HIV-associated sensory neuropathies
Introduction Peripheral neuropathy has emerged as the most common neurological complication of HIV infection [1–4]. There are several discrete types of HIV-associated neuropathy, which can be classified according to the timing of their appearance during HIV infection, their etiology and whether they...
Պահպանված է:
| Հիմնական հեղինակներ: | , , , , |
|---|---|
| Ձևաչափ: | Revisão |
| Լեզու: | անգլերեն |
| Հրապարակվել է: |
2002
|
| Առցանց հասանելիություն: | https://doi.org/10.1097/00002030-200211080-00002 |
| Ցուցիչներ: |
Ավելացրեք ցուցիչ
Չկան պիտակներ, Եղեք առաջինը, ով նշում է այս գրառումը!
|
| Ամփոփում: | Introduction Peripheral neuropathy has emerged as the most common neurological complication of HIV infection [1–4]. There are several discrete types of HIV-associated neuropathy, which can be classified according to the timing of their appearance during HIV infection, their etiology and whether they are primarily axonal or demyelinating (Table 1). Some represent a consequence of HIV infection producing neuropathological damage [e.g., distal symmetrical polyneuropathy (DSP)], while others are related to opportunistic pathogens [e.g., cytomegalovirus (CMV) polyradiculitis]. An increasingly common group is that which occurs as a result of treatment toxicity [e.g., toxic neuropathy from antiretroviral drugs (TNA) and lactic acidosis syndrome].Table 1: Peripheral nervous system (PNS) involvement in HIV infection.HIV-associated sensory neuropathies The HIV-associated sensory neuropathies include both DSP due to HIV infection per se, and TNA associated with the use of dideoxynucleoside reverse transcriptase inhibitors (NRTI), particularly zalcitabine (ddC), stavudine (d4T) and didanosine (ddI). These two conditions are phenotypically identical, and together form the commonest neurological disorder affecting people with HIV/AIDS. This review shall focus primarily on these neuropathies as they form the bulk of peripheral nervous system disease encountered in HIV clinical practice today. Inflammatory demyelinating polyradiculoneuropathies Acute and chronic inflammatory demyelinating polyradiculoneuropathies (AIDP and CIDP, respectively) occur with increased frequency in people with HIV, particularly in the early stages of infection, and probably represent autoimmune phenomena [5]. The clinical and electrophysiological features, as well as the response to immunomodulating therapies are indistinguishable from those seen in non-HIV associated AIDP and CIDP. However the classic finding of ‘dissociation albuminocytologique', where cerebrospinal fluid (CSF) protein is elevated but the CSF is acellular, may be absent, because typically in HIV-associated IDP, there is a lymphocytic pleocytosis in addition to elevated CSF protein. Mononeuropathy multiplex Mononeuropathy multiplex (MM) can occur both early in infection, due to dysimmune or vasculitic mechanisms [6,7], and also in advanced AIDS associated with a low CD4 cell count, usually from opportunistic pathogens such as CMV and varicella-zoster virus (VZV) [8]. MM can also occur secondary to hepatitis B and C viruses, particularly when there is an associated cryoglobulinemia [9]. Electrophysiological studies are helpful in diagnosis, often demonstrating multiple, asymmetric mononeuropathies that cannot be localized to typical sites of entrapment. A nerve/muscle biopsy is usually necessary to distinguish between etiologies that require very different therapies (e.g., immunosuppression for vasculitis versus ganciclovir therapy for CMV neuritis). Progressive polyradiculopathy Progressive polyradiculopathy presents characteristically with lumbosacral pain, ‘saddle’ anesthesia, a rapidly progressive flaccid paraparesis and urinary retention [10,11]. The most common cause is CMV infection occurring in the advanced stages of immunodeficiency, and it is important to note that early anti-CMV therapy may prevent a neurologically devastating outcome. Other causes of this syndrome are VZV infection, neurosyphilis and leptomeningeal lymphoma [12,13]. HIV-associated sensory neuropathies (HIV-SN) Clinical features of HIV-SN The symptoms of HIV-SN are dominated by pain. The pain is typically bilateral, of gradual onset, and described as ‘aching', ‘painful numbness', or ‘burning’ [14,15]. It is usually most severe on the soles of the feet, and is typically worse at night. Patients often have hyperalgesia (lowered pain threshold) and allodynia (pain induced by normally non-noxious stimuli such as rubbing). The feet are tender to touch, wearing shoes is painful, and the gait becomes ‘antalgic'. In a typical length-dependent fashion, the dysesthesias ascend proximally up the lower extremities over months, and may begin to involve the fingertips at around the same time as they reach the mid-leg level. Characteristically, ‘weakness’ is a rare presenting symptom, and objective weakness is absent or confined to the intrinsic foot muscles. Examination usually reveals absent or reduced ankle reflexes in addition to distal sensory loss. Electrophysiological studies, which predominantly test the function of large, myelinated fibers, usually show an axonal, length-dependent, sensory, polyneuropathy. DSP DSP was recognized as a common neurological manifestation of AIDS even before HIV had been identified as the organism underlying this syndrome. A cross-sectional study of hospitalized patients with AIDS at San Francisco General Hospital in the early 1980s revealed that 13 of 37 (35%) had clinical and electrophysiological evidence of DSP [16]. Studies among ambulatory patients with HIV described prevalence rates less than half of this [17], and a study of airforce recruits who were incidentally found to be HIV infected noted DSP to be present in only 12 of 798 people, all of whom had CD4 cell counts of < 200 × 106/l [18]. Indeed, the common finding among the various early studies is that DSP is primarily a complication of late stage HIV disease with advanced immunosuppression. Incidence data from the Multicenter AIDS Cohort Study (MACS) in the pre-highly active antiretroviral therapy (HAART) era estimated an annual incidence of HIV-SN of 7% in those with CD4 cell counts < 200 × 106/l [19]. Studies have revealed an association between viral set point and the subsequent development of HIV-SN. Data from MACS revealed that, in addition to those with a lower CD4 cell count being at higher risk of neuropathy, individuals with HIV RNA >10 000 copies/ml had a 2.3-fold greater risk of HIV-SN than those with < 500 copies/ml [20]. Until recently, it had been assumed that HIV infected children were spared the development of DSP. However, a cross-sectional study has now shown that, while symptoms are, in general, less severe than those observed in adults the prevalence of DSP is similar in children, at 34% [21]. The increasing rates of DSP seen with advancing HIV disease suggest that the pathology develops gradually, and that subclinical or silent nerve damage may be present in many people with HIV who are not yet symptomatic from DSP. This concept is supported by the finding that pathological abnormalities in peripheral nerves are detectable in virtually all patients dying of AIDS [22]. Similarly, elevated thermal thresholds are seen in people with HIV infection even in the absence of overt neuropathy [23]. An early prospective study demonstrated abnormal electrophysiological studies in 50 (89%) of 56 HIV-infected patients, of whom only 21 (42%) had symptoms to suggest DSP [24]. In another study, subclinical peripheral nerve dysfunction was present by quantitative sensory testing in 36% of patients with AIDS or AIDS-related complex [25]. Pathology of DSP DSP is characterized by distal degeneration of long axons [26]. This pattern is often termed ‘dying back', due to the observation that the distal regions of the fibers degenerate first, with centripetal progression. The density of small and large myelinated fibers, and in particular, of unmyelinated fibers is reduced [21,22]. Punch skin biopsies reveal evidence of reduced intra-epidermal nerve fiber density in the distal leg, suggesting prominent involvement of small, unmyelinated fibers (Fig. 1) [27]. This pattern resembles the profile of fiber reduction in diabetic and amyloid neuropathy, which also have prominent small sensory fiber loss. Dorsal root ganglion (DRG) neuronal loss has also been shown in DSP, but the reduction is more modest than the distal axonal loss [28]. Furthermore, selective degeneration of the gracile tracts in patients with DSP, characterized by loss of axons and myelin sheaths in the cervical and upper thoracic cord, has been described [29]. This finding represents degeneration within the centrally directed extensions of the sensory neurons, and is the central nervous system's counterpart of the dying back process seen in the peripheral nerve.Fig. 1.: Skin biopsy from a normal control (a) and a HIV patient with DSP (b). Note the decreased number of epidermal nerve fibers and formation of nerve fiber swellings (b) (bar, 50 μm).Immunopathological studies in DSP have shown prominent macrophage activation with the local release of proinflammatory cytokines in areas of axonal degeneration (Fig. 2). Moreover, there has been consistent demonstration of increased frequency of Nageotte nodules. Nageotte nodules are compact areas of proliferation of satellite cells that frequently accompany DRG neuronal loss from any cause. DRG inflammatory infiltrates are seen (Fig. 3) comprised mainly of lymphocytes and activated macrophages, with concomitant immunostaining for pro-inflammatory cytokines such as tumor necrosis factor (TNF)-α, interferon-γ and interleukin-6 [30–32].Fig. 2.: Macrophage infiltration of peripheral nerve as observed by immunocytochemistry with anti-CD68 antibodies (a) Cross section. (b) Longitudinal section. (Bar, 50 μm.) Reproduced with permission from [26].Fig. 3.: DRG pathology in HIV neuropathy is characterized by foci of macrophage-lymphocytic infiltration (a) and Nagoette nodules (b). Nagoette nodules are compact areas of proliferation of satellite cells that frequently accompany DRG neuronal loss from any cause. Infiltration by activated macrophages is demonstrated by immunostaining with anti-CD68 antibodies (c). (Bar, 50 μm.) Reproduced with permission from [26].What causes the multifocal macrophage activation in peripheral nerves and DRG in DSP? The answer is not yet known, and there are two complementary and not necessarily mutually exclusive hypotheses. In one theory, it is postulated that a mild degree of distal axonal degeneration occurs because of nutritional deficiencies, alcohol exposure, substance abuse [33] or other nonspecific factors. As in other types of Wallerian-like degeneration, macrophages would then be recruited into the affected nerves. However, in HIV infection, these macrophage responses to the axonal degeneration are ‘hyperactive', resulting in multifocal inflammation in the nerve and DRG. In the second theory, circulating activated monocytes and pro-inflammatory cytokines enter DRG and peripheral nerves in excessive numbers through a ‘leaky’ blood–nerve barrier. This ingress of activated cells leads to further release of chemokines and cytokines with subsequent axonal and DRG neuronal injury (Table 2).Table 2: Theories on the etiology of distal symmetrical polyneuropathy.The role of HIV in DSP In contrast to the consistent high levels of macrophage activation in both the peripheral nerve and DRG of patients with DSP, productive HIV replication in these tissues appears to be sparse and confined to monocytes/macrophages. Brannagan et al. demonstrated HIV infection of DRG neurons using PCR in situ hybridization studies (ISH) in patients with and without neuropathy [34]. Others, by contrast, have shown that HIV is localized predominantly in perivascular inflammatory cells and the nodules of Nageotte [30,32]. ISH has demonstrated rare HIV-1 replication within peripheral nerve and DRG macrophages in both patients with DSP and in HIV-positive patients without neuropathy. This paucity of local HIV replication, together with abundant macrophage activation in DSP resembles the situation in HIV dementia and vacuolar myelopathy, in that proinflammatory cytokines rather than direct effects of HIV are thought to mediate much of the neurotoxicity [35]. However, there is some evidence that HIV or viral gene products may be directly neurotoxic. In a transgenic mice model expressing the entire HIV genome in neurons under the transcriptional control of a neurofilament promoter, axonal degeneration was observed in the sciatic nerves of 50% of the transgenic mice, in the absence of inflammatory infiltrates [36]. There is evidence for the relevance of ongoing systemic viral replication to DSP, with plasma HIV-1 RNA levels correlating directly with the severity of pain [37]. The pathogenesis of pain in DSP The degeneration of rostral gracile tracts as well as distal sensory axons, has suggested that the primary pathology of DSP could be at the level of the DRG. It is hypothesized that the aberrant inflammatory response results in ‘sick’ DRG neurons and subsequent ‘dying-back’ axonal degeneration. This process however sheds little light on the prominence of pain in DSP, which is often out of proportion to the degree of nerve fiber loss. Two main theories, which are not mutually exclusive, have been proposed, with multifocal inflammation in nerve and DRG being the initial prerequisite step in both (Fig. 4).Fig. 4.: Hypothetical model of pathogenesis of pain in DSP. (1) Injury of peripheral nerve fibers due to multifocal inflammation and secreted macrophage activation products results in abnormal spontaneous activity of neighboring uninjured nociceptive fibers (`peripheral sensitization'). (2) Furthermore, the aberrant inflammatory response in DRG leads to alterations in neuronal sodium and calcium channel expression and ectopic impulse generation. (3) This results in central remodeling within the dorsal horn due to A-fiber sprouting and synaptic formation with pain fibers in lamina II, and maintenance of neuropathic pain (`central sensitization').The first, or ‘peripheral', hypothesis is that the neuropathic pain associated with DSP is derived from spontaneous activity of uninjured C (nociceptive or pain) fibers after injury in neighboring peripheral nerve fibers. Wu et al. [38] made this observation after ligation and transection of the L5 spinal nerve in the rat. Furthermore, Wallerian degeneration can lead to altered properties in adjacent intact nociceptive afferents [39]. The prominence of the macrophage inflammation in DSP has suggested that the local release of proinflammatory cytokines sensitizes nerves. Indeed, in animal models, endoneural injection of TNF-α into the sciatic nerve has been shown to sensitize nociceptive fibers and to produce neuropathic pain [40]. The second or ‘central’ hypothesis is that alterations in DRG neuronal sodium and calcium channel expression and function result in abnormal processing of pain after peripheral nerve injury [41]. Central remodeling within the dorsal horn of the spinal cord is thought to play an important role in the maintenance of neuropathic pain [42]. Several studies have shown that peripheral nerve injury in animal models is followed by sprouting of Aβ-fibers in central terminals and formation of new synaptic contacts outside their normal termination zones into the lamina II of the dorsal horn, an area that normally receives unmyelinated nociceptive fiber input [43–45]. This plasticity and fiber organization in the dorsal horn is probably modulated by several factors, including the induction of growth factors, growth factor receptors and cytokines [46,47]. Finally, a direct role for viral proteins such as gp120 in the genesis of the neuropathic pain of DSP has been suggested by studies showing that gp120 induces allodynia when injected into rat paws [48], when delivered epineurally to the rat sciatic nerve [49] and when administered intrathecally [50]. Miller et al. has shown that DRG neurons express a wide variety of chemokine receptors, including those that are thought to act as receptors for gp120. Furthermore, this group demonstrated that gp120 produces excitatory effects on cultured DRG nociceptor neurons and stimulates the release of substance P. However, the hypothesis that gp120 acts directly on sensory neurons to elicit pain is based on the still unproven assumption that gp120 is shed from infected macrophages in sufficient amounts to affect surrounding cells. TNA With the introduction of combination antiretroviral therapy in the mid 1990s, the incidence of most neurological complications of HIV has fallen dramatically [51]. HIV-SN, however, has actually become more prevalent in the last decade. This increase began at around the same time as dideoxynucleosides (particularly ddC) were introduced into clinical practice. In the MACS, the annual incidence of HIV-SN was noted to have significantly and progressively increased over the time period 1988–1992 to 2.81% among all CD4 cell count groups, and 7% in those with a CD4 cell count of < 200 × 106/l. Multivariate analysis showed that neither CD4 cell count nor viral load was an independent predictor of HIV-SN, as had been true in several earlier studies. However, the use of antiretroviral therapy was independently associated with an increased risk of neuropathy [19]. In a moderately immunosuppressed cohort (average CD4 cell count of 160 × 106/l), many of whom were on dual dideoxynucleoside therapy, Schifitto et al. have recently charted extremely high rates for the development of symptomatic sensory neuropathy; 25% at 1 year and 52% at 2 years of follow-up [52]. The NRTI, ddC, ddI and d4T, have well-recognized peripheral neurotoxicity. TNA was the dose-limiting toxicity of all these agents from the time they were introduced into phase 1 clinical trials, with the incidence of TNA relating to both the dose and duration of drug exposure. ddC is the most potent neurotoxin among these drugs. TNA was seen in all patients given ddC in doses > 0.04 mg/kg/day, 80% of those receiving 0.04 mg/kg/day, and a third of those receiving lower dose therapy [53]. Similarly, in a randomized trial involving d4T, TNA was seen within 2–4 months of beginning treatment, in 31% of patients on 2 mg/kg/day, 15% of those on 1 mg/kg/day, and in 6% of those receiving 0.5 mg/kg/day [54]. In another trial, SN developed in 34% of patients receiving low-dose ddC (2.25 mg/day), but in only 4% of comparable patients treated with zidovudine (ZDV) alone [55]. TNA is also seen with ddI. This finding was common with the high doses used in initial dose ranging trials, but is much less frequent with the currently recommended doses [56]. An initial trial of ddI and d4T used in combination found that this combination was safe, and associated with a low incidence of neurotoxicity [57]. However, this study was conducted in relatively immunocompetent patients, who were therefore at intrinsically low risk of HIV-SN. Data from several years of experience with dideoxynucleoside use in our own clinic has shown that the risk of TNA is synergistic for regimens containing combination ddI and d4T when compared to either drug used alone [58] (Fig. 5). The co-administration of hydroxurea with dideoxynucleosides also increases the risk of TNA [58,59]. Risk factors for TNA identified in clinical trials using dideoxynucleosides include pre-existing neuropathy from any cause (e.g., due to diabetes, vitamin B12 deficiency and heavy alcohol consumption), older age, poorer nutrition, and more advanced HIV disease [60,61].Fig. 5.: Incidence rates of TNA/100 person years in 1116 patients, receiving one of five antiretroviral regimens and followed by the Johns Hopkins AIDS Service.With the increasing use of combination antiretroviral therapy (often including dideoxynucleosides) for long periods of time and in patients with long standing HIV infection, it is likely that TNA will become an increasingly important problem for people living with HIV. Indeed, Cherry et al. have recently shown that the current, cross-sectional prevalence of neuropathy even in relatively immunocompetent, ambulatory HIV infected patients (average CD4 cell count > 400 × 106/l) is 44% in one clinic. This finding represents a significant increase from the 14% prevalence described in the same clinic in 1993. Furthermore, having ever used one or more dideoxynucleoside agents was the most important risk factor for neuropathy in this cohort [62]. Clinical features of TNA The clinical symptoms, physical examination findings and electrophysiological studies in TNA are very similar to those of DSP [63,64]. The distinguishing clinical feature is the temporal association between initiating one or more dideoxynucleoside NRTI and the onset or worsening of sensory symptoms. The timing of onset of TNA has ranged from 1 week to 6 months after NRTI initiation in clinical trials, depending on the NRTI and the dose administered. Symptomatic improvement over weeks to months has been reported to occur in about two-thirds of patients after discontinuation of the precipitating drug, often preceded by an initial period of worsening symptoms, or ‘coasting’ [55]. The onset of TNA is typically more acute than the onset of DSP, and pain may be more prominent. However, in many cases, these two conditions are indistinguishable. The failure of at least one-third of cases of TNA to improve upon cessation of the precipitating NRTI increases the difficulty of distinguishing TNA from DSP. The pathogenesis of TNA It has been hypothesized that NRTI toxicity is mediated by mitochondrial dysfunction [65] (Fig. 6). Complications of NRTI therapy resemble the clinical manifestations of inherited mitochondrial diseases, which include peripheral neuropathy, hepatic steatosis, lipodystrophy, lactic acidosis, ‘ragged red fiber’ myopathy (ZDV) and pancreatitis (ddI, d4T) [66]. Furthermore, mitochondrial ultrastructural abnormalities have been noted in affected tissues [67,68]. Mitochondrial dysfunction has also been suggested by the demonstration of an increased serum lactate level in patients with TNA [69]. Moreover, hyperlactatemia and muscle pain following low-level exercise are common in patients receiving NRTI [70].Fig. 6.: Various proposed mechanisms of mitochondrial toxicity by the NRTI. They include ‘late’ mechanisms such as (1) mitochondrial DNA depletion by DNA polymerase-γ inhibition, and ‘early’ mechanisms leading to impairment of oxidative phosphorylation such as (2) interference with the activity of respiratory complexes forming the electron transport chain (I–V), (3) adenylate kinase (AK) activity and (4) the function of the ADP/ATP translocator (T).As all the NRTIs have been shown in vitro to inhibit mitochondrial DNA polymerase-γ, the host enzyme responsible for mitochondrial DNA synthesis, mitochondrial toxicity may result from mitochondrial DNA depletion [71,72]. However, Cui and coworkers showed that although the observed toxicity of ddI and ddC on neuron-like PC-12 cells correlated with inhibition of mitochondrial DNA synthesis, no change in mitochondrial DNA content was observed by d4T at doses that caused toxic effects on neurites [73]. This finding suggests that NRTI mitochondrial toxicity may be mediated by mechanisms other than mitochondrial DNA depletion. NRTI exert rapid cellular toxicity by directly inhibiting mitochondrial bioenergetic function in a tissue-specific fashion [74]. ZDV inhibits NADH-linked respiration and NADH-cytochrome c reductase activity in isolated rat skeletal muscle, brain and liver mitochondria [75]. Furthermore, ZDV inhibits adenylate kinase and the ADP/ATP translocator in isolated liver mitochondria, resulting in early impairment of oxidative phosphorylation [76–78]. In cultured human muscle cells, ZDV reduces the activity of SDH, a complex II protein that is encoded by nuclear rather than mitochondrial DNA [79]. ddC induces cardiotoxicity rapidly in rats, and this disorder is associated with decreased activity of respiratory complexes, but not with mitochondrial DNA depletion [80]. Furthermore, no correlation was demonstrated between the ability of nucleoside analogs to increase lactate production and their potency in mitochondrial DNA depletion [81]. Development of strategies to treat or prevent TNA has been hampered by the lack of a reproducible animal model. A model of ddC neurotoxicity in New Zealand White (NZW) rabbits has been reported [82,83]. In contrast to the axonal pathology seen in HIV-infected individuals, this neuropathy was described as being predominantly demyelinating with prominent involvement of motor fibers. Another group [84] failed to reproduce TNA in NZW rabbits despite using high doses of ddI and d4T for 24 weeks. We were unable to establish a rodent model of TNA. Despite daily intraperitoneal administration for 8 months of up to 100 mg/kg of ddC to rats, no effect on unmyelinated nerve fibers in the foot pad was observed (J.C. McArthur, unpublished data). Furthermore, in a pre-licensing study, primates receiving high doses of d4T for 1 year did not develop TNA [85]. These difficulties in establishing an animal model of TNA may reflect the concomitant need for HIV infection, with resultant macrophage infiltration into peripheral nerve, for manifestation of NRTI-induced neurotoxicity. This hypothesis is supported by analogous attempts to establish an animal model of NRTI-induced cardiomyopathy. HAART administration resulted in mitochondrial cardiomyopathy and elevated plasma lactate in HIV transgenic mice but not in wild-type mice [86]. How might HIV infection predispose individuals to TNA? Possible explanations include a ‘double-hit’ hypothesis, whereby DRG neurons are first damaged/sensitized by an aberrant inflammatory response associated with HIV infection, and are then further compromised by NRTI-induced mitochondrial toxicity. A possible contributing factor is that the perturbation of the blood–nerve barrier that is suspected to exist in HIV-infected patients might result in easier access of circulating cytokines or NRTI to DRG neurons and subsequent neurotoxicity [87]. The importance of DSP in the pathogenesis of TNA TNA is phenotypically similar to DSP. In addition, TNA is consistently found to be more common in patients with any degree of pre-existing DSP in clinical trials of dideoxynucleosides. It has thus been hypothesized that the development of TNA is due to ‘unmasking’ of clinically silent pre-existing neuronal damage due to HIV. In this model, HIV infection, and consequent abnormalities of the peripheral nervous system are prerequisites for the development of TNA. Treatment of HIV-SN Current treatment of HIV-SN is primarily symptomatic as there are no proven regenerative therapies to reverse the underlying pathogenesis. Antiretroviral treatments have been demonstrated to improve thermal thresholds in those who show virological responses [88]. In the case of TNA, typically the suspected offending NRTI is either stopped or its dose decreased, thus risking loss of virologic control. Furthermore, despite drug withdrawal, a substantial proportion of patients continue to experience debilitating pain. The use of pain modifying agents Pain modifying agents have been used with varying success to manage neuropathic pain in other conditions, such as diabetic neuropathy. A number of recent studies have been performed to investigate the efficacy of some of these therapies in the symptomatic control of HIV-SN. Some degree of pain relief has been reported in small, open label trials of both topical 5% lidocaine gel [89] and the anticonvulsant drug gabapentin [90], but controlled data are lacking. Results of placebo-controlled trials of pain modifying therapies have generally been disappointing, with most agents (including amitryptiline, mexilitine, topical capsaicin and acupuncture) shown to be either ineffective or not more effective than placebo in relieving pain [91–94]. To date, the only symptomatic therapies shown to be effective in relieving the pain of HIV-SN in randomized, placebo-controlled trials are lamotrigine and recombinant human nerve growth factor (NGF) [95–97]. Lamotrigine Lamotrigine is an anticonvulsant agent that blocks voltage-sensitive sodium channels, and inhibits glutamate and aspartate release. In a small randomized, double-blind study in the treatment of HIV-SN, lamotrigine use was associated with a substantial decrease in pain compared with placebo [95]. These data have now been extended to a much larger group of subjects (227), demonstrating that the therapeutic effect is seen primarily in patients with TNA rather than DSP [96]. This difference is as yet unexplained. An advantage of lamotrigine over some other anticonvulsant drugs in the setting of HIV is that it is metabolized by glucuronic acid conjugation rather than the cytochrome P450 system, and therefore its use is not limited by interactions with protease inhibitors. Although rash was a common adverse effect of lamotrigine in the original study affecting five of the 20 patients, no difference was found in the incidence of rash between those taking lamotrigine and placebo in the large-scale trial. Recombinant human NGF NGF is a neurotrophin that is critical for the development and maintenance of small, unmyelinated sensory fibers [98]. A multicenter, 18-week, double-blind, placebo-controlled study evaluated the safety and efficacy of two doses (0.1 μg/kg and 0.3 μg/kg subcutaneously twice weekly) of recombinant human NGF in 270 patients with HIV-SN (AIDS Clinical Trials Group 291) [97]. Both doses of NGF were associated with major improvements in average and maximum daily pain scores compared with placebo. However, there was no improvement in objective signs of neuropathy severity as assessed by neurologic examination, quantitative sensory testing and epidermal nerve fiber density. The most frequent adverse event was injection site pain. The findings of the extended 48-week open-label phase of ACTG 291 were similar to those of the double-blind study [99]. Despite the apparent usefulness of NGF in relieving neuropathic pain in HIV-SN, there are no plans currently to develop this agent for clinical use. There are, however, a number of other potential neuroregenerative agents under investigation for possible use as human therapeutics, including the neuroimmunophilin ligands [100] and prosaposin [101]. Conclusions HIV-SN is the most prevalent neuropathy associated with HIV/AIDS, and is now the commonest neurological complication of HIV infection. Studies performed prior to the availability of antiretroviral therapy documented HIV-SN (in this case DSP) affecting over one-third of AIDS patients. With the introduction of NRTI (particularly the dideoxynucleosides, ddC, ddI and d4T), a second form of HIV-SN, TNA, became apparent. Several authors have described an increasing prevalence of HIV-SN in the face of declining rates of almost all other neurological complications of HIV since the introduction of combination antiretroviral therapy in the mid-1990s. It is likely that TNA accounts for a large part of this increase, resulting from the increasing exposure of people living with HIV to potentially neurotoxic antiretroviral agents. DSP and TNA are phenotypically identical. Thus it has been hypothesized that dideoxynucleosides may be ‘unmasking’ clinically silent HIV-mediated neuropathy, and that some degree of DSP underlies cases of TNA. This possibility is supported by the fact that TNA occurs more frequently in those with any degree of pre-existing clinical DSP. We propose that HIV infection results in multifocal aberrant inflammation and macrophage activation in the peripheral nerve and DRG, akin to the microglia/macrophage activation characteristic of HIV dementia and vacuolar myelopathy. This process leads to loss of both large myelinated and unmyelinated fibers in a ‘dying back’ pattern. The use of dideoxynucleosides may exacerbate or magnify this process by causing mitochondrial toxicity of DRG neurons. The development of spontaneous activity in uninjured nociceptive fibers, changes in the patterns of expression of DRG neuronal ion channels, and remodeling within the dorsal horn leads to persistent neuropathic pain. There is a need for improved methods of determining the degree of subclinical DSP in an individual prior to commencing potentially neurotoxic antiretroviral agents, and of monitoring patients for early mitochondrial dysfunction once they commence therapy, in order to better describe the individual's risk of developing TNA. Furthermore, the establishment of a robust animal model of TNA is urgently needed to allow the development of preventive and therapeutic strategies, and the screening of novel antiretroviral agents for inherent neurotoxicity. Sponsorship: S.C.K., C.A.P., A.H., J.C.M. are supported by NS26643, AI 3, NS35619, RR00522, Johns Hopkins University Center for AIDS Research and the Blaustein foundation. C.L.C. is supported by NHMRC. |
|---|