Here, we outline a method of applying existing machine learning (ML) approaches to aid citation screening in an on-going broad and shallow systematic review of preclinical animal studies. The aim is ...to achieve a high-performing algorithm comparable to human screening that can reduce human resources required for carrying out this step of a systematic review.
We applied ML approaches to a broad systematic review of animal models of depression at the citation screening stage. We tested two independently developed ML approaches which used different classification models and feature sets. We recorded the performance of the ML approaches on an unseen validation set of papers using sensitivity, specificity and accuracy. We aimed to achieve 95% sensitivity and to maximise specificity. The classification model providing the most accurate predictions was applied to the remaining unseen records in the dataset and will be used in the next stage of the preclinical biomedical sciences systematic review. We used a cross-validation technique to assign ML inclusion likelihood scores to the human screened records, to identify potential errors made during the human screening process (error analysis).
ML approaches reached 98.7% sensitivity based on learning from a training set of 5749 records, with an inclusion prevalence of 13.2%. The highest level of specificity reached was 86%. Performance was assessed on an independent validation dataset. Human errors in the training and validation sets were successfully identified using the assigned inclusion likelihood from the ML model to highlight discrepancies. Training the ML algorithm on the corrected dataset improved the specificity of the algorithm without compromising sensitivity. Error analysis correction leads to a 3% improvement in sensitivity and specificity, which increases precision and accuracy of the ML algorithm.
This work has confirmed the performance and application of ML algorithms for screening in systematic reviews of preclinical animal studies. It has highlighted the novel use of ML algorithms to identify human error. This needs to be confirmed in other reviews with different inclusion prevalence levels, but represents a promising approach to integrating human decisions and automation in systematic review methodology.
Background
Antiepileptic drugs have been used for treating different types of neuropathic pain, and sometimes fibromyalgia. Our understanding of quality standards in chronic pain trials has improved ...to include new sources of potential bias. Individual Cochrane reviews using these new standards have assessed individual antiepileptic drugs. An early review from this group, originally published in 1998, was titled 'Anticonvulsants for acute and chronic pain'. This overview now covers the neuropathic pain aspect of that original review, which was withdrawn in 2009.
Objectives
To provide an overview of the relative analgesic efficacy of antiepileptic drugs that have been compared with placebo in neuropathic pain and fibromyalgia, and to report on adverse events associated with their use.
Methods
We included reviews published in theCochrane Database of Systematic Reviews up to August 2013 (Issue 7). We extracted information from each review on measures of efficacy and harm, and methodological details concerning the number of participants, the duration of studies, and the imputation methods used, in order to judge potential biases in available data.
We analysed efficacy data for each painful condition in three tiers, according to outcome and freedom from known sources of bias. The first tier met current best standards ‐ at least 50% pain intensity reduction over baseline (or its equivalent), without the use of last observation carried forward (LOCF) for dropouts, an intention‐to‐treat (ITT) analysis, in parallel group studies with at least 200 participants lasting eight weeks or more. The second tier used data from at least 200 participants where one or more of the above conditions were not met. The third tier of evidence related to data from fewer than 200 participants, or with several important methodological problems that limited interpretation.
Main results
No studies reported top tier results.
For gabapentin and pregabalin only we found reasonably good second tier evidence for efficacy in painful diabetic neuropathy and postherpetic neuralgia. In addition, for pregabalin, we found evidence of efficacy in central neuropathic pain and fibromyalgia. Point estimates of numbers needed to treat for an additional beneficial effect (NNTs) were in the range of 4 to 10 for the important outcome of pain intensity reduction over baseline of 50% or more.
For other antiepileptic drugs there was no evidence (clonazepam, phenytoin), so little evidence that no sensible judgement could be made about efficacy (valproic acid), low quality evidence likely to be subject to a number of biases overestimating efficacy (carbamazepine), or reasonable quality evidence indicating little or no effect (lamotrigine, oxcarbazepine, topiramate). Lacosamide recorded such a trivial statistical superiority over placebo that it was unreliable to conclude that it had any efficacy where there was possible substantial bias.
Any benefits of treatment came with a high risk of adverse events and withdrawal because of adverse events, but serious adverse events were not significantly raised, except with oxcarbazepine.
Authors' conclusions
Clinical trial evidence supported the use of only gabapentin and pregabalin in some neuropathic pain conditions (painful diabetic neuropathy, postherpetic neuralgia, and central neuropathic pain) and fibromyalgia. Only a minority of people achieved acceptably good pain relief with either drug, but it is known that quality of life and function improved markedly with the outcome of at least 50% pain intensity reduction. For other antiepileptic drugs there was no evidence, insufficient evidence, or evidence of a lack of effect; this included carbamazepine. Evidence from clinical practice and experience is that some patients can achieve good results with antiepileptics other than gabapentin or pregabalin.
There is no firm evidence to answer the important pragmatic questions about which patients should have which drug, and in which order the drugs should be used. There is a clinical effectiveness research agenda to provide evidence about strategies rather than interventions, to produce the overall best results in a population, in the shortest time, and at the lowest cost to healthcare providers.
Pain and the global burden of disease Rice, Andrew S C; Smith, Blair H; Blyth, Fiona M
Pain (Amsterdam),
04/2016, Letnik:
157, Številka:
4
Journal Article
A distal symmetrical sensory peripheral neuropathy is frequently observed in people living with Human Immunodeficiency Virus Type 1 (HIV-1). This neuropathy can be associated with viral infection ...alone, probably involving a role for the envelope glycoprotein gp120; or a drug-induced toxic neuropathy associated with the use of nucleoside analogue reverse transcriptase inhibitors as a component of highly active anti-retroviral therapy. In order to elucidate the mechanisms underlying drug-induced neuropathy in the context of HIV infection, we have characterized pathological events in the peripheral and central nervous system following systemic treatment with the anti-retroviral agent, ddC (Zalcitabine) with or without the concomitant delivery of HIV-gp120 to the rat sciatic nerve (gp120+ddC). Systemic ddC treatment alone is associated with a persistent mechanical hypersensitivity (33% decrease in limb withdrawal threshold) that when combined with perineural HIV-gp120 is exacerbated (48% decrease in threshold) and both treatments result in thigmotactic (anxiety-like) behaviour. Immunohistochemical studies revealed little ddC-associated alteration in DRG phenotype, as compared with known changes following perineural HIV-gp120. However, the chemokine CCL2 is significantly expressed in the DRG of rats treated with perineural HIV-gp120 and/or ddC and there is a reduction in intraepidermal nerve fibre density, comparable to that seen in herpes zoster infection. Moreover, a spinal gliosis is apparent at times of peak behavioural sensitivity that is exacerbated in gp120+ddC as compared to either treatment alone. Treatment with the microglial inhibitor, minocycline, is associated with delayed onset of hypersensitivity to mechanical stimuli in the gp120+ddC model and reversal of some measures of thigmotaxis. Finally, the hypersensitivity to mechanical stimuli was sensitive to systemic treatment with gabapentin, morphine and the cannabinoid WIN 55,212-2, but not with amitriptyline. These data suggests that both neuropathic pain models display many features of HIV- and anti-retroviral-related peripheral neuropathy. They therefore merit further investigation for the elucidation of underlying mechanisms and may prove useful for preclinical assessment of drugs for the treatment of HIV-related peripheral neuropathic pain.
Summary Background Existing treatments for postherpetic neuralgia, and for neuropathic pain in general, are limited by modest efficacy and unfavourable side-effects. The angiotensin II type 2 ...receptor (AT2 R) is a new target for neuropathic pain. EMA401, a highly selective AT2 R antagonist, is under development as a novel neuropathic pain therapeutic agent. We assessed the therapeutic potential of EMA401 in patients with postherpetic neuralgia. Methods In this multicentre, placebo-controlled, double-blind, randomised, phase 2 clinical trial, we enrolled patients (aged 22–89 years) with postherpetic neuralgia of at least 6 months' duration from 29 centres across six countries. We randomly allocated 183 participants to receive either oral EMA401 (100 mg twice daily) or placebo for 28 days. Randomisation was done according to a centralised randomisation schedule, blocked by study site, which was generated by an independent, unmasked statistician. Patients and staff at each site were masked to treatment assignment. We assessed the efficacy, safety, and pharmacokinetics of EMA401. The primary efficacy endpoint was change in mean pain intensity between baseline and the last week of dosing (days 22–28), measured on an 11-point numerical rating scale. The primary efficacy analysis was intention to treat. This trial is registered with the Australian New Zealand Clinical Trials Registry, number ACTRN12611000822987. Findings 92 patients were assigned to EMA401 and 91 were assigned to placebo. The patients given EMA401 reported significantly less pain compared with baseline values in the final week of treatment than did those given placebo (mean reductions in pain scores −2·29 SD 1·75 vs −1·60 1·66; difference of adjusted least square means −0·69 SE 0·25; 95% CI −1·19 to −0·20; p=0·0066). No serious adverse events related to EMA401 occurred. Overall, 32 patients reported 56 treatment-emergent adverse events in the EMA401 group compared with 45 such events reported by 29 patients given placebo. Interpretation EMA401 (100 mg twice daily) provides superior relief of postherpetic neuralgia compared with placebo at the end of 28 days of treatment. EMA401 was well tolerated by patients. Funding Spinifex Pharmaceuticals.
Background
This review is an update of 'Topical capsaicin (high concentration) for chronic neuropathic pain in adults' last updated in Issue 2, 2013. Topical creams with capsaicin are used to treat ...peripheral neuropathic pain. Following application to the skin, capsaicin causes enhanced sensitivity, followed by a period with reduced sensitivity and, after repeated applications, persistent desensitisation. High‐concentration (8%) capsaicin patches were developed to increase the amount of capsaicin delivered; rapid delivery was thought to improve tolerability because cutaneous nociceptors are 'defunctionalised' quickly. The single application avoids noncompliance. Only the 8% patch formulation of capsaicin is available, with a capsaicin concentration about 100 times greater than conventional creams. High‐concentration topical capsaicin is given as a single patch application to the affected part. It must be applied under highly controlled conditions, often following local anaesthetic, due to the initial intense burning sensation it causes. The benefits are expected to last for about 12 weeks, when another application might be made.
Objectives
To review the evidence from controlled trials on the efficacy and tolerability of topically applied, high‐concentration (8%) capsaicin in chronic neuropathic pain in adults.
Search methods
For this update, we searched CENTRAL, MEDLINE, Embase, two clinical trials registries, and a pharmaceutical company's website to 10 June 2016.
Selection criteria
Randomised, double‐blind, placebo‐controlled studies of at least 6 weeks' duration, using high‐concentration (5% or more) topical capsaicin to treat neuropathic pain.
Data collection and analysis
Two review authors independently searched for studies, extracted efficacy and adverse event data, and examined issues of study quality and potential bias. Where pooled analysis was possible, we used dichotomous data to calculate risk ratio and numbers needed to treat for one additional event, using standard methods.
Efficacy outcomes reflecting long‐duration pain relief after a single drug application were from the Patient Global Impression of Change (PGIC) at specific points, usually 8 and 12 weeks. We also assessed average pain scores over weeks 2 to 8 and 2 to 12 and the number of participants with pain intensity reduction of at least 30% or at least 50% over baseline, and information on adverse events and withdrawals.
We assessed the quality of the evidence using GRADE and created a 'Summary of findings' table.
Main results
We included eight studies, involving 2488 participants, two more studies and 415 more participants than the previous version of this review. Studies were of generally good methodological quality; we judged only one study at high risk of bias, due to small size. Two studies used a placebo control and six used 0.04% topical capsaicin as an 'active' placebo to help maintain blinding. Efficacy outcomes were inconsistently reported, resulting in analyses for most outcomes being based on less than complete data.
For postherpetic neuralgia, we found four studies (1272 participants). At both 8 and 12 weeks about 10% more participants reported themselves much or very much improved with high‐concentration capsaicin than with 'active' placebo; the point estimates of numbers needed to treat for an additional beneficial outcome (NNTs) were 8.8 (95% confidence interval (CI) 5.3 to 26) at 8 weeks and 7.0 (95% CI 4.6 to 15) at 12 weeks (2 studies, 571 participants; moderate quality evidence). More participants (about 10%) had average 2 to 8‐week and 2 to 12‐week pain intensity reductions over baseline of at least 30% and at least 50% with capsaicin than control, with NNT values between 10 and 12 (2 to 4 studies, 571 to 1272 participants; very low quality evidence).
For painful HIV‐neuropathy, we found two studies (801 participants). One study reported the proportion of participants who were much or very much improved at 12 weeks (27% with high‐concentration capsaicin and 10% with 'active' placebo). For both studies, more participants (about 10%) had average 2 to 12‐week pain intensity reductions over baseline of at least 30% with capsaicin than control, with an NNT of 11 (very low quality evidence).
For peripheral diabetic neuropathy, we found one study (369 participants). It reported about 10% more participants who were much or very much improved at 8 and 12 weeks. One small study of 46 participants with persistent pain following inguinal herniorrhaphy did not show a difference between capsaicin and placebo for pain reduction (very low quality evidence).
We downgraded the quality of the evidence for efficacy outcomes by one to three levels due to sparse data, imprecision, possible effects of imputation methods, and susceptibility to publication bias.
Local adverse events were common, but not consistently reported. Serious adverse events were no more common with active treatment (3.5%) than control (3.2%). Adverse event withdrawals did not differ between groups, but lack of efficacy withdrawals were somewhat more common with control than active treatment, based on small numbers of events (six to eight studies, 21 to 67 events; moderate quality evidence, downgraded due to few events). No deaths were judged to be related to study medication.
Authors' conclusions
High‐concentration topical capsaicin used to treat postherpetic neuralgia, HIV‐neuropathy, and painful diabetic neuropathy generated more participants with moderate or substantial levels of pain relief than control treatment using a much lower concentration of capsaicin. These results should be interpreted with caution as the quality of the evidence was moderate or very low. The additional proportion who benefited over control was not large, but for those who did obtain high levels of pain relief, there were usually additional improvements in sleep, fatigue, depression, and quality of life. High‐concentration topical capsaicin is similar in its effects to other therapies for chronic pain.
Background
Although often considered to be lacking adequate evidence, nonsteroidal anti‐inflammatory drugs (NSAIDs) are widely used in the management of neuropathic pain. Previous surveys found 18% ...to 47% of affected people reported using NSAIDs specifically for their neuropathic pain, although possibly not in the United Kingdom (UK).
Objectives
To assess the analgesic efficacy of oral NSAIDs for chronic neuropathic pain in adults, when compared to placebo or another active intervention, and the adverse events associated with its use in clinical trials.
Search methods
We searched CENTRAL, MEDLINE, and EMBASE from inception to 29 May 2015, together with reference lists of retrieved papers and reviews, and an online trials registry.
Selection criteria
We included randomised, double‐blind studies of two weeks duration or longer, comparing any oral NSAID with placebo or another active treatment in chronic neuropathic pain.
Data collection and analysis
Two review authors independently searched for studies, extracted efficacy and adverse event data, and examined issues of study quality. We did not carry out any pooled analysis.
Main results
We included two studies involving 251 participants with chronic low back pain with a neuropathic component or postherpetic neuralgia; 209 of these participants were involved in a study of an experimental NSAID not used in clinical practice, and of the remaining 42, only 16 had neuropathic pain. This represented only third tier evidence, and was of very low quality. There was no indication of any significant pain reduction with NSAIDs. Adverse event rates were low, with insufficient events for any analysis.
Authors' conclusions
There is no evidence to support or refute the use of oral NSAIDs to treat neuropathic pain conditions.
Cannabinoids, cannabis, and cannabis-based medicines (CBMs) are increasingly used to manage pain, with limited understanding of their efficacy and safety. We summarised efficacy and adverse events ...(AEs) of these types of drugs for treating pain using randomised controlled trials: in people of any age, with any type of pain, and for any treatment duration. Primary outcomes were 30% and 50% reduction in pain intensity, and AEs. We assessed risk of bias of included studies, and the overall quality of evidence using GRADE. Studies of <7 and >7 days treatment duration were analysed separately. We included 36 studies (7217 participants) delivering cannabinoids (8 studies), cannabis (6 studies), and CBM (22 studies); all had high and/or uncertain risk of bias. Evidence of benefit was found for cannabis <7 days (risk difference 0.33, 95% confidence interval 0.20-0.46; 2 trials, 231 patients, very low-quality evidence) and nabiximols >7 days (risk difference 0.06, 95% confidence interval 0.01-0.12; 6 trials, 1484 patients, very low-quality evidence). No other beneficial effects were found for other types of cannabinoids, cannabis, or CBM in our primary analyses; 81% of subgroup analyses were negative. Cannabis, nabiximols, and delta-9-tetrahydrocannabinol had more AEs than control. Studies in this field have unclear or high risk of bias, and outcomes had GRADE rating of low- or very low-quality evidence. We have little confidence in the estimates of effect. The evidence neither supports nor refutes claims of efficacy and safety for cannabinoids, cannabis, or CBM in the management of pain.
Study participants with HIV-associated sensory polyneuropathy (HIV-SN) had higher plasma triglyceride concentrations, depression, anxiety, catastrophizing scores, and prevalence of insomnia than HIV ...participants without HIV-SN.
HIV-associated sensory neuropathy (HIV-SN) is a frequent complication of HIV infection and a major source of morbidity. A cross-sectional deep profiling study examining HIV-SN was conducted in people living with HIV in a high resource setting using a battery of measures which included the following: parameters of pain and sensory symptoms (7day pain diary, Neuropathic Pain Symptom Inventory NPSI and Brief Pain Inventory BPI), sensory innervation (structured neurological examination, quantitative sensory testing QST and intraepidermal nerve fibre density IENFD), psychological state (Pain Anxiety Symptoms Scale-20 PASS-20, Depression Anxiety and Positive Outlook Scale DAPOS, and Pain Catastrophizing Scale PCS, insomnia (Insomnia Severity Index ISI), and quality of life (Short Form (36) Health Survey SF-36). The diagnostic utility of the Brief Peripheral Neuropathy Screen (BPNS), Utah Early Neuropathy Scale (UENS), and Toronto Clinical Scoring System (TCSS) were evaluated. Thirty-six healthy volunteers and 66 HIV infected participants were recruited. A novel triumvirate case definition for HIV-SN was used that required 2 out of 3 of the following: 2 or more abnormal QST findings, reduced IENFD, and signs of a peripheral neuropathy on a structured neurological examination. Of those with HIV, 42% fulfilled the case definition for HIV-SN (n=28), of whom 75% (n=21) reported pain. The most frequent QST abnormalities in HIV-SN were loss of function in mechanical and vibration detection. Structured clinical examination was superior to QST or IENFD in HIV-SN diagnosis. HIV-SN participants had higher plasma triglyceride, concentrations depression, anxiety and catastrophizing scores, and prevalence of insomnia than HIV participants without HIV-SN.
The EQIPD framework for rigor in animal experiments aims to unify current recommendations based on evidence behind their rationale and was prospectively tested for feasibility in multicenter animal ...experiments.