To systematically review randomized, controlled trials that investigated the efficacy of acupressure for the management of symptoms.

Databases searched were CINAHL, MEDLINE, and PubMed.

Search keywords were acupressure, clinical trial, human, and/or randomized.

Studies were included in the review if

• They were randomized, controlled trials published from January 1, 2000 to January 31, 2010
• They were published in English
• They used acupressure as the sole intervention for one group
• There were four or more studies of the efficacy of acupressure for that particular symptom.

Studies were excluded from the review if they

• Had sample sizes of less than 30 patients
• Used auricular or hand pressure, reflexology, shiatsu, and electronic or magnetic devices
• Were unpublished studies and abstracts.

In total, 108 references were screened.

Each experimental study was evaluated for quality using the Cochrane risk of bias (RoB) tool. Articles were evaluated for the presence of each of the six domains, and one point was assigned for each domain present. Scores ranged from zero to six, with higher values indicating higher quality and less risk for bias. All trials were evaluated by two authors, and agreement between them was 100%. A significant likelihood of bias was found in the evaluation.  

Only six randomized, controlled trials included were performed to determine the efficacy of acupressure on the reduction of fatigue and improvement of sleep in adults in various populations. Fatigue and insomnia were grouped together because those investigations were typically studied simultaneously.  
 

• The final number of studies included was 43.
• The total sample size was 5,021 patients across all studies of all symptoms. For studies that specifically focused on fatigue and insomnia, the sample size was 211 patients.
• The sample range across all studies was 36 to 739 for all studies of all symptoms.
• Multiple diagnoses and symptoms were included:  nausea, pregnancy, chemotherapy, radiation, surgery, motion sickness, pain, dysmenorrhea, labor, back pain, fracture, trauma, dyspnea, asthma, chronic obstructive pulmonary disease (COPD), bronchiectasis, endstage renal disease (ESRD), and insomnia.

Patients were undergoing the active treatment phase of care.

Only three studies conducted included a measure of fatigue. Six studies concluded that acupressure was effective in improving fatigue and reducing insomnia. Multiple symptoms were discussed in topic areas; findings from other symptom reviews were not included in the summary.

The review did not provide rigorous support for the use of acupressure for the efficacy of symptom management.

• Significant bias existed according to the Cochrane RoB tool. 
• Lack of fidelity to the intervention confounded the results and may have added bias to the studies.

Well-designed randomized, controlled trials are needed to determine the utility and efficacy of acupressure to manage various symptoms in several patient populations. Issues exist when looking at fatigue and insomnia concurrently, such as determining whether fatigue is an intervening variable for insomnia or an outcome variable of insomnia or whether insomnia is an intervening variable for fatigue. A conceptual framework is needed to guide how concurrent or symptom clusters are studied.

To assess the efficacy and safety of an herbal decoction including Astragali Radix in improving appetite and body weight and changing cytokine levels in patients with advanced cancer experiencing anorexia

The herbal decoction consisted of a combination of herbs including Astragali Radix made in the hospital pharmacy every week. The decoction was administered 30 minutes after meals, three times a day for three weeks. The decoction was stopped for treatment refusal, the aggravation of anorexia or weight loss in the midst of therapy, or physician decision. Assessments of appetite were performed at baseline and then weekly until termination of the study. Cytokines were checked at baseline and at the third week.

• The study reported on 11 patients.
• Mean patient age was 59.8 years (range = 42–76 years).
• The sample was 72.7% male and 27.3% female.
• Patients were diagnosed with lung cancer, gastric cancer, or colorectal cancer.
• Use of chemotherapy or radiation therapy was permitted throughout the study period.
• Exclusion criteria included use of tube feedings or parenteral nutrition; simultaneous treatment with adrenal corticosteroids, androgens, progestational agents, or other appetite stimulants; brain metastasis; pregnancy or lactation; and mechanical obstruction of the alimentary tract.

• Single site
• East-West Neomedical Center at Kyung Hee University in Seoul, Republic of Korea

• Patients were undergoing multiple phases of care.
• The study has clinical applicability for end-of-life and palliative care.

A prospective, phase II trial design was used.

Toxicity was assessed as well as tumor response using Recist criteria.

• History and physical exam
• Visual analog scale (VAS) for anorexia
• Body weight
• Skinfold thickness
• Cytokine levels: IL1β, IL-6, TNF-α, and lactate

Primary endpoint: VAS anorexia scores before and after the herbal decoction were significantly improved (p = 0.008). Secondary endpoints: 55% of patients had improvements in both anorexia and body weight at the third week (p = 0.009). The correlation of improvement of anorexia and body weight at the best point was also significantly positive (p = 0.003). The changes in skinfold thickness of the triceps, subscapula, and suprailiac did not significantly change (p = 0.182). The cytokines did not significantly change. Lactate and lactate dehydrogenase (LDH) levels were significantly worse at three weeks. Eight patients were administered the full treatment of the herbal decoction, three patients had the decoction for two weeks, one patient had treatment stopped due to toxicity of concomitant chemotherapy, and two patients had disease progression. Grades 1/2 diarrhea occurred in two patients. Grade 1 dyspepsia occurred in two patients. Grade 1 alanine transaminase abnormality occurred in one patient. Grade 2 neutropenia and grade 2 anemia occurred in two patients who also received concomitant chemotherapy.

This small study demonstrated potential for a combination herbal product to improve anorexia in patients with advanced cancers. Although patients showed significant improvements in anorexia scores and weight gain, cytokine and lactate and LDH values did not improve significantly and in some cases worsened, which is contrary to current thinking regarding the chemical picture of anorexia. Results of this study should be interpreted with caution.

• The study had a small sample size, with less than 30 participants.
• The study was not randomized, so placebo effect cannot be ruled out.
• Cytokines and VAS scores were not evaluated at the same intervals (cytokines were checked at baseline and the third week, and anorexia VAS scores were evaluated at baseline and weekly thereafter). Cytokine changes may have been missed. Changes in anorexia may not have been affected by cytokines.
• The study did not meet its projected accrual number of 30 patients and closed early because of an extended recruitment process. Sample size may have been too small to notice a change in the short period assessed.

Anorexia is a common symptom in patients with cancer. There are some drugs such as corticosteroids and megestrol acetate that can help to control anorexia. Complementary and alternative medicines may help treat cancer-related symptoms. Nurses need to remain knowledgeable about potential medications and complementary agents that may be useful in managing patients’ symptoms.

To test the efficacy of doxepin hydrochloride in reducing radiotherapy-induced oral mucositis pain

Arm 1: Doxepin oral rinse 10 mg/ml x 2.5 ml = 25 mg diluted with sterile or distilled water for a total of 5 ml. Arm 2: Placebo rinse prepared in similar manner with flavored syrup added. Study dose was prepared by unblinded nurse or pharmacist. Blinded personnel administered the rinse. On day 1, the patients swished for one minute, gargled, and spit. On day 2, patients received the opposite rinse. Patients completed symptom questionnaires at set time intervals during the hour following administration and then again at home at two and four hours. No analgesics or oral rinses were allowed 60 minutes prior to and following the study doses. Patients returned the next time their oral pain was ≥ 4/10 and received their second study dose, after which they were unblinded and given the option to continue use of doxepin rinses prn. Patient-reported questionnaires with numerical pain analog scale (0–10) done at BL, 5, 15, 30, 60, 120, and 240 minutes after the rinse on days 1 and 2 to measure pain, taste, stinging or burning, and drowsiness.

• N = 129 
• MEAN AGE = Doxepin-Placebo, 62 years (range = 39–93 years), Placebo-Doxepin, 60 years (range 37–86 years)
• MALES: 80%, FEMALES: 20%
• KEY DISEASE CHARACTERISTICS: Patients 18 years or older with head and neck malignancy currently undergoing radiotherapy, with or without chemotherapy, with a minimum planned dose of 50 Gy, including one third of the oral cavity mucosa using 1.6–2.2 Gy per fraction currently experiencing oral mucositis without infection and mouth pain rated 4 or higher (0–10 scale) and ECOG 0-2.
• OTHER KEY SAMPLE CHARACTERISTICS: No known allergies to doxepin or TCAs, use of TCAs or MAO inhibitors two weeks prior, no active or untreated herpes simplex virus or oral candidiasis, and no untreated narrow angle glaucoma or urinary retention within six weeks of registration. Patients were stratified according to sex, concurrent sensitizing chemotherapy, and age (< vs ≥ 60 years of age). Patients had to receive an oral examination with confirmed oral mucositis and without oral infection to be enrolled. 

• SITE: Multi-site
• SETTING TYPE: Outpatient
• LOCATION: 26 radiation centers in the United States

• PHASE OF CARE: Active antitumor treatment
• APPLICATIONS: Elder care, palliative care

Randomized, double-blind, placebo-controlled, phase III trial with crossover phase followed by continued use of active agent

• Questionnaires were based on the Oral Mucositis Daily Questionnaire and the Oral Mucositis Weekly Questionnaire-Head.  
• CTCAEv.4.0 was used for toxicities.

Primary outcome: AUC for mean mouth and throat pain reduction was greater for doxepin compared with placebo (–4.4 difference p < .001). Crossover data revealed similar findings with –2.3 difference p < .001.

Largest placebo-controlled trial to date for oral mucositis pain with positive results

• Doxepin can cause drowsiness, caused more stinging and burning, and had a worse taste than placebo. Patients may have suspected what arm they were receiving because of these side effects, and it may have affected their decision about continuing with the doxepin rinses rather than efficacy alone. 
• Lack of longitudinal comparison data beyond one or two doses.
• Minimal data was collected in the continuation phase.
• No comparison with magic mouthwash. 
• Further studies are needed to confirm results. Prevention of mucositis was not looked at in this study, which is a much needed indication that should be researched.   

This is a promising option for mucositis pain although there are concerns about how the solution will be made, the cost, the ease of use, the side effects, and adherence to the rinse. Additional questions exist concerning whether or not this data can also be applied to patients receiving chemotherapy or those who have cancers other than head and neck cancer.

To identify and review all relevant data to determine whether lactulose or polyethylene glycol (PEG) is more effective in treating chronic constipation.

Databases searched were Medline, Embase, CINAHL, and the Cochrane Central Register of Controlled Trials. Bibliographies of studies and conference proceedings  were hand searched.

Search keywords were fecal impaction, chronic constipation, delayed bowel movement, obstipation, irregularity, polyethylene glycol, ethylene glycol, PEG, ethylene oxide, PEG 3350, lactulose, and disaccharide.

Studies were included in the review if they

• Were randomized controlled trials comparing the use of lactulose and PEG
• Reported on adult patients or children diagnosed with chronic constipation or fecal impaction.

One hundred three initial references to January 24, 2008, were identified. A final set of 10 trials was included in this review. Studies were evaluated for various sources of bias and overall methodological quality. Trials were conducted in different countries in a variety of settings.

• The 10 trials included a total of 868 participants.
• 322 were adults and 546 were children.

• In most studies, assessment of study quality by investigators yielded highly variable results and little agreement of bias issues.
• In five studies, the overall mean difference in stool frequency per week was 0.65 (95% confidence interval [CI] [0.15, 1.15], p = 0.01) in favor of PEG.
• Among three studies in children, the mean difference in stool frequency per week was 1.57 (95% CI [0.36, 2.77], p = 0.011) in favor of PEG.
• In two studies that measured stool form and consistency, the mean difference was 0.89 (95% CI [0.43, 1.45], p = 0.00015) in favor of PEG.
• In one study of adults and two studies in children, the odds ratio for relief of abdominal pain was 2.09 (95% CI [1.26, 3.44], p = 0.004) in favor of PEG.
• Across three studies, patients on PEG had much less need for additional products to manage constipation (p = 0.000078).

• Findings suggest that PEG may be more useful than lactulose in both children and adults for improvement in constipation; however, the evidence is insufficient to draw firm conclusions because of the small number of studies and degree of heterogeneity found.
• Studies had a number of methodological issues.
• This review involved studies that did not include an oncology population.

One study included a sample of patients on methadone maintenance, which may have some relevance for constipation associated with opioid use.

To evaluate the effectiveness of OROS hydromorphone in reducing the use of medication for breakthrough pain (BTP) in Korean patients with chronic cancer pain

Patients were stabilized on the existing pain regimen for three days. After stabilization, pain medication was converted to OROS hydromorphone at a dose equivalent to oxycodone in effectiveness. Dose equivalence was calculated using the ratio 2.5:1 controlled-release oxycodone to hydromorphone hydrochloride. The minimum starting dose of hydromorphone was 8 mg, which was provided in the form of 8, 16, or 32 mg tablets. Patients took the medication at 8 am each morning, swallowing tablets whole, with water, and without chewing, dividing, or crushing. Immediate-release hydromorphone was used as the rescue medication at a dose equaling 10%–15% of OROS hydromorphone. No other opioid medications were permitted postconversion to OROS hydromorphone. Patients were allowed to use nonopioid and adjuvant analgesics. On day 7, if the patient required more than four rescue doses per day, medication was increased in increments of 8 or 16 mg. Investigators collected data at baseline and on days 7 and 14.

• The sample was composed of 114 patients.
• Mean patient age was 58.2 years (SD = 11 years).
• Of all patients, 67% were male and 33% were female.
• The sample comprised multiple cancer sites. Of all cancer diagnoses, 84% had metastasized.

• Multisite
• Outpatient
• South Korea

• Phases of care: multiple phases of care
• Clinical application: palliative care

Prospective open-label phase IV study

• Pain diary, to record frequency of BTP and its subtypes
• Mean of medication frequency of BTP (Investigators calculated the mean by averaging the use of BTP medications for the three days prior to the study visit.)
• Korean Brief Pain Inventory, 0–10 scale, to measure pain intensity
• Scale, 0–100, to measure pain relief
• Scale, 0–10, to measure pain-related interference with general activities, mood, ability to walk, normal work activities, relationships with others, sleep, and enjoyment of life
• European Organization for Research and Treatment of Cancer Core Quality of Life questionnaire

Compared to the mean of BTP frequencies at baseline, the mean of BTP frequencies decreased significantly (p < 0.001) on days 7 and 14. Also compared to baseline, BTP subtypes decreased on days 7 and 14. Compared to baseline measures, measures of pain intensity at its worst (p < 0.001) and average pain intensity (p < 0.005) decreased on day 14. Compared to baseline measures, measures of physical functioning improved significantly (p = 0.015) on day 14. Emotional, cognitive, and social functioning were stable after drug conversion (p > 0.05). At the end of the trial, 61.2% of patients rated the medication as effective and 91.2% of patients reported mild or moderate adverse events. (Authors reported that 32% of adverse events were considered related to study treatment.) Of all patients, 88% preferred the study regimen to their previous regimen, based on ease of use.

OROS hydromorphone was effective in reducing pain intensity in patients with chronic cancer pain. Physical function improved over 14 days while cognitive and emotional functioning remained intact. Although most patients reported adverse events, the majority of events were mild or moderate and unrelated to the study drug. In regard to ease of use, the overwhelming majority of patients preferred the study regimen to their previous regimen.

The study had risks of bias due to no control group, no blinding, and no random assignment.

Because this study showed that OROS hydromorphone therapy could decrease pain severity and that patients found the medication easy to use, clinicians should consider the drug in the treatment of patients with chronic cancer-related pain. The results of this study warrant a randomized, blinded trial.

STUDY PURPOSE: To investigate the effects of massage therapy on cancer pain

TYPE OF STUDY: Meta-analysis and systematic review

• FINAL NUMBER STUDIES INCLUDED = 12  
• TOTAL PATIENTS INCLUDED IN REVIEW = 559
• SAMPLE RANGE ACROSS STUDIES: Not evaluable
• KEY SAMPLE CHARACTERISTICS: Varied types of cancer and cancer treatments

PHASE OF CARE: Multiple phases of care
 
APPLICATIONS: Palliative care

Massage was associated with lower pain (SMD = –1.25, p = 0.0001). Subgroup analysis according to type of intervention, massage, aromatherapy massage, or foot reflexology showed positive effect of each type, though few studies in each subgroup existed. PEDro scores indicated that only three studies were low quality, and the rest were high quality; however, risk of bias evaluation varied widely and showed that most studies had moderate to high risk of bias. Effects of massage were more positive in those studies with higher risk of bias.

Findings show that massage therapy has positive benefits for cancer-related pain.

• High heterogeneity
• Sample sizes within studies werenot provided
• Given total patients across 12 studies, some of these must have been small. 
• No differentiation between acute or chronic pain

Massage therapy is effective in reducing pain among patients with cancer. This low-risk intervention can be helpful in pain management.

STUDY PURPOSE: To evaluate the effects of herbal medicine as adjunctive treatment for cancer-related pain

• FINAL NUMBER STUDIES INCLUDED = 24 (four included in the meta-analysis)
• TOTAL PATIENTS INCLUDED IN REVIEW = 4,045
• AVERAGE SAMPLE SIZE: 70 patients (One Chinese study included 2,466 patients.)
• KEY SAMPLE CHARACTERISTICS: Various tumor types

The results of this meta-analysis showed an overall standard mean difference in favor of the herbal intervention across four studies (SMD = -0.51, p < 0.05). In a second meta-analysis, the risk ratio reported was 1.6. The direction of this risk ratio was not clear in the results reported. A meta-analysis was done for studies using herbal interventions in combination with traditional cancer therapy versus only traditional cancer therapy rather than treatment aimed specifically at pain control. Types of in interventions included botanicals, processed animal products, and processed minerals. In some studies, herbal interventions were compared to chemotherapy, aspirin, NSAIDs, Tylenol, diets, placebos, or unnamed analgesic drugs. Most studies showed a high risk of bias according to a study evaluation. A single study that was of high quality across all aspects showed no effect for the intervention.

Although the quantitative analysis showed a statistic significance, the comparisons studied for pain relief did not appear to be appropriate. Although herbal medicine may have better results on pain than no intervention or low-level analgesics, the efficacy of herbal medicine for pain should compare results with those obtained from other interventions shown to be effective for cancer-related pain.

Most studies were of low quality. Many study comparison groups were not adequate because well-recognized, effective interventions for pain were not used in the controls. The results reported in the meta-analysis of risk ratios were unclear. The methods described the use of an odds ratio analysis, but a risk ratio was reported, and it was not clear which study group was considered at-risk.

The findings of this analysis suggested that traditional herbal medicine may have a role in cancer-related pain control. However, the included studies did not provide strong evidence. The potential effects of herbal medicine for pain management remain unclear. Additional well-designed research, including comparisons to pain interventions with strong supportive evidence, are needed to determine any benefits of herbal medicine.

STUDY PURPOSE: To evaluate evidence of the effectiveness of education interventions in pain management

• FINAL NUMBER STUDIES INCLUDED = 27 with ten randomized, controlled trials included in the meta-analysis
• TOTAL PATIENTS INCLUDED IN REVIEW = 2,380
• SAMPLE RANGE ACROSS STUDIES: 30–1,256 patients
• KEY SAMPLE CHARACTERISTICS: Not provided

The most frequent educational used tool was a booklet. A few studies involved the discussion of self-control and relaxation, but most of the content was description of pain management and misconceptions regarding the use of opioids. An analysis of seven randomized, controlled trials' (using the Brief Pain Inventory) effects on severe pain showed an SMD –0.34 (CI = 95%, –0.55, –0.13, p = 0.001). The SMD of average pain was –0.73 (CI = 95%, –0.64, –0.15, p = 0.002). The effects on pain in studies using the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire showed no significant effect.

This analysis suggests that educational interventions have a positive effect on pain control.

• High heterogeneity in the analysis
• Diverse educational interventions
• The analysis included a few studies that also incorporated more psychoeducational types of interventions; however, these were included with simple educational booklet use in the meta-analysis.
• Most of the studies had an undetermined risk of bias.

The findings of this meta-analysis need to be viewed with some caution given the heterogeneity found in the analysis, the risk of bias determined. and the lack of differentiation between the education provided via booklets or other materials and the psychoeducational interventions that also included counseling, discussion, and relaxation techniques. However, it is clear that patients who need pain management also need education for self-management. Yet it is not clear if simple education alone is sufficient to facilitate effective patient self-management with analgesics.

To describe pain relief in patients with spinal metastasis treated with the CyberKnife^® Robotic Radiosurgery System, and to analyze the factors associated with a relapse in pain after pain relief

Radiosurgery was performed using CyberKnife^®, a unique, image-guided, frameless stereotactic radiosurgery delivery system. The CyberKnife^® is a lightweight 6-MV linear accelerator (LINAC) attached to the end of a robotic arm with six axes of freedom. An MRI or PET-CT scan was obtained of all patients prior to treatment to evaluate the extent of spinal lesions, and the MRI was fused with a planning CT to better contour the target volume. An appropriate immobilization device was used to treat patients in the supine position. The clinical target volume (CTV) and the planning target volume (PTV) were the same, a 1–2 mm margin around the gross tumor volume. The delivered dose was kept at less than 13 Gy in a single fraction and less than 21.4 Gy in three fractions. The total dose of radiation delivered to the spinal cord was 15–35 Gy in one to five fractions. Pain was measured before radiosurgery, “soon after” radiosurgery, and then every one to three months (last measurement was done by telephone). Analgesic use was measured before and after treatment.

• N = 57
• AGE: Not reported 
• MALES: 54%, FEMALES: 46%
• KEY DISEASE CHARACTERISTICS: All patients had one or more spinal lesions related to metastatic cancer. Primary tumors included breast, liver, lung, cervix, colorectum, biliary tract, stomach, prostate, and others. 
• OTHER KEY SAMPLE CHARACTERISTICS: Seventy percent of the lesions caused moderate to severe pain before radiosurgery (six lesions did not cause pain). All patients were on opioid analgesics. Fourteen patients had received previous external beam radiotherapy for their spinal lesions; the median duration between radiosurgery and prior radiotherapy was 10 months. Three patients had a previous laminectomy or corpectomy and were then receiving radiosurgery for disease progression.

• SITE: Not stated/unknown  
• SETTING TYPE: Outpatient  
• LOCATION: Although not explicitly stated, it is assumed that the study was conducted in Korea at either the department of radiation oncology at Konyang University or Ajou University or both.

• PHASE OF CARE: Mutliple phases of care
• APPLICATIONS: Palliative care 

Retrospective pre/post design

Pain was measured using a 0–10 Visual Analog Scale (VAS). Pain relief was considered a reduction in the pain score by three levels, and pain progression was considered an increase in the pain score by three levels. No use of analgesics or a score of 0 or 1 was considered complete relief. The latency to pain relief was calculated beginning from the first day of radiosurgery. The duration of pain relief was measured from the first day of pain relief to the time of pain progression or until the last follow-up. Analgesic use was recorded before and after treatment.

Pain relief was obtained in 88% of the painful lesions (59 of 67 lesions), and complete relief was obtained in 51% (34 of 67 lesions). No radiation-induced myelopathy occurred. Those patients (n = 6) who did not have painful lesions prior to radiosurgery remained pain-free without disease progression. For the majority of patients who received pain relief from treatment, pain relief was obtained within seven days of beginning radiosurgery. The median duration of pain relief was 3.2 months (range = 1–30 months). Factors associated with pain relapse were the presence of spinal cord compression (p = 0.001) and an Eastern Cooperative Oncology Group (ECOG) performance status of ≥ 2 (0 = 0.01).

Although the sample size was small, this study found similar results as previous studies examining the use of spinal radiosurgery for pain relief. CyberKnife^® is an effective method for treating pain related to spinal metastasis. Long-term follow-up of patients treated with spinal radiosurgery is needed to determine the duration of pain relief and any long-term effects related to treatment.

• Small sample (< 100)
• Risk of bias (no control group)
• Risk of bias (no blinding): No blinding of those reviewing radiological records; no blinding of patients  
• Risk of bias (no random assignment): Not a randomized, controlled trial; retrospective review, so there is a risk of bias
• Risk of bias (no appropriate attentional control condition)  
• Findings not generalizable: Not a randomized, controlled trial; findings similar to other studies
• Other limitations/explanation: Patient demographics not reported; difficult to generalize

The use of spinal radiosurgery in treating pain related to metastatic cancer is an important palliative care treatment for pain relief as patients tolerate it well and there appear to be no side effects. Although CyberKnife^® treatment is not available at all hospitals and medical centers, nurses should be aware of this effective pain relief intervention.

TOTAL REFERENCES RETRIEVED: 1,027

• FINAL NUMBER STUDIES INCLUDED = 16
• TOTAL PATIENTS INCLUDED IN REVIEW = 1,120 
• SAMPLE RANGE ACROSS STUDIES: 6–48 patients
• KEY SAMPLE CHARACTERISTICS: Mean age was 49 years; various tumor types; cancer therapies consisted of chemotherapy, radiotherapy, chemoradiotherapy, and hematopoietic stem cell transplantation

PHASE OF CARE: Multiple phases of care

Outcome mineral derivatives included zinc (n = 549, seven studies), calcium phosphate (n = 227, three studies), povidone-iodine (n = 228, two studies), and selenium (n = 116, two studies). The severity of OM was measured across variable OM grading systems. In 13 studies, individuals in treatment groups (n = 958) experienced peak OM less than controls (g = −0.47,95% CI −0.7 to −0.2, p = 0.0006). In five studies, time to OM onset was significantly delayed in treatment versus controls (g = −0.51, 95% CI−0.8 to −0.2, p = 0.0002), but the mean duration of OM, pain incidence, and analgesic use was not significantly different. 

The general positive effect trend suggested that individuals taking mineral derivatives during cancer therapies were less likely to experience peak OM. However, the significant bias and heterogeneity in this analysis indicated the need for additional methods because of diverse protocols and novel recordings (serum mineral levels and cell signals) in estimating a uniform true effect. The decision analysis favored selenium.

This review was limited in recommending a definitive mineral derivative to alleviate OM. Limitations included high heterogeneity implicated by variable conditions (i.e., different protocols, diverse cancer therapies). In addition, a placebo effect may have undermined nonblinded studies at a high risk of bias.

Future trials should consider serum levels and computer simulations when designing mineral tolerance thresholds to weigh benefits and harms.