Do the potential benefits outweigh the risks? An update on the use of ziconotide in clinical practice

Ziconotide is a selective and potent blocker of N‐type voltage‐gated calcium channels. It was approved by the Food and Drug Administration in 2004 and by the European Medicines Agency in 2005 for the treatment of severe chronic pain in patients needing intrathecal analgesia (ITA). The aim of this paper is to provide a practitioner‐oriented, educational, narrative, up‐to‐date review on the use of ziconotide in clinical pain medicine. Of special concern regarding safety is the partial incongruity between dosing statements in the Summary of Product Characteristics and novel low‐dosage, slow uptitration recommendations. Even though ziconotide has obvious advantages compared to opioids, pain practitioners pondering the use of ziconotide nonetheless have to balance its proved potential analgesic effect against its neurological side effects, with special consideration being given to dosing and neuropsychiatric dangers. Using a seesaw analogy, the paper discusses what factors pain physicians should weigh in when considering ziconotide as ITA drug, the non‐opioid advantages of ziconotide being counterbalanced by its potential psychiatric side effects. Ziconotide is an important part of the armamentarium of modern interventional pain medicine. If ITA is deemed necessary, ziconotide is a rational alternative, at least in chronic (neuropathic) non‐cancer pain. However, in many European countries, ziconotide treatment is only available in a few (if any) centres. The safety profile of ziconotide is not fundamentally more worrying than that of opioids or cannabinoids; it is just different. This paper provides a concise, up‐to‐date and clinically‐oriented summary of the use of ziconotide in clinical practice, not least concerning safety and dosage issues.


Background
Research on the potential pharmacological use of venom-based medicinal products is ongoing (Gorson and Holford, 2016;Netirojjanakul and Miranda, 2017). Most drugs used in western medicine have a natural origin, and animal venoms have in recent years emerged as an important source of potential new drugs (Robinson et al., 2017), for example in the field of pain medicine where toxins from snakes, spiders or marine snails are being investigated (Diochot et al., 2012;Rigo et al., 2017).
In 2004 and 2005, respectively, the synthetic conotoxin ziconotide was approved by the Food and Drug Administration and by the European Medicines Agency for the treatment of severe chronic pain in patients needing intrathecal analgesia (ITA) (Schmidtko et al., 2010). Ziconotide is a synthetic version of the hydrophilic conotoxin x-MVIIA from the venom of the Pacific fish-hunting marine snail Conus Magus (Lyseng-Williamson and Perry, 2006).
Ziconotide is a selective and potent blocker of Ntype voltage-gated calcium channels (VGCC). Its analgesic effect is thought to be mediated by blockage of presynaptic VGCC on primary nociceptive afferents in Rexed laminae I and II of the dorsal horn, leading to subsequent lessened release of pronociceptive neurotransmitters and neuropeptides (Schmidtko et al., 2010;Pope et al., 2017). Although ziconotide is a first-in-class drug in virtue of being a blocker of VGCC, the well-known gabapentinoid analgesics gabapentin and pregabalin modulate the same VGCC by binding its accessory a 2 d subunit (Burgess and Williams, 2010;Patel et al., 2017a), suggesting that combining oral gabapentinoids with intrathecal ziconotide might perhaps be of value (Patel et al., 2017a).
Ziconotide is a highly hydrophilic 25 amino acids polypeptide with a molecular weight of 2639 Da, making it 10 times 'bigger' than the alkaloid morphine (Miljanich, 2004;Pope and Deer, 2013; Web References 1 and 2). Being rapidly degraded by peptidases when it reaches the systemic circulation, ziconotide has to be infused directly into the cerebrospinal fluid (CSF), where its median terminal half-life is about 4.5 h (Wermeling et al., 2003;Schmidtko et al., 2010).
For many years, morphine was the gold standard for ITA. Despite the widespread off-label use of other opioids, local anaesthetics, or clonidine, the place of morphine as first choice ITA-drug was never really challenged before the apparition of the non-opioid ziconotide, which has now for a decade been considered a first-line ITA-drug by the Polyanalgesic Consensus Conference (PACC) (Deer et al., 2007(Deer et al., , 2012(Deer et al., , 2017b. No ITA-drug has been as thoroughly investigated as ziconotide, including three pivotal randomized controlled trials (RCTs) (Staats et al., 2004;Rauck et al., 2006;Wallace et al., 2006) and many open-label studies in diverse study populations Ver Donck et al., 2008;Wallace et al., 2008;Deer et al., 2009Deer et al., , 2018Saulino et al., 2009;Webster et al., 2009;Raffaeli et al., 2011;Alicino et al., 2012;Dupoiron et al., 2012). However, despite algorithms such as the ones published by PACC (Deer et al., 2017b), the optimal use of ziconotide and other ITA drugs in clinical practice is not self-evident. Indeed, it is important to bear in mind that pharmacological treatment is only a small part of the management of chronic non-cancer pain, and that ITA itself in many ways is a last resort when it comes to the pharmacological treatment of chronic non-cancer pain (Turk et al., 2011).
Based on (1)  , the aim of this study is to provide a practitioner-oriented, educational, narrative, up-to-date review on the use of ziconotide in clinical pain medicine. Given the fact that the Summary of Product Characteristics (SPC) is in part incongruent with the latest dosing recommendations (Deer et al., 2017a; Web Reference 3), such an update seem warranted from a safety point of view.

Literature search methods
In addition to previous knowledge of the literature, a PubMed search using the word 'ziconotide' yielded 369 counts (28 December 2017), the yearly distribution being depicted in Fig. 1. The 5 years limit function was applied to the search, yielding a total of 88 articles whose abstract were screened to ensure upto-dateness concerning recent published clinical data. New pre-clinical papers were not considered relevant to the aim of this study.

Is ziconotide really that effective in clinical practice?
Although there is one small study on acute postoperative pain (Atanassoff et al., 2000), it is important to remember that ziconotide is approved for the treatment of chronic pain. The three pivotal RCTs underlying the approval of ziconotide (Staats et al., 2004;Rauck et al., 2006;Wallace et al., 2006) have recently been reviewed in a meta-analysis (Brookes et al., 2017). All three studies defined a responder as a patient experiencing ≥30% reduction of visual analogue scale pain intensity (VASPI) at follow-up, and the meta-analysis of (Brookes et al., 2017) yielded a pooled odds ratio (95% confidence interval) of 2.77 (1.37-5.59), hence favouring ziconotide over placebo.
The number needed to treat (NNT), which is the inverse of the absolute risk reduction (Edelsberg and Oster, 2009), is an alternative effect size measure (McGough and Faraone, 2009). Based on the figures tabulated by (Brookes et al., 2017), the NNT for 30% reduction in VASPI was calculated by the present author to be 5.56. However, it is important to remember that in two of the studies (Staats et al., 2004;Rauck et al., 2006), dosages were much higher than those described in the subsequent SPC (Web Reference 3), whereas in (Rauck et al., 2006) (with dosages corresponding to the subsequent SPC) the proportion of responders was not significantly different between groups (although statistical significance was noted for VASPI mean percentage improvement from baseline to week 3). Hence, the clinical validity of the effect sizes described above should be critically pondered. This is especially important when considering that recent starting dosage recommendations are more cautious than even the SPC (see below) (McDowell and Pope, 2016; Web Reference 3). Given the obvious clinical fact that ITA is much more of a last resort than a first-line option in the armamentarium of the pain physician, and given the high NNT for analgesics in general in chronic pain conditions (as exemplified for instance by (Finnerup et al., 2015) for neuropathic pain), it is perhaps no surprise that there are many non-responders. As aptly expressed by Brookes et al. (2017), ziconotide 'should not be seen as a panacea'.
Although case reports and small case series can give valuable information and should therefore be published (recent examples are the papers by de la Calle Gil et al., 2015;Heifets et al., 2013;Horazeck et al., 2015;Lanzillo et al., 2016;Lux, 2010;Narain et al., 2015;Obafemi and Roth, 2013;Patel et al., 2017b;Phan and Waldfogel, 2015;Pozzi et al., 2014;Russo et al., 2015;Staquet et al., 2016;Voirin et al., 2016), 'real life' data from larger cohorts such as the Registry of Intrathecal Ziconotide Management (PRIZM) (Deer et al., 2018) are more valuable. A recent interim analysis of the PRIZM study showed that, after 12 months of treatment, 34.8% of patients on ziconotide experienced ≥30% pain intensity reduction (Deer et al., 2018). However, and most importantly when it comes to generalizability, only 23 of 93 enroled patients reported data at 12 months, the authors stating the following 'possible reasons' for this high frequency of missing values: study discontinuation; missed assessment; recently enroled patients not yet reaching that time point. High frequencies of treatment discontinuation have indeed been observed by other investigators in similar open-label studies Wallace et al., 2008;Webster et al., 2009;Raffaeli et al., 2011). All in all, although the analgesic effect of ziconotide has been demonstrated in RCTs, the magnitude of its long-term effect size in clinical practice is difficult to ascertain. That there is a subset of patients who experience long-term clinically significant relief while on ziconotide is however undeniable.

Is ziconotide safe to use?
Opioid-like adverse effects such as respiratory depression, tolerance, dependency or hormonal changes have not been described for ziconotide (Schmidtko et al., 2010;Webster, 2015). However, this obvious advantage is counterbalanced by the fact that ziconotide has a narrow therapeutic window with numerous neurological side effects (see below) as well as a substantial time lag after dose titration (probably because of hydrophilicity-related slow diffusion into neural tissue) (Schmidtko et al., 2010). Hence, the adage 'start low, go slow' is important to bear in mind for this drug (Brookes et al., 2017), and the present writer routinely follows the advice of not augmenting the dose more often than once a week (Fisher et al., 2005). Dosing issues are addressed in more detail below.
According to the SPC, the most common adverse reactions reported in clinical trials were dizziness (42%), nausea (30%), nystagmus (23%), confusional state (25%), gait abnormal (16%), memory impairment (13%), blurred vision (14%), headache (12%), asthenia (13%), vomiting (11%) and somnolence (10%); most adverse reactions were mild to moderate in severity and resolved over time (Web Reference 3). Due to the fact that the RCTs used much higher dosages than are nowadays recommended (see below), it is also of great interest to ponder registry open-label reports of side effects such as those from the Italian registry of ziconotide with mean doses after 6 months ranging from 3.2 to 5.1 mcg/day depending on diagnosis; these data are summarized in Fig. 2.
Of the numerous neurological side effects that have been described, the most feared are the neuropsychiatric ones, which are thought to be associated with rapid titration and high dosages (Sanford, 2013). Suicidality, psychosis and confusion are of particular concern (Maier et al., 2011;Phan and Waldfogel, 2015;Webster, 2015). Hence, on the (unproven but sensible) assumption that a patient's prior psychiatric disorders may be a predisposition for dangerous neuropsychiatric adverse effects, the patient's psychological state should be carefully evaluated before considering ziconotide treatment (Deer et al., 2012). At the very least, a history of psychosis should be seen as a contraindication (Deer et al., 2012;Brookes et al., 2017). Low-dose strategies have substantially mitigated, but not eliminated, the risk of serious neuropsychiatric side effects (Deer et al., 2012). If a severe adverse reaction occurs, ziconotide infusion should and can be stopped immediately without withdrawal effects (Kress et al., 2009).

Weighing potential benefits against risks
Bearing in mind both the shown efficacy of intrathecal ziconotide infusion in short RCTs and its potential neuropsychiatric dangers, pain practitioners must balance the one against the other in their 'tailormade' handling of the individual patient (Fig. 3). In the decision-making process, the following factors should at least be weighed in.

Patient selection
Although it is probably true that ITA should no longer been seen exclusively as a last resort salvage therapy (Deer et al., 2017b), it is nevertheless still the case that it is only suitable for a minority of chronic pain patients. This being a review of the use of ziconotide in particular and not on the use of ITA in general, the following will focus on the particulars of ziconotide. If ITA is indeed indicated, which patients should be treated with ziconotide, as opposed to other ITA-drugs?

Cancer pain
The latest PACC recommendations (Deer et al., 2017b) emphasize that cancer stage and life expectancy should be taken into consideration when selecting ITA medication for cancer pain, and the following three categories of cancer patients are defined: (1) Category 1: Patient with imminent death or with relatively short life expectancy (2) Category 2: Stable or slowed disease, with high likelihood of recurrence or progression (3) Category 3: Cancer in partial remission or cured, with residual chronic pain Although monotherapy with ziconotide or morphine has the strongest evidence base, the PACC recommendations recognize that titratability is of paramount importance for Categories 1 and 2, and that there is significant evidence for using an opioid AE bupivacaine in this population. Confirming this assertion, and from a European perspective, there are several case series reporting the use of an external pump connected to an externalized intrathecal catheter or to a subcutaneous port (i.e. a much less invasive procedure than implanting a pump) (Sjoberg et al., 1991(Sjoberg et al., , 1994Nitescu et al., 1995;Dahm et al., 2000;Mercadante et al., 2007;B€ ackryd and Larsson, 2011). In advanced cancer and/or in the end-of-life setting, the slowness of titration inherent in ziconotide treatment seems to be a clear drawback compared to opioids and local anaesthetics such as bupivacaine. All in all, it is especially important to individualize the ITA treatment for Categories 1 and 2, that is, to base the treatment on disease progression and in collaboration with the oncology team (Bruel and Burton, 2016).

Non-cancer pain
For Category 3 patients (i.e. including 'cancer survivors') and for patients with non-cancer pain, monotherapy with ziconotide or morphine are firstline options (Deer et al., 2017b). However, the PACC panel stated as a consensus point that 'unless contraindicated, ziconotide should be the first drug selected in the population of non-cancer patients discussed in this consensus' (Deer et al., 2017b). This is true whether the pain is nociceptive or neuropathic. However, it has to be remembered that in the two non-cancer RCTs (Rauck et al., 2006;Wallace et al., 2006), about 75% of the patients had a neuropathic pain condition.
The latest PACC offer detailed recommendations and multiple algorithms and tables (Deer et al., 2017b). In Fig. 4, an overview algorithm is proposed, synthesizing the above-given information in the simplest possible manner. Needless to say, Fig. 4 is no substitute for the PACC or other more in-depth documents.

Trialing procedures
As part of the clinical decision-making process, preimplantation trialing with the planned analgesic seems rational (Deer et al., 2017a). For ziconotide, this may be performed either using an external pump for a limited period (perhaps 1-4 weeks) or by a bolus injection (Burton et al., 2010). Bolus injection is much simpler and seems to have become the preferred trialing technique (Deer et al., 2017a). However, because of surprisingly low rates of analgesic effect in an open-label bolus study, the rationale of trialing ziconotide by bolus injection has been questioned (B€ ackryd et al., 2015). There are, however, open-label studies that report good analgesia by bolus injections (Mohammed et al., 2013;Pope and Deer, 2015). This discrepancy is difficult to explain. The majority view seems to be that bolus injection is a valuable method, but whether this is right or not is arguably a matter for debate. Trialing by external pump infusion has been shown to be feasible (Ver Donck et al., 2008), and this method gives the patient more time to assess efficacy. A case Figure 2 Ziconotide-related adverse events recorded in the Italian registry of ziconotide. Based on tabulated data in (Raffaeli et al., 2011). report on the external pump trialing method is presented in Fig. 5. At the very least, a pain diary should be part of the evaluation.
Studies have shown that there is limited rostral spread of ITA-drugs within the IT space, and there is therefore nowadays a consensus (despite no hard evidence) that catheter tip placement congruent with the dermatomal area of pain is important (Deer et al., 2017b).

Dosing issues
As per the SPC, the start dose is 2.4 lg/day (Web Reference 3). If 1-2 weeks of infusion at that dose does not give ≥30% pain reduction, the present writer considers the trial to be negative and a permanent pump will therefore not be implanted. Such strict trialing generates a number of negative trialing issues, but given the potential neuropsychiatric side effects of ziconotide (which appear in large part to be dose-related Sanford, 2013), there is arguably a lot to commend for such a cautious primum non nocere approach. It is also important to note that 2.4 lg/day is at the upper end of what is nowadays recommended as start dosage (i.e. 0.5-2.4 lg/day) (Deer et al., 2017a).
In case of emergency, a ziconotide infusion can be stopped immediately without withdrawal effects (Kress et al., 2009). Based on the above-mentioned CSF half-life time (median 4.5 h), ziconotide should be practically cleared from CSF 24 h after treatment interruption (i.e. after approximately 5 half-lives). However, the resolution of adverse effects takes much more time, that is, there is a time lag between CSF pharmacokinetics and the pharmacodynamics of ziconotide (Smith and Deer, 2009;Pope et al., 2017).
The median time to onset of the most commonly reported adverse events has ranged from 3 to 9.5 days (Smith and Deer, 2009). Hence, increasing doses no more often than at least once a week makes sense (Fisher et al., 2005); however, this is much slower than what the SPC allows (Web Reference 3). Concerning the magnitude of dose increments, dose increments of up to 1.2 lg/day seem safe, and this low-dose practice is consistent with the recent publication from Prusik et al. (2017).
The maximum dosage as per the SPC is 21.6 lg/ day; however, the SPC also states that 'approximately 75% of patients who respond satisfactorily to treatment require a dose ≤9.6 lg/day' (Web Reference 3). If a strict trailing strategy is followed as described above, high doses should rarely be needed.
Whether the novel nocturnal flex bolus dosing described by Pope and Deer (2015) is advantageous or not will have to be shown in future studies.

Psychiatric evaluation before and during treatment
Before initiating ziconotide treatment, the patient's psychological state should be carefully evaluated. There is an (unproven but sensible) assumption that a patient's prior psychiatric disorders may be a predisposition for dangerous neuropsychiatric adverse effects (Deer et al., 2012). At the very least, a history of psychosis should be seen as a contraindication (Deer et al., 2012;Brookes et al., 2017).
Continual psychiatric review is mandatory when a patient is on ziconotide, and both patients and relatives should be educated in this regard. Psychotic features and suicidal thoughts should be especially sought for.

Conclusion
The current US 'opioid epidemic' and its catastrophic consequences, as well as the weak evidence base for the use of opioids in chronic pain conditions, illustrates how difficult it is to 'hijack' the endogenous opioid system without creating harm (Ballantyne and Sullivan, 2017;Clauw, 2017). How to use opioids is a problem facing every pain physician, and the European Pain Federation has recently published a position paper on the appropriate use of opioids in chronic pain (O'Brien et al., 2017). In this context, some researchers have described cannabis-based drugs as a possible alternative, for example for 'harm reduction' (Lau et al., 2015). Concerning cannabisbased analgesia per se, an overview of systematic review has recently been published (Hauser et al., 2018), and there seems to be a gap between public belief and available medical evidence (Fitzcharles et al., 2014).
But does the future perhaps belong not to opiumbased or even cannabis-based analgesics but to venom-based drugs? If the age of indiscriminate use of opium-based analgesics is passing, has the age of venom-based products begun? Only time will tell. For the time being, there is but one animal toxin based analgesic on the market and, with all its drawbacks, ziconotide is nonetheless an important part of the armamentarium of modern interventional pain medicine. If ITA is deemed necessary, ziconotide is a rational alternative, at least in chronic (neuropathic) non-cancer pain. However, in many European countries (e.g. the Nordic countries or the UK Brookes et al., 2017), ziconotide treatment is only available in a few centres.
Will future pain physician look back on ziconotide as the prototypical example of a new era of venombased analgesics? Perhaps, but this will presuppose new and easier administration routes, for example using 'nanocontainers' (Anand et al., 2015;Manda et al., 2016). All in all, the psychiatric side effects of opioids, cannabinoids and ziconotide are a sobering reminder of how difficult it is to interfere safely with receptors in the central nervous system. In that respect, the safety profile of ziconotide is not fundamentally more worrying than that of opioids or cannabinoids; it is just different.