Dopamine receptor-blocking antipsychotics, first introduced into clinical practice in 1952, were hailed as a panacea in the treatment of a number of psychiatric disorders. However, within 5 years, this notion was to be shattered by the recognition of both acute and chronic drug-induced movement disorders which can accompany their administration. Tardive syndromes, denoting the delayed onset of movement disorders following administration of dopamine receptor-blocking (and also other) drugs, have diverse manifestations ranging from the classic oro-bucco-lingual dyskinesia, through dystonic craniocervical and trunk posturing, to abnormal breathing patterns. Although tardive syndromes have been an important part of movement disorder clinical practice for over 60 years, their pathophysiologic basis remains poorly understood and the optimal treatment approach remains unclear. This review summarises the current knowledge relating to these syndromes and provides clinicians with pragmatic, clinically focused guidance to their management.
- SUPRANUCLEAR PALSY
- MOVEMENT DISORDERS
Statistics from Altmetric.com
Movement disorders developing as a direct consequence of the administration of dopamine receptor-blocking neuroleptic drugs were first reported in 1957, 5 years after their introduction into psychiatric practice. The year 1964 saw the first collective description of these movement disorders as a ‘tardive’ (from the latin tardus, meaning late) phenomenon,1 reflecting their delayed onset following medication administration, in contrast to ‘acute’ dystonic reactions, which also follow dopaminergic blockade. This term was rapidly adopted, and in the following decades, a flurry of publications were to expand the phenotypic spectrum of the disorder. Concurrently, theories aiming to explain disease pathogenesis began to emerge, and several therapeutic strategies were explored.
This review provides physicians with a pragmatic, clinically based platform with which to approach tardive syndromes. In addition, we explore recent developments in our understanding of disease pathophysiology, discuss how to approach treatment of tardive syndromes and try to dispel some commonly held myths.
The nosology of tardive syndromes is plagued by inconsistent use of descriptive language. The term ‘tardive dyskinesia’, when first introduced, was intended to subsume the range of diverse movements that can emerge in a delayed fashion following long-term neuroleptic administration. However, more recently, a less confusing approach which classifies tardive movements according to their clinical phenomenology has been promoted, and will be used in this review. Accordingly, we use ‘tardive syndrome’ as the umbrella term for any/all potential tardive movement disorders but reserve ‘tardive dyskinesia’ as a descriptor of a specific clinical entity, namely the characteristic oro-bucco-lingual choreiform movements (see The ‘typical’ tardive syndrome).
The scale of the problem
Tardive syndromes are a predictable, sometimes permanent, disabling consequence of medication administration. They occur predominantly in the psychiatric population, where they exacerbate the burden of social stigma and are linked to poorer quality of life and increased morbidity and mortality.2 3 Antipsychotic drugs are by far the most common offenders, though numerous others have also been implicated (table 1).
Tardive syndromes affect 20%–50% of patients receiving neuroleptic drugs.12 Advancing age is the most robust risk factor, with incidence increasing from 5% per annum in those aged under 40 years to 12% or more per annum in older age groups.12–14 The risk increases cumulatively with duration of exposure and medication dose, with a cumulative incidence rate of 20%–25% after 5 years of exposure.15-17 Note however that the medication compliance rate in patients with schizophrenia is around 50%, so these figures may well be an underestimate.18
Numerous other factors may further increase the risk, including history of an affective disorder, previous organic brain damage, diabetes mellitus, female sex (oestrogen perhaps being protective premenopausally) and race.19 Indeed, disease prevalence is lower on average in Asians (roughly 20%) and higher on average in African–Americans compared with Caucasians.19–21
The pathophysiological basis of tardive syndromes remains poorly understood, as reflected in the large number of theories purporting to explain the delayed development of these movement disorders.
The earliest theory to gain popular acceptance was the so-called dopamine receptor hypersensitivity theory. This suggested that dopamine-blocking neuroleptics led to compensatory upregulation and/or hypersensitivity of postsynaptic dopamine (particularly D2) receptors.22 23 Hypersensitivity of these receptors, which are expressed on indirect pathway medium spiny neurones and are inhibitory, would have the net effect of pallidal and subthalamic nucleus disinhibition, producing abnormal hyperkinetic movements.22 This hypothesis was largely based on clinical observations, such as the greater likelihood of tardive syndromes in patients receiving potent D2 blockers and the apparent improvement in tardive dyskinesia with additional dopaminergic blockade, as well as on some animal studies.22 23 However, evidence in humans for such alterations is lacking. There is no correlation between in vivo striatal D2 receptor ligand binding assessed by positron emission tomography and the severity of tardive dyskinesia. Equally, postmortem examinations have not demonstrated significant differences in D2 receptor numbers in those with and without tardive syndromes.22 Moreover, this theory does not explain why many patients do not recover after they stop the offending medication; if the only problem were receptor upregulation/hypersensitivity, one would expect this to normalise following drug withdrawal.
An alternative hypothesis is that tardive syndromes actually represent a neurodegenerative disorder of striatal interneurones induced by oxidative stress. This theory, which is supported by animal and human neuropathological studies,24 25 holds that dopaminergic receptor blockade causes increased dopamine turnover and oxygen free radical production by monoamine oxidase.22 These free radicals are thought to be toxic to striatal interneurones, causing gliosis within the basal ganglia, thus explaining why the symptoms persist after stopping the medication. However, the significant and sustained improvement that sometimes follows deep-brain stimulation for tardive syndromes might argue against this idea.
A further theory implicates damaged or dysfunctional striatal gamma-aminobutyric acid (GABA)ergic neurones in the pathogenesis of tardive dyskinesia. These neurones synapse on the soma of medium spiny neurones, providing potent feedforward inhibition, balancing activity in the direct and indirect basal ganglia pathways, and providing surround inhibition.22 23 Selective lesioning of these neurones produces dyskinesia.26 Long-term D2 agonism, in theory, could potentially damage GABAergic interneurones via glutamate-mediated excitotoxicity and increased oxidative stress from dopamine turnover.27
Finally, altered N-methyl-D-aspartate (NMDA)-mediated synaptic plasticity may provide a unifying theory. Antipsychotics are known to influence NMDA receptor-mediated synaptic plasticity. In this setting, patterns of abnormal neurotransmission, for example, secondary to D2 receptor hypersensitisation could be abnormally potentiated, perpetuating a cycle of abnormal sensorimotor integration and abnormal tardive movements.22
Of course, not everyone who is exposed to neuroleptic drugs develops a tardive syndrome, implying that other, possibly genetic factors are at play, conferring increased vulnerability to tardive syndromes. Genome-wide association studies have identified some potential candidate genes, though their relevance to clinical practice remains unclear.28
Making the diagnosis: the devil is in the detail
This section describes the typical (or perhaps simply better recognised) and less typical presentations of tardive syndromes. One must be mindful however that individual components of the syndrome rarely occur in isolation, but rather generally coexist to greater or lesser degrees (though one may be dominant). A confident diagnosis often depends on identifying multiple movement phenomena that are compatible with a tardive syndrome. Thus, an important part of the evaluation involves not only identifying a movement of potentially tardive aetiology, but actively searching for the presence of other compatible abnormalities. Failing to notice clues, such as a fidgety patient (akathisia) who sighs deeply (respiratory dyskinesia) and moves his legs back and forth during the consultation (stereotypies), can rapidly lead one down the wrong diagnostic path.
Although diagnostic criteria for tardive syndromes have been developed (table 2), only three questions matter in clinical practice:
Is there a history of taking a dopamine receptor-blocking or other tardive syndrome-causing drug, as prescription medication, over-the-counter/traditional remedies or poisoning?
What is the temporal relationship of taking this drug intake to the onset of the movement disorder?
Is the clinical phenomenology compatible with a tardive syndrome (see below)?
The ‘typical’ tardive syndrome
‘Classic’ tardive dyskinesia involves stereotyped choreoathetoid movements predominantly involving the lips, tongue and perioral region. The movements often predominate in the lower face, with frontalis involvement being unusual. Patients often move the tongue in a writhing motion inside the mouth, are prone to frequent rapid tongue protrusion (‘flycatcher tongue’) and pushing of the tongue against the inside of the cheek, creating a bulge (‘bonbon sign’). Chewing and/or grimacing movements, lip smacking and puckering are typical. This may be accompanied by low amplitude choreiform movements of the distal limbs, the so-called ‘piano player dyskinesia’, resembling finger movements on piano keys.29 Patients are often unaware of these involuntary movements, though those involving the lips and tongue may cause problems with feeding.
Tardive dyskinesia is usually accompanied by one or more of the following tardive phenomena:
This is an uncomfortable sense of inner restlessness, requiring the affected individual to repeatedly move about in order to ease the unpleasant sensation. Movements can include rocking in one place when seated, marching when standing, repetitively scratching or rubbing, or just appearing generally ‘fidgety’ during the consultation.
These are patterned, purposeless, repetitive and somewhat ritualistic movements that may appear as truncal rocking, pelvic thrusting, to-and-fro leg movements, hand-wringing or crossing/uncrossing of the legs. They may outwardly resemble akathisia but are not accompanied by inner restlessness.
As with most tardive syndromes, tardive dystonia adopts distinct phenomenological characteristics which are easily identified by the trained observer. The disorder frequently involves the craniocervical region, manifesting as retrocollis. Dystonia may extend to the trunk as opisthotonic posturing, while in the arms, abduction, internal rotation and wrist flexion is the classically adopted posture.29 Blepharospasm may also emerge. In contrast to other tardive syndromes, tardive dystonia is particularly common in young men aged around 40 years.24 Remission is also less likely than with tardive dyskinesia, particularly with drug exposure beyond 10 years.24
The following tardive disorders are less well defined, with only a handful of reported cases.
This rare disorder manifests as multiple motor and verbal tics that emerge after exposure to dopamine receptor-blocking agents. The tics generally resemble those of primary tic disorders, exhibiting suppressibility, build-up of tension before the tic and release of tension afterwards.30 31
This was first proposed as an entity in a 1992 report of five patients with a 3–5 Hz postural and action greater than rest tremor but without parkinsonism.32 Although similar to parkinsonian tremor, tardive tremor is distinguished by its postural and kinetic (rather than rest) predominance, its coarse disabling nature, its lack of levodopa responsiveness and its occasional improvement with further dopaminergic blockade or tetrabenazine.32 33 The syndrome generally persists despite withdrawal of dopamine receptor-blocking agents.
This describes brief, upper-limb predominant postural myoclonic movements that are said to result from long-term dopaminergic blockade.34 35 However, there is only very limited literature on this entity, which should therefore be interpreted with caution.34 35
This is a poorly characterised and non-uniform phenomenon, with gait disturbances having been described as ‘dancing’ (multiple short steps followed by a long step) or ‘duck-like’ (broad based with short stride length and some steppage features). Other abnormalities include walking with initial floor contact with toes rather than heels, spastic qualities and abnormal arm swing.36
Some ‘atypical’ presentations
Patients with tardive syndromes not infrequently exhibit other less recognised, but nonetheless characteristic features that point towards the diagnosis. Among these, the most important are respiratory phenomena, tardive Pisa syndrome and withdrawal emergent dyskinesia.
First described in 1964, respiratory dyskinesia involves periodic disturbances of ventilatory rate, rhythm and amplitude, sometimes with ventilatory pauses or forced inspiration against a closed glottis.37 Patients may complain of dyspnoea or dysphonia, or may be seen to huff, grunt, gasp or take short, rapid breaths.38 These phenomena often accompany other more classic tardive motor features.
Tardive Pisa syndrome
This phenomenon, predominantly affecting older women, describes a drug-induced persistent truncal dystonia manifesting as tonic lateral flexion, occasionally with slight rotation.39 The ‘laterally leaning patient’ is an important clue to a tardive aetiology.
Withdrawal emergent dyskinesia
This syndrome is considered a variant of tardive dyskinesia, which generally develops after either abruptly stopping or significantly reducing the dose of neuroleptic medications.40 It predominantly affects children and usually manifests as generalised chorea (as opposed to the facial-predominant movements observed in classic tardive dyskinesia). It is usually self-limiting and resolves after days to weeks.40
Tardive oculogyric crises
Oculogyric crises were originally described as being characteristic of encephalitis lethargica, although now they are more commonly associated with medication-related acute dystonic reactions (as well as dopamine synthesis pathway defects). However, oculogyric crises can also rarely develop as a tardive phenomenon in patients chronically exposed to antipsychotic medications.41 42 Tardive oculogyric crises often accompany other tardive motor phenomena and may go unrecognised. They are sometimes associated with transient recurrences of psychiatric symptoms, including anxiety, auditory hallucinations and bizarre behaviour.41
Tardive pain syndromes
A variety of tardive pain syndromes have also been described, temporally associated with neuroleptic use and often responding to standard tardive syndrome treatments. Examples include tardive oral pain, which describes an uncomfortable, often burning sensation in the mouth and lips, and painful genital syndrome, with similar affliction of the genital region.43
Bruxism, of either the grinding or mixed grinding-clenching type, may develop as a side effect of long-term neuroleptic exposure. It probably represents a forme fruste of tardive oromandibular dystonia.44 A striking feature of the syndrome is noise production, sometimes sufficiently severe to annoy roommates. The movements disappear during sleep.
Assessing the severity of tardive syndromes
Before prescribing dopamine receptor-blocking drugs, clinicians should strive to document the presence or absence of abnormal involuntary movements. While both physician and nurse-led standardised assessment tools (such as the abnormal involuntary movement scale and ScanMove instrument, respectively) may not always be practical in the busy clinical setting,45 46 a focused examination is nevertheless important. It was recognised over 140 years ago that psychiatric patients may exhibit stereotypies, chorea or abnormal facial grimacing as a result of their disease—failure to document this before treatment may lead to these later being misattributed to a drug effect.47 48
It has also been suggested that some older people develop spontaneous movements of the face as part of normal ageing. Whether this is true or merely represents the emergence of facial or craniocervical dystonic syndromes with age is yet to be resolved.
FACTS AND FALLACIES
Myth number 1: Second-generation antipsychotics, with their lower D2 binding affinity, have reduced the incidence of tardive syndromes
This has been a particularly contentious issue and it is difficult to make a definite statement in either direction. What can be said with certainty is that the introduction of second-generation antipsychotics has not done away with tardive syndromes. Rather, due to rapid uptake in their prescription, including off-label use for mood disorders and sleep, ironically they may have contributed further to the problem.
While some studies suggest that the incidence of tardive syndromes with second-generation antipsychotics is not vastly dissimilar from that of their first-generation counterparts,10 49 the largest literature review to date, involving 34, 555 patients treated with antipsychotics across 56 studies, found an annualised incidence rate of 2.98% with second-generation antipsychotics versus 7.7% with first-generation antipsychotics, supporting the claim that second-generation antipsychotics may indeed carry a lower risk.50 A recent large meta-analysis of 57 studies on tardive syndromes also supported this.9
Myth number 2: Prolonged exposure to a causative drug is necessary to be at risk of tardive syndromes
Although, as detailed above, the cumulative risk of tardive syndromes increases year-on-year and most patients develop the disorder after at least 1–2 years of drug exposure,23 24 there are reports of its occurrence after just a single dose of neuroleptic. Prolonged drug exposure is therefore not always necessary.
Myth number 3: Some neuroleptics are safe
The recognition that first-generation (‘typical’) antipsychotics were associated with a number of extrapyramidal side-effects prompted the development of newer compounds, termed ‘atypical’ antipsychotics, which were supposedly defined by the absence of extrapyramidal symptoms at therapeutic doses. Numerous mechanistic differences of these newer compounds, including effects on serotonergic signaling, more rapid dissociation from the D2 receptor, limbic selectivity and in the case of aripiprazole, partial dopaminergic agonism were posited as the reason behind their more favourable side effect profiles.
While it is true that not every neuroleptic has the same propensity to cause tardive syndromes, none is devoid of risk. All classes of antipsychotics can produce tardive syndromes.20 51 Nevertheless, newer ‘atypical’ agents probably carry about half the risk of producing later tardive syndromes as compared with their ‘typical’ counterparts.9 Furthermore, it is important to remember that it is not just neuroleptics that are implicated in the development of tardive syndromes (table 1).
Differential diagnoses not to miss, and how to spot them
Differentiating spontaneous from drug-induced movement disorders in patients with psychiatric illness can be a challenging endeavour. Nonetheless, it is imperative to give adequate thought to excluding important differential diagnoses that can present with the combination of psychiatric disease and abnormal movements,29 and particularly the following conditions:
As a trinucleotide repeat expansion disorder with the cardinal manifestations of chorea, psychiatric disease and cognitive decline, Huntington’s disease is one of the most important differential diagnoses of tardive dyskinesia. Psychiatric disease (often requiring neuroleptic treatment) can precede the development of hyperkinetic movements in this condition by several years. Inexperienced observers can therefore easily misdiagnose such hyperkinetic movements as tardive.
In this setting, there are some particularly helpful clinical clues52 including
Hyperkinetic movements: In tardive dyskinesia, these movements tend to be stereotyped and semi-purposeful, as opposed to the random, flowing movements of chorea that typify Huntington’s disease.
Topographical distribution: In tardive syndromes, the movements are predominately lower facial and axial, manifesting as retrocollis and opisthotonus. In contrast, patients with Huntington’s disease often have significant limb chorea, which is unusual in tardive syndromes. Hyperkinetic movements of the frontalis muscle are also common in Huntington’s disease but uncommon in tardive syndromes.
Eye movements: Eye movement disorders are often a prominent, early feature of Huntington’s disease. They can involve disorders of saccadic initiation, broken pursuits and gaze impersistence. However, in tardive dystonia, the eye movements are generally normal. Thus, a careful oculomotor examination is an important part of the evaluation of all tardive syndromes.
Motor impersistence (of grip, tongue protrusion or gaze fixation): This is a classic feature of Huntington’s disease but is very uncommon in tardive dyskinesia, and therefore a valuable clinical sign.
Other features: Akathisia and opisthotonus strongly suggest tardive syndromes. Conversely, a family history suggesting dominant inheritance and caudate atrophy on MR scan of brain would suggest Huntington’s disease.
Anti-NMDA receptor encephalitis
Several autoimmune movement disorders can have co-existent neurobehavioural features, which are extensively reviewed elsewhere.53 Anti-NMDA receptor encephalitis in particular however, could be confused with tardive dyskinesia, due to the prominent stereotyped perioral dyskinesia that typifies the disorder. The condition presents differently depending on age: children have more ‘neurological’ (seizures, movement disorders) phenotypes, while adults tend to present with neurobehavioural syndromes, which can be mistaken for psychosis.54 Sometimes, the neuropsychiatric features require neuroleptic treatment, creating an additional pitfall in the diagnostic pathway. A ‘full house’ of symptoms, including autonomic dysfunction, generally develops within 1 month.54 Clinical suspicion should prompt testing for the causative antibody in serum and cerebrospinal fluid.
This condition should always be kept in the differential diagnosis of any movement disorder, especially in patients under the age of 40 years (though late presentations are reported). Psychiatric symptoms are common in Wilson’s disease, and perioral movements are also classic. However, they tend to assume a more dystonic quality, frequently producing risus sardonicus. Dysarthria and drooling are also common in Wilson’s disease, but unusual in tardive dyskinesia.
This hyperkinetic movement disorder affects 15% of the edentulous population,55 manifesting as stereotyped, choreiform perioral and lip movements which bear striking resemblance to tardive oro-bucco-lingual dyskinesia. It presents in people with partial or complete edentulism, and often resolves or significantly improves with the introduction of dentures to the mouth. Its pathogenesis is thought to relate to altered sensory feedback from oral structures as a result of malocclusion.
This primary dystonic disorder mostly affects women in their 50s and 60s, being characterised by the combination of blepharospasm and oromandibular dystonia. Differentiation from tardive conditions on purely clinical grounds can be particularly difficult; hence, a history of exposure to dopamine receptor-blocking agents is critical to explore thoroughly in the history.
The management of tardive syndromes should incorporate three key aspects.
First, prevention is always better than cure. As such, medications with documented potential for inducing tardive syndromes should be used at the lowest possible dose for the shortest period of time possible. This may of course not always be possible.
Second comes the question of medication withdrawal. In actual fact, the evidence that withdrawing the offending drug significantly alters the natural history of tardive syndromes is not as strong as one might think.56 Nevertheless, this is an intuitive move in clinical medicine—remove the thing that is causing the problem. Most movement disorder physicians would therefore advocate stopping the offending dopamine receptor-blocking agent, or at least changing it to a drug with less potential for tardive phenomena, if possible. The alternative drug of choice in this setting is often clozapine, both due to its proven efficacy in the treatment of and its lower risk of inducing tardive syndromes.57–59 Close consultation with psychiatric services is necessary before embarking on such a course of action. It is also important to realise that tardive symptoms may initially worsen following neuroleptic drug withdrawal and that equally the symptoms may be suppressed by switching to a more potent dopamine receptor-blocking agent.60
Finally comes the question of symptomatic treatments for tardive syndromes. Numerous agents have been trialled in this regard, with varying evidence for their effectiveness.
As mentioned earlier, tardive syndromes are often a complex medley of different movement disorders, and approaches that may work for one movement may worsen another. It is therefore important to adopt a tailored approach, focused on addressing the issue that primarily bothers the patient; generally, this will be either tardive dyskinesia or tardive dystonia.
Concerning tardive dyskinesia, the mainstay of medical treatment resides around the administration of vesicular monoamine transporter-2 (VMAT-2) inhibitors (tetrabenazine, deutetrabenazine, valbenazine—the latter two being the only Food and Drug Administration-approved drugs for the treatment of tardive dystonia), which act through presynaptic dopamine depletion. The main side effects of these medications are the development of reversible parkinsonism, as well as dose-dependent mood changes, particularly in the elderly; the side effect profiles of deutetrabenazine and valbenazine appear significantly more favourable.61
Other compounds worth mentioning include amantadine, which has shown antidyskinetic properties in multiple controlled and uncontrolled studies, and is supported by American Academy of Neurology guidelines for short-term treatment of tardive dyskinesia. Propranolol has surprisingly good data to support its use, though this is likely due to its effect of increasing neuroleptic drug concentrations.47 Clonazepam also appears effective, though in the randomised controlled trial setting it appeared to lose its efficacy after 5–8 months and thus can only be tentatively recommended for short-term use. Several antioxidants have also been trialled but data on their efficacy are largely inconclusive.60 Other options such as additional dopaminergic blockade, for example, with haloperidol, have proven efficacy in reducing tardive dyskinesia, at least in the short term. However, this comes at the cost of an increase in akinetic rigid syndromes. Furthermore, there are insufficient data on the long-term effects of such approaches, and given that these agents have great propensity to cause tardive syndromes, additional potent dopaminergic blockade is not recommended as a treatment for these conditions.60
Botulinum toxin is an effective option for tardive dystonia.23 Trihexyphenidyl can also improve dystonic syndromes, though occasionally at the cost of worsening dyskinesia.
Functional neurosurgery is gaining increasing recognition as a treatment for both tardive dyskinesia and dystonia. Indeed, pallidal deep-brain stimulation can be greatly beneficial, and early referral to a centre with experience in this procedure should be encouraged in refractory or debilitating cases.62 Physicians may be reluctant to recommend this procedure due to the risk of worsening underlying psychiatric comorbidity, though in practice, this is seldom an issue, especially with pallidal targets.62 Pallidotomy can also be considered in poor surgical candidates.
Tardive akathisia can be equally bothersome, but there is little evidence regarding its optimal treatment. Clonidine, moclobemide and benzodiazepines as well as electroconvulsive therapy have been used in some instances, with varying degrees of success.63–66 Tardive pain syndromes often respond to VMAT-2 inhibitors, though other options such as electroconvulsive therapy have been used.43 Withdrawal emergent dyskinesia often settles spontaneously over a few weeks without treatment. Severe symptoms can however be managed by reintroduction of the offending drug, followed by a slower taper.
In an ideal world, patients developing tardive syndromes would have their causative neuroleptic treatment stopped. Then, and only then, could the true reversibility of the syndrome be assessed. However, the nature of psychiatric disease means that ongoing treatment is often needed, making it difficult to assess the outcomes of tardive syndromes. Predictors of poor outcome appear similar to those of developing tardive syndromes in the first place and include advanced age, longer duration of antipsychotic treatment and greater cumulative dose.67 Once established, the severity of tardive syndromes often fluctuates over time, though in a significant proportion, the tardive syndrome fails to resolve.56 68
Tardive syndromes can comprise many characteristic movement disorders; each needs to be carefully sought in suspected cases
Clozapine is the drug of choice for patients with tardive syndromes who require ongoing neuroleptic treatment
Vesicular monoamine transporter-2 (VMAT-2) inhibitors, such as tetrabenazine, deutetrabenazine and valbenazine, are the best medical treatment options for tardive dyskinesia
Pallidal deep-brain stimulation is an effective treatment option in refractory or debilitating tardive syndromes
Contributors All authors were involved in the conception and design of the work, data acquisition and interpretation, drafting the work, revising it critically for important intellectual content and the final approval of the version to be published.
Funding This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient consent for publication Not required.
Provenance and peer review Commissioned; externally peer reviewed by Oliver Bandmann, Sheffield, UK.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.