In most of the studies assessed, the indications for implantable devices were weakly defined, according to the terms used, previous analgesic treatments, and outcomes. The definition of refractory as “a condition in which a patient has failed to respond to multiple therapies provided by a specialist (pain or palliative care physician) or at sufficient doses for adequate durations” [3] was almost never reported, regardless of the positive outcomes.

In the first large multicenter controlled randomized study, an IDDS improved clinical success in pain control, reduced pain, significantly relieved common drug adverse effects, and improved survival in patients with “refractory cancer pain” in comparison with patients receiving a comprehensive medical management [20]. However, no definition of refractory pain was provided. Thus, patients with severe (but not refractory) pain (about 7/10) were randomized to receive a comprehensive pain management or intrathecal therapy and comprehensive pain management. It is unclear what was considered a comprehensive pain management and who was involved, for example, whether the medical management included different trials of opioids to optimize opioid analgesia and/or symptomatic treatment of adverse effects. Of interest, the changes in pain intensity were similar in both groups, with a decrease from mean 7.81 and 7.57 to 4.76 and 3.67 in patients receiving medical treatment alone and intrathecal therapy plus medical treatment, respectively. Initial opioid dosages, expressed as oral morphine equivalents (OME), were similar in the two groups (272 mg/day and 250 mg/day, respectively). By 4 weeks, dosages were slightly increased up to 290 mg/day in patients receiving medical treatment alone, and decreased to 50 mg/day in the other group. This means that just a minimal increment of systemic opioid doses allowed good pain control. On the other hand, intrathecal dosages largely increased up to 2 mg/day intrathecally, which is equivalent to about 600 mg/day of OME when using a ratio of 1:300. Thus, the sum of opioid amounts was largely superior in patients receiving the intrathecal treatment. Of interest, the authors disclosed that medical management was unexpectedly effective, underlining that just a minimal increase of systemic opioid doses reproduced an analgesia similar to that of an implantable system. Sixteen patients (16%) had adverse events related to implanted pump or related procedures. There were three explantations and nine pump revisions.

In another well-numbered randomized controlled study comparing intrathecal hydromorphone and morphine spinal delivery, the indications for intrathecal therapy were as follows: “(failed?) standard treatment and opioid rotation according to the guidelines, an OME dosage greater than 200 mg/day with unsatisfactory analgesia or patients with intolerable opioid-related adverse effects, an average pain intensity ≥ 5.0, and breakthrough pain episodes ≥ 3 times a day” [6]. It is unclear which guidelines authors refer to in terms of indications for spinal therapy. Moreover, information regarding the changes in OME after implantation were not reported. The starting infusion rate was based on baseline OME (conversion ratio 1:300). The equivalence of hydromorphone to morphine was set to 0.15:1. The 233 patients were on a mean dosage of 360 mg/day of OME. After intrathecal implantation, the pain intensity in the two groups significantly decreased from baseline values. Clinical success rate (defined as percentage of patients with pain relief rate ≥ 50%) was 72.8% and 74.0% in hydromorphone and morphine groups, respectively. The average starting daily infusion dosages were 0.276 mg/day and 1.55 mg/day for intrathecal hydromorphone and morphine, respectively. A lower rate of dose increase in the hydromorphone group was reported. Of interest, the occurrence of breakthrough pain (BTP) was used as a signal to modify the daily dose of the intrathecal infusion the day after, but this does not represent a treatment for BTP. Hydrophilic molecules such as morphine and hydromorphone have a slow onset of analgesia when given intrathecally, therefore needing many hours to move rostrally to provide an effect. Thus, the use of hydrophilic opioids by the intrathecal PCA route is not justifiable from a mere pharmacologic point of view.

The third randomized controlled trial was a multicenter, double-blind, and placebo-controlled study performed in a mixed population of patients with cancer and acquired immunodeficiency syndrome (AIDS) to assess the efficacy of intrathecal ziconotide [7]. To be eligible, patients had to have a pain score of 50 mm on a visual analog scale despite a regimen of systemic or intrathecal analgesics in the last 3 days before enrollment. Not all patients were taking opioids (98–95%). Mean opioid dosages were 300 mg/day of OME for the ziconotide group, but 600 mg/day for the placebo group. This suggests that the placebo group was receiving double the OME dose of the ziconotide group. Ziconotide produced a better reduction of pain score and more adverse effects in comparison with placebo.

A multicenter, prospective, open-label clinical study was undertaken in patients with “pain from cancer that was uncontrolled by conventional delivery routes including oral, intravenous, intramuscular, subcutaneous, and transdermal dosing or experienced pain that could only be controlled at doses that caused intolerable side effects” [20]. Of the 119 patient implants, 76 (64%) reported “intractable pain” as the reason for implant, 19 (16%) reported adverse effects to systemic opioids, and 24 (20%) reported both intractable pain and adverse effects. The mean pain intensity score was 6.2, with a systemic opioid dosage (intramuscular muscular equivalents?) of a median of 70 mg/day (in OME it should be about 200 mg/day). The median average daily dosage of intrathecal morphine was 1.8 mg/day, which increased steadily to 5.1 mg/day at 4 months (which should be equivalent to 500 mg/day of OME). Pain levels decreased of 31% (from 6.1 to 4.2). For “opioid complications,” such as sleep disorders, drowsiness, constipation, and nausea, there was a dramatic shift to lesser levels of severity. Serious adverse effects, better-defined complications, were device related (n = 7) and procedure related (n = 55). Indeed, it is curious to note from a terminological point of view that opioids produced complications while devices produced adverse effects.

In a study performed in an acute palliative care unit, the clinical response to a combination of intrathecal morphine and levobupivacaine was evaluated in 55 patients with advanced cancer. Patients were highly opioid tolerant, being previously treated with multiple opioid trials unsuccessfully. Initial intrathecal morphine dose was calculated from the previous opioid consumption using a morphine oral–intrathecal ratio of 100:1. The following doses of both drugs were modified during the treatment according to the clinical needs and balanced with adverse effects. Complete data with adequate follow-up until death were available for 45 patients.

Statistical differences in pain intensity were found at the different time intervals examined until death. Moreover, statistical decreases in the intensity of drowsiness and confusion were found until 1 month after starting intrathecal therapy, as well in daily intrathecal morphine doses, with a threefold increase at time of hospital discharge. OME significantly decreased at all the intervals examined until death. Early complications included mild bleeding in two patients, without consequences, headache in four patients, bladder catheterization in six patients, reoperation for bleeding or changes of catheter position in four patients, unrelated death in one patient, and stroke in one patient. Late complications included local infection in two patients, and discontinuation of intrathecal therapy due to spinal compression. Thus, in patients who had received multiple trials of opioids and routes of administration, the intrathecal treatment started with an oral–intrathecal morphine conversion ratio of 100:1, and local anesthetics at the most convenient clinical doses provided a long-term improvement of analgesia, with a decrease in adverse effects and opioid consumption until death [9].

In a study aimed at measuring the incidence of ziconotide-related adverse events seen with a low starting dosage and slow titration in a multidrug protocol, eligible patients had “intractable pain” rated above 6/10 on a numerical analog scale while on high-dose opioid therapy (more than 200 mg/day of OME). Seventy-seven patients were included. Adverse events were recorded in 57% of them, and moderate adverse events occurred in 51%. Adverse events required treatment discontinuation in 9% of cases, of which 5% were serious. Patients experienced a significant and lasting decrease in pain intensity (by 48%) in response to intrathecal analgesic therapy that included ziconotide [21].

In a period of 5 years, 98 pumps were implanted during the study period. Forty died before data collection opportunity, had gone to home hospice care, or were too ill to perform survey. A total of 58 patients, including 50 with a complete dataset and 8 with an incomplete dataset, remained in the study population. Patients were included if they had pain “refractory to conventional therapies,” from either suboptimal pain control or unacceptable adverse effects prohibiting further dose escalation [22]. Different drugs were used, including opioids, local anesthetics, clonidine, baclofen, and ziconotide. A mean dose of 8 mg of intrathecal morphine was used (equivalent to 800–2400 mg of OME, depending on the ratio used of 1:100 or 1:300). A significant post-implant decrease of OME was found. Forty patients were excluded as they died before data collection, went to hospice care, or were too ill. Mild complications, principally due to local anesthetics, were reported. One patient required explant. Of interest, 95% of patients reported the presence of BTP. BTP was better controlled with patient-controlled analgesia through the pump than with unspecific conventional drugs. This is unrealistic, as no drug produces a rapid analgesia through the intrathecal route. Hydrophilic substances such as morphine introduced into the intrathecal space have a slow onset of analgesia when given intrathecally, as they need to move rostrally to provide an effect [23, 24]. Thus, the bolus simply corresponds to an increase in daily dosage.

Fifty-three patients with “intractable” cancer-related pain were recruited in a prospective cohort study. Patients were eligible if they “failed to respond to conservative management on at least 200 mg/day of OME with a pain intensity rated 7/10 or did not tolerate lower opioid doses due to serious toxicity [25]. Although at baseline patients were reported to receive a mean of 452 mg of OME, prior to intrathecal therapy patients were given a mean OME dosage of about 60 mg/day. An electronic patient-controlled intrathecal analgesia pump was used to deliver intrathecal morphine. The initial morphine dose was calculated from the baseline opioid use with an oral–intrathecal morphine conversion ratio of 300:1. The basal intrathecal morphine dosage, combined with ropivacaine, was 1.69 mg/day, and increased up to 3 mg/day at last follow-up. This would correspond to 300 mg/day of OME (calculating an intrathecal–oral ratio of 1:100). Systemic opioids were discontinued. No device-related complication was reported. They concluded that intrathecal therapy allowed for rapid and highly effective pain relief with less toxicity in comparison with conservative medications, although this was not a comparative study.

Ninety-three patients with pancreatic cancer received IDDS. Patients with implant suffered from severe pain before implantation (median presurgical pain intensity 8/10, despite a median 360 mg/day of OME). Different options with opioids were offered before, including oral morphine, intravenous morphine, oral oxycodone, intravenous oxycodone, oral hydromorphone, and transdermal fentanyl. IDDS was associated with significant pain relief at 1 week, 1 month, and 3 months, with severe pain decreasing from 89.2% before surgery to 4.5%, 6.7%, and 10.3% after after 1 week, 1 month, and 3 months, respectively. Global complication rate was consistent with published literature. In this study, patients were offered IDDS on the basis of the presence of “severe pain and/or or unacceptable adverse effects by maximum tolerated dose of pain medication” [26]. A standard opioid titration using several opioids was first trialed. The mean basal pain intensity was 8/10. The previous OME dosages were 300 mg/day. Intrathecal analgesia was provided by a multitude of drugs, including morphine, ropivacaine, ziconotide, and clonidine, of which doses were not reported. Minor complications were reported. About 80% of patients had a pain reduction of more than 50%. Of interest, in the study period of 11 years, about 20% of patients with pancreatic cancer had refractory pain treated with intrathecal analgesia, meaning that 1/5 of patients with pancreatic pain had pain “refractory” to systemic analgesic therapy.

In a prospective observational study of 51 patients, the indication for an intrathecal pump implant was “a moderate to severe pain” in those patients for whom an intrathecal pump implant was planned, according to individual judgment of a pain physician [27]. They were receiving an average dosage of 375 mg/day of OME. At 4 and 8 weeks post-implant, data were available for only 32 and 24 patients, respectively, due to transition to hospice, inability to follow up due to geographic constraints, intensive care unit admission, declining study participation, inability to attend follow-up appointments, and death. At 4 and 8 weeks the intrathecal morphine dosages were 4.73 and 5.34 mg/day, respectively. Pain and symptom intensity, as well as constipation, improved.

Authors found that intrathecal therapy was associated with a marked reduction in serum opioid concentrations, with the majority of patients having undetectable serum levels, suggesting that reducing serum opioid concentrations in patients with cancer may improve bowel function, fatigue, and immune function.

A total of 173 patients with “refractory cancer pain” were assessed, 93% with stage IV disease. The pre-implant median daily OME was 240 mg. The post-implant median OME was 0 mg, with 82.6% of patients discontinuing systemic opioids completely. Mean OME decreased by 94% following IDDS implant, and all patients who continued to use systemic opioids required a lower OME compared with pre-implant. No data regarding changes in pain intensity were provided [28].

In a prospective study, 33 patients were enrolled for implanting an intrathecal pump because of “refractory cancer-related pain” [29]. Authors showed that the implantation of intrathecal morphine pump relieved the pain. Pain intensity decreased from a mean of 7.7 to about 3.8 after implantation. However, no systemic or intrathecal opioid doses were reported. Patients treated with family integrated care had significantly more pain relief than those in the traditional group, of which the definition was unclear. Similarly, depression and anxiety scores significantly improved and fewer adverse events occurred with family integrated care.

A study aimed at examining the effectiveness of intrathecal opioid therapy in regard to different pain mechanisms. Forty-three patients with cancer were treated by continuous intrathecal opioid therapy due to “the failure of conventional pain therapy or because of intolerable side effects of conventional pharmacotherapy.” Intrathecal morphine dosages ranged from 4–5 mg/day to 14.8 mg/day, providing significant pain relief for 2–6 months. In the long term, patients with neuropathic pain, however, benefitted much less than patients with nociceptive pain. Operative complications or those related to the medication pump were observed in five patients. No definition of intractable pain or opioid consumption was reported [30].

A patient-controlled intrathecal analgesia for the treatment of BTP was assessed to reduce the need for breakthrough opioids and improve the patient perception of pain. Both total and breakthrough dosing of conventional opioid medications significantly decreased following the initiation of intrathecal analgesia with patient-controlled intrathecal analgesia. Of interest, 10 patients were excluded as they died within 1 month, leaving 33 patients for analysis. No information regarding the intractable pain condition was given [31].

In 131 patients who had been implanted, 2.8% developed infectious complications. No information on refractory pain condition or opioid consumption was reported [32].

In another study, 160 patients were assessed. The median pain score decreased from a median of 7.1 at time of implantation to 5.0 at 1 month postimplantation. Median survival was 65 days. Five patients had pumps explanted due to infection, with a median time to pump extraction of 28 days [33]. No data regarding the pain conditions of refractoriness were provided.

A total of 220 patients with cancer with “refractory pain” were considered eligible for intrathecal analgesia. Patients with dose increases and no increases during the titration phase were analyzed. OMEs were 315 mg/day. About 60% had poor performance status. Intrathecal therapy was considered more effective when patients were implanted at an early stage of the disease. Authors claimed data as important in the context of the opioid epidemic crisis, indirectly suggesting that IDDS may be useful in preventing opioid misuse, without any proof in this regard [34].

One study looked at establishing a safe and efficacious fixed bupivacaine dosing algorithm in intrathecal pumps for patients with cancer. A generic “cancer related pain refractory to systemic opioids” was given as indication for IDDS [35].

In another study, a review chart of 173 patients was included, 93% of them with stage IV disease. The pre-implant median daily OME was 240 mg. Mean OME decreased by 94% following IDDS implant, and all patients who continued to use systemic opioids required a lower OME compared with pre-implant. No information about the indications for implant was provided [36].

A cohort of 50 oncological patients who had IDDSs was assessed. Median OMEs decreased from preoperative 503 mg/day to 105 mg/day at 6 months. Due to malignant mortality, 27 patients did not make it to the 6-month postoperative timepoint. Twenty-four patients experienced adverse effects after intrathecal pump implantation. No specific information regarding an “intractable pain condition” was reported. Of interest, some of them appeared opioid naive[29].