A Narrative Review of Pain in Pediatric Oncology: The Opioid Option

Abstract

Opioid therapy is the mainstay for managing pain in pediatric oncology. This narrative review describes the current literature regarding opioids for pediatric cancer pain. The review explores the multifaceted landscape of opioid utilization in this population, including the role of opioids in certain clinical circumstances, modalities of opioid delivery, unique opioids, outpatient and at-home pain management strategies, and other key concepts such as breakthrough pain. This review highlights the importance of individualized dosing and multimodal approaches to enhance efficacy and minimize adverse effects. Drawing from a wide range of evidence, this review offers insights to optimize pediatric oncology pain management.

FormalPara Key Points

Opioids are essential for pain management in patients with pediatric cancer, but for best outcomes, their usage needs individualized dosing and monitoring for adverse effects.

Combining opioids with other medications can enhance pain relief while minimizing side effects, especially in complex pain scenarios such as chronic neuropathic pain or acute treatment-related pain episodes.

1 Introduction

Opioid therapy remains the mainstay of pain management in children with cancer. The topics captured in this review are “pieces of the puzzle” reflecting the role of opioids in pain treatment for specific clinical circumstances in pediatric oncology, from general epidemiology of pain in pediatric cancer [1,2,3,4,5,6], to specific pain-related clinical circumstances such as hematopoietic stem cell transplant (HSCT) [7,8,9,10,11,12,13] and mucositis [14,15,16,17,18,19,20,21,22,23,24,25]. Special consideration was given to sections on opioids for pain during palliative care (PC) and end-of-life (EOL) [26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49] as well as for neuropathic pain (NP) [36, 50,51,52,53,54,55,56,57,58,59,60,61,62].

Modalities of opioid delivery are reviewed, including patient-controlled analgesia/nurse-controlled analgesia (PCA/NCA) [8, 18, 63,64,65,66,67], long-acting opioids [68,69,70], transdermal fentanyl [71,72,73], and morphine infusions [25, 74, 75], as are opioids with specific roles in pain management: methadone [42, 43, 59, 76,77,78,79,80,81,82], buprenorphine [27, 83], and nalbuphine [84].

A large portion of retrieved references addressed outpatient pain management; these are analyzed in sections based on clinical circumstances. Several articles focused on pain management at home, either as data collection for analgesia quality [85], PCA utilization at home [66], or attitudes and barriers [86].

Important concepts in pain management are explored: breakthrough pain [87, 88], opioid rotation [89], opioid-sparing interventions including ketamine infusions [90,91,92], lidocaine infusions [93], bisphosphonates [94], opioid-related side effects [95,96,97,98,99,100], and opioid pharmacogenetics [101]. The theme of opioid misuse/abuse in pediatric oncology patients and in survivors of childhood cancer is also captured [102,103,104,105,106,107].

The sources analyzed in this narrative review expand from the early 1980s, when Miser et al. first reported on morphine infusions in pediatric oncology [25, 74], to the most recent research. Most of the evidence regarding the role of opioids in pediatric cancer pain originated from high-resource countries; nevertheless, research from low- and middle-income countries (LMICs) contributed to understanding the place of opioids in LMICs settings, based on resource limitations and cultural variables [108,109,110,111].

2 Literature Search Methodology

Narrative reviews are a popular methodology for describing themes across a body of literature [112]. This narrative review was undertaken to explore the current intersections of opioids in pain management for pediatric populations in oncologic settings. A systematic search of literature was executed in July 2023 in PubMed/MEDLINE, Scopus, and Embase. The primary search concepts included pediatrics, a comprehensive list of opioids, and any cancer, filtered for English-only articles, without limits on publication date range. Specific pain-related search terms were not utilized, as this increased the search sensitivity and yielded about 83,000 results. Inclusion criteria were specific to pediatric patients and pain related to cancer or cancer treatment. Any opioids, alone or in combination, were included. Outcomes of interest focused primarily on pain intensity, duration, and relief. Review articles, gray literature, and case series or case reports with n ≤ 5 were excluded. The search strategy used for PubMed/MEDLINE is available in Appendix 1.

The search results were exported to EndNote (v. 20), deduplicated, and imported into Covidence, a web-based tool for facilitating the screening and extracting processes. Following further deduplication in Covidence, 9256 abstracts were reviewed. Covidence was used for a two-phase review, beginning with title and abstract review guided by the primary search concepts mentioned previously. Due to the large number of results, the researchers proceeded with having one reviewer per title/abstract. However, a 10% spot check was performed by two authors (D.L.A. and T.M.H.) evaluating both the excluded and the included title/abstract groups and 100% agreement was found. The second phase was a full-text review of 1125 manuscripts, each reviewed by two authors, before advancing to data extraction. Data extraction was completed for 153 articles included in this review. A second literature search was performed in March 2024, which identified four additional references. Overall, 157 records have been included in this multipart review. This process is illustrated in Fig. 1. The total of records included in this manuscript is 111; 46 articles about the topics of procedural pain and acute postoperative pain will be reviewed in a subsequent manuscript.

Fig. 1

PRISMA flow diagram illustrating literature search and selection process

3 General Epidemiology of Pain in Pediatric Oncology

The epidemiology of outpatient pain treatment in pediatrics was described based on over 30 million administrative claims in children 0–16 years old. Leading pain-related diagnostic categories were surgery, trauma, and orthopedic conditions. The prevalence of malignancy as a pain-related diagnosis was low, both in commercially insured (2.1%) and Medicaid-insured children (1.7%). Over 50% of children did not receive outpatient prescription pain medications. For patients prescribed opioids (morphine, hydrocodone, oxycodone, and oxymorphone), immediate-release opioids were most used; extended-release opioid use was rare [1].

In a retrospective study of 90 children evaluated by a pain service, mean (SD) age 11.4 ± 4.1, range 2–17 years, pain was cancer-related in 81.2%, with the most common diagnosis leukemia/lymphoma; 90% received opioids and 28.9% used dual opioids. Mean (SD) doses (mg/day) were reported for tramadol (129.0 ± 97.9, range 12–380), morphine (14.8 ± 11.3, range 1–52), and for transdermal fentanyl (mcg/h) (33.2 ± 21.6, range 12–75). The mean initial pain scores (PS) decreased from 5.2 ± 1.7 to 1.5 ± 0.7 (p < 0.001). No pain management complications were noted in 93% [4].

Opioid utilization during hospitalization has been explored for children with acute myeloid leukemia (AML) [2] and sarcoma [3]. In 1600 children newly diagnosed with AML (42 pediatric hospitals), the prevalence of exposure to analgesics overall, nonopioids, and opioids was 95.2%, 84.7%, and 77.7%, respectively. The proportion of opioid-exposed patients increased with age, without gender, race, or insurance-based differences. Rates of opioid use (opioid days/1000 inpatient days) appeared modestly higher for females, patients < 1 year or ≥ 10 years of age, and white versus Asian patients. The oral route was utilized less commonly (44.5%) than the intravenous (IV) route (68.8%). Opioid utilization was highest during induction I and intensification II, reflecting higher acuity at diagnosis, invasive diagnostic procedures, and treatment-related complications [2]. In a study of 188 patients newly diagnosed with sarcoma (rhabdomyosarcoma, 56.4%), 19.7% presented with pain (most commonly osteosarcoma), and 89.2% of them received analgesics; only ten patients were treated with morphine, for a median of 7 days (range 1–20) using standard dosing [3].

Two articles report on results from a national quality improvement (QI) program for pediatric oncology pain management [5, 6]. In one study, the change in providers’ knowledge and attitude was evaluated based on a sample of 224 patients (median age, 9 years) over 2265 treatment days. The educational interventions led to lower utilization of painful routes of drug administration and mixed opioid agonists–antagonists, increased knowledge of NP treatment, lower PS, and fewer episodes of severe pain. The reported IV morphine equivalent dose (MED) (mg/kg/day) were mean (SD), 0.65 (0.71), median (range) 0.38 (0.02–4.22) initially; MED did not change significantly postintervention, nor did the number of days on opioid [5]. In the second manuscript, the authors explored the application of the WHO analgesic ladder in a prospective analysis of 2265 treatment days in 224 patients (median age, 9 years, range 0.2–32.1), indicating that the most frequently administered nonopioid analgesics were dipyrone and acetaminophen in the World Health Organization (WHO) step 1, tramadol was almost the only opioid used in WHO step 2, and morphine was utilized on most treatment days as part of WHO step 3. Strong opioids were combined with a nonopioid analgesic on 41% of the treatment days. The mean IV MED were reported as 0.034 mg/kg/h and 10 mcg/kg/h for tramadol. The authors indicated a lack of evidence that a combination of an opioid with a nonopioid is more effective (based on PS) than opioid therapy alone in inpatient pediatric oncology pain treatment [6].

Advancing knowledge of opioid pharmacogenetics is expected to inform the development of personalized pain therapies and avoid treatment failure and/or adverse effects. Crescioli et al. hypothesized that response to opioid therapy in patients with pediatric cancer may vary based on pharmacogenetics. Their observational study indicated that certain single nucleotide polymorphisms could be markers of pain intensity and opioid requirements in patients with pediatric cancer [101].

4 Clinical Circumstances

4.1 Hematopoietic Stem Cell Transplant

In six cohort studies [7, 8, 10,11,12,13] and one qualitative case study [9], opioids used for pain management in children (n = 914) undergoing HSCT included morphine [8,9,10,11,12], fentanyl [10,11,12], and hydromorphone [8, 10, 12], primarily via PCA, for acute pain related to mucositis [8, 10, 12] or graft-versus-host disease [8].

While opioid dosing varied, relatively high doses were necessary for extended periods, particularly in young children. Notable age-related differences were observed in one study reporting PCA initiation around day +3, with average MED (mg/kg/day) as 21 and 16, over durations of 40 and 25 PCA-days, in children younger and older than 6 years, respectively [10]. Another study reported 2–22 days (mean 17.65) of opioid utilization, with MED (mg/kg/day) 0.09–1.88, which was noted to be 3–5.6 times higher than adult doses [9]. Dunbar et al. reported bolus doses of morphine (20 mcg/kg) and hydromorphone (3 mcg/kg), with 8-min lockout intervals, for a PCA duration averaging 19 days [8]. In a study of 230 patients with high-risk neuroblastoma, those treated with carboplatin/etoposide/melphalan (CEM) for autologous stem cell rescue required opioids for a significantly longer duration compared with those treated with busulfan/melphalan (BuMel), especially in the first 30 days after transplant (mean 13.77 versus 9.90 days, p < 0.0001) [13]. Similarly, another study of 578 patients with HSCT with high-risk neuroblastoma found that the duration of opioid use was significantly shorter in the BuMel group (median 9 days) versus the CEM (median 15 days) and tandem groups (median 23.5 days; p < 0.0001) [7]. These two studies highlight that the level of toxicity associated with the conditioning regimen is a key factor that can impact the duration of opioid use post-HSCT.

Opioid tapering practices, PS, and withdrawal symptoms were examined in 45 HSCT recipients, including 12 children aged 7–18 years. Children received significantly higher MED (mean 24.95 mg/kg) than adults during the taper, primarily for mucositis pain, with an average PS of 1–2 on a 0–5 scale, during the first 10 days. Wide variations occurred in tapering, ranging from 67% decreases to 14% increases, with sparse nursing documentation [12]. Opioid-related adverse effects were generally minimal, with no reported cases of respiratory depression, overdose, or opioid misuse [8]. However, withdrawal symptoms were commonly reported during opioid tapering, peaking around taper days 2–6 [12].

One qualitative case study highlighted challenges in managing severe and persistent pain in children undergoing HSCT. Despite representing the foundation of pharmacological pain management, opioids were often insufficient in achieving adequate analgesia, due to the multifactorial nature of pain and the complexities of post-HSCT complications. Multimodal analgesia, combining opioids with adjuvants such as ketamine and clonidine, was utilized. Evidence-based guidelines are necessary to lead to consistent practices [11].

4.2 Mucositis

Mucositis, a common adverse effect of chemotherapy and radiation, involves gastrointestinal mucosa ulceration, pain, compromised oral hydration, nutrition, and infection risks [113]. We identified 12 studies (Table 1) focused on opioids for mucositis pain in pediatric oncology [14,15,16,17,18,19,20,21,22,23,24,25]; eight investigated IV (including PCA), subcutaneous, or topical opioids [14,15,16, 18, 20, 21, 23, 25], and four focused on nonopioid agents and used opioids for breakthrough pain [17, 19, 22, 24].

Table 1 Opioids for chemotherapy-related mucositis pain in pediatric cancer patients

All patients with grade ≥ 3 mucositis required IV opioid therapy for analgesia in a study that investigated the clinical characteristics of children and adolescents post-HSCT [16]. A study of pediatric patients with Ewing sarcoma and esophagitis grade ≥ 2 indicated neutropenia and esophageal radiation therapy as risk factors; effective analgesia was attained with oral hydrocodone, fentanyl patch, and immediate-release morphine, with decreased opioid need within 2–5 days [14].

A prospective evaluation in 16 children and young adults (median age 18 years) with pain from esophagitis/mucositis reported good analgesia with IV/subcutaneous morphine infusions, median dose 0.11 mg/kg/h, over 8 days, but dose-limiting toxicity was noted [25]. Two studies reported PCA delivery for post-HSCT mucositis-related pain [15, 20]. In a prospective randomized study (n = 20), of morphine PCA versus continuous infusion (CI), PCA dosing was bolus 15 mcg/kg every 10 min, with basal infusion added at night equal to the hourly average over the prior 16 h. The CI group started at 15 mcg/kg/h after loading with 45 mcg/kg, with titration by staff as needed. Significant differences were observed in mean cumulative morphine use, as 4.94 and 12.17 mg/kg for PCA and CI, respectively (p < 0.01). There were no significant differences in PS or adverse effects between groups. The PCA group reported less sedation and difficulty concentrating [20].

In a double-blind, crossover study, mucositis-related pain was treated by PCA in ten post-HSCT children; morphine and hydromorphone were alternated between days 1–3 and 4–6, then reverted to the initial opioid on days 7–9. Bolus doses were up-titrated per protocol, by 25–100%, up to four doses per hour. Basal rates increased similarly if the pain persisted. Morphine and hydromorphone groups (defined based on the initial opioid) had similar PS, nausea/vomiting, sedation, and pruritis scores, with no significant differences in efficacy or side effects [15].

Two studies investigated the addition of ketamine to morphine PCA/NCA [18, 23]. Of 33 children aged 0.3–13.6 years (mean 5.1) with mucositis pain treated with morphine PCA boluses, 27 had NCA and 6 had PCA. Due to insufficient analgesia by day 6 (median, day 4), ketamine was added to morphine and infusions lasted 5–32 days (median 16.8). No statistical difference was noted between average morphine use pre- and post-ketamine addition. Ketamine addition led to significant reductions in the proportion of PS ≥ 4, in PCA and NCA groups. Side effects of nausea/vomiting and pruritus were not statistically significant after ketamine addition. There were no reports of significant adverse effects in either group [18].

In a retrospective study of 100 severe mucositis episodes in 82 children (3–14 years), treatment modalities included morphine only PCA or NCA, morphine with ketamine post-PCA initiation, or morphine with ketamine from the outset. Morphine bolus was 20–40 mcg/kg every 5–20 min, with infusion 0–40 mcg/kg/h. Morphine analgesia was insufficient in 26%, more common in older females (median age 12 versus 7 years). Side effects were minimal and similar between groups. Adding ketamine to morphine PCA correlated with reduced morphine use and improved PS in children with escalating morphine requirements [23].

In a retrospective study, 34 children (median age 13.5 years) with severe mucositis received IV tramadol 1 mg/kg every 6 h (maximum 400 mg/day). Unrelieved pain prompted tramadol 2 mg/kg every 6 h. If pain persisted, morphine 0.1 mg/kg IV every 2 h was administered. Patients requiring more than four rescue doses/24 h received morphine PCA in addition to tramadol. Tramadol alone successfully treated 63% of episodes; 28% required IV morphine rescue. Patients needing morphine had higher PS (p < 0.0001), particularly with grades 3–4 mucositis. Side effects were milder in those receiving tramadol alone [24].

Analgesia and absorption characteristics of oral topical morphine were investigated in 12 children (ages 2–17 years) with chemotherapy-induced oral mucositis, receiving an aqueous solution (1 or 2 mg/mL) via atomizing spray, initially 4 mg/kg, adjusted for subsequent patients, administered every 3 h for three doses. Children received acetaminophen and supplemental opioids (morphine PCA or oral morphine/oxycodone) as needed. Six out of seven children experienced reduced PS with topical morphine doses of 0.025–0.4 mg/kg, and three required additional morphine IV or oxycodone. Side effects included a brief burning or itching sensation in the mouth. In the absorption study, five patients received a single dose of 0.05 mg/kg topical morphine, and plasma concentrations remained below the limit of quantification for up to 180 min postadministration [21].

Three prospective RCTs explored non-opioid approaches for oral mucositis prevention or treatment, with opioids reserved for uncontrolled pain. Caphosol, a calcium/phosphate solution resembling saliva, was evaluated in 19 patients with pediatric cancer. Although peak PS were similar between caphosol and placebo groups, patients in the caphosol group experienced pain for a longer duration, required higher morphine peak doses (mean 0.89 mg/kg versus 0.77 mg/kg; p = 0.583), and used morphine for a significantly longer period (15.5 versus 9.1 days; p = 0.035) [22]. Laser photobiomodulation was compared with placebo sham laser in 101 pediatric patients with chemotherapy-related mucositis. While baseline PS were similar, self-reported PS were significantly lower in the laser group and the analgesic consumption was not significantly different [17]. Oral cryotherapy feasibility during chemotherapy was assessed in 53 patients with HSCT. Children with severe mucositis received significantly more opioids than those with lower-grade mucositis in both duration and dose (mean 13 days and 8.8 mg/kg versus 5 days and 1.9 mg/kg; p < 0.001) [19].

4.3 The Role of Opioids in Neuropathic Pain Conditions in Pediatric Oncology

The narrative review search identified 14 articles (Table 2) relevant to the role of opioids for NP in pediatric oncology [36, 50,51,52,53,54,55,56,57,58,59,60,61,62]. Mechanisms and etiologies for NP in children with cancer are diverse. A retrospective review (n = 160) reported NP in 16% of children with cancer (excluding leukemia/brain tumor), mean (SD) age 11.8 (4) years, range 5–18 years, most frequently with diagnosis of osteosarcoma (38%). Causes of NP were compression of a nerve/root/spinal cord (35%), limb-sparing (LS) and amputation surgery (28%), and chemotherapy/vincristine (19%). Gabapentin was the first line of treatment for NP (85%); nevertheless, opioid administration became more common with disease progression (p < 0.05), and good or partial responses to treatment were reported in 73% [50].

Table 2 Opioids for neuropathic pain in pediatric oncology patients

The use of opioid therapy has been reported for a rare complication of chemotherapy in pediatric oncology patients, which involves a neuropathic type of pain as part of the plantar erythrodysesthesia syndrome (PPES). In a retrospective investigation of 22 patients treated with high-dose methotrexate or cytarabine who experienced PPES, 82% required treatment with opioids for pain, and 23.1% of episodes required admission for parenteral pain management. Although rare in children, PPES symptoms recurred during subsequent courses of chemotherapy in half of the patients, and more than 25% of subsequent chemotherapy courses were complicated by PPES, indicating a high risk of recurrence; older age and genetic variables are suggested associations with PPES within the pediatric population [62].

The literature review identified several NP circumstances: (1) acute NP episodes during immunotherapy with chimeric 14.18 antibodies for high-risk neuroblastoma [51,52,53,54]; (2) chronic NP post-LS and amputation surgeries for bone malignancies [56,57,58,59, 61]; (3) vincristine/chemotherapy-related NP, during treatment for acute lymphoblastic leukemia (ALL) [55, 60]; and (4) NP in the EOL context [36].

4.3.1 Acute NP during Immunotherapy with Chimeric 14.18 Antibody Infusions for Neuroblastoma

Chimeric 14.18 antibody therapy is well-established in protocols for high-risk neuroblastoma, usually as three 4-day cycles, which can induce intense, acute episodes of NP, often necessitating opioid PCA or NCA. In a series of 16 patients, mean age 4.3 years (range < 2–10), morphine PCA/NCA was administered as 0.05 mg/kg load, 0.02–0.04 mg/kg/h infusion, 0.02–0.04 mg/kg on demand, with mean (range) morphine consumption (mg/kg/day) 0.46 (0.15–1.39), and highest mean utilization on day 1 (0.45). The mean daily PS in the range of 0–5 (higher on days 1 and 4), mostly 0–1.3 on all days, is indicative of good analgesic efficacy [51].

Two studies investigated multimodal plans, including dexmedetomidine, hydromorphone, and gabapentin [52] or ketamine, morphine, and gabapentin [53]. The efficacy of dexmedetomidine, hydromorphone, and gabapentin in six children, median age 3.5 years (range 2–12), undergoing chimeric14.18 chemotherapy, was reported based on infusions of hydromorphone 2–8 mcg/kg/h and dexmedetomidine 0.1–0.6 mcg/kg/h. The median hydromorphone use was 2.9 mcg/kg/h (equivalent to morphine 15–20 mcg/kg/h), range 2.0–4.7. The highest and lowest median use (mcg/kg/day) were 80.5 and 35.0 during cycles 1 and 3, respectively. For one patient, moderate-to-severe pain was reported during half of treatment days [52].

In a multimodal regimen of gabapentin (15 mg/kg/day), ketamine infusion (0.2 mg/kg/h), and morphine PCA/NCA, during three cycles of chimeric 14.18 antibody infusions (72 infusion days), in six patients, median age, 4.3 years (range 3.0–7.3), morphine PCA parameters were 0.01 mg/kg/h infusion and 0.02 mg/kg boluses. During cycles 1, 2, and 3, respectively, median (range) daily morphine consumption (mg/kg/day) was 0.21 (0.12–0.38), 0.26 (0.03–0.45), and 0.29 (0.23–0.47). Mean PS for cycles 1, 2, and 3, respectively, were 0.5 [confidence interval (CI) 0.3–0.6], 0.4 (CI 0.2–0.5), and 0.3 (CI 0.2–0.5), and low side effect rates were reported [53].

Morgan et al. evaluated the impact of opioid exposure during the first two courses (8 days) of antibody chemoimmunotherapy on subsequent neuroblastoma response to chemotherapy, by examining the relationship between cumulative opioid consumption (mg/kg) and degree of tumor reduction per two measures (primary tumor volume and total tumor burden per Curie score). In 36 children with mean age (SD) 4.2 years (3.57), opioid PCA data for morphine (89%), hydromorphone (8.3%), and fentanyl (2.8%) were reported as cumulative MED over 8 days of therapy (mg/kg) [mean (SD) 7.02 (6.86), median (IQR) 4.71 (3.49–7.96)]. By both measures of tumor response, higher opioid consumption correlated with a subsequent lesser degree of tumor reduction after chemotherapy, though not statistically significant [54].

4.3.2 Chronic NP Post-LS and Amputation Surgeries for Bone Malignancies

Among oncological surgical interventions, LS and amputation pose the risk of chronic postoperative NP. This review captured five studies reporting opioid therapy details for chronic post-LS or amputation [56,57,58,59, 61], either retrospectively [61], or prospectively [56]; postamputation phantom limb pain (PLP) and the role of mirror therapy (MT) [58] and amputation as palliative intervention and pain outcomes [57]. One study retrospectively investigated the characteristics and treatment of chronic postsurgical NP depending on the postoperative management with continuous epidural versus peripheral nerve infusions [59]. All five studies present opioid consumption data as an outcome measure for pain management.

In 150 patients with osteosarcoma (86.1%), Ewing’s sarcoma (9.3%), and other bone cancers (4.6%), median age, 14 years (range 6–21), post-LS pain was treated with oral short-acting and long-acting opioids (60.3% and 48.3%, respectively), IV opioids via PCA or non-PCA (58.9% and 93.4%, respectively), and methadone (4%). In addition to opioids, an algorithm for NP included anticonvulsants, tricyclic antidepressants, and methadone, added sequentially [61].

In a multi-institutional prospective investigation of 37 patients with osteosarcoma, median age 13.3 years (range 6.8–20.2), the incidence of NP in LS and amputation groups, respectively (81% and 83%), duration of NP (weeks, mean/SD, 7.2/8.4 and 4.9/4.0), duration of treatment, and NP-specific medications dose regimens were similar between groups. Most patients (71%) were treated with both opioids and NP-specific medications; four patients were treated with opioids only (11%). Mean (SD) duration of opioid treatment (weeks) in 30 patients (31 surgeries) was 8.4 (7.3) (median 6, range 0.1–28). The mean (SD) IV opioid dose, as MED (mg/kg/day), at initiation and maximum dose was 0.9 (2.6) and 14.7 (2.8), respectively. Methadone for NP postsurgically, mean starting/maximum dose (mg/kg/day) in four patients was 0.3/0.3 [56].

Pain outcomes were investigated postamputation performed for palliation within the last year of life in 12 patients with metastatic osteosarcoma (median age 13, range 7–20 years). The MED (mg/kg/day) reported at 1 week preamputation and 1 week and 1, 3, and 6 months postamputation, showed a trend toward higher values at 1 week [median 0.2 (range 0–7.6); p = 0.6) and 3 months [median 0.2 (range 0–0.5); p = 0.3) postsurgery versus presurgery [median 0.1 (range 0–0.5)]. Daily mean PS was significantly lower at 1 week [median 3 (range 0–6); p = 0.03] and 3 months [median 0 (range 0–8); p = 0.02) postsurgery than at 1 week presurgery [median 5.5 (range 0–10)] [57].

In 18 patients, median age of 13 years (range 8–24) with osteosarcoma as the most common diagnosis (66.7%), who developed PLP postamputation, pain outcomes were examined for patients treated with medications versus those receiving additional MT. Opioid treatment with hydrocodone, hydromorphone, oxycodone, codeine, morphine, and fentanyl, reported as mean (SD) MED (mg/kg/day), in MT versus standard group, was 0.81 (0.99) and 0.33 (0.31), respectively; methadone was used for five patients. While PLP incidence 1-year postamputation was 11.1% versus 66.7% in the MT versus standard group, no positive effect of MT was detected on the opioid MED consumption [58].

Based on the perioperative intervention of epidural versus peripheral nerve infusions, in 150 patients post-LS or amputation for lower extremity malignancy, NP was reported as incidence (23.8%), duration (median days, 38 for epidural versus 50 for nerve blocks), and NP-specific medication doses. Methadone was part of a NP-directed regimen, escalated step-wisely, from gabapentin only (94.5%), to addition of amitriptyline (24.2%) and methadone (12.7%). Methadone dosing [mean (SD) (mg/kg/day)] was 0.2 (0.1) at initiation of treatment and 0.4 (0.2) at maximum dosing, over a mean (SD) duration (days) of 89.5 (74.3) [59].

4.3.3 Vincristine Chemotherapy-related NP during Treatment of ALL

In two studies focused on vincristine-related NP in children with ALL, the investigators retrospectively explored the incidence and treatment of pain [55] and subsequently developed a prospective randomized controlled study comparing outcomes of treatment arms of gabapentin versus placebo, in which the use of morphine was a pain outcome measure [60]. In the retrospective study including 498 patients, mean age 7.9 years (range 1.2–19.2), 34.9% developed NP and some experienced recurrent NP episodes. Of 180 NP episodes with treatment data available, 62.2% (112) and 37.8% (68) were treated with gabapentin or opioids, respectively. The selection of treatment with gabapentin or opioids was not influenced by the PS at the time of diagnosis of NP (p = 0.91) [55].

The prospective randomized study evaluated the efficacy of gabapentin 20 mg/kg/day versus placebo, based on PS and opioid dosing for pain during NP episodes, as open-label oral morphine 0.15 mg/kg, every 2 h, as needed. The mean (SD) opioid doses as MED (mg/kg/day), for breakthrough pain, were 0.26 (0.43) in the gabapentin group (25 patients, 432 days) and 0.15 (0.22) in the placebo group (24 patients, 411 days; p = 0.15). The failure to demonstrate the superiority of gabapentin versus placebo was attributed to the low gabapentin dose regimen, lack of dose escalation, and short duration of follow-up (21 days) [60].

4.3.4 Specific Considerations for NP in the End-of-Life Context

Studies indicated that the substantial opioid dose escalation in the EOL context is associated with circumstances that involve NP-generating mechanisms of nerve/root/plexus or spinal cord compression [35, 36], and contribute to making pain opioid-resistant, leading to escalation of multiple lines of NP-directed medications, massive doses of opioid [35], methadone utilization [43], and regional analgesia/block techniques [39].

4.4 The Role of Opioids in End-of-Life and Palliative Care

The search identified 24 articles (Table 3) relevant for opioid-based pain management during PC and EOL in pediatric oncology [26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49]. Opioids retain a major role in ensuring quality analgesia during PC and EOL as part of multimodal analgesia. In this review, some studies explore opioids for pain and their clinical value; others explore nonopioid modalities for pain at EOL, such as regional blocks, and opioid consumption is reported as an outcome measure. Still other studies report on opioids in the context of palliative sedation therapy (PST) for intractable symptoms at the EOL.

Table 3 Opioids for end-of-life pain management in pediatric oncology patients

4.4.1 Opioid Use Comparisons Between Oncology Diagnostic Groups

The use of morphine continuous infusions for pain at EOL in pediatric oncology was reported as early as 1980 in a retrospective report of eight patients, aged 3–16 years, at rates of 0.025 to 2.6 mg/kg/h (from 0.8 to 80 mg/h), median dose 0.04–0.07 mg/kg/h, over 1–16 days. Side effects were noted as constipation, drowsiness, and respiratory depression [26].

In a large retrospective study of 70 children receiving cancer terminal care, only 11% did not require treatment with analgesics and 10% of those receiving analgesics used non-opioids. Adequate analgesia was reached in 86% and 72% at home and in-hospital, respectively. Analgesia remained inadequate in 26% and 20% of leukemia and solid tumors groups, respectively, and in two patients with brain tumors. Methadone was the most used oral opioid and morphine was the preferred parenteral opioid by intermittent dosing or infusion. The wide overall morphine use range (mg/kg/day) was 0.18–55. Doses of methadone (mg/kg/day) were 0.27–0.89 orally and 0.34–0.46 IV; buprenorphine dose was 13.4–18 mcg/kg/day sublingually. Oral opioid utilization was limited due to insufficient analgesia, inability to swallow pills, decreased consciousness, or nausea [27].

Variations in opioid needs at EOL based on oncology diagnoses in children (n = 165), mean age 8.9 years (range from 4 months to 19 years) were noted, as patients with solid tumors (ST) (non-CNS) received higher dose opioids and multiple opioids especially during the last month of life. Opioids used in 147 patients included morphine (n = 123), diamorphine (n = 95), fentanyl (n = 19), and combinations of 2 or more (n = 73), with median oral MED (mg/kg/day) 3.67 (range 0.09–1500). Opioid rotation was noted in patients who had more pain and received significantly higher maximum doses (p = 0.002). Dose escalation was observed, with the median monthly maximum dose (mg/kg/day) rising from 2.1 at study entry to 4.4 at death (p < 0.001). The need for high doses of opioids (n = 34) was defined as > 20 mg/kg/day oral MED (mostly in patients with ST, 59%), and in some cases it exceeded 50 mg/kg/day (n = 11, also mostly in patients with ST, 73%) [28].

The use of opioids at EOL in children with brain tumors, median age 8.3 years (range from 14 months to 17 years), is reportedly less common than in other pediatric oncology subpopulations based on a retrospective investigation (n = 39); 51.6% received morphine with a median of 14.5 days (range 1–584) [29]. In a retrospective comparison of children with hematologic malignancies (n = 15) versus ST (n = 50), median age of 12.2 years (range 2.1–22.4), 33 of 42 patients treated with morphine PCA were in the ST group. Consistent with other studies, children with ST diagnoses (versus hematological malignancies) more frequently experience difficult-to-treat pain at EOL and receive more aggressive opioid therapy [30]. In a similar retrospective study, the EOL care characteristics of children and adolescents (n = 99), median age 9.8 years (range 0.3–24.3), with diagnoses of brain tumor (n = 36), ST (n = 44), or hematologic malignancies (n = 20) were compared. Pain was the most prevalent symptom and required opioid treatment (n = 91) [31].

A retrospective population-based study (n = 1659, median age, 13 years) reported 93% use of opioids during the last week of life in hematologic (56%), ST (29.9%), CNS 14.1%), and post-HSCT (26%) populations. Opioid exposure increased from 76% at 1 week to 82% at 1 day before death [adjusted odds ratio (aOR) of 1.55, p < 0.001]. When comparing opioid consumption based on oncological diagnosis, patients post-HSCT demonstrated increased use of all symptom management medications including opioids (aOR of 1.34) and patients with solid tumors had higher odds of opioid (aOR of 1.39) and benzodiazepine (aOR of 1.24) treatment than patients with CNS malignancies (aOR of 0.50 and 0.65), respectively [32].

Two studies noted the details of therapy with transdermal fentanyl in PC [33, 34]. In an observational cohort study, 75% reported good or very good satisfaction with fentanyl patches used for 15 days with a 3-month follow-up (n = 41, median age 10.5 years, range 2.6–18.8) based on prior oral morphine > 30 mg/day for ≥ 48 h. Mild side effects reported were drowsiness and constipation ~50% each, nausea/vomiting ~45%, dry mouth ~33%, and itching ~30% [33]. In a retrospective study, 13 patients ages 3–18 years with progressive malignant disease (mostly relapsed ST) previously on oral morphine > 30 mg/day, having side effects or poor compliance with oral medication, were treated with transdermal fentanyl for 6 h to 112 days; 11 of 13 achieved adequate analgesia without side effects [34].

4.4.2 Massive Opioid Dose Escalation at End of Life

In a landmark retrospective study of 12 children (5 months to 20.5 years) with malignancy (mostly ST) receiving massive doses of hydromorphone or morphine infusions during terminal care, Collins et al. reported IV MED (mg/kg/h) > 3, range 3.8–518, of duration 6–240 days and described a rapid logarithmic escalation in the last 2 weeks of life for most patients. Though massive opioid infusions (> 3 mg/kg/h) were rare at EOL (12 of 199), a ST diagnosis appeared as a predictive factor (p < 0.01), along with tumor extension to nerve roots/plexus/peripheral nerve or spinal cord compression (p < 0.01). In addition to massive opioid doses, 5 of 12 patients required continuous sedation, and 3 required epidural or subarachnoid infusions [35].

Consistent with Collins et al., the association between ST diagnosis, neuropathic mechanism of pain, and rapid increase of opioid needs at EOL was reported in a retrospective investigation of 18 children (age, 0.5–19 years), with ST (n = 8), brain tumors (n = 7), and leukemia/lymphoma (n = 3). The increase in mean (SD) morphine (mg/kg/day) consumption within the last 72 h of life was statistically significant, from 155.0 ± 408.0 to 254.2 ± 551.0 (p < 0.005). The greatest increase was noted in patients with NP, from 231.2 ± 488 to 379.7 ± 647 (p = 0.009); patients without NP had stable opioid needs [36]. Anecdotally, the authors noted that methadone therapy initiated within 7 days of NP diagnostic may mitigate the opioid escalation at EOL, suggesting the benefit of NP-directed medications [36].

Opioid consumption based on opioid PCA self-titration has been evaluated [37, 38]. In eight patients (median age 8.5 years, range 3–17), with diagnoses of leukemia, brain tumor, and ST, MED (mg/kg/day) range was 0.32–18.76 (median 0.66–2.21) over median 9 days (range 1–50). Significantly increased MED was noted at PCA initiation and during the last week of life. Median delivered and undelivered bolus requests/day were 7.5–21 and 0–4.5, respectively; 39 PCA parameter changes were made on 22 occasions. Median daily mean PS were low (0–3). These findings support PCA delivery as both titratable and effective to achieve analgesia at EOL [37].

A larger retrospective study (n = 43, age median 4.7 years, range 1.2–24) of patients with leukemia/lymphoma (59%) and ST/brain tumor (41%), reported increasing rates of PCA utilization in 63%, 88%, and 100% of patients for the last 14, 7, and 3 days, respectively, with a significant increase in MED (mg/kg/day) during the last 2 weeks of life (p < 0.001). While in the overall group the mean (SD) increased from 10.7 (17.9) to 19 (25.8) from day 14 before death to the day of death, higher opioid needs were observed in patients younger than 13 years versus adolescent young adults (AYA) and in those with ST diagnoses versus leukemia/lymphoma. The highest mean (SD) PS was recorded on the day before death as 6.7 (2.3), and the increase in mean PS during the last 2 weeks was significant (p < 0.001) [38].

4.4.3 Regional Blocks and Impact on Opioid Consumption at End of Life

Regional blocks at EOL have been utilized to improve analgesia and reduce opioid exposure and opioid-related side effects [39,40,41]. Eleven terminal pediatric malignancy patients with ST (9) and hematological malignancy (2), age 5 months to 16.3 years, were treated with epidural or subarachnoid infusions or neurolytic blocks for indications of limiting side effects of systemic opioids (nausea, respiratory depression, and clouded sensorium), NP unresponsive to opioids, or thoracocenteses of malignant pleural effusions and intrapleural chemotherapy. Systemic opioid dosing (IV MED, mg/kg/h) at institution of regional interventions was 0.02–3.9. Epidural and subarachnoid infusions included morphine, hydromorphone, or fentanyl for 3 days to 7 weeks. Satisfactory analgesia was reported in all cases; complications included dural puncture headache and mild respiratory depression [39].

Two retrospective studies utilizing similar methodologies report analgesia outcomes (PS and opioid consumption) with regional interventions in pediatric oncology patients at EOL, from the same institution, a decade apart [40, 41]. The first study described ten patients (age 4.4–21.3 years), nine with ST and one with lymphoma, who had 14 devices (epidural and peripheral nerve catheters) for 3–81 days. A total of 12 of 13 catheters provided improvement by at least one analgesic efficacy criterion: improved mean PS at 24 h and decreased opioid requirement at 24 h and 5 days. The neuraxial infusions included local anesthetic and opioids (fentanyl, hydromorphone). Typical contraindications for indwelling catheters (spinal metastasis, vertebral fracture, thrombocytopenia, and fever) were superseded by palliative care needs, and no complications were noted [40]. The second study included 27 patients, mean age 14 years, range 2–26, of which 81.5% had ST, most often osteosarcoma, treated with continuous nerve block (CNB) catheters or single-shot nerve blocks (SSB) during the last 3 months of life. The mean opioid MED (mg/kg/day) based on PCA use, before regional intervention, 24 h post, and at day 5 were considered as analgesia outcome measures in addition to PS. Among patients receiving CNB (mostly epidural catheters) with concurrent opioid PCA, 37% had decreased PCA utilization at 24 h and 37% at 5 days after catheter placement. Of patients with SSB (mostly celiac plexus block) using opioid PCA, 25% and 17% had decreased or stable PCA requirement at 24 h respectively, and 17% had decreased opioid use at 5 days, suggesting, unsurprisingly, that continuous catheter analgesia delivery lasts longer than single doses [41].

4.4.4 Methadone for Pain Management at the End of Life

In a retrospective review of pediatric oncology patients with advanced disease (median age, 8.9 years, range 2.6–18.9), with ST (n = 13), hematologic malignancies (n = 3), and brain tumor (n = 1), methadone was utilized as the primary opioid analgesic at EOL. Of 17 patients treated for 925 patient days (range 1–199, median 36), 94% remained on methadone until death, without significant adverse events. Conversion ratios of methadone/morphine had a wide range, from 1:2 to 60:1 [42].

Data regarding methadone for cancer pain in 20 pediatric patients during the last 30 days of life (ST/brain tumor, n = 15; leukemia/lymphoma, n = 5) include treatment duration, median 32 days (range 2–323), and dose range (mg/kg/day) at 4 weeks, 2 weeks, and 24 h before death, as 0.09–7.76. Indications for methadone were nociceptive and mixed nociceptive/neuropathic pain in 60% and 40%, respectively. Methadone discontinuation due to concerns regarding increased QTc interval or arrhythmia episodes was not needed [43].

4.4.5 Opioids and Palliative Sedation Therapy at End of Life

Palliative sedation therapy (PST) has been utilized in pediatric oncology as a measure of last resort when facing intractable symptoms at EOL [44,45,46,47]. In a retrospective review of at-home PST for 21 terminally ill children with cancer, median age 13.5 years, midazolam was most used, while concurrently providing morphine for pain management. The symptoms leading to PST were pain (n = 13), dyspnea (n = 9), and anxiety (n = 5). The dose of morphine steadily increased or remained stable throughout PST in ten and nine children, respectively. Morphine dose (mg/kg/h) was median (25–75th percentile): on the initial day of sedation: 0.05 (0.02–0.07), on the last day: 0.06 (0.04–0.08), maximum dose day: 0.07 (0.04–0.14), and mean dose (mg/kg/h): 0.06 (0.04–0.10) [44].

The opioid MED (mg/kg/h), based on continuous infusions of morphine, hydromorphone, fentanyl, and meperidine, in nine terminally ill pediatric oncology patients, mean age, 6 years (range 1–18), was reported as range, at initiation (0.11–13), maximum (0.11–46), final dosing (0.11–34), and ratios maximum/initial (1–81). Propofol PST was initiated for inadequate analgesia with opioid infusions and sedatives and massive MED (> 3 mg/kg/h) utilization [45]. End-of-life characteristics of PST as a comparison between children with brain tumors (median age 9 years) versus sarcomas (median age 17 years) were examined retrospectively in 26 patients. The number of patients treated with opioids or details of opioid consumption were not provided; a typical morphine dosage of 0.5–5 mg/h was described [46].

Opioid consumption was compared retrospectively during PST for children, median age 8.5 years (range 1–22), for refractory symptoms caused by ST in brain tumors and sarcoma groups. Median morphine doses (mg/kg/h) were 0.024 (range 0.002–0.06) and 0.09 (range 0.015–0.13), respectively, for brain tumors (n = 15) and sarcoma (n = 11) groups. At admission, cumulatively during admission, and on the day of death, the morphine dose ranges were, for brain tumors versus sarcoma groups, 0–0.02 versus 0–30, 0–0.08 versus 0.015–0.4, and 0–0.17 versus 0.02–1.6, respectively [47].

4.4.6 Populational Data Studies of Opioid Consumption at the End of Life

Population-based studies have explored the characteristics of opioid prescribing at EOL in patients with cancer, in one study based on data from opioid prescriptions by pharmacy dispensations in Nova Scotia, Canada, in all age groups [48], and in another study as opioid prescriptions during the last week of life among a cohort of pediatric oncology patients who died during hospitalization [49]. In the Canadian study, data regarding the pediatric population are reported as an age group including young adults (< 30 years, n = 48, representing < 1% of n = 11,498). Overall opioids prescribed during the last year of life included: hydromorphone, morphine, fentanyl, oxycodone, acetaminophen- or acetylsalicylic acid-opioid combinations, codeine, methadone, meperidine, pentazocine, dextropropoxyphene, and buprenorphine. For those < 30 years, opioid prescriptions were found for n = 37 patients, representing < 1% of the total opioid prescriptions of 6186. The adjusted average number of prescriptions per person per year, for n = 48 patients, was 11.4 (95% CI 10.1–12.8), and the adjusted mean morphine equivalent per day dispensed per person, for n = 37, was 53.9 (95% CI 36.5–79.6) [48].

The pediatric population-based study [49] included patients 0–24 years with neoplasm diagnoses, who died during hospitalizations and received opioids in the last week of life (n = 1466). Comparisons were made between patients on daily opioids versus less than daily, during the last week of life. Overall, 56% received opioids daily while hospitalized during the last week of life, with substantial variations across 33 hospitals (range 0–90.5%); the hospital-level effect accounted for 10.5% of the variance in daily opioids (p < 0.001). Patient variables were identified. Based on private insurance versus Medicaid/governmental insurance, 63.4% versus 51.9% received daily opioids (p < 0.001). Based on age and diagnosis, patients 10–14 years and diagnoses of leukemia/lymphoma were significantly most likely to receive daily opioids when compared with younger age and diagnosis of brain tumors, respectively. Regarding the duration of hospitalization, patients who died within 3 days of admission were markedly more likely to receive daily opioids when compared with patients hospitalized for a week or longer [49].

A third populational-based study by Prozora et al. included 1659 patients up to 21 years, from a network of 100 academic medical centers (n = 1659), and reported the EOL characteristics of opioid exposure, noting progressing increases in the percentage prescribed opioids from one week before death to the day of death, in oncological diagnoses categories of hematologic, ST, CNS, and post-HSCT [32].

4.5 Breakthrough Pain

Two studies evaluated the management of breakthrough pain in 53 pediatric patients, mean age 13 years (range 5–18), with leukemia, ST, or CNS tumors [87, 88]. In a study of rapid onset fentanyl sublingually (mean oral MED 16.2 ± 17.0 mg/kg/day), lozenge (mean oral MED 18.5 ± 29.5 mg/kg/day), and nasal spray (mean oral MED 3.5 mg/kg/day), fentanyl lozenges appeared to be relatively safe and effective in children as young as 5 years and might be considered as alternative to oral opioids for breakthrough pain in pediatric cancer [87]. Friedrichsdorf et al. explored breakthrough cancer pain management with PCA or oral oxycodone in 28 pain episodes, noting a significantly higher risk of breakthrough pain in children aged 7–12 versus 13–18 years. Most children’s pain was successfully managed with PCA boluses, suggesting efficacy for breakthrough pain [88]. Both studies indicate the efficacy of rapid-onset opioids for pediatric cancer breakthrough pain, offering viable alternatives to oral opioids.

4.6 At-Home Use

Three studies investigated aspects of at-home use of opioids. In a study of 161 surveys of parents/caregivers of children with cancer, concurrent with pain diaries for one month and qualitative interviews with parents, morphine was the most used pain medication. Parents often undertreated pain due to uncertainty about analgesic medications and opioid-related stigma. A preference for nonpharmacologic treatments led to suboptimal pain management [86]. In a study of 45 parent-child pairs during a 14-day pain diary assessment, among children with cancer aged 4–17 years, 40% received at least one analgesic at home, predominantly orally or topically. Children experienced pain-related difficulties in activities, and parental misconceptions about analgesic use were associated with lower administration rates [85]. In a retrospective evaluation of at-home PCA for PC in 33 patients with pediatric cancer, opioids provided good pain control in most cases; although, 3 patients required readmission due to uncontrolled pain [66].

4.7 Opioid Misuse and Abuse

In a focused review, a risk assessment was conducted in 38 AYA with cancer pain, median age 19 years (range 16–29), referred for chronic pain management and expected to require chronic opioid therapy (COT). Using the SOAPP-R tool and a diagnostic interview, 39.5% were identified as high risk for opioid abuse, and five patients displayed red flag behaviors (seeking medications from multiple providers, reporting lost/stolen prescriptions or pills, resisting changes to medication regimens, medication noncompliance such as purposeful oversedation, hoarding drugs or illicit drug use). The authors recommended safeguards for patients at high risk (limited number of opioid prescribers, prescribing medications with lower potential for diversion, and cognitive behavioral therapy). Education about the risks of opioid use and nonpharmaceutical options for chronic pain management should be offered to all patients receiving COT, regardless of risk category [102].

One cohort study investigated opioid misuse in pediatric patients undergoing cancer treatment. Adolescent young adults (AYA) (n = 94), mean age 16.3 years (range 12–28) with leukemia/lymphoma and ST, had their opioid regimens reviewed and were assessed for psychosocial risk factors associated with opioid misuse. The overwhelming majority of patients required COT (≥ 90 days) during their cancer treatment; 60–75% presented opioid abuse/misuse risk factors including personal or family history of substance use disorder (SUD) or mental health diagnoses of depression, conduct disorder, ADHD, and post-traumatic stress disorder [104].

Four studies specifically investigated substance misuse/abuse in childhood cancer survivors [103, 105,106,107]. In a retrospective study of patients with sarcoma (10–26 years), new persistent opioid use was found in 14% during the first year after completion of cancer treatment [103]. A large database study of childhood cancer survivors ages ≤ 21 years (n = 8969) compared the prevalence of opioid prescriptions, potential misuse, and SUD, in survivors versus peers without cancer. Cancer survivors filled a higher proportion of opioid prescriptions, regardless of sociodemographic factors, compared with matched peers. Misuse of opioids was two times higher in survivors versus peers but was not statistically significant for SUD [107]. Following the CDC 2016 opioid prescribing guidelines for chronic pain [114], a similar matched cohort study of childhood cancer survivors (mean/SD age 13.7 ± 6.2 years) evaluated opioid use/misuse according to four indicators: high daily opioid dose (≥ 100 MED), multiple opioid prescription overlap of ≥ 7 days, opioid and benzodiazepine overlap of ≥ 7 days, and opioid dose escalation defined as ≥ 50% increase in monthly average MME twice/year. A reduction in opioid prescriptions and potential misuse/SUD was noted after the guidelines’ publication, with a greater reduction seen in childhood cancer survivors, demonstrating improved prescribing practices [106]. Another long-term, retrospective study evaluated 1208 adult childhood cancer survivors at least 10 years from diagnosis and at least 18 years of age at assessment. The mean age was 33.6 (SD 7.9) years, and mean time since diagnosis was 24.7 (SD 7.9) years. Pain status and opioid and/or marijuana use were self-reported at baseline and a follow-up evaluation (mean interval 4.2 years). Overall, 8.5% of patients reported using an opioid for at least 30 days and 11.9% reported having used marijuana at baseline and/or at follow-up. Persistent or increased significant pain was associated with higher odds of opioid use but not marijuana use. Survivors reporting anxiety or depression were more likely to report both opioid and marijuana use, suggesting that self-medication of emotional distress may be a mitigating factor [105].

4.8 Low- and Middle-Income Countries Publications

This literature search identified four publications from LMICs (not otherwise categorized by clinical circumstances); one focused on barriers to pain management [108], and three examined the application of WHO pain management guidelines [109,110,111]. Access to opioids emerged as a significant challenge in pediatric oncology pain management in LMICs, compounded by limited financial resources, insufficient pain management education, and cultural perceptions about pain. Rural, public hospital-based, and lower socioeconomic status children encounter notable disparities in pain relief [108]. The application of WHO pain management guidelines in rural India was explored retrospectively in children (aged 4–18 years) with hematologic malignancies during 290 admissions, of which 32.1% were pain related. Statistically significant age-related differences in pain medication choices were found. Acetaminophen (step 1) was predominantly used in younger children (83%), and tramadol (Step 2) was more common in older children (58%) (p < 0.001). Strong opioids (step 3) were administered in 9.7% of admissions [109]. In a study in children with ALL in India, step 1 analgesia provided effective analgesia in 31%, step 2 in 54%; step 3 (strong opioids) were needed in 15%, particularly for neuropathic and bone pain [110]. A study examining the application of WHO guidelines for cancer pain in 184 pain episodes in children and adolescents with cancer in Brazil reported satisfactory pain relief in 97% of episodes using morphine, with no severe side effects noted; however, psychological dependence was observed in 2%. Morphine was used during 111 admissions for 2758 patient days, with mean doses (mg/kg/day) increasing during admission, on days 1, 7, and final day, from 1.8 to 2.1 and 3.9, respectively, for 1–197 days (mean 26). Morphine doses were escalated in patients during EOL care [111].

5 Specific Opioids

5.1 Methadone

Ten studies, eight of which were retrospective cohort studies, have described the use of methadone (n = 262) for a variety of pediatric cancer pain indications, including acute [76,77,78,79,80], chronic [43, 59, 77, 78], neuropathic [43, 76,77,78, 80, 81], and EOL pain [42, 43, 76, 82] (Table 4). Methadone has also been reported in opioid weaning and prevention of withdrawal [76].

Table 4 Methadone for pediatric cancer pain

Methadone dosing varied across studies, with initial dose ranges (mg/kg/dose) of 0.05–3.8 (most commonly 0.1) [59, 76, 78,79,80,81]. The maximum daily methadone doses (mg/kg/day) were reported as 9.4 in one case [76] but were typically 0.05–0.5 [43, 59, 77, 79]. Methadone was administered every 12 h in two studies [78, 81] and every 4 h in one study [79]. A cohort study of 29 patients featured various dosing intervals, with most receiving methadone every 6–8 h (some as frequently as every 4 h or as infrequently as every 24 h) [82]. Routes of administration included oral [43, 76, 78, 81, 82], IV intermittent [43, 76] and continuous [43], and nasogastric [76]. The duration of methadone therapy was highly variable, from 1 day to more than a year [42, 43, 59, 76,77,78,79,80, 82]. Overall, the examined studies demonstrated the use of methadone at various doses for pediatric cancer pain management. Careful individualized dosing with gradual titration is recommended to achieve optimal analgesic effects while minimizing side effects.

Methadone appears to provide effective analgesia in pediatric cancer pain. Improvements in PS following methadone initiation were noted in several studies. Anghelescu et al. reported that among 14 patients with documented maximum PS before and after starting methadone, 9 showed a reduction in PS, and 7 had complete pain resolution [76]. Davies et al. noted challenges in obtaining consistent PS due to young age and EOL setting. However, the proportion of positive pain observations increased from pre- to post-methadone in 55% of evaluable patients, and parents reported improved analgesia in 94% of cases [42]. Palat et al. documented good analgesia on the pain scale in five of nine evaluable patients [80]. Several studies did not report specific PS, yet authors noted qualitativelyqualitatively effective pain control with methadone in 24 of 28 patients [82] and over 21 of 22 methadone courses [79].

The side effect profile of methadone is favorable. Mild sedation, nausea, and constipation are reported most frequently; significant adverse events are uncommon. One study examined the effects on cardiac repolarization, finding QTc prolongation in 16% of patients [81]; however, no arrhythmias were observed. Another study focused on QTc prolongation found mean QTc longer during methadone treatment versus baseline, but methadone dosage and duration were not correlated with QTc prolongation [77].

5.2 Buprenorphine

Buprenorphine has been reported for chronic cancer pain [83] and EOL cancer pain [27] in pediatric patients. Sixteen patients with pediatric cancer received transdermal buprenorphine for moderate to severe chronic cancer pain, with initial dose 8.75–35 mcg/h based on weight, mean dose 32.6 ± 14.78 mcg/h (0.5–1.8 mcg/kg/h). Buprenorphine provided effective pain relief, with mean PS decreasing significantly from 6.25 at baseline to 1.38 after 60 days of treatment (p < 0.001). The quality of sleep, activity level, and other quality of life parameters also improved significantly. Common side effects of nausea, vomiting, and constipation were generally mild and managed with supportive medications [83]. Sirkia et al. reported that 12 out of 70 terminally ill patients with pediatric cancer received sublingual buprenorphine at a dose of 13.4–18 mcg/kg/day. While not focused on buprenorphine specifically, this study provides context on dosing for severe EOL cancer pain in children [27].

5.3 Nalbuphine

In a prospective study comparing morphine and nalbuphine for pediatric mucositis-related pain, 96 patients ages 4–12 years, treated for 14 days, received mean (SD) doses (mg/kg/h) of nalbuphine 0.07–0.08 ± 0.03–0.05 and morphine 0.03–0.04 ± 0.02–0.04. On day 2 of therapy, patients receiving morphine reported lower mean PS (4.38 ± 2.9) compared with nalbuphine (6.22 ± 2.4; p < 0.05); this was not significantly different on subsequent days of therapy. Nalbuphine recipients experienced fewer side effects of pruritis, constipation, and urinary retention compared with morphine (p < 0.05), with less frequent withdrawal symptoms after discontinuation (p < 0.05). Overall, nalbuphine demonstrates safety and efficacy as an analgesic for pediatric oncological pain [84].

5.4 Transdermal Fentanyl

Three studies reported on the efficacy of transdermal fentanyl for pain in 88 pediatric oncology patients, each utilizing unique dosing approaches. Hiyama et al. tailored doses based on patient age [72], while Collins et al. administered doses of 25–300 mcg/h [71]. Othman et al. dosed according to weight, with those under 15 kg and experiencing moderate pain receiving 12 mcg/kg/h, and those over 15 kg or experiencing severe pain receiving 25 mcg/kg/h; IV morphine was used during the initial patch application and for breakthrough pain [73].

The indications for transdermal fentanyl varied from acute pain [71] to chronic cancer pain [73]. All studies reported effective pain relief and no respiratory depression was reported. One patient experienced withdrawal with the IV dosing change to transdermal, which was corrected with a patch dose increase [71].

5.5 Long-Acting Morphine

While morphine is considered a gold standard in the treatment of moderate to severe pain, long-acting oral morphine may be difficult to use in pediatric populations due to dosage availability limitations. A study assessed controlled-release oral morphine in 60 children (age 3 months to 18 years) with leukemia, ST, or aplastic anemia for palliation at EOL [68]. When analgesia was insufficient with acetaminophen or weak opioids, patients received long-acting morphine 0.2–2.3 mg/kg/dose (median 0.8) twice daily for a mean of 14 days (range 1–370), which was found effective and acceptable by patients and caregivers and especially convenient for at-home treatment. The most appropriate starting dose was 1 mg/kg twice daily with a median maximum dose of 1.5 mg/kg (range 0.4–41.5), with a short-acting opioid for breakthrough pain. The most common side effects were drowsiness, constipation, nausea, and itching. Sometimes, families were unaware that the tablets should not be crushed, indicating that education of caregivers is essential [68].

Population pharmacokinetics of oral morphine and morphine glucuronides, provided as immediate-release liquid or controlled-release tablets for cancer pain associated with either hematologic malignancy or ST, were assessed in 40 children, median age 11.4 years (range 1.7–18.7) [69]. Eighteen children received immediate-release liquid, median dose 5 mg (range 2–20) every 4 h, and 22 children received sustained-release tablets, median dose 30 mg (range 4–460) every 12 h. A median of 4 blood samples per child was taken to assess pharmacokinetics parameters. Higher PS were recorded in children who had average morphine concentrations < 12 ng/mL. Significant pain was present in 30% of the children; PS tended to increase toward the end of the dosing intervals, with no relationship with plasma morphine or metabolite levels. Children younger than 11 years had significantly higher clearance and volume of distribution of morphine and metabolites compared with adults and older children. This study supported an initial dose of 1.5–2 mg/kg/day to attain adequate plasma morphine concentrations in children [69].

Both age-dependent effects and adverse effects of long-acting morphine were evaluated in 95 pediatric patients with cancer, aged 1–19 years (median 9.3) The average oral equivalent morphine dose during treatment was 2.1 ± 2 mg/kg/day (range 0.4–0.8). Children younger than 7 years of age had the highest average dose (2.6 mg/kg/day, SD 2.8). Median PS tended to decrease during the individual recorded periods. Initially, about 25% of patients reported adverse effects such as constipation, sedation, and itching, but they decreased throughout therapy, and there were no reports of severe symptoms [70].

5.6 Morphine Infusions

Three studies reported on the efficacy of morphine infusions in 115 pediatric and young adult oncology patients, by IV [25, 75] or subcutaneous route [74]. Some reported median dosing based on indication (0.2 mg/kg/h for tumor pain and 0.11 mg/kg/h for mucositis), with prolonged duration in patients with tumor pain (range 2–154 days) [25]; others reported an initial bolus, median, 50 mcg/kg, and infusion 20–50 mcg/kg/h with boluses of 20–50 mcg/kg for breakthrough pain every 20–30 min, followed by a subsequent 25% increase in the infusion [75]. Continuous subcutaneous morphine infusions were reported at a median dose of 0.06 mg/kg/h (range 0.025–1.79) for acute pain [74].

Twenty (76.9%) patients reported good or excellent pain control when provided with continuous IV morphine infusions [25]. During 85 morphine infusions, patients receiving continuous morphine for pain after major surgery had more breakthrough pain episodes than others during day 1 of morphine treatment even though their continuous infusion was almost twice as high (0.64 versus 0.36 mg/kg; p < 0.001); of those patients, 92% had ST diagnoses [75]. Subcutaneous morphine was noted to result in achieving satisfactory pain control regardless of the cause of pain [74].

All studies reported mild adverse effects of drowsiness and respiratory depression; one study reported nausea, vomiting, and constipation were most noted [75]. Two patients required discontinuation due to apnea (n = 1) and nightmares/hallucinations (n = 1) [75]. Overall, continuous morphine infusion appears to be a viable option for pain management; however, it is important to consider dose-limiting side effects and utilize prophylactic antiemetics and/or laxatives.

5.7 Opioid PCA

Six studies evaluated the safety and/or efficacy of opioid PCA for pediatric cancer pain management (Table 5); one was a prospective clinical study [67]. Of 2029 patients evaluated in these studies, ages were 4–18 years, except for one study which included from newborn to young adults [63]. Opioids included fentanyl [63,64,65,66,67], morphine [8, 63,64,65,66], and hydromorphone [8, 63,64,65,66]. One study also reported the use of methadone via PCA [66].

Table 5 Patient-controlled analgesia for pain in pediatric oncology

The predominant indication reported was acute pain [8, 64, 66, 67]. Dosing was standard across studies. Ruggiero et al. reported fentanyl as 1 mcg/kg/h with bolus 1 mcg/kg and 7-min lockout interval, with a mean total dose delivered during 48 h as 16.3 mcg/kg [67]. Another study reported a mean MED of 2.13 mg/kg/day with a wide range (0.5–24) and a mean duration of 33.7 days (range 1–150 days) [66]. PCA duration ranged from under 10 to over 30 days, mean 19 days [8]. For morphine PCA, standard bolus doses were 0.02 mg/kg [8, 63,64,65] with lockout intervals from 8 min [8] to 15 min [63,64,65]. Hydromorphone bolus doses ranged from 0.003 mg/kg [8] to 0.004 mg/kg [63,64,65] with the same lockout intervals as reported for morphine. The remaining studies evaluating fentanyl used a starting bolus dose of 0.5 mcg/kg with a 15-min lockout interval [63,64,65].

In the studies that reported outcomes of PS, there was a significant lowering of PS noted before PCA compared with 4 h after initiation (p < 0.001) and at end of therapy (p < 0.001) [67]. No differences were noted between mean PS at the time of initiation of outpatient PCA in the groups of patients discharged from inpatient status versus the group having a new outpatient PCA started (3.2 versus 3.43, p = 0.71) [65]. However, this study also noted significantly higher mean PS in the group of patients who died when compared with the group who changed outpatient status (transfer to another institution or hospice or readmission) and in the group discontinuing PCA due to required changes in pain management (4.29 versus 3.8 versus 1.83, respectively; p < 0.0001) [65]. Mherekumombe et al. reported PCA as being an effective mode of analgesia with a mean PS of 5 (range 2–10) [66]. PCA was reported to be safe and effective in 95% (n = 37) of patients; PS were not reported [8].

Two studies compared PCA to PCA by proxy [63, 64]. Complication rates were consistently low in both clinician-proxy and parent-proxy as compared with standard PCA (0.96% and 0.62% versus 1.94%) [64] and authors concluded that PCA by proxy was a safe method of pain management in pediatric patients with cancer pain. Across both studies, adverse effects with PCA were rare, occurring in less than 1.5% of patients [63, 64]. A study of outpatient utilization of PCA reported an overall complication rate of 0.36%, with no patients experiencing concurrent respiratory and neurological complications [65]. Other side effects reported included itching, nausea/vomiting [66, 67], rash [67], sedation, and urinary retention [66]. Overall, PCA and PCA by proxy appear to be safe and effective in managing pediatric cancer pain.

5.8 Opioid Rotation

In an evaluation of opioid rotation and its therapeutic value (n = 162), during 397 admissions, mean age 7.7 years (range 1–15.8), up to 14% of patients treated with opioids required at least one opioid rotation, mostly prompted by excessive side effects (86.7%). Other reasons included tolerance to opioids and poor pain control without opioid tolerance (6.7%, each). Pre- and post-rotation PS did not change significantly (remained at 2.4), and the mean MED (mcg/kg/h) did not significantly differ from prerotation (66.5) versus postrotation (72.1). Opioid rotation appears to be of therapeutic value as a management strategy for dealing with excessive side effects, while not significantly changing PS or MED [89].

5.9 Opioid-Sparing Treatments

Five studies reported on opioid-sparing interventions for pediatric cancer pain treatment in 148 patients. Three studies evaluated the addition of ketamine to opioids [90,91,92]. Finkel et al. described the addition of low-dose ketamine in children with intractable, terminal cancer pain either for inadequate pain control or serious opioid side effects and reported improvement in PS [92]. Anghelescu et al. evaluated the effects of ketamine in addition to opioid PCA. The oncology group had significant decreases in opioid consumption as compared with during and before ketamine treatment (median MED, mg/kg/day, 2.2 versus 2.6; p = 0.03) and after and before ketamine treatment (1.7 versus 2.6; p = 0.03) [90]. Courade et al. reported that the addition of low-dose ketamine continuous infusion reduced the mean PS from 6.7 to 4.3 (p < 0.001) from day 1 to day 3, with 56% of patients achieving a reduction of at least 2 points within 48 h after ketamine initiation [91].

Anghelescu et al. reported on bisphosphonate therapy [94] and the use of lidocaine infusions [93] to reduce opioid consumption. The retrospective study evaluating the effects of bisphosphonate in patients with sarcomas or bone metastases reported a significant decrease in opioid consumption at days 4–8, 11–12, and in week 3 after the first bisphosphonate administration. Pain outcomes at 2 weeks did not appear to significantly improve, but opioid consumption was decreased at several time points during the first 3 weeks. [94].

In a retrospective study evaluating the effects of IV lidocaine infusions on opioid consumption, patients, median age 15 years (range 3–21), demonstrated a MED consumption (mg/kg/day) one day after the infusion significantly lower than during the infusion (1.15 versus 1.26 mg/kg/day; p = 0.01). Mean PS was significantly lower one day after the infusion versus the day before the infusion (6.07 versus 6.8; p < 0.001) [93]. All studies reported success with opioid-sparing interventions as measured by decreased opioid consumption and some noted lower PS.

6 Opioid Side Effects

Six studies focused specifically on opioid side effects in pediatric patients with cancer [95,96,97,98,99,100]. An investigation of mental status change in 60 children treated for cancer indicated that opioids (not specifically identified) accounted for the highest number of cases of somnolence, stupor, confusion, and delirium, caused by medications, though about 30% had more than one cause for their symptoms, usually a combination of medications or a metabolic anomaly [95].

A case series described six children aged 1–6 years (median 2) with reversible, generalized movement disorders (sudden muscle twitches, dystonia, sudden proximal limb movements, and intermittent conjugate gaze deviation), during critical care admission. Confounding variables included systemic or intrathecal chemotherapy and concurrent medications benzodiazepines and opioids. Regarding opioids, four children received fentanyl for ≥7 days, one had been changed to morphine infusion 8 days before movement symptoms, and one was 1-day post-fentanyl wean. Midazolam and fentanyl infusions were the common denominator and the most likely reason for the movement disorder [99].

The prevalence of constipation in children with cancer was assessed in two studies [97, 98]. Data were collected from 48 hospitals, including children 1–21 years (median 5.9) with ALL who had a diagnosis of constipation. In 4647 patients with ALL, the most common GI complaint was constipation in 1576 children (33.9%; CI 32.6–35.3%). The possible effect of opioid use defined as fentanyl or nonfentanyl opioid was investigated; 92% of patients received an opioid at some point: 21.6% received fentanyl, 33.5% had < 2 days of a nonfentanyl opioid, and 45.0% had > 2 days of a nonfentanyl opioid. As compared with patients without any opioids, the use of nonfentanyl opioids for ≥ 1 day [OR of 1.33 (95% CI 1.03–1.71); P = 0.0262], as well as the use of nonfentanyl opioids for ≥ 2 days [OR of 1.80 (95% CI 1.41–2.29); p < 0.0001] were associated with increased risk of constipation. The use of fentanyl was not found to be significantly associated with constipation (p = 0.32). On multivariable analysis, use of nonfentanyl opioids ≥ 2 days was found to be independently associated with increased odds of constipation (p < 0.001). There was a significant association between constipation in induction therapy with the use of nonfentanyl opioids [97]. Using the same database, the incidence of constipation in patients 0–21 years (median 9.3 years) with ST (CNS, bone, lymphoma, kidney, liver, retinal, abdominal, and adrenal) was assessed. In 13,275 patients, constipation was the most common GI complaint (64.7%); 45% of admissions included an opioid: 4.5% fentanyl, 15.2% < 2 days of a nonfentanyl opioid, and 25.4% > 2 days of a nonfentanyl opioid. Among admissions with constipation, 50.2% received at least one opioid, whereas 43.8% of admissions without constipation received at least one opioid. Extended use of nonfentanyl opioids (> 2 days) was more common in admissions with a constipation diagnosis compared with those without (33.1% versus 23.5%; p < 0.0001) [98].

The safety and efficacy of adjuvant stimulants dextroamphetamine or methylphenidate in adolescents with cancer and opioid-related somnolence were assessed in 11 patients aged 12–20 years (mean 15.5). The mean MED was 0.183 mg/kg/h (range 0.034–0.625). Stimulants were administered orally daily or twice daily. Mean initial doses for methylphenidate (n = 7) were 14.6 mg (range 7.5–25) and 6.9 mg (range 5–7.5) for dextroamphetamine (n = 4). Decreased somnolence or improved family interaction were noted in five patients. The mean duration of stimulant use was 10.3 days (range 1–27 days). The authors concluded that stimulants may help counteract opioid-related somnolence [100].

Adverse effects and toxicities due to complex chemotherapy, disease, medications, or drug interactions were assessed over 7 years based on the CHOP cancer registry database. Of 1713 pediatric patients with cancer, 326 had ≥ 1 adverse drug reaction, mostly in those with leukemia (52.1%) and neuroblastoma (28.5%), and aged 2 to 12 years. Patients who experienced toxicity (n = 319) received fentanyl (63%), oxycodone (59.3%), and midazolam (71.5%). One drug interaction, fentanyl-midazolam, was a key finding: their coadministration within 30 min increased the risk of hepatotoxicity [96].

7 Clinical and Research Implications

Pain in pediatric oncology is multifaceted and complex, and patients are likely to experience several types of pain during cancer therapy and possibly into survivorship. Each pain clinical circumstance will necessitate tailored treatment plans, depending not only on the cause and mechanism of pain but also on individual co-morbidities and other individual patient factors. All clinical circumstances of moderate-to-severe pain will likely call for opioids as part of the multimodal pain treatment plan in pediatric oncology. While nonopioids represent the first of the building blocks of pain management, opioids retain a crucial role in certain clinical circumstances, as detailed below. We acknowledge that nonpharmacological interventions (behavioral/psychological and physical interventions) represent an essential part of the triad of pain management, in addition to medications, yet they are not included in the focus of this review.

The use of opioids in pediatric cancer pain management, particularly in scenarios such as HSCT and mucositis, presents several significant implications. In HSCT, opioids such as morphine, fentanyl, and hydromorphone are commonly administered via PCA to address acute pain related to mucositis or graft-versus-host disease. However, the variability in opioid dosing, especially the need for high doses over prolonged periods in young children, emphasizes the importance of individualized dosing strategies and careful monitoring to mitigate adverse effects and withdrawal symptoms during tapering. Additionally, the challenges in achieving adequate pain relief with opioids alone underscore the necessity of exploring multimodal analgesia approaches integrating adjunctive medications such as ketamine and clonidine.

The place of opioids in treatment plans for NP in pediatric oncology varies based on clinical entity. In managing acute NP episodes induced by chimeric 14.18 antibody infusions for high-risk neuroblastoma, opioid PCA has played a vital role. Multimodal approaches combining medications such as gabapentin, dexmedetomidine, and ketamine with opioids have shown promise in improving analgesic efficacy and minimizing side effects. However, studies also underscore the importance of cautious opioid use due to concerns regarding tumor response to chemotherapy in neuroblastoma cases. In the context of chronic NP following limb-sparing or amputation surgeries for bone malignancies, opioids remain essential for pain management; although, anticonvulsants and tricyclic antidepressants are essential first-line NP-directed lines of therapy. Methadone has emerged as a valuable option in this regard, holding a special place among opioids for NP therapy.

Opioids are integral to ensuring quality analgesia in palliative and EOL care for pediatric oncology patients as part of multimodal pain management strategies. Variations in opioid needs based on oncology diagnosis have been observed at EOL, with patients with ST often requiring higher doses and multiple opioids, especially in the last month of life. Moreover, opioid consumption tends to increase significantly toward EOL, particularly in patients with ST diagnoses and NP.

These findings emphasize the need for tailored approaches, continuous monitoring, and further research, particularly multicenter studies, to optimize opioid use and mitigate adverse effects in managing pain across various pediatric oncology contexts. The topics of pain in pediatric oncology contexts of procedural pain and acute postoperative pain are explored in a subsequent review article.

Appendix 1: PubMed/MEDLINE search strategy

(“analgesics, opioid” [MeSH Terms] OR opioid OR opiate OR meperidine OR Codeine OR Fentanyl OR Hydrocodone OR Oxycodone OR Oxymorphone OR Morphine OR tramadol OR methadone OR buprenorphine OR hydromorphone OR tapentadol OR pentazocine OR butorphanol OR dextromethorphan OR loperamide OR OxyContin OR Percocet OR Palladone OR Vicodin OR dolophine OR Demerol OR norco OR Duragesic OR actiq OR dilaudid OR suboxone OR Subutex OR Nucynta OR opana OR talwin OR stadol OR Alfentanil OR Diamorphine OR Levorphanol OR Nalbuphine OR Oliceridine OR Remifentanil OR Sufentanil OR Tapentadol)

AND

(pediatric*[tiab] OR child*[tiab] OR adolescen*[tiab] OR youth*[tiab] OR toddler*[tiab] OR infan*[tiab] OR teen*[tiab] OR prepubescen*[tiab] OR pubescen*[tiab] OR "Pediatrics"[Mesh])

AND

(cancer OR tumor OR chemother* OR neuropath* OR mucositis OR oncolog* OR "bone marrow" OR “neoplasms” [MeSH Terms] OR malignan* OR neoplas* OR oncolog* OR leukemia OR glioma OR blastoma OR osteosarcoma OR tumor OR lymphoma OR carcinoma)