These concerns are generally unfounded if proper dose selection and clinical assessment are appropriate. (Thorns 2000, Morita 2001) Of far greater concern to the individual is inadequate analgesia leading to persistent pain and suffering near the end of life, particularly in cancer patients. (Portenoy 2006, Steindal 2011) The Institute of Medicine has reinforced the need to improve primary pain management skills for all clinicians caring for patients nearing the end of life. (IOM 2014)

The emergency department is the most accessible point of entry for many individuals with chronic, life-threatening diseases. Emergency clinicians are therefore often confronted with the need to manage symptoms and other concerns that arise at the end of life. Furthermore, patients, and their caregivers, may benefit from initiating goals-of-care discussions in the ED, and these discussions may have a significant impact on both the in-hospital trajectory and prognosis. As in many other clinical domains, emergency physicians have expanded their scope of practice to better serve this growing patient population by augmenting their palliative care skills, seeking specialized training through fellowships and other focused programs, and by partnering with local palliative care specialists. The patient- and family-oriented palliative care priorities apply to patients at every stage of life; implementation of ED-based palliative care initiatives therefore broadly improves the quality of emergency care. (Lamba 2014)

There is much room to improve the care of patients approaching death. Family reports suggest that half of patients who die in the hospital have moderate to severe pain more than 50% of the time, (SUPPORT 1995) and cancer patient caregivers report 61% prevalence of “very distressing” pain. (Constantini 2009) Risk factors for the undertreatment of pain at the end of life include a history of substance abuse, older age, minority status, and physician discordance in the perception of pain. (Cleeland 1994, Higginson 2012)



The first and most important consideration for successful pain management is an adequate assessment, including an understanding of the cause of pain, prior to selecting interventions. There are two broad classes of pain that are readily distinguished and have markedly different treatment requirements: Nociceptive pain is caused by immediate tissue threat or injury, such as in cancer progression into previously healthy tissue. Somatic nociceptive fibers are highly myelinated, causing rapid transmission of pain impulses with discrete localization of pain. Visceral nociceptive fibers are less myelinated and transmit more diffuse, poorly localized symptoms such as cramping. In visceral abdominal cancers, both systems may be activated–visceral pain from, for example, bowel involvement, and somatic pain from invasion of the parietal pleura. Neuropathic pain, on the other hand, may develop from neurotoxic effects of chemotherapeutics, metabolic microvascular conditions such as diabetes, infectious diseases such as HIV, or direct neoplastic invasion of neural tissue. It may be felt as a numb or tingling sensation, hypersensitivity, or abnormally severe and altered response to minor stimulation known as allodynia.  

Once the likely cause of pain is determined, its severity should be repeatedly assessed to gauge the adequacy of analgesia and facilitate safe titration to the relief of pain or the development of intolerable adverse effects that limit further dose escalation. A patient’s ability to quantify their distress determines the most appropriate method of assessment; no pain scale has been shown to be superior to others, so the tool most acceptable or applicable to the patient should be used. For some patients able to communicate, a subjective response to adequate or inadequate control may be sufficient. For others, an objective assessment may be more helpful. Commonly used and equally valid pain scales for communicative patients include the Visual Analogue Scale (VAS), the Numerical Rating Scale (NRS), the Wong-Baker FACES scale, and the Verbal Rating Scale (VRS).  All of these use the anchors of “no pain” to “worst pain” to help quantify changes. (Jensen 2003) Patients may differ in their response to one scale or another, some being more comfortable with the 0 to 10 NRS, whereas others may only be able to point to one of the FACES expressions, and still others may be limited to only “mild, moderate, or severe.”  Whichever is most useful should continue to be used to optimize longitudinal assessment of a patient’s pain trajectory. For patients with severe dementia, non-verbal pain scales may be more useful, such as PAINAD. (Warden 2003) For patients in critical condition, either conscious or unconscious, the Critical Care Pain Observation Tool (CPOT) is well-validated and widely used. (Gelinas 2006) At the end of life, CPOT can be particularly helpful in looking for signs of distress related to pain, and for determining adequacy of response to interventions.

Once an appropriate assessment tool is chosen for the patient, whether numeric or descriptive, the same scale should be used among clinicians to assess analgesic effect based on the patient’s response to treatment and the pharmacokinetics of the agent(s) used. Reassessments focus on both analgesic efficacy and adverse effects. The goal is the patient’s report that they are comfortable and do not wish to have more pain medication. For patients only able to use a verbal scale, “better” or “enough” may be sufficient to communicate their level of response. For critical patients, alleviation of non-verbal signs of distress, for example, a CPOT score less than 3, suggests an adequate response.



Environmental influences such as excessive light, noise, strong smells, or direct physical manipulation may undermine efforts to provide comfort to the patient and should be controlled to the degree possible. Placing the patient in a quiet room with familiar caregivers and family, turning off or adjusting device alarm settings, and turning monitors to “comfort mode” allow attention to be focused on the patient rather than their evolving physiological decline.

Goals specific to palliation should be established and regularly reassessed. For example, if maintaining consciousness is an equal priority to adequate pain control, this requires a different approach than when maximal pain control is paramount.

A pain crisis occurs when the patient develops severe, uncontrolled pain that causes the patient or family severe distress. A pain crisis requires immediate and rapid medication titration, generally using multimodal analgesia, to alleviate symptoms. (Moryl 2008)


Route of administration

For patients in the hospital with severe pain, The oral or rectal routes are not optimal, as delayed onset limits effective titration. (DeSandre 2009) For patients in an acute care setting, intravenous (IV) administration is most effective. If IV access is unavailable, the subcutaneous (SQ) route is a reasonable alternative, and intramuscular administration is discouraged as it is more painful  and provides no analgesic or pharmacokinetic advantage over SQ delivery. If IV or SQ administration is not desired, orally administered liquid morphine is commonly used and effective; peak effect is reached in approximately 60 minutes. The rectal route is feasible and has similar absorption characteristics to oral administration, but is often impractical and uncomfortable. Fentanyl is available in various formulations, including a transmucosal lozenge (lollipop) and dissolving buccal tablets, which are indicated for breakthrough pain in opioid tolerant individuals. All formulations are associated with a specific risk evaluation and mitigation strategy ( that outlines specific criteria for use. Transdermal fentanyl patches are commonly used for long term management of chronic pain, but due to their delayed onset of action they are not appropriate for the treatment of breakthrough pain. Fentanyl and ketamine can be delivered via atomizer to the nares, and fentanyl can be inhaled via conventional or breath-actuated nebulizer for both pain and air hunger at the end of life. If fentanyl is not available, lesser evidence supports the use of nebulized morphine in this context. (Pavis 2002, Fitzgibbon 2003, Zeppetella 2000)

Evidence supports direct opioid placement into exposed ulcers such as decubiti, which are known to have mu receptors and may respond fairly rapidly. One approach to topical application of opioids for malignant ulcers or painful decubiti is to create a gel dressing of 10 mg IV morphine solution in 8 grams of hydrogel, and apply to the ulcer in a thin layer one to three times per day. (Graham 2013)



Opioids are considered first line therapy for patients with acute severe pain.  The most commonly used parenteral opioids in the acute care setting include morphine, hydromorphone, and fentanyl. The analgesic efficacy of all three drugs is similar when administered in equianalgesic doses; efficacy correlates more with dose and interval than drug choice. For terminally ill patients in pain, analgesic selection is based on palliative goals, altered physiology at the end of life, and, most importantly, the patient’s prior experience with efficacy and adverse effects of the individual agents.

Morphine and hydromorphone are often used parenterally at the end of life. Both achieve peak effect in 5-10 minutes when given IV, or 20-30 minutes when given SQ. (Table 1) We recommend reassessment of efficacy and adverse effects every 15 minutes when morphine or hydromorphone is given as an intravenous bolus. For patients with renal or hepatic impairment, lower initial dosing is recommended with extended titration intervals. Morphine is particularly prone to accumulating toxic metabolites in patients with severe renal failure, and alternative agents are favored in this group.

Table 1: Time to Maximal Concentration of Specific Opioids

see for more pharmacokinetics info

Opioid Route of Delivery Time to Peak Effect
Morphine/Hydromorphone Oral 1h
Intravenous 15m
Subcutaneous 30m
Fentanyl Intravenous 5m

For acute care providers, fentanyl has several advantages for managing pain at the end of life. Intravenous fentanyl reaches peak effect in less than five minutes, allowing rapid titration; has minimal renal clearance, making it well suited for patients with kidney failure; and is less likely than commonly used alternatives to cause or worsen hypotension. However, IV fentanyl has a duration of action of only 30-60 minutes, and so is best suited to rapid titration followed by a continuous infusion. As mentioned, fentanyl may also be administered transdermally, intranasally, and by inhalation, as appropriate.

Patients may be considered opioid tolerant if in the week prior to presentation they have consistently been exposed to at least 60 mg oral morphine per day (equivalent to 25 mcg transdermal fentanyl/hour, 40 mg oral oxycodone/day, or 8 mg oral hydromorphone/day). All other patients should be considered opioid naïve.  For patients near the end of life in a pain crisis who are opioid naive, doses as low as 4 mg IV morphine or 0.5 mg IV hydromorphone with reassessment every 15 minutes may be sufficient. Using rapidly-titrated small doses is more likely to limit side effects, to which the dying patient may be more susceptible, while optimizing efficacy. If, after repeating the dose, pain is still severe (>7/10 NRS) and the patient remains alert and without distressing side effects, the dose may be doubled once and then repeated at this dose, and this rapid-titration process is continued until adequate analgesia is achieved, the patient is no longer alert, or adverse effects such as nausea or delirium supervene. In the majority of patients, adequate analgesia without severe side effects should be achievable within 60 mins. (Harris 2003) If an acceptable effect cannot be achieved with the chosen opioid, switching to an alternative opioid or adding a non-opioid such as ketamine may produce better results.  

Once the acute pain crisis has been managed and the patient is comfortable, a continuous analgesic infusion may initiated using a 4-hour reassessment of the total dose required to achieve adequate analgesia without significant adverse effects, and dividing that dose into an hourly rate. For example, if 20 mg morphine is required over 4 hours, then the hourly rate would be 5 mg/hour. If the patient is able to use a patient-controlled analgesic (PCA) device, it may be programmed to start the demand dose at 1-3 mg morphine with a lockout of 10-15 minutes and no continuous rate initially if opioid naive. In most patients, self-administering an opioid using these parameters would cause somnolence (and therefore an inability to push the demand button) before overdose complications occur. Patients with sleep apnea or significant lung disease should be started at lower doses with longer lockout intervals, and all patients recently started on continuous or PCA opioid infusions require continued close monitoring for efficacy and adverse effects. (Caraceni 2012) PCAs should only be initiated and titrated by those trained in their use.

Patients with end-of-life pain who are on high dose opioids at home and present with severe pain present a particular challenge. For these opioid-tolerant patients, a four-step fentanyl titration can provide rapid control based on prior opioid dosing. (Table 2 and Figure 1) Such assessments and calculations are complicated, and may require the assistance of a physician or pharmacist with expertise in pain management.

Step 1: Initial bolus of fentanyl IV, based on 10% of prior 24 hour IV morphine equivalent use. For example, if the patient has taken a total equivalent of 180 mg of oral morphine in the past 24 hours, the IV equivalent would be 1/3 the oral amount or 60 mg IV morphine. 10% of this amount is 6 mg IV morphine, which is equivalent to 60 mcg IV fentanyl. If daily morphine equivalence for an opioid tolerant patient cannot be calculated, fentanyl titration may reasonably be initiated at 25-50 mcg.

Step 2: Reassess in 5 minutes. If no severe side effects or excessive somnolence, and patient is still distressed or requesting further analgesia (pain > 4/10), repeat initial dose.

Step 3: Reassess in 5 minutes. If no severe side effects or excessive somnolence, and patient is still distressed or requesting further analgesia (pain > 4/10), double the original fentanyl dose.

Step 4: Reassess in 5 minutes. If no severe side effects or excessive somnolence, and patient is still distressed or requesting further analgesia (pain > 4/10), repeat step 3 dose.

Many patients will achieve adequate analgesia using this 4-step protocol in 30 minutes. (Soares 2003) If ineffective, consider adding a non-opioid adjuvant such as ketamine to augment analgesia. It is recommended to discuss the use of ketamine as an effective analgesic with the patient or family prior to administration, as it may contribute to terminal delirium, and may require combined dosing with benzodiazepines.


Table 2: Equianalgesic Dosing of Specific Opioids
Opioid Route of Delivery Equianalgesic Dose
Morphine PO 30 mg
IV 10 mg
Oxycodone PO 20 mg
Hydromorphone PO 7.5 mg
IV 1.5 mg
Fentanyl IV 100 mcg

see online calculators for more conversion information:
NYC Health MME Calculator
AMDG Calculator
Practical Pain Management Calculator


When rapidly titrating opioids to treat pain at the end of life, ideally symptoms are relieved with a single opioid or opioid alternative. Co-titration of other sedating medications, such as benzodiazepines, complicates reassessment and increases the risk of serious adverse effects. A cautious approach centered on small, frequent doses and frequent reassessment minimizes the likelihood of dangerous adverse effects, but individual response characteristics may be unpredictable. If the patient is not in the final stages of dying or if the goals for critical intervention remain unclear and a dangerous opioid overdose has been determined, then intervention may be necessary. Hypoventilation causing decreased level of consciousness, agitation, or distress from hypoxia that is thought to be caused by overmedication requires prompt but judicious intervention, especially in opioid-tolerant patients. The opioid antagonist naloxone reliably reverses dangerous hypoventilation but can precipitate withdrawal and associated pain and distress, which may be difficult to overcome. Unless hypoventilation is thought to be causing immediately dangerous consequences (bradycardia or malignant arrhythmia), a gentle approach to reversal, starting with aliquots of 0.04 mg IV, will avert overdose harms without precipitating a distressing (and sometimes dangerous) withdrawal syndrome. (Figure 2)

The goal is not a fully awake and alert patient, rather a patient who is breathing adequately, maintaining their airway, and at least partially responsive to voice or gentle painful stimuli. Once an adequate reversal response is achieved, the patient should be carefully observed for at least 90 minutes, to verify that dangerous overdose effects do not recur as the naloxone is metabolized. If repeat dosing of naloxone is required, the patient should be moved to a clinical setting capable of continuous cardiorespiratory monitoring; a continuous infusion of very-low-dose naloxone may be required in these patients. (Boyer 2012)


Adjuvant analgesics

A variety of non-opioid analgesics may augment (or replace) the efficacy of opioids. Non-steroidal antiinflammatory drugs (NSAIDs) and acetaminophen are useful supplemental analgesics. NSAIDs are particularly effective in visceral inflammatory pain, but require a careful harm:benefit consideration in patients with renal failure or gastrointestinal bleeding. Patients who are suffering from chronic life-threatening illness may also benefit from the use of adjuvant analgesics such as gabapentin or a tricyclic antidepressant for the treatment of neuropathic pain; these agents are poorly studied as therapies for acute pain because of their limited parenteral availability.

Subdissociative-dose ketamine can provide remarkably effective analgesia in palliative care patients, especially opioid tolerant patients whose pain may be difficult to control with additional opioids. The dose is 0.1-0.3 mg/kg IV as a loading dose given over 15-30 minutes, followed by an intravenous infusion at 0.1-0.3  mg/kg/hour. The goal is a calm and comfortable patient who doesn’t have bothersome or distressing psychoperceptual effects due to the ketamine; this can usually be achieved with frequent reassessments and titration. Subdissociative-dose ketamine does not require cardiorespiratory monitoring, though effective ketamine analgesia may lead to important adverse effects from co-delivered opioids. (Shlamovitz 2013, Winegarden 2016, Prommer 2012)

There are a variety of palliative procedures that may offer significant relief for certain types of terminal pain; for example, bony pain may be best addressed through palliative radiation and visceral pain of cholangiocarcinoma and pancreatic cancer may be very effectively treated by a celiac plexus block. The efficacy of these specialized procedures underscores the benefit of early involvement of pain medicine and palliative care in the management of patients with refractory pain at the end of life.


Palliative Sedation

Refractory pain is pain that cannot be controlled, usually because therapeutic efforts are associated with intolerable side effects. Refractory pain presumes that other symptoms which could be labeled as pain, such as dyspnea, agitation, delirium, and anxiety are properly addressed. (Cherny 1994) If it is determined that all appropriate efforts to alleviate the patient’s pain have failed and future efforts would be ineffective, it is appropriate to consider using sedating medications to alleviate suffering; this is known as palliative sedation. Following an informed consent process including the patient or appropriate proxy, sedating medications such as benzodiazepines are administered to achieve a condition where conscious suffering is no longer evident. The level of sedation is proportionate to the level of distress, and this approach is often a planned temporary intervention to determine if the patient’s suffering can be improved through the procedure. The intent of palliative sedation must always be solely to alleviate suffering. Given that patients being considered for palliative sedation have refractory symptoms in the late stages of a terminal illness, the use of palliative sedation would not be expected to alter the timing or mechanism of dying. (AAHPM 2014)


Patients entering an acute care setting in pain at the end of life often are experiencing suffering that may be multifactorial, including physiologic/nociceptive distress, psychosocial distress, and spiritual distress. A careful evaluation of the etiology and characteristics of pain using frequent reassessments of symptoms is the foundation for rapid control. Opioids should be considered first-line therapy for severe pain at the end of life. With careful consideration of the patient’s physiological challenges and the characteristics of available drugs, successful pain control can usually be achieved rapidly and safely. Optimal pain control often requires a multimodal approach with non-opioid adjuvants. When appropriately dosed and targeted therapies have failed to alleviate suffering from refractory pain in the dying patient, palliative sedation may be considered.


The editors thank Ashley Shreves, MD and Michael Turchiano, MD for their thoughtful reviews of the manuscript.

The authors report no relevant conflicts of interest.



American Academy of Hospice and Palliative Medicine. Statement on palliative sedation. 2014; (accessed February 18, 2017).

Boyer EW. Management of opioid analgesic overdose. NEJM 2012; Jul 12;367(2):146-155.

Caraceni A, Hanks G, Kaasa S, et al. Use of opioid analgesics in the treatment of cancer pain: evidence-based recommendations from the EAPC. The Lancet Oncology 2012; 13.2: e58-e68.

Cherny NI and Portenoy RK. Sedation in the management of refractory symptoms: guidelines for evaluation and treatment. J Palliat Care. 1994; 10(2):31.

Cleeland CS, Gonin R, Hatfield AK, et al. Pain and its treatment in outpatients with metastatic cancer. NEJM 1994; 330(9):592-596.

Constantini M, Ripamonti C, Beccaro M, et al. Prevalence, distress, management, and relief of pain during the last 3 months of cancer patients’ life. Results of an Italian mortality follow-back survey. Ann Oncol. 2009;20(4):729-735.

Desandre PL, Quest TE. Management of cancer-related pain. Emerg Med Clin Noth Am. 2008 May; 27(20:179-194.

Fitzgibbon D, Morgan D, Dockter D, et al. : Initial pharmacokinetic, safety and efficacy evaluation of nasal morphine gluconate for breakthrough pain in cancer patients. Pain 2003, 106: 309– 315. 46.

Gelinas C, Fillion L, Puntillo KA, et al. Validation of the critical-care pain observation tool in adult patients. Am J Crit Care 2006;15(4):420-427

Graham T, Grocott P, Probst S, et al. How are topical opioids used to manage painful cutaneous lesions in palliative care? A critical review. Pain 2013;154(10): 1920-1928.

Greco MT, Roberto A, Corli O, et al. Quality of cancer pain management: an update of a systematic review of undertreatment of patients with cancer. J Clin Oncol 2014: 32(36):4149-4154.

Grudzen CR, Richardson LD, Hopper SS, et al. Does palliative care have a future in the emergency department? Discussions with attending emergency physicians. J Pain Symptom Manage. 2012;43(1):1-9.

Harris JT, Suresh Kumar K, Rajaqopal MR. Intravenous morphine for rapid control of cancer pain. Palliative Medicine 2003; Apr;17(3):248-56.

Higginson IJ, Gao W. Opioid prescribing for cancer pain during the last 3 months of life: associated factors and 9-year trends in a nationwide United Kingdom cohort study. J Clin Oncol 2012;30(35):4373-2379.

IOM (Institute of Medicine). 2014. Dying in America: Improving quality and honoring individual preferences near the end of life. Washington, DC: The National Academies Press.

Jensen MP. The validity and reliability of pain measures in adults with cancer. J Pain 2003;4:2-21.

Lamba S, DeSandre PL, Todd KH, et al. Integration of palliative care into emergency medicine: the Improving Palliative Care in Emergency Medicine (IPAL-EM) collaboration. J Emerg Med. 2014;46(2):264–70.

Morita T, Tsunoda J, Inoue S, et al. Effects of high dose opioids and sedatives on survival in terminally ill cancer patients. J Pain Symptom Manage 2001;21:282-289

Moryl N, Coyle N, Foley KM. Managing an acute pain crisis in a patient with advanced cancer, “this is as much of a crisis as a code.” JAMA 2008. 26;299(12):1457-1467.

Pavis H, Wilcock A, Edgecombe J, et al. Pilot study of nasal morphine-chitosan for the relief of breakthrough pain in patients with cancer. J Pain Symptom Manage 2002, 24: 598– 602.

Portenoy RK, Sibirceva U, Smout R, et al. Opioid use and survival at the end of life: a survey of a hospice population. J Pain Symptom Manage;32(6): 532-540.

Prommer EE. Ketamine for pain: an update of uses in palliative care. J Palliat Med. 2012 Apr;15(4):474-83.

Shlamovitz GZ, Elsayem A, Todd KH. Ketamine for palliative sedation in the emergency department. J Emerg Med. 2013 Feb;44(2):355-7.

Smith AK, Fisher J, Schonberg, MA, et al. Am I doing the right thing? Provider perspectives on improving palliative care in the emergency department. Ann Emerg Med. 2009;54(1):86-93.

Soares LG, Martins M, Uchoa R. Intravenous fentanyl for cancer pain: a “fast titration” protocol for the emergency room. Journal of Pain and Symptom Management 2003 Sep;26(3):876-881.

Steindal S, Bredal L, Sørbye L, et al. Pain control at the end of life: a comparative study of hospitalized cancer and noncancer patients. Scandinavian Journal of Caring Sciences [serial online]. December 2011;25(4):771-779.

The SUPPORT Principal Investigators. A controlled trial to improve care for seriously ill hospitalized patients: the study to understand prognoses and preferences for outcomes and risks of treatments (SUPPORT). JAMA 1995;274:1591-1598

Thorns A, Sykes N. Opioids use in the last week of life and implications for end-of-life decision-making. Lancet 2000;356(9227):398-399

Warden V, Hurley AC, Volicer L. Development and psychometric evaluation of the Pain Assessment in Advanced Dementia (PAINAD) scale. J Am Med Dir Assoc. 2003 Jan-Feb;4(1):9-15.

Winegarden J, Carr DB, Bradshaw YS. Intravenous Ketamine for Rapid Opioid Dose Reduction, Reversal of Opioid-Induced Neurotoxicity, and Pain Control in Terminal Care: Case Report and Literature Review. Pain Med. 2016 Apr;17(4):644-9.

Zeppetella G: An assessment of the safety, efficacy, and acceptability of intranasal fentanyl citrate in the management of cancer-related breakthrough pain: a pilot study. J Pain Symptom Manage 2000, 20: 253– 258.