Treatment of Distal Colitis with Local Anaesthetic Agents

Authors


Author for correspondence: Håkan Ahlman, Department of Surgery, Sahlgrenska University Hospital, University of Göteborg, Göteborg, Sweden (fax +46 31 82 21 04, e-mail hakan.ahlman@surgery.gu.se).

Abstract

Abstract: The results of clinical and experimental studies on topical treatment of distal colitis with local anaesthetic agents are summarized. The original observation was an adrenergic hyperinnervation of the inflamed mucosa (hyperinnervation hypothesis). In order to silence local nervous reflexes, the mucosa was treated topically with 2% lidocaine gel. The clinical results are promising and no side effects have been observed. The relapse rate is relatively high and related to the duration of treatment. In studies of experimental colitis a potential antagonism between harmful adrenergic nerves (vasoconstrictor substances and proinflammatory cytokines) and mucosa-protective visceral afferents (antiinflammatory cytokines) in the pathogenesis of colitis is intriguing. Other studies have emphasized the importance of neutrophils for causing damage to the colon epithelium (neutrophil hypothesis) and local anaesthetics have potent effects on several steps of the inflammatory response in addition to the nervous blockade.

Ulcerative colitis is a mucosal inflammation, which starts in the rectum and extends into the colon. Localized forms of ulcerative colitis are proctitis and proctosigmoiditis (distal colitis), which seem to increase in frequency in patients with ulcerative colitis. In different series in 40–70% of ulcerative colitis patients the distal colon is involved in the disease. The age of onset of proctosigmoiditis is in general 10 years higher (30–40 years) than seen in other forms of inflammatory bowel disease (Farmer 1987). The mucosal immune response in ulcerative colitis involves activation of both humoural and cellular immune mechanisms with B- and T-cell responses and antibody formation against dietary factors, bacteria and auto-antigens (MacDonald & Murch 1994). No specific pathogen has been isolated, but the normal enteric flora itself may be of importance for development of disease (Sartor 1997). Despite large research efforts neither a specific antigen triggering a specific antibody response, nor a specific immune cell clone have been defined as causative of inflammatory bowel disease. It is obvious that certain individuals can have genetic susceptibility for inflammatory bowel disease, but it seems reasonable to suggest that several risk factors interact, i.e. that the disease has a polygenetic background (Schreiber 1997).

Histopathological findings

On inspection, the diseased mucosa displays hyperaemia, granularity, friability and ulcerations. The acute inflammatory response is associated with accumulation of neutrophils around deeper portions of the crypts. Chronic inflammation, involving the lamina propria, develops in less than half of the patients and may lead to flattening of the mucosa with branching of the crypts and decreased mucosal height. Widely spread distorted crypts can also be found in the inflamed lamina propria (Riddell 1988). A diffusely inflamed and atrophic mucosa is characteristic of quiescent ulcerative colitis. The course of the disease is difficult to predict, but the risk of proximal spread of proctosigmoiditis decreases with time (Farmer 1987).

Treatment strategies

The conventional treatment of distal colitis varies, but most patients are treated topically with corticosteroid enemas and sulphasalazine perorally, or topically with 5-aminosalicylic acid (5-ASA) compounds. Relapsing disease is thought to represent an imbalance between pro-inflammatory mediators (e.g. TNF-α) and anti-inflammatory mediators (e.g. IL10). New therapeutic principles may thus be pharmacological or immunological inhibition of pro-inflammatory mediators (phosphodiesterase inhibition, TNF-α antibodies) or enhancement of anti-inflammatory mediators (recombinant IL10). Transcriptional control of pro-inflammatory genes can be exerted by elements of the JAK-STAT family and NF-κ B (p65), which are upregulated in inflammatory bowel disease. Potential therapeutic interventions may thus include antisense oligonucleotides and cell-permeable peptides (Schreiber 1997).

The hyperinnervation hypothesis

In the pathophysiology of ulcerative colitis an autonomic imbalance between sympathetic and parasympathetic nerves was early suggested, i.e. an adrenergic preponderance might produce both increased epithelial turnover and vasoconstriction resulting in mucosal damage (Tutton & Helme 1974). Also the observed beneficial effects of nicotine on inflammatory bowel disease support this suggested autonomic imbalance (Calkins 1989). In our initial studies of patients with proctitis (Björck et al. 1989) we observed an adrenergic hyperinnervation of the submucosa, muscularis mucosae and mucosa, also reported by Kyösola et al. (1977) Hyperinnervation with neuropeptide Y-containing nerves and marked infiltration of the mucosa with certain subsets of T-lymphocytes were also observed (Björck et al. 1992). Our hypothesis was that: 1) Excessive release of noradrenaline and neuropeptide Y causes postcapillary vasoconstriction that leads to oedema and congestion of the mucosa; 2) An adrenergic drive on epithelial stem cells leads to increased cell proliferation and stress-ulcer formation (Tutton & Helme 1974); and 3) A disturbed innervation dysregulates the immune system, i.e. mucosal nerves may release homing factors for T-lymphocytes (Felten et al. 1985).

Clinical results with lidocaine treatment

In order to silence hyperactive local nervous reflexes we treated the diseased mucosa topically with 2% lidocaine gel (400 mg twice daily), resulting in symptomatic relief and restored mucosal integrity in the majority of patients. The best treatment results in our consecutive series of 100 patients (Björck et al. 1992) were seen in those with limited disease, i.e. all patients with proctitis (n=28) responded within 3–12 weeks despite the fact that two thirds of the patients had prior failure with other forms of medical therapy. However, relapses after cessation of treatment were seen in 68% within the first two years (fig. 1). In a small French series of therapy-resistant proctitis the majority of patients were symptom-free within two weeks and healed endoscopically within 3–7 weeks (Vilotte et al. 1991). Encouraging results were also achieved in our patients with proctosigmoiditis (n=49) with 83% response rate after longer treatment periods (6–34 weeks). Less than half of these patients recurred within 16 months after interruption of treatment. Seventeen patients with left-sided colitis displayed symptomatic relief and improved histology after 2–4 months with 23% recurrence within the first year after cessation of treatment. In one small Japanese series of left-sided ulcerative colitis, ulcers were healed and the number of mucosal erosions was reduced already after 2 weeks of lidocaine treatment (Taoka et al. 1996). The possibility to topically treat diseased mucosa in total colitis was considered to be low and we have only treated very few patients with total colitis (fig. 1).

Figure 1.

Characterization of the 100 patients reported in Björck et al. (1992): A. Duration of symptoms before lidocaine treatment, B. Duration of lidocaine treatment, C. Observation time from onset of treatment to last endoscopy, D. Recurrence rate during the time of observation (p=proctitis, p.s.=proctosigmoiditis, l.c.=left-sided colitis and t.c.=total colitis)

There were no complaints about adverse effects despite continuous therapy over several months. Many patients reported prompt relief of symptoms. The discharge of blood and mucus gradually diminished within the first 7–10 days of treatment and tenesmus symptoms were clearly reduced. All patients with proctitis responded to lidocaine treatment with the disappearance of symptoms and gradual improvement of mucosal histology. In proctosigmoiditis 41 out of 49 patients showed clinical remission. Two were non-responsive: one developed fulminant colitis and the other responded to subsequent cortisone therapy. Six patients reported symptom relief, but did not show objective histological improvement according to the study criteria (reduction of lymphocytes and polydendritic cells, normalized crypts, and reduced hyperinnervation). Rectal application of 300 mg lidocaine generates maximal plasma concentrations after 1 hour with total elimination within 8 hours (de Boer et al. 1979). Using the 2% gel (400 mg lidocaine) the maximum plasma levels ranged between 0.5–1.9 mg/l in patients with proctitis studied two hours after application of gel.

The mucosal and submucosal hyperinnervation varied between patients, probably reflecting the severity of inflammation. Already early in the disease, i.e. at diagnosis, the number of adrenergic and neuropeptide Y-nerves was increased with enlarged varicose nerve terminals and thickened nerve bundles. Most prominent in refractory disease was the additional hyperplasia of substance P- and vasoactive intestinal peptide-immunoreactive nerve fibres and terminals. The total innervation pattern was assessed by an S-100-antiserum (fig. 2), which also labelled the polydendritic antigen-presenting cells abundant around the crypts of diseased mucosa (Björck et al. 1992). With healing and restitution of the epithelium, the polydendritic cells disappeared from the mucosa. T lymphocytes (OKT4+ and OKT8+) invaded the lamina propria and epithelial lining in the diseased state and were part of the crypt inflammation and crypt abscesses. With successful treatment these lymphocytes gradually disappeared (fig. 3). Treatment was stopped when the number of lymphocytes was normalized. The number of plasma cells in the lamina propria was reduced in parallel with the number of lymphocytes (Björck et al. 1992).

Figure 2.

Mucosal biopsy from a patient with left-sided colitis incubated with anti-S-100. The patient had been on topical lidocaine treatment for 2 weeks. A dense network of immunoreactive Schwann cell sheaths, which accompany the majority of nerve terminals, can be seen around crypts. Arrows indicate subepithelial accumulation of polydendritic antigen-presenting cells, also immunopositive for S-100. Bar=50 μm.

Figure 3.

Mucosal biopsies from a patient with proctosigmoiditis before (A) and after (B) 3 weeks of topical lidocaine treatment. Biopsies were incubated with monoclonal antibodies against OKT 8+ lymphocytes. In diseased state the mucosa was massively invaded by these lymphocytes, but normalized after treatment. Bar=50 μm.

Possible mode of action of lidocaine

The mechanism of action of lidocaine in distal colitis may be multimodal. We initially proposed adrenergic/peptidergic hyperinnervation as an early event in the pathogenesis of mucosal inflammation, but we cannot exclude that the hyperinnervation is secondary to inflammation. Conceiveably hyperactive nervous reflexes are inhibited by lidocaine, but the agent may also exert a direct influence on immune functions, e.g. lymphocyte maturation and proliferation and macrophage migration (Ramus et al. 1983; Dickstein et al. 1985). Also, secretion of cytokines may be influenced by lidocaine. Parameters reflecting the degree of inflammation (infiltration of T lymphocytes and plasma cells and presence of polydendritic antigen-presenting cells) were easy to monitor. The presence of polydendritic cells subepithelially during active disease may indicate a low-resistance epithelium with antigen leakage into the lamina propria. One function of plasma cells is to secrete and cover the luminal surface with IgA bound to “secretory component” to protect the intestine from pathogens (Mayer 1990); this defense function seems to be enhanced in active disease and normalizes during treatment. Also more general morphological features of the diseased mucosa, i.e. low epithelial height, reduced mucus stores in goblet cells, short and branched crypts, could be restituted after successful treatment with lidocaine for 4 weeks.

The cause of the inflammation in ulcerative colitis is still obscure, but lidocaine treatment represents a new principle, which may interfere with several pathogenetic mechanisms. Its mode of action in relation to pro- or antiinflammatory cytokines and to the transcripton of proinflammatory genes has not been investigated. Lidocaine treatment had a high success rate without apparent side effects and can be used to treat patients resistant to other therapy. The relapse rate after treatment is, however, as high as after conventional treatment, but may relate to the duration of treatment (Björck et al. 1992). Lidocaine treatment inhibits transmittor release not only from efferent autonomic nerves, but also from afferent fibers (axon reflexes), which may release, e.g. substane P and/or CGRP.

Lidocaine in experimental colitis

Beneficial effects of local, or systemic, lidocaine were seen in experimental colitis induced by trinitrobenzene sulfonic acid (McCafferty et al. 1994). These authors found that lidocaine, given 30 min. before trinitrobenzene sulfonic acid, dose-dependently could reduce the severity of colitis and also reduce the number of granulocytes at 24 hr and 7 days after induction of colitis. Similar effects were seen if lidocaine was given after the induction of colitis. In a later study these authors (McCafferty et al. 1997) combined ablation of primary afferents by capsaicin treatment in neonate animals, or chemical sympathectomy, with trinitrobenzene sulfonic acid challenge and lidocaine treatment. Their results strongly favour a pathogenetic role of adrenergic nerves, since sympathectomy alone reduced the severity of colitis. Lidocaine treatment gave full mucosal protection of the sympathectomized animals. On the other hand, the primary afferents may have a protective role, since inhibition of these afferents by capsaicin treatment (adult rats) increased the damage score. Lidocaine administration in these capsaicin-treated rats still attenuated the macroscopic damage (McCafferty et al. 1997). Mazelin et al. (1998) found that capsaicin, locally applied on the vagus nerve (vagal deafferentiation), enhanced the inflammatory response in trinitrobenzene sulfonic acid colitis also in favour of a protective role of substance P-containing vagal afferents. Reinshagen et al. (1996) showed that capsaicin-sensitive nerves had a protective effect in the acute and subacute phases of experimental trinitrobenzene sulfonic acid colitis, which was diminished in later stages of chronic inflammation (fig. 4). The neuronal hypothesis may be one explanation for the relapse rate observed, since after lidocaine withdrawal it is likely that the innervation gradually resumes pretreatment activity.

Figure 4.

The neuronal hypothesis indicates a pathogenic role of vasocontrictor substances (NA, neuropeptide Y (NPY)) and homing factors for T-lymphocytes released from adrenergic nerves, which also may enhance the release of pro-inflammatory cytokines. On the other hand, visceral afferents (mucosa-protective nerves) may inhibit the release of proinflammatory cytokines and enhance the release of anti-inflammatory cytokines from lymphocytes.

In another rat model of colitis, induced by treatment with peroral dextran sulfate for 7 days, the permeation of Evans blue from the lumen into the gut wall of proximal and distal colonic loops was studied (Björck et al. 1997). Dextran sulfate is a large and negatively charged molecule with low absorption from the luminal surface. One interesting feature of dextran sulfate-induced colitis is the invasion of polydendritic cells shortly (1–2 days) after ingestion of dextran sulfate, which differs from other rat colitis models, e.g. acetic acid. Dextran sulfate treatment caused a rapid permeation of Evans blue into the colonic mucosa. During induction of colitis potential therapeutic agents (lidocaine, mesalazine, prednisolone or sucralfate) were applied topically daily for mucosal protection. All these agents have been used in the treatment of patients with ulcerative colitis. Mesalazine (5-amino salicylic acid) inhibits the synthesis of prostaglandins and platelet-activating factor and also reduces binding of chemotactic peptides to neutrophils. Corticosteroids influence the maturation of lymphocytes, the synthesis of lymphokines besides the inhibition of cyclooxygenase activity and reduce the amount of nuclear NF-κ B via cytoplasmic inhibitors and thereby down-regulate proinflammatory gene transcription (Schreiber 1997). Sucralfate was chosen, since it is chemically related to dextran sulfate and has been shown to be effective against ulcers in experimental colitis induced by acetic acid (Zahavi et al. 1987). The mucosal changes were evaluated with respect to the abundance of substance P- and neuropeptide Y-nerves, invasion of polydendritic cells and mucin-containing goblet cells. In the proximal colon a significant inhibition of Evans blue permeation was observed after treatment either with lidocaine, prednisolone, or sucralfate. However, in the distal colon only lidocaine had such protective effect (table 1). Increased numbers of substance P- and neuropeptide Y-immunoreactive nerve fibres were seen in regions with crypt abnormalities; and goblet cells in these regions were devoid of mucin. The invasion of polydendritic cells was similarly reduced by lidocaine, prednisolone or sucralfate in the proximal colon, and by lidocaine or prednisolone, but not by sucralfate, in the distal colon. We therefore conclude that influx of antigens from the lumen is essential for induction of colitis, reflected by the appearance of polydendritic cells. This influx is effectively prevented by lidocaine treatment in both the proximal and distal colon. The nerve hyperplasia may be secondary to luminal challenge with antigens and mucosal inflammation. The discrepancy between increased permeation and absent polydendritic cell response in the distal loops after prednisolone may reflect separate actions of steroids on gut epithelium and immune cells (Björck et al. 1997).

Table 1. 
Proximal colonDistal colon
  1. Permeation of Evans blue into the gut wall from the lumen in proximal and distal colonic loops after 2 hr exposure in controls and DSS-treated animals. During 7 days of dextran sulfate (DSS) treatment the animals were daily treated topically with potentially protective agents (lidocaine, mesalazine, prednisolone, or sucralfate). The enhanced permeation after DSS treatment was prevented by all drugs except mesalazine in the proximal loop, while only lidocaine had protective effect in the distal loop.

1.Controls96±19 g Evans blue/g
colon tissue×10−7
221±30 g Evans blue/g
colon tissue×10−7
n=6
2.DSS182±31350±36n=6
3.Lidocaine+DSS53±8214±29n=8
4.Mesalazine+DSS128±34323±41n=7
5.Prednisolone+DSS68±29306±58n=6
6.Sucralfate+DSS78±43352±46n=6
Group 1 vs. 2 P<0.05Group 1 vs. 2 P<0.05 
Group 2 vs. 3 P<0.01Group 2 vs. 3 P<0.01 
Group 2 vs. 5 and 2 vs. 6 P<0.05

The exact mode of action of dextran sulfate in experimental colitis is not known. From histopathological studies inflammation seems to be secondary to crypt damage (Cooper et al. 1993). From our experimental studies we would suggest that the primary event is the opening of tight junctions followed by an increase in permeation of the epithelium. Lidocaine could prevent this effect probably via both its membrane-stabilizing properties and via the nervous blockade. The observed substance P- nerve hyperplasia may parallel an up-regulation of this neuropeptide as seen in Clostridium difficile toxin A-enteritis, which also is associated with mucosal hyperemia and increased leakage (Pothoulakis et al. 1995).

The neutrophil hypothesis

Other studies have emphasized the importance of the neutrophils for causing damage to the colonic epithelium, resulting in impaired barrier function. Neutrophils produce potentially harmful substances, e.g. eicosanoids, proteinases and free oxygen radicals, which may disrupt tight junctions and lead to mucosal ulcer formation (Nash et al. 1991; Babbs 1992; Nusrat et al. 1997). Circulating leukocytes are attracted to the tissue damage site; the activated leukocytes will first adhere and roll along the endothelium before migration into tissue. The interaction between leukocytes and endothelial cells is promoted by cytokines, eicosanoids and adhesion molecules, e.g. integrins, selectins and members of the immunoglobulin superfamily (Springer 1994, Ley 1996). Local anaesthetic agents reversibly block nerve impulse propagation by preventing influx of Na+ ions (Strichartz 1976), but may in addition influence several other cell functions, e.g. collagen synthesis, cell motility, NK-cell mediated cell lysis, adhesion, leukocyte phagocytosis, lysosomal function and histamine release from mast cells, and modulate the inflammatory response (Jönsson 1996). The latter properties may be due to additional blockade of calcium-regulated K+ channels. The involvement of neutrophils in colitis has, however, been challenged. In rats made neutropenic (95% reduction of circulating neutrophils and 85% reduction of tissue myeloperoxidase activity) the major pathologic features of colitis developed essentially as in rats with normal neutrophil counts (Buell & Berin 1994).

Ropivacaine and colitis

Ropivacaine is a long-acting local anaesthetic of the amide type (structure similar to bupivacaine) with high safety profile developed for nerve blocks and epidural analgesia (McClure 1996). Ropivacaine has been suggested as an alternative to lidocaine in the treatment of distal colitis. It undergoes oxidative metabolism in the liver by cytochrome 450 enzymes and 1% is excreted unchanged in the urine (Ekström & Gunnarsson 1996). Its metabolism can be dose-dependently inhibited by certain antimicrobial agents (Jokinen et al. 2001).

The cardiac and CNS toxicity of ropivacaine has been studied in man. Continuous epidural analgesia with ropivacaine (20 mg/hr) results in low concentration levels (0.88 mg/l) (Ericksen et al. 1996). Knudsen et al. (1997) reported CNS toxicity at a mean concentration of 2.2 mg/l after intravenous infusion in volunteers. After nerve blockades considerably higher levels (2.9–3.6 mg/l) have been observed in patients without side effects (Hickey et al. 1990; Wulf et al. 1999). Cardiac arrythmia has been reported in one patient with an estimated concentration of 7.5 mg/l (Ruetsch et al. 1999). Thus, there seem to be no single critical concentration threshold for CNS or cardiovascular toxicity.

In 1996 Arlander et al. tested ropivacaine in a pharmacokinetic study of 12 patients with distal colitis using ropivacaine topically (200 mg×2). The mean peak plasma concentrations (Cmax) were 0.99–1.37 mg/l after treatment for 1–14 days. The observed decrease of Cmax and area under plasma concentration-time curve may be due to decreased absorption and/or changed metabolism. The median time of Cmax was 2 hr and the mean half-life 2.7 hr. Clinical symptoms of colitis were rapidly reduced. Although the duration of treatment was short and only the most severely diseased site assessed, there were significantly reduced endoscopic scores in treated patients parallelled by improved mucosal scores in blinded biopsies. No side effects were seen. Thus, ropivacaine as topical treatment over two weeks does not accumulate and carries low risk for systemic effects.

The antiinflammatory properties of ropivacaine were studied by vital microscopy in the hamster cheek pouch vasculature after challenge with leukotrienes. Ropivacaine abolished leukocyte rolling and adherence to the endothelium and markedly reduced vascular permeability. In vitro ropivacaine also inhibited the TNF-α up-regulated expression of CD11b/CD18 (Martinsson et al. 1997a). Furthermore, ropivacaine in vitro reduced the release of arachidonic acid derivatives from leukocytes. This effect was pronounced for compounds generated via the lipooxygenase pathway (leukotrienes, prostaglandins) and did not influence cytoprotective substances like prostacyclin. The effects were superior to 5-amino salicylic acid, but inferior to steroids (Martinsson et al. 1997b). In cell culture ropivacaine caused a dose-dependent membrane depolarisation and inhibited the proliferation of human fibroblasts, endothelial cells and epithelial cells. The drug concentration range (100 μM-1mM) used was similar to that seen in the colonic wall of patients treated rectally with the drug (Martinsson et al. 1993).

Topical treatment with ropivacaine (once daily during 1 week) was tested in rat experimental colitis, induced by a single administration of trinitrobenzene sulfonic acid. Such treatment resulted in much less pronounced morphological changes of the treated mucosa and decreased myeloperoxidase activity. Topical treatment with ropivacaine had more pronounced protective effect than steroids, when mucosal damage was scored, while 5-amino salicylic acid had no significant effect. Each of the three agents used reduced myeloperoxidase activity (Martinsson et al. 1999).

Local anaesthetics as topical treatment of distal colitis have given new clues to the pathophysiology of ulcerative colitis and can clearly improve symptoms. The mode of action may relate to silencing of hyperactive nerves, which in turn modulate the immune system. A potential antagonism between adrenergic nerves and visceral afferents in the pathogenesis of ulcerative colitis is intriguing. Local anaesthetics also have beneficial effects on several steps of the inflammatory response.

Acknowledgements

This work has been supported by grants from the Swedish MRC (2207, 5220) and the Ulf Widengren Memorial Fund.

Ancillary