Ketamine Infusions for Chronic Pain

A Systematic Review and Meta-analysis of Randomized Controlled Trials

Vwaire Orhurhu, MD, MPH; Mariam Salisu Orhurhu, MD, MPH; Anuj Bhatia, MD, FRCPC; Steven P. Cohen, MD

Disclosures

Anesth Analg. 2019;129(1):241-254. 

In This Article

Results

Search Results

From the initial 696 records identified through searching databases and other sources, we removed duplicate records and then screened the remaining 467 relevant publications. Among these, 437 were excluded because the publications focused on the perioperative period or failed to report our primary outcome. Thirty full-text articles were assessed for eligibility, with 23 of these excluded because they did not meet inclusion criteria (Figure 1). Seven randomized control trials[23–29] consisting of 211 patients (108 in the IV ketamine groups and 103 in comparator groups; 1 trial had a crossover design with the subjects receiving each of the trial infusions sequentially after the washout period) were included for the systematic review (Table 1), and data from all of these were subjected to meta-analysis.

Figure 1.

Flow chart demonstrating the identification and selection of articles for the systematic review and meta-analysis.

Participants and Interventions

The median (range) sample size of the 7 randomized control trials was 24 participants (range, 19–60). The median (range) age of participants in the 7 randomized control trials was 48 years (range, 41.9–71 years). All trials evaluated chronic pain syndromes, but there was wide variation in type of pain, distribution, and etiology. Two studies (on phantom limb pain[26] and post–spinal cord injury pain[27]) enrolled patients with predominantly neuropathic pain according to the International Association for the Study of Pain definition.[11] Two studies enrolled patients with pain syndromes considered to have a mix of neuropathic and nociceptive characteristics[30,31] (ie, complex regional pain syndrome types I and II[24] and cancer-related pain[25]), and 3 studies[23,28,29] enrolled patients with nociplastic[23,28] or nociceptive[29] (ie, nonneuropathic) pain. In 2 of these, the conditions treated are believed to derive from disorders involving central pain processing rather than discrete nerve injury (complex regional pain syndrome type I[23] and fibromyalgia[28]), while in the other (limb ischemia[29]), pain is predominantly nociceptive. The exclusion criteria based on participants' age and comorbidities were fairly similar in all studies, although the study on fibromyalgia[28] did not state reasons for exclusion of eligible participants. IV ketamine was used as a primary analgesic in all studies. Two studies included the use of opioids in the intervention and control arms throughout the trial,[25,29] subjects in both arms of another study received gabapentin,[27] and 1 study used midazolam and clonidine in both arms.[24]

Ketamine Infusion Regimens

A wide range of infusion protocols was investigated in the studies included in this review.[23–29] The median (range) duration of ketamine infusion for all 7 studies was 5 hours (range, 0.5–100 hours), and the median number of days over which ketamine was administered, continuously or intermittently, was 1 day (range, 1–10 days). The median (range) dose of ketamine during the infusion using a 70-kg patient as the reference was 0.35 mg·kg−1 (0.23–0.6 mg·kg−1). For data available on analgesic effects at the end of the second week after infusion, 3 of 7 studies were included.[23,24,27] One study used the S(+)-isomer of ketamine at an infusion dose range of 0.07–0.43 mg·kg−1·hour−1.[23]

Risk of Bias Assessment

Overall, the risk of bias was considered high in 4 trials[23,27–29] (Figure 2). These trials did not adequately describe the procedures for blinding participants and personnel. The majority of participants in the only 2 studies to assess blinding were able to correctly guess their treatment allocation.[23,28] This suggests that blinding was ineffective in the other trials.[23,24,27–29] Detection bias was high in 2 of the studies.[27,29] Quality of reporting was fair in 6 trials[23–28] and good in 1 trial.[29] There was no difference between the 2 reviewers (V.O. and M.S.O.) for assessment of risk of bias in any study.

Figure 2.

Risk of bias summary. Green circles with "+" sign indicate low risk; red circles with "−" sign indicate high risk; and blank boxes indicate unclear risk.

Primary Outcome

Pain Scores. Six[23,24,26–29] of the 7 studies reported the lowest pain score between 48 hours and 2 weeks after treatments on a 10-point numerical rating scale, and data from these studies were included in the meta-analysis. The study by Salas et al,[25] which reported no benefit from ketamine in patients with refractory cancer pain, could not be included because it did not provide the variance (SDs) for the mean pain scores. Meta-analysis of the data from these trials revealed a significant reduction in pain scores favoring ketamine over standard or control comparative treatments (mean difference, −1.83 points; 95% CI, −2.35 to −1.31 points; P < .0001; I2 = 48.5%; Figure 3).[23,24,26–29]

Figure 3.

Forest plot of the effects of ketamine infusion on maximum reduction in mean numeric rating scale (0–10) pain score. df indicates degrees of freedom; D+L, DerSimonian and Laird.

Dose–Response. We conducted a subgroup analysis to determine whether the effect on our primary outcome differed depending on the dose of ketamine (high versus low), with the high-dose threshold designated as a cumulative dose exceeding 400 mg. Three studies that reported our primary outcome treated patients with high-dose ketamine,[23,24,27] with 2 of these[23,27] reporting lower pain scores in the ketamine as compared to the control groups. A meta-analysis of data from these high-dose ketamine trials demonstrated a significant reduction in pain scores with ketamine compared to placebo (mean difference, −2.11 points; 95% CI, −2.87 to −1.35 points; P < .0001; I2 = 69.2%). Three of the studies that reported our primary outcome of interest treated patients with low-dose ketamine,[26,28,29] with 1 study reporting benefit.[29] Meta-analysis of data from the low-dose ketamine trials revealed an analgesic benefit for IV ketamine compared to the placebo group (mean difference, −1.30 points; 95% CI, −2.01 to −0.59 points; P = .0001; I 2 = 0.0%). When the primary outcome for both categories was compared using meta-regression, the use of high-dose ketamine was not significantly different from the use of low-dose ketamine (P = .213) (Supplemental Digital Content, Figure 1, http://links.lww.com/AA/C794).

Efficacy in Different Types of Pain.We performed a subgroup analysis comparing analgesic efficacy of ketamine infusions in participants with neuropathic and mixed neuropathic–nociceptive pain syndromes (spinal cord injury, phantom limb pain, and complex regional pain syndrome types I and II)[24,26,27] to primarily nociceptive (ischemic limb pain)[29] or nociplastic pain syndromes (complex regional pain syndrome type I and fibromyalgia).[23,28] Three studies investigated patient populations that had either neuropathic[26,27] or mixed neuropathic–nociceptive pain,[24] with 1 study[27] reporting a significant reduction in pain scores with ketamine infusion. Meta-analyses of these data demonstrated a significant reduction in pain scores with ketamine compared to placebo control groups (mean difference, −1.75 points; 95% CI, −2.08 to −1.43 points; P < .00001; I2 = 0.0%). For the nonneuropathic (nociceptive or nociplastic) pain category, 1[23] of the 3 trials[23,28,29] reported lower pain scores favoring IV ketamine. This trial enrolled patients with a nociplastic pain syndrome (complex regional pain syndrome type I).[23] Meta-analysis of these data revealed a significant reduction in pain scores with IV ketamine (mean difference, −1.97 points; 95% CI, −3.04 to −0.90 points; P < .00001; I 2 = 69.5%). However, there was no significant difference when the primary outcomes for both groups were compared using meta-regression (P = .720) (Supplemental Digital Content, Figure 2, http://links.lww.com/AA/C794).

We also performed a subgroup analysis comparing studies that enrolled participants with complex regional pain syndrome compared to those whose participants did not have complex regional pain syndrome. Two trials investigated patients with complex regional pain syndrome[23,24] with 1[23] revealing lower pain scores with ketamine as compared to placebo control groups. Meta-analysis of these data demonstrated a significant reduction in pain with IV ketamine compared to control (mean difference, −2.38 points; 95% CI, −3.53 to −1.23 points; P < .0001; I2 = 34.9%). Four trials investigated patients without complex regional pain syndrome.[26–29] Two[27,29] of these studies, which enrolled patients with neuropathic pain after spinal cord injury[27] and ischemic limb,[29] reported lower pain scores favoring IV ketamine compared to placebo control group, with the studies performed for phantom limb[26] and fibromyalgia[28] pain reporting a lack of efficacy. Meta-analysis of these data demonstrated a significant reduction in pain scores with ketamine (mean difference, −1.71 points; 95% CI, −2.01 to −1.41 points; P = .001; I 2 = 0.0%). When the primary outcomes for both categories were compared using meta-regression, there was no significant difference between groups (P = .079) (Supplemental Digital Content, Figure 3, http://links.lww.com/AA/C794).

Efficacy With Adjunct Medications Versus Stand-Alone Therapy. We performed a subgroup analysis comparing studies using only ketamine[23,24,28] to studies that used analgesic adjuncts (opioids, gabapentin, and calcitonin) in the group receiving ketamine infusion.[26,27,29] Medications administered with the expressed purpose to prevent adverse effects (eg, midazolam and low-dose clonidine) were not considered analgesics.[24] When the primary outcome for both categories were compared using meta-regression, there were no significant differences (P = .127) (Supplemental Digital Content, Figure 4, http://links.lww.com/AA/C794).

Figure 4.

Adverse effects associated with the use of intravenous ketamine. A, Forest plot displaying risk of nausea and 95% CIs with IV ketamine compared to placebo. B, Forest plot displaying risk of psychotomimetic effect and 95% CIs with IV ketamine compared to placebo. C, Forest plot displaying risk of headache and 95% CIs with IV ketamine compared to placebo. D, Forest plot displaying risk of tiredness and 95% CIs with IV ketamine compared to placebo.

Efficacy of Ketamine at Different Time Points. Three trials reported pain scores at 2 weeks posttreatment.[23,24,27] Two of these found lower pain scores in the ketamine group compared to placebo.[23,27] Meta-analysis of these data revealed a significant reduction in pain scores favoring IV ketamine (mean difference, −2.23 points; 95% CI, −2.59 to −1.87 points; P < .001; I2 = 0.0%) (Supplemental Digital Content, Figure 5, http://links.lww.com/AA/C794). Three trials[23,24,27] reported pain scores 4 weeks after treatment, with 1[23] finding lower pain scores in the ketamine group. Meta-analysis of these data revealed a nonsignificant reduction in pain scores with IV ketamine compared to placebo (mean difference, −0.74 points; 95% CI, −1.88 to 0.41 points; P = .208; I2= 58.6%) (Supplemental Digital Content, Figure 6, http://links.lww.com/AA/C794). Three trials[23,24,28] reported pain scores at 8 weeks (2 months) posttreatment, with 1[23] reporting lower pain scores in the IV ketamine group. Meta-analysis of these data yielded a nonsignificant reduction in pain scores with ketamine compared to placebo (mean difference, −0.68 points; 95% CI, −1.75 to −0.40 points; P = .174; I2 = 48.2%) (Supplemental Digital Content, Figure 7, http://links.lww.com/AA/C794). Two trials[23,24] reported pain scores 12 weeks (3 months) after treatment, with neither reporting lower pain scores in the IV ketamine group. Meta-analysis of these data revealed a nonsignificant reduction in pain scores with IV ketamine compared to placebo (mean difference, −0.55 points; 95% CI, −1.50 to 0.39 points; P = .251; I 2 = 0.0%) (Supplemental Digital Content, Figure 8, http://links.lww.com/AA/C794). In summary, administration of ketamine resulted in a significant reduction of pain scores when compared to placebo at 2 and 8 weeks (but not 4 weeks) after treatment.

Secondary Outcomes

Positive Response. Three of the 7 studies in this review reported the proportion of patients achieving a positive analgesic response (as designated by the authors) with ketamine compared to placebo.[25,26,28] One study used a ≥30% reduction in pain as the criterion for a positive outcome,[25] whereas the other 2 used a ≥50% reduction as the threshold.[26,28] A meta-analysis of these studies demonstrated analgesic superiority for IV ketamine compared to placebo control based on the study-designated definition of success[25,26,28] (19.4% vs 51.3%; relative risk, 2.43; 95% CI, 1.10–5.40; P = .029; I 2 = 0.0%) (Supplemental Digital Content, Figure 9, http://links.lww.com/AA/C794).

Adverse Effects. Common ketamine-related adverse outcomes reported in the trials included nausea, vomiting, psychotomimetic effects, headache, fatigue, and sedation. Compared to placebo, the ketamine group had significantly higher relative risk of nausea (relative risk, 3.52; 95% CI, 1.74–7.14; P < .00001; Figure 4A) and psychotomimetic effects (relative risk, 5.92; 95% CI, 2.95–11.89; P < .00001; Figure 4B). The relative risk of headache (relative risk, 1.26; 95% CI, 0.67–2.34; P = .475; Figure 4C) and tiredness (relative risk, 2.16; 95% CI, 0.64–7.33; P = .218; Figure 4D) was higher among the ketamine group, but the differences were not statistically significant. Nausea and vomiting were also higher in the ketamine group, but this was reported by one study. The incidence of these side effect measures is reported in Supplemental Digital Content, Table 2, http://links.lww.com/AA/C794.

Heterogeneity. For the primary outcome, the I2 statistic was 45.9% for the meta-analysis of IV ketamine use from 6 randomized control trials. These results indicate the presence of a moderate degree of heterogeneity. Several characteristics of these studies may have contributed to heterogeneity including differences in the types of pain conditions and patient populations, time periods for assessing the primary outcome, and variations in the ketamine infusion regimen and comparators explored. To explore the sources of heterogeneity, we conducted subgroup analyses using meta-regression and a sensitivity analysis. We found no significant differences between the following groups: mixed or neuropathic versus nociceptive or nociplastic pain, complex regional pain syndrome versus non–complex regional pain syndrome, inpatient setting (n = 5 studies)[23,25–27,29] versus outpatient setting (n = 2 studies),[23,24] studies with >60% females versus those with less, and studies with a median age >46 years versus those under 46 years. A sensitivity analysis was performed by sequentially removing each individual trial and evaluating how it affected the pooled estimate of the primary outcome, but this process failed to find a significant difference (Supplemental Digital Content, Figure 10, http://links.lww.com/AA/C794).

Publication Bias. To elucidate this possibility, we performed Begg and Egger tests, with the nonsignificant P values (P = .380 and .423, respectively) suggesting an absence of publication bias. We did not display a funnel plot because the studies in our analysis did not meet the minimum recommended number of studies required for interpreting a funnel plot (usually 10).[18,32]

Grade Evidence. We provided evidence as per GRADE recommendations[22] for the primary outcome, secondary outcomes, and complication profile based on risk of bias, inconsistencies, indirectness, imprecision, magnitude of effect size, evidence of dose–response, and other methodological shortcomings. Based on these considerations, the grade of evidence for the analgesic efficacy of IV ketamine in chronic pain was determined to be low (Table 2). For similar reasons, the grade of evidence for categorical response (≥30% or ≥50% pain relief) and efficacy at 2 weeks after the infusion was also deemed to be low. Our analysis suggests that we have limited confidence in the analgesic effect of IV ketamine in chronic pain patients. However, the true effect with IV ketamine use may be significantly different from the proposed estimates in our study.

Comments

3090D553-9492-4563-8681-AD288FA52ACE

processing....