Safe and Effective Management of Pain in People with CKD : Clinical Journal of the American Society of Nephrology

Journal Logo


Safe and Effective Management of Pain in People with CKD

Davison, Sara N.; Rathwell, Sarah

Author Information
CJASN 14(11):p 1551-1553, November 2019. | DOI: 10.2215/CJN.11140919
  • Free

For patients with CKD, pain is common, often severe, and associated with poor health-related quality of life (QOL), increased psychosocial distress, and increased health services utilization. Analgesics, including opioids, play an important role in pain management and may be required to achieve adequate pain relief.

The United States has seen an unprecedented increase in the prescribing and dose of opioids. This has been associated with an increase in opioid-related hospitalizations and deaths, higher rates of addiction and social dysfunction, and cognitive impairment and falls in the elderly. The rates of harm correlate directly with dose. Given the high burden of pain, the increase in opioid use has also been seen in dialysis patients. A recent study across the United States showed that 64% of 153,758 patients on dialysis in 2010 received opioids, a third of whom (23%) received an opioid prescription for ≥90 days (1).

The pharmacologic management of pain is challenging for patients with CKD, in large part because of the altered pharmacokinetics and pharmacodynamics of analgesics with reduced kidney function. It is not surprising, therefore, that the current prescribing patterns of opioids for patients on dialysis are also associated with dose-dependent increase in mortality, altered mental status, falls, and fractures (1,2). In this issue, Novick et al. (3) present important data from a community-based CKD cohort showing that people who received a prescription for opioids were at a higher risk of death and hospitalization compared with propensity score–matched patients who received nonsteroidal anti-inflammatory drugs (NSAIDs).

Neither opioids nor NSAIDs are benign treatments. Like all treatment decisions, potential risks must be balanced carefully with potential benefits. The literature lacks clear evidence that analgesics relieve chronic pain in the long term, and there are no studies in CKD that look at clinical outcomes important to patients, such as QOL or functionality with long-term analgesic use. Therefore, we need to carefully consider the potential risks. In this editorial we will discuss the results from this study (3) within the context of recommendations for the pharmacologic management of pain in CKD, which appear underutilized yet are critical to mitigate opioid use and dose.

Novick et al. (3) showed that opioids as currently prescribed were associated with higher risk of death and hospitalization than NSAIDs in patients with mild kidney failure (eGFR of 80 ml/min per 1.73 m2). The risks of death and hospitalization increased for all patients prescribed both opioids and NSAIDS as kidney function declined. However, for patients with stage G4 CKD, the risks associated with opioid prescription were only clearly higher compared with those prescribed NSAIDs when considering high-dose prescriptions (≥60 morphine milligram equivalents [MME]). In G4 CKD, the hazard of death in patients prescribed NSAIDs appears to approach that of patients receiving the lower-dose opioid (1–59 MME). Results were similar in the sensitivity analyses. A limitation is the low event rate for G4 CKD in this cohort.

This pattern of risk was even more pronounced in terms of hospitalization. At higher eGFRs, opioid prescription, regardless of dose, was associated with greater risk compared with NSAIDs. However, with more advanced stages of CKD, the association of opioid use compared with NSAID use on the risk of hospitalization diminished, perhaps because of the greater risk of NSAIDS with more severe kidney failure. Only high-dose opioid prescription was associated with a higher hazard of hospitalization at an eGFR of 40 ml/min per 1.73 m2, and the hazard ratios for both opioid doses appeared to merge with the NSAID group as eGFR levels approached 20 ml/min per 1.73 m2. This trend is mirrored throughout multiple sensitivity analyses.

The association of opioid prescription with risk of death and hospitalization is unlikely to be entirely causal and almost certainly reflects some bias with opioid prescription being a marker of more severe illness. The authors attempted to reduce this bias by propensity score matching. Unfortunately, they were not able to match for severity of comorbid diseases, a clear confounder. In addition, propensity score matching does not always decrease the variability between groups for key confounders (4); in fact, it can increase bias, known as the “propensity score matching paradox” (4,5). There are a few issues in this study with the propensity matching that may have influenced results. First, sample size and precision can be decreased when covariates used to build the scores are not associated with the outcome variable (5,6). It was not specified that the independent variables in the propensity model were associated with risk of death and hospitalization in this data set. Additionally, their use of a fixed caliper width performs worse in terms of optimizing sample size and precision than one calculated from the SD of the logit of the propensity score (7). This may have pruned patients from the original cohort, increasing the variability of the subsequent estimates. Furthermore, the authors did not stratify Cox regression on the matched pairs, a better model for propensity-matched observations than the unstratified version (5). Although it is unclear whether these statistical issues had any significant effect on the overall findings, they should be taken into consideration for future research in defining risks of analgesic therapy. Further, cause of hospitalization and death were not ascertained, and increased risks associated with opioids may have reflected issues related directly to the pain syndrome itself, such as wound care or amputation for severe peripheral vascular disease, in addition to issues associated with opioid use such as mental illness, substance abuse, and toxicity.

Perhaps the most important consideration is that these findings are in the context of analgesic prescribing that appears inconsistent with current recommendations. The pharmacologic management of pain in patients with CKD, and the role analgesics, including opioids, play in the overall approach to pain treatment, has been extensively reviewed recently (8). These guidelines highlight strategies critical to mitigate the use of opioids. First, it is important that patients and care providers have realistic expectations. Complete relief of pain is often not possible, especially with chronic pain; the goal of therapy is to relive the pain to a tolerable level, allowing for acceptable function and QOL. Furthermore, analgesics should not be the sole treatment as in isolation they are unlikely to succeed. Rather, they should augment nonpharmacologic therapies such as exercise, stretching, and massage, or behavioral therapies such as cognitive behavioral therapy and relaxation techniques, when these other therapies have been insufficient to meet patient-specific treatment goals. In addition, pain can be intensified by coexisting depression, anxiety, and sleep disorders. Assessing and treating these conditions may ameliorate perceived pain.

However, many patients will require analgesics. Current recommendations advocate for a cautious stepwise approach to analgesics that considers the nature of pain and the judicious use of adjuvant and nonopioid analgesics (8). Many pains experienced by patients with CKD have a neuropathic component that is poorly responsive to both NSAIDs and opioids, often requiring high doses that are associated with unacceptable toxicity. For pain that has a neuropathic component, it is important to target this first with adjuvant therapy such as gabapentin, to prevent inappropriate opioid use. For neuropathic pain inadequately treated with an adjuvant or for nociceptive pain, acetaminophen should be trialed before considering opioids. Only if pain remains inadequately treated should an opioid be added to acetaminophen with or without the adjuvant. These strategies do not appear to be optimized. In almost 154,000 patients on dialysis, only 3.2% were prescribed solely nonopioid analgesics; 11.0% were prescribed an opioid and analgesic, and 51.4% were prescribed an opioid only (1). If an opioid is considered necessary, careful selection is required. Given the minimal changes in pharmacokinetics in kidney failure, hydromorphone, fentanyl, methadone. and buprenorphine are considered the safest. However, the opioids most commonly prescribed for ≥90 days were hydrocodone, oxycodone, propoxyphene, and tramadol (1), none of which are recommended for use in advanced CKD because of their high risks of adverse events (8). Only 1.9% of patients received a chronic prescription for an opioid that is considered safer for use in these patients. These factors likely add substantially to the risks associated with opioid prescribing. In this study, initial prescriptions for the more appropriate opioids buprenorphine and methadone were not included, as they were assumed to be prescribed for addiction.

Dosing needs to be titrated cautiously given the narrow therapeutic window for many patients. Even in the general population, there is no standard “optimal” dose. The optimal dose for any given patient is that dose that treats the pain adequately without unacceptable adverse effects. Patients using opioids will always require careful ongoing clinical assessment to balance the risks and benefits. Lastly, before starting an opioid, the risk for aberrant opioid-related behavior should be assessed using a tool such as the Screener and Opioid Assessment for Patients with Pain-Revised (9). Those categorized as high-risk of abusive drug-related behavior would benefit from referral to a pain specialist for opioid prescribing.

Findings from this study should not be used to avoid using opioids to treat pain when necessary in this population, but rather to work toward more effective and safe practices that includes appropriate use of acetaminophen, adjuvants, and topical NSAIDs before considering opioids. Occasionally, systemic NSAIDs may provide greater pain control and potentially fewer side effects than other medications, especially opioids, particularly among patients with mild reductions in eGFR. The risks of NSAIDs include progression of CKD, gastrointestinal bleeding, and cardiovascular morbidity and mortality. The elderly are at risk for NSAID-associated agitation, depression, anxiety, paranoia, delirium, and hallucinations. NSAIDs are best reserved for specific indications of acute rather than chronic pain, limiting their use to the lowest effective dose and the shortest duration.

There is a clear need for high-quality prospective trials that evaluate the effectiveness of these best-evidence recommendations on the outcomes important to patients, which should include overall symptom burden, QOL, and function, as well as survival and hospitalization.


Dr. Davison and Ms. Rathwell have nothing to disclose.

Published online ahead of print. Publication date available at

See related article, “Associations of Opioid Prescriptions with Death and Hospitalization across the Spectrum of Estimated GFR,” on pages .


1. Kimmel PL, Fwu CW, Abbott KC, Eggers AW, Kline PP, Eggers PW: Opioid prescription, morbidity, and mortality in United States dialysis patients. J Am Soc Nephrol 28: 3658–3670, 201728935654
2. Ishida JH, McCulloch CE, Steinman MA, Grimes BA, Johansen KL: Opioid analgesics and adverse outcomes among hemodialysis patients. Clin J Am Soc Nephrol 13: 746–753, 201829674340
3. Novick T, Surapaneni A, Shin JI, Alexander G, Inker L, Wright E, Chang A, Grams M: Associations of opioid prescriptions with death and hospitalizations across the spectrum of estimated GFR. Clin J Am Soc Nephrol 14: 1581–1589, 2019
4. King G, Nielsen R: Why propensity scores should not be used for matching. Polit Anal 1–20, 2019
5. Austin PC: A critical appraisal of propensity-score matching in the medical literature between 1996 and 2003. Stat Med 27: 2037–2049, 200818038446
6. Austin PC, Grootendorst P, Anderson GM: A comparison of the ability of different propensity score models to balance measured variables between treated and untreated subjects: A Monte Carlo study. Stat Med 26: 734–753, 200716708349
7. Austin PC: Some methods of propensity-score matching had superior performance to others: Results of an empirical investigation and Monte Carlo simulations. Biom J 51: 171–184, 200919197955
8. Davison SN: Clinical pharmacology considerations in pain management in patients with advanced kidney failure. Clin J Am Soc Nephrol 14: 917–931, 201930833302
9. Varney SM, Perez CA, Araña AA, Carey KR, Ganem VJ, Zarzabal LA, Ramos RG, Bebarta VS: Detecting aberrant opioid behavior in the emergency department: A prospective study using the screener and Opioid Assessment for Patients with Pain-Revised (SOAPP®-R), Current Opioid Misuse Measure (COMM)™, and provider gestalt. Intern Emerg Med 13: 1239–1247, 201829502329

chronic kidney disease; pain; chronic renal insufficiency

Copyright © 2019 by the American Society of Nephrology